Language：Japanese   Publisher：帝京大学  

researchmap

Language：English   Publishing type：Research paper (scientific journal)  

A drone swarm is a robotic architecture having multiple drones cooperate to accomplish a mission. Nowadays, heterogeneous drone swarms, in which a small number of gateway drones (GWs) act as protocol translators to enable the mixing of multiple swarms that use independent wireless protocols, have attracted much attention from many researchers. Our previous work proposed Path Optimizer — a method to minimize the number of end-to-end path-hops in a remote video monitoring system using heterogeneous drone swarms by autonomously relocating GWs to create a shortcut in the network for each communication request. However, Path Optimizer has limitations in improving communication quality when more video sessions than the number of GWs are requested simultaneously. Path Coordinator, which we propose in this paper, achieves a uniform reduction in end-to-end hops and maximizes the allowable hop satisfaction rate regardless of the number of sessions by introducing the cooperative and synchronous relocation of all GWs. Path Coordinator consists of two phases: first, physical optimization is performed by geographically relocating all GWs (relocation phase), and then logical optimization is achieved by modifying the relaying GWs of each video flow (rerouting phase). Computer simulations reveal that Path Coordinator adapts to various environments and performs as well as we expected. Furthermore, its performance is comparable to the upper limits possible with brute-force search.

DOI： 10.1587/transcom.2022EBP3091

researchmap

Language：English   Publishing type：Research paper (international conference proceedings)  

DOI： 10.1109/ICNC57223.2023.10074541

researchmap

Language：English   Publishing type：Research paper (scientific journal)  

Heterogeneous drone swarms are large hybrid drone clusters in which multiple drones with different wireless protocols are interconnected by some translator drones called GWs. Nowadays, because inexpensive drones, such as toy drones, have become widely used in society, the technology for constructing huge drone swarms is attracting more and more attention. In this paper, we propose an autonomous GW mobility control algorithm for establishing stabilized and low-delay communication among heterogeneous clusters, assuming that only GWs are controllable and relocatable to ensure the flexible operationality of drone swarms. Our proposed algorithm is composed of two independent sub algorithms - the Link Stabilizer and the Path Optimizer. The Stabilizer maintains the neighbor links and consists of two schemes: the neighbor clustering based on relative velocities and the GW velocity calculation using a kinetic model. The Optimizer creates a shortcut to reduce the end-to-end delay for newly established communication by relocating the GW dynamically. We also propose a conceptual protocol design to implement this algorithm into real-world drone swarms in a distributed manner. Computer simulation reveals that the Stabilizer improved the connection stability for all three mobility models even under the high node mobility, and the Optimizer reduced the communication delay by the optimal shortcut formation under any conditions of the experiments and its performance is comparable to the performance upper limit obtained by the brute-force searching.

DOI： 10.1587/transcom.2021WWP0010

researchmap

Language：English   Publishing type：Research paper (scientific journal)  

DOI： 10.1587/transcom.2020NVP0005

Scopus

researchmap

Language：English   Publishing type：Research paper (international conference proceedings)  

DOI： 10.34385/proc.63.d1-5

researchmap

Language：English   Publisher：arXiv preprint, arXiv:2010.15572v2 [cs.NI]  

In recent years, more and more applications use ad-hoc networks for local M2M communications, but in some cases such as when using WSNs, the software processing delay induced by packets relaying may not be negligible. In this paper, we planned and carried out a delay measurement experiment using Raspberry Pi Zero W. The results demonstrated that, in low-energy ad-hoc networks, processing delay of the application is always too large to ignore; it is at least ten times greater than the kernel routing and corresponds to 30% of the transmission delay. Furthermore, if the task is CPU-intensive, such as packet encryption, the processing delay can be greater than the transmission delay and its behavior is represented by a simple linear model. Our findings indicate that the key factor for achieving QoS in ad-hoc networks is an appropriate node-to-node load balancing that takes into account the CPU performance and the amount of traffic passing through each node.

DOI： 10.48550/arXiv.2010.15572

researchmap

Language：English   Publishing type：Research paper (international conference proceedings)  

researchmap

近年，アドホックネットワークは，IoT分野のみならず，自動車走行支援のITSや次世代電力網スマートグリッドに代表されるCPSなど， これからの社会基盤を担うであろう注目株な各技術において，積極的な活用が模索されている． このようなアプリケーションの多様化を背景に，異種アドホックネットワーク間相互接続の需要は拡大が見込まれる． また昨今，通信主体が人からモノへ変わったことで，ネットワークインフラには一層厳しい遅延保証が要求される傾向にあり， アドホックネットワークにおいても，通信遅延低減は主要な課題となりうる． 本研究では，プロトコルトランスレータ (GW) による異種アドホックネットワーク間の直接相互接続を想定し， ネットワーク中のフローに許容遅延制約が存在するとしたときの，GW配置最適化による許容遅延満足率最大化について提案する． 提案手法の性能をシミュレーションで評価し，さらにリンク切断予測の性能向上効果についても考察した．

researchmap

researchmap

researchmap

researchmap

researchmap

researchmap

ネットワーク内のマルウェア感染端末検出には，ブラックリストに基づくシグニチャ型の検出手法が一般的である．しかしマルウェア多様化のため，近い将来この手法では未知マルウェア検出が困難になると推測される．そこで本稿では未知マルウェア検出を目指し，機械学習を用いたアノマリ型の検出手法を提案する．DNSクエリ/応答のログ情報から抽出した特徴ベクトルの時系列データを，RNNを用いて分散表現に変換し，そのクラスタ分析を行なった．実験により，送信元IPアドレスの挙動が2クラスに分類され，少数側のクラスタがいずれも特異的なクエリ送信をしていることを確認した．

researchmap

HTTP/2 では，単一接続上での要求多重化やサーバプッシュによりページロード時間 （PLT） 短縮を図っている．しかし，複数の接続を利用するマルチオリジン構成の Web ページにおいては，十分な PLT 短縮効果が得られない．これに対し，ページ書き換えに基づく接続集約により，多重化効率を向上する手法が提案されているが，スクリプトから動的に要求を送信する Web ページにおいては，この手法では完全な接続集約を行えない．本研究では ServiceWorker 用い，動的な Web ページにおいて完全な接続集約とクロスオリジンサーバプッシュを実現し，従来手法比最大 40% の PLT 短縮を達成した．

researchmap

Other Link： http://orcid.org/0000-0002-1423-3337

researchmap

researchmap

Language：English   Publishing type：Research paper (scientific journal)  

DOI： 10.1587/transcom.2018EBP3368

Scopus

researchmap

Language：English   Publishing type：Research paper (international conference proceedings)  

DOI： 10.1109/PIMRC.2018.8580986

researchmap

Language：English   Publishing type：Research paper (scientific journal)   Publisher：Institute of Electronics, Information and Communication, Engineers, IEICE  

DOI： 10.1587/transcom.2017NRP0012

Scopus

researchmap

Language：English   Publishing type：Research paper (scientific journal)  

researchmap

Language：English   Publishing type：Research paper (scientific journal)  

DOI： 10.1364/JOCN.10.000209

researchmap

Language：English   Publishing type：Research paper (international conference proceedings)   Publisher：IEEE Computer Society  

DOI： 10.1109/COMPSAC.2017.190

Scopus

researchmap

Language：English   Publishing type：Research paper (scientific journal)  

researchmap

Language：English   Publishing type：Research paper (international conference proceedings)   Publisher：Institute of Electrical and Electronics Engineers Inc.  

DOI： 10.1109/ICC.2017.7997202

Scopus

researchmap

Language：English   Publishing type：Research paper (international conference proceedings)   Publisher：Institute of Electrical and Electronics Engineers Inc.  

DOI： 10.1109/IWQoS.2017.7969176

Scopus

researchmap

Language：Japanese   Publishing type：Research paper (scientific journal)  

researchmap

Language：English   Publishing type：Research paper (scientific journal)  

DOI： 10.1587/transfun.E99.A.1518

researchmap

Language：English   Publishing type：Research paper (international conference proceedings)  

Web of Science

researchmap

Language：English   Publishing type：Research paper (international conference proceedings)  

DOI： 10.1145/2984356.2985236

Web of Science

researchmap

Language：English   Publishing type：Research paper (international conference proceedings)   Publisher：Institute of Electrical and Electronics Engineers Inc.  

DOI： 10.1109/CQR.2014.7152634

Scopus

researchmap

Language：English   Publishing type：Research paper (international conference proceedings)  

DOI： 10.1109/RNDM.2015.7325235

Web of Science

researchmap

Publishing type：Research paper (international conference proceedings)   Publisher：IEEE  

DOI： 10.1109/GLOCOM.2014.7417660

researchmap

Other Link： https://dblp.uni-trier.de/db/conf/globecom/globecom2015.html#MiyataBYK15

Language：English   Publishing type：Research paper (international conference proceedings)  

DOI： 10.1109/NGMAST.2015.42

Web of Science

researchmap

Publishing type：Research paper (international conference proceedings)   Publisher：ACM  

DOI： 10.1145/2795381.2795384

researchmap

Other Link： https://dblp.uni-trier.de/db/conf/mobiarch/mobiarch2015.html#TagamiYSY15

Language：English   Publishing type：Research paper (scientific journal)  

DOI： 10.1587/transcom.E97.B.2706

researchmap

Language：English   Publishing type：Research paper (international conference proceedings)  

DOI： 10.1109/RNDM.2014.7014947

Web of Science

researchmap

Language：English   Publishing type：Research paper (international conference proceedings)  

Web of Science

researchmap

Language：English   Publishing type：Research paper (scientific journal)  

DOI： 10.1587/transfun.E95.A.2470

Web of Science

researchmap

Language：English   Publishing type：Research paper (scientific journal)  

DOI： 10.1587/transcom.E95.B.1953

Web of Science

researchmap

Language：Japanese   Publishing type：Research paper (scientific journal)  

researchmap

Language：English   Publishing type：Research paper (international conference proceedings)  

DOI： 10.1109/ICASSP.2012.6288123

Web of Science

researchmap

Language：English   Publishing type：Research paper (scientific journal)  

DOI： 10.1587/transcom.E95.B.832

researchmap

Language：Japanese   Publisher：（社）電子情報通信学会  

researchmap

Language：English   Publishing type：Research paper (international conference proceedings)  

DOI： 10.1109/ICIP.2011.6116216

Web of Science

researchmap

Language：English   Publishing type：Research paper (scientific journal)  

researchmap

Language：English   Publishing type：Research paper (international conference proceedings)  

Web of Science

researchmap

Language：English   Publishing type：Research paper (scientific journal)  

DOI： 10.1109/JLT.2008.928958

Web of Science

researchmap

Language：English   Publishing type：Research paper (scientific journal)  

DOI： 10.1587/transcom.E92.B.34

Web of Science

researchmap

Language：English   Publishing type：Research paper (scientific journal)  

DOI： 10.1587/transcom.E92.B.68

Web of Science

J-GLOBAL

researchmap

Other Link： http://dblp.uni-trier.de/db/journals/ieicet/ieicet92b.html#journals/ieicet/MisumiNCSY09

Language：English   Publishing type：Research paper (scientific journal)  

DOI： 10.1093/ietcom/e91-b.3.722

researchmap

Language：English   Publishing type：Research paper (international conference proceedings)  

Web of Science

researchmap

Language：English   Publishing type：Research paper (international conference proceedings)  

Web of Science

researchmap

Language：English   Publishing type：Research paper (international conference proceedings)  

Web of Science

researchmap

Language：English   Publishing type：Research paper (international conference proceedings)  

Web of Science

researchmap

Language：English   Publishing type：Research paper (international conference proceedings)  

DOI： 10.1109/ICC.2007.51

Scopus

researchmap

Language：English   Publishing type：Research paper (international conference proceedings)  

Web of Science

researchmap

Language：English   Publishing type：Research paper (scientific journal)  

DOI： 10.1002/ecja.20253

Scopus

researchmap

Language：Japanese   Publishing type：Research paper (scientific journal)  

researchmap

Language：English   Publishing type：Research paper (scientific journal)  

DOI： 10.1093/ietcom/e89-b.5.1534

Web of Science

researchmap

Language：English   Publishing type：Research paper (scientific journal)  

DOI： 10.1093/ietcom/e89-b.2.373

researchmap

Language：English   Publishing type：Research paper (international conference proceedings)  

Web of Science

researchmap

Language：English   Publishing type：Research paper (international conference proceedings)  

DOI： 10.1109/GLOCOM.2005.1577676

researchmap

Language：English   Publishing type：Research paper (international conference proceedings)  

researchmap

Language：English   Publishing type：Research paper (scientific journal)  

researchmap

Language：English   Publishing type：Research paper (international conference proceedings)  

researchmap

Language：English   Publishing type：Research paper (scientific journal)  

DOI： 10.1093/ietcom/e88-b.3.1128

researchmap

Language：Japanese   Publishing type：Research paper (scientific journal)  

researchmap

Publishing type：Research paper (international conference proceedings)   Publisher：IEEE Computer Society  

DOI： 10.1109/ICDE.2005.173

researchmap

Other Link： https://dblp.uni-trier.de/db/conf/icde/icdew2005.html#FujisakiMKYSI05

Language：English   Publishing type：Research paper (scientific journal)  

researchmap

Language：English   Publishing type：Research paper (international conference proceedings)  

researchmap

Language：English   Publishing type：Research paper (international conference proceedings)  

researchmap

Language：English   Publishing type：Research paper (international conference proceedings)  

researchmap

Language：English   Publishing type：Research paper (scientific journal)  

DOI： 10.1002/ecja.1134

Scopus

researchmap

Language：Japanese   Publishing type：Doctoral thesis  

researchmap

Language：English   Publishing type：Research paper (scientific journal)   Publisher：John Wiley and Sons Inc.  

DOI： 10.1002/(SICI)1520-6424(200007)83:7<15::AID-ECJA2>3.0.CO;2-3

Scopus

researchmap

Language：English   Publishing type：Research paper (international conference proceedings)   Publisher：IEEE  

Scopus

researchmap

J-GLOBAL

researchmap

J-GLOBAL

researchmap

J-GLOBAL

researchmap

J-GLOBAL

researchmap

J-GLOBAL

researchmap

J-GLOBAL

researchmap

J-GLOBAL

researchmap

Language：Japanese   Publisher：工学院大学総合研究所  

CiNii Books

CiNii Research

researchmap

J-GLOBAL

researchmap

Language：Japanese   Publisher：工学院大学総合研究所  

CiNii Books

CiNii Research

researchmap

Language：Japanese   Publisher：工学院大学総合研究所  

CiNii Books

CiNii Research

researchmap

J-GLOBAL

researchmap

Language：Japanese   Publisher：東京 : 電子情報通信学会  

CiNii Research

researchmap

Other Link： https://ndlsearch.ndl.go.jp/books/R000000004-I028877363

Language：English  

CiNii Books

CiNii Research

researchmap

Language：Japanese  

researchmap

Language：Japanese  

CiNii Books

researchmap

Language：Japanese  

CiNii Books

researchmap

Language：Japanese  

CiNii Books

researchmap

Language：Japanese  

CiNii Books

researchmap

Language：Japanese  

CiNii Books

researchmap

Language：Japanese   Publisher：工学院大学総合研究所  

CiNii Books

CiNii Research

researchmap

Language：Japanese  

researchmap

Language：Japanese  

researchmap

Language：Japanese  

CiNii Books

researchmap

Language：Japanese  

CiNii Books

researchmap

Language：Japanese  

CiNii Books

researchmap

Language：Japanese   Publisher：The Institute of Electronics, Information and Communication Engineers  

CiNii Books

researchmap

Language：Japanese   Publisher：工学院大学総合研究所  

CiNii Books

CiNii Research

researchmap

Language：Japanese  

CiNii Books

researchmap

Language：Japanese   Publisher：The Institute of Electronics, Information and Communication Engineers  

CiNii Books

researchmap

Language：Japanese   Publisher：The Institute of Electronics, Information and Communication Engineers  

CiNii Books

researchmap

Language：Japanese  

researchmap

Language：Japanese   Publisher：The Institute of Electronics, Information and Communication Engineers  

Recently, mobile traffic has been increasing rapidly. in this situation, we need to shift frequency band used commonly now to alternative one with less congestion. Today, wireless access gate system in millimeter wave (MMW) band, which has less congestion, is proposed. However, bandwidth is limited and many users share backbone network in real networks. In this paper, we propose a method of distribution for file content to operate the system efficiently in the situation where backbone network has limited bandwidths. Finally, proposed method reduces required time to download content and enabled users to relax mobility restriction in the following situation.

CiNii Books

researchmap

Language：Japanese   Publisher：The Institute of Electronics, Information and Communication Engineers  

A P2P network has a problem: it is vulnerable to attacks on hub nodes. To solve this problem, a bimodal degree network has been studied that is robust to both hub node attacks and node departures and we proposed the distributed method to construct such a network. However, we haven't research the nature of the network that our method constructs. Therefore, we research the nature of the network that our method constructs by computer simulations in this paper.

CiNii Books

researchmap

Language：Japanese   Publisher：The Institute of Electronics, Information and Communication Engineers  

The traffic amount of the streaming communication is increasing every year by the spread of VoIP and live streaming application. Streaming transmission has the severe restrictions in allowable delay, compared with the data transmission, and various network designs are proposed up to the present. For example, flow scheduling which controls band distribution of traffic flows is proposed. However, a conventional scheduling method does not consider the allowable delay of each flow. In some cases, the packet of the flows which has enough room for allowable delay is processed with priority, and there may be a lot of packets which exceed of allowable delay. In this paper, we propose the optimal weight distribution method to maximize the number of processing packets in allowable delay, based on queuing theory and Lagrange multiplier method in the network where each flow has different of allowable delay We examined the effectiveness of the proposed method under various traffic conditions by numerical analysis. As a result, we showed that the packets are effectively processed in allowable delay even when the network is heavily congested.

CiNii Books

researchmap

Language：Japanese   Publisher：一般社団法人電子情報通信学会  

CiNii Books

researchmap

Language：Japanese   Publisher：The Institute of Electronics, Information and Communication Engineers  

CiNii Books

researchmap

Language：Japanese   Publisher：The Institute of Electronics, Information and Communication Engineers  

CiNii Books

researchmap

Language：Japanese   Publisher：The Institute of Electronics, Information and Communication Engineers  

CiNii Books

researchmap

Language：Japanese   Publisher：工学院大学総合研究所  

CiNii Books

CiNii Research

researchmap

Language：Japanese   Publisher：The Institute of Electronics, Information and Communication Engineers  

In recent years, a Pure P2P network has attracted much attention because it does not require expensive central servers and can construct a large-scale network. However, a P2P network has a problem that it is vulnerable to attacks on hub nodes. For this problem, a bimodal degree distribution network which is tolerant to both attacks on hub node and node departures has been studied. However, the construction method of a bimodal degree distribution network in the previous studies is not targeted to a large-scale Pure P2P. Therefore, we study the construction method of a bimodal degree distribution network for a large-scale Pure P2P.

CiNii Books

researchmap

Language：Japanese   Publisher：The Institute of Electronics, Information and Communication Engineers  

We propose a threshold configuration and holding time limitation for general telephone calls under trunk reservation control in emergency situations. Our proposed method aims to accept more general calls while required emergency calls are still accepted. With our method, the threshold is configured in consideration of the ratio of traffic intensities which are estimated in advance. Computer simulation with a queueing system model of a telephone exchange showed that the general call-blocking rate was much lower than that without threshold relaxation or trunk reservation, while the emergency call-blocking rate was reduced to almost zero unless the sum of estimated traffic intensity was nearly equal to the number of lines. Moreover, we confirmed that the threshold was appropriately configured by using the proposed method, from the comparison between the result of the proposed method and that obtained by using the fixed value of a threshold.

CiNii Books

researchmap

Language：Japanese   Publisher：The Institute of Electronics, Information and Communication Engineers  

Recently, LR-PON (Long Reach-Passive Optical Networks), in which the distance between OLT and ONUs is long, may become widely used. In LR-PON, Dynamic Bandwidth Allocation (DBA) method is required because Round Trip Time (RTT) is large and the idle interval occurs. However, almost conventional studies assume that RTT is small. Moreover, the derived mean packet delay in LR-PON is approximation formula. In this paper, we propose the derivation of the mean packet delay for LR-PON. First, we derive the mean packet delay for IPACT without RTT, more simply compared to the conventional derivation. Then, we derive the mean packet delay for LR-PON by using M/G/1 model. We compare the numerical result with simulation result and confirm the effectiveness of our derivation.

CiNii Books

researchmap

Language：Japanese   Publisher：The Institute of Electronics, Information and Communication Engineers  

We propose a threshold relaxation and holding time limitation for general telephone calls under trunk reservation control in emergency situations. Our proposed method consists of a threshold relaxation strategy and two holding time limitation strategies and it aims to accept more general calls while required emergency calls are still accepted. Computer simulation with a queueing system model of a telephone exchange shows that the proposed method increases the number of accepted general calls while still enabling the required number of emergency calls to be accepted. We also investigate the necessary number of reserved lines for guaranteeing emergency calls.

CiNii Books

researchmap

Language：Japanese   Publisher：The Institute of Electronics, Information and Communication Engineers  

Offloading 3G/LTE mobile traffic onto other networks such as Wi-Fi is a major challenge that mobile carriers face. Mobile traffic offloading onto 2.4 GHz band is widely practiced, but currently the lack of spectrum resources has become a serious problem. To solve this problem, 60 GHz millimeter wave band has a great attention since it requires no license and can provide wireless data transfer at more than 1 Gbps. However, it can cover only a few meters. The authors proposed a millimeter wave access network which delivers the prefetched content to a client using a millimeter wave. This network uses Content-Centric Networking (CCN) to aggregate the millimeter wave spot and the mobile network seamlessly. The realization of this network requires the prediction of a client's movement and placement of content on a gate before the client arrives at the gate. This paper formulates a condition for the success of content preparation and evaluates it using the measured value of a prototype of the millimeter wave device. The evaluation result shows that how far this network needs form the gate to the client to prepare content. Moreover, this paper evaluates the relationship between the data size of millimeter wave communication and the performance of offload.

CiNii Books

researchmap

Language：Japanese   Publisher：The Institute of Electronics, Information and Communication Engineers  

The packet transmission of upstream in EPON can be divided into data interval and reservation interval. In reservation interval, each ONU requests transmission time to OLT by using REPORT message. The OLT receiving REPORT message calculates transmission time for each ONU. Then, each ONU can reserve the resources by receiving GATE message from OLT. The traditional method is a kind of this reservation method. However, if the RTT is large, GATE message may not arrive within a cycle time. In this situation, resource efficiency decreases because the occurrence of idle time. In order to solve this problem, IPACT has been proposed. In IPACT, a reservation interval is allocated after own ONU' data interval. However, waiting time of packets may be large depending on the arriving timing of the packets. In this paper, we propose DR-MPCP method which delay the transmission timing of the REPORT message. We model our DR-MPCP by using M/G/1 queue model and show the effectiveness of our method.

CiNii Books

researchmap

Language：Japanese   Publisher：The Institute of Electronics, Information and Communication Engineers  

In order to improving QoS of livestreaming application High-functional nodes (HN) that performs ARQ between the relay nodes is proposed. On the other hand, the further improved QoS is expected by duplicating packets and transmitting. However, duplication of packets may cause increase of traffic, few as much as possible the number of packet duplications are desirable. Thus, we propose a determining the number of duplication method being satisfied with the smaller number of duplications of the demand QoS, and the validity was shown. In our method, we notice expectation of delay for transmission in one link. And, we duplicate in most efficient link. In numerical analysis, our method show good performance in even many links network. Furthermore, we analyze our method in various condition networks. In analysis, our method show little increase of traffic, in the case where delay of each link is long, or the case where there is little allowable delay.

CiNii Books

researchmap

Language：Japanese   Publisher：The Institute of Electronics, Information and Communication Engineers  

CiNii Books

researchmap

Language：Japanese   Publisher：The Institute of Electronics, Information and Communication Engineers  

CiNii Books

researchmap

Language：Japanese   Publisher：The Institute of Electronics, Information and Communication Engineers  

On current networks, intermediate nodes merely forward packets. Then, communications are established between source and destination nodes through shortest path from the point of delay or packet loss rate. However, various protocols are applied to communications between source and destination nodes in usual. Therefore, QoS in communications may be improved when protocol processings are applied between not only end nodes but also some intermediate nodes. In this presentation, "High-functional Node" is denned as node which can process various protocols in contrast to legacy node which forwards packets only. Examples of high-functional node are SDN node or node of Overlay Network. And, high-functional network is also denned.as network which is constructed by high-functional nodes. Then, our challenge is presented to a new routing approach on high-functional network.

CiNii Books

researchmap

Language：Japanese   Publisher：The Institute of Electronics, Information and Communication Engineers  

On current networks, intermediate nodes merely forward packets. Then, communications are established between source and destination nodes through shortest path from the point of delay or packet loss rate. However, various protocols are applied to communications between source and destination nodes in usual. Therefore, QoS in communications may be improved when protocol processings are applied between not only end nodes but also some intermediate nodes. In this presentation, "High-functional Node" is defined as node which can process various protocols in contrast to legacy node which forwards packets only. Examples of high-functional node are SDN node or node of Overlay Network. And, high-functional network is also defined.as network which is constructed by high-functional nodes. Then, our challenge is presented to a new routing approach on high-functional network.

CiNii Books

researchmap

Language：Japanese   Publisher：The Institute of Electronics, Information and Communication Engineers  

Recently, high-functional mobile devices (ex. smartphone, tablet PC) are immedeately spreading. They can access the Intenet by wireless access network (3G, LTE, WiMAX) whose frequency band is less than micro- meter band. However, according to the widespread of smartphone and alarge-capacity contents, wireless demand exceeds the spectrum capacity. To solve this, wireless access gate system which is designed to shift traffics occured in micro-meter wave band wave to in millimeter wave bandis proposed. In this system ,when user wants to get contents in internet, contents server has to distribute the contents to some gates before user passes by mobility predction. So we consider effects of an error in prediction on this system. In this paper, we reveal effects of mobility prediction on offload rate and a backbone network traffic volume and effectiveness of a prediction method which uses the speed of mobile terminal at previous gate.

CiNii Books

researchmap

Language：Japanese   Publisher：The Institute of Electronics, Information and Communication Engineers  

Recently, high-performance mobile devices are rapidly spreading, and more and more people use a large-volume content such as a streaming video by using them. Meanwhile, when we download a streaming video, we can't use high speed network in many cases because of bandwidth limitations in wireless access network (3G, LTE). To solve this problem, wireless access gate system which shifts traffic from micrometer wave band to millimeter wave band is proposed. We propose a content distribution system to devices on wireless access gate system, which places data pieces of a content to access gates from a content server when a backbone network has enough room. In this paper, we suggest a continuous content distribution system and evaluate the system.

CiNii Books

researchmap

Language：Japanese   Publisher：The Institute of Electronics, Information and Communication Engineers  

To improve QoS of live streaming applications, high-functional nodes which have packet-recovering functions in addition to conventional node functions have been studied. Additionally, for constructing the appropriate path under high-functional nodes network which enable to communicate under high transmission success rate in live streaming applications, the path construction method "Delay Propagation Method" was proposed. However, Delay Propagation Method has a problem that cannot construct the appropriate path under networks including a specific circuit. In this paper, we firstly proposed a new path construction method which find the appropriate path even though Delay Propagation Method cannot construct. This proposed method find the appropriate path by repetition of two processes. One of them is removing links from the Topology which is a part of circuits, and the other one is applying Delay Propagation Method. Finally, we evaluate the effectiveness of the proposed method. As a result, we show even if the number of repetition in the proposed method is small, it sufficiently improve transmission success rate in live streaming applications.

CiNii Books

researchmap

Language：Japanese   Publisher：The Institute of Electronics, Information and Communication Engineers  

Conventional study of per-node priority assignment for live streaming flow, for simplicity, assumed that the number of available priority classes at each priority control node is unlimited so as to accommodate each flow to each different class. However, the assumption is not suitable for real network in terms of scalability and installing cost of the priority control in the network. Therefore, in this study, we propose per-node priority assignment method for live streaming under limited number of priority classes environment. We evaluated effectiveness of the proposed method by numerical analysis and the result showed that the method can find near optimum solution under small network environment. In addition, we could guess that the method could also find near optimum solution in the large network environment.

CiNii Books

researchmap

Language：Japanese   Publisher：The Institute of Electronics, Information and Communication Engineers  

CiNii Books

researchmap

Language：Japanese   Publisher：The Institute of Electronics, Information and Communication Engineers  

CiNii Books

researchmap

Language：Japanese   Publisher：The Institute of Electronics, Information and Communication Engineers  

CiNii Books

researchmap

Language：Japanese   Publisher：The Institute of Electronics, Information and Communication Engineers  

CiNii Books

researchmap

Language：Japanese   Publisher：The Institute of Electronics, Information and Communication Engineers  

Under emergency conditions, a threshold control method, which reserves several telephone lines to accommodate emergency telephone calls, is used against the congestion in telephone exchange system. In threshold control, it is desirable to set the appropriate threshold which can accommodate the necessary number of emergency calls and as many general calls as possible. However, in order to accommodate more general calls, we consider that we should appropriately limit the holding time of general calls. In this paper, as a basic study of limiting the holding time of general calls under threshold control, we model the telephone exchange system as a M_1M_2/MD/s/s queueing system and calculate blocking rate of both emergency calls and general calls in different traffic conditions by computer simulation. From the comparison between the results and the blocking rate without limiting general call holding time, we show the efficiency of limiting general call holding time in different traffic conditions. Moreover, we show the relation between the efficiency of limiting general call holding time and the ratio of the arriving call type. Furthermore, we also discuss the relation between the blocking rate of general calls and traffic intensity or number of telephone lines.

CiNii Books

researchmap

Language：Japanese   Publisher：一般社団法人電子情報通信学会  

CiNii Books

researchmap

Language：Japanese   Publisher：The Institute of Electronics, Information and Communication Engineers  

CiNii Books

researchmap

Language：Japanese   Publisher：The Institute of Electronics, Information and Communication Engineers  

Recently, high-functional mobile devices (ex. smartphone, tablet PC) are immedeately spreading. They can access the Intenet by wireless access network (3G, LTE, WiMAX) whose frequency band is less than micro- meter band. However, according to the widespread of smartphone and alarge-capacity contents, wireless demand exceeds the spectrum capacity. To solve this, wireless access gate system which is designed to shift traffics occured in micro-meter wave band wave to in millimeter wave bandis proposed. In this system ,when user wants to get contents in internet, contents server has to distribute the contents to some gates before user pass by mobility predction. So we consider effects of an error in prediction on this system. In this paper, we reveal effects of mobility prediction accuracy on offload rate and a backbone network traffic volume. From the results of computer simulation, we could get a guiding principle on designing a mobility prediction method.

CiNii Books

researchmap

Language：Japanese   Publisher：The Institute of Electronics, Information and Communication Engineers  

Per-node priority assignment to flows is effective to increase number of flows whose packets can arrive at destination node within allowable delay of the flow. However, conventional study of per-node priority assignment dose not discuss a case which heterogeneous flows are accommodated in network. We propose per-node priority as- signment method to heterogeneous flows based on characteristic of optimum solution. The effectiveness of proposed method is confirmed by numerical analysis. Numerical analysis shows that proposed method whose bandwidth threshold is low can find near optimum solution in small network, and can find good solution in large network too. Additionally, conventional method which is equal to propose method whose bandwidth threshold is zero can find good solution too when heterogeneous flows are accommodated in network.

CiNii Books

researchmap

Language：Japanese   Publisher：The Institute of Electronics, Information and Communication Engineers  

On general content delivery networks, replica allocation is effective in reducing propagation delay when users ob- tain contents. And contents formatted into file type are supposed to be delivered under a reliable transmission control protocol, e.g. Transmission Control Protocol. In this case, round trip time dominates delivery transmission rate between a server that deliver a requested content and a user that requests the content, and the delay for delivering contents occurs. We call it delivery transmission delay. Assuming that contents are delivered under a reliable transmission control protocol, we proposed a content replication algorithm that considers the popularity per file size of content to reduce delivery transmission delay. On the other hand, the effect of network scale on performance of the proposed method is not evaluated. In this paper, we evaluate the effect of the number of servers, the number of contents, and storage capacity of servers on performance of the proposed method. The proposed method provides better performance constantly than other heuristics under any network scale condition.

CiNii Books

researchmap

Language：Japanese   Publisher：The Institute of Electronics, Information and Communication Engineers  

CiNii Books

researchmap

Language：Japanese   Publisher：The Institute of Electronics, Information and Communication Engineers  

Per-node priority assignment to each streaming flow can increase number of flows which are able to arrive at destination node within allowable delay of the flow (delay-satisfied flow) in network whose communication resource is limited. In this work,we first introduced end-to-end delay distribution of packets in streaming flow accommodated in network which is constructed by priority control nodes. Then we defined delay allowance to each flow which means difference between allowable delay of the flow and α percentile of end-to-end delay probabilitydistribution of packets in the flow. Second,we proposed per-node priority assignment method for flows to increasethe number of flows whose delay allowance are not negative (satisfied flows in terms of delay). Finally,effectiveness of proposal method is shown by numerical analysis. Analysis shows that proposal method can find near optimal solution.

CiNii Books

researchmap

Language：Japanese   Publisher：The Institute of Electronics, Information and Communication Engineers  

CiNii Books

researchmap

Language：Japanese   Publisher：The Institute of Electronics, Information and Communication Engineers  

CiNii Books

researchmap

Language：Japanese   Publisher：The Institute of Electronics, Information and Communication Engineers  

CiNii Books

researchmap

Language：Japanese   Publisher：The Institute of Electronics, Information and Communication Engineers  

On general content delivery networks, replica allocation is effective in reducing propagation delay when users obtain contents. We proposed a content replication algorithm that considers the popularity per file size of content. On the other hand, delivering large-sized contents causes large transmission delay. Thus, replication algorithms should consider transmission delay to reduce download time for obtaining contents by users. In this paper, we assume that contents are delivered under reliable transmission control protocols. In this case, round trip time dominates content transmission rate between a server that deliver a requested content and a user that requests the content. Under this assumption, the effect of proposed method on transmission delay is evaluated, and proposed method provides better performance than a method that preferentially replicates popular and large-sized contents in large-scale simulation conditions.

CiNii Books

researchmap

Language：Japanese   Publisher：The Institute of Electronics, Information and Communication Engineers  

Under emergency condition, call admission control is required to accommodate more general telephone calls while accommodating the minimum necessary number of emergency telephone calls. Moreover, in order to increase the number of accommodated general telephone calls, it is important to encourage general telephone users to reduce holding time. In this paper, we propose the optimal trunk reservation method which maximizes the number of accommodated general telephone calls, while accommodating the minimum necessary number of emergency telephone calls. Then, we compare the blocking rate by using the proposed method with one by using the conventional method, and analyze the characteristic of blocking rate when the holding time reduce. By this comparison, we show that the number of general telephone calls can be increased, or guarantee the minimum necessary number of accommodated emergency telephone calls can be achieved by using our method. Moreover, by this characteristic analysis, we can reveal the effective traffic conditions that the blocking rate of general telephone calls decrease by reducing the holding time.

CiNii Books

researchmap

Language：Japanese   Publisher：The Institute of Electronics, Information and Communication Engineers  

On general content delivery networks, replica allocation is effective in reducing propagation delay when users get contents. On the other hand, the storage capacity limitations do not allow servers to store all contents. Then, a conventional method that considers popularities of contents has been proposed. However, the method performance is insufficient because of not considering file sizes. In this paper, we propose a novel content replication algorithm that considers the popularities per file size to reduce propagation delay. And, we show this method provides better performance than the conventional method in any simulation condition.

CiNii Books

researchmap

Language：Japanese   Publisher：The Institute of Electronics, Information and Communication Engineers  

CiNii Books

researchmap

Language：Japanese   Publisher：The Institute of Electronics, Information and Communication Engineers  

CiNii Books

researchmap

Language：Japanese   Publisher：The Institute of Electronics, Information and Communication Engineers  

Flow convergence on particular nodes causes various problems such as degradation of QoS. Then, we defined the degree of flow convergence and propose a routing algorithm to reduce convergence on a node determined to have the maximum degree of convergence (referred to as "the maximum convergence degree"). And we showed that the maximum convergence degree and network delay can be reduced by using our proposed method. However, too many resources are required for our previous method because it calculates a new route every time a flow is accommodated. Then, we propose a new reducing flow convergence routing method of restricting the number of selectable route in order to reduce the resources required for selecting a route. Finally, we show that the maximum convergence degree and network delay can be reduced by using our proposed method.

CiNii Books

researchmap

Language：Japanese   Publisher：The Institute of Electronics, Information and Communication Engineers  

High-functional nodes that have packet recovery processes such as retransmission are contribute to decrease total transmission number. However routing methods for networks including high-functional nodes have not been considered. Then, we propose routing method for decreasing total transmission number in high-functional node networks. In our method, we construct virtual high-functional node networks whose link cost is expectation of total transmission number, and configure paths with Dijkstra's algorithm. As a result of the evaluation by the computation simulation, we showed that our method decrease total transmission number more than an existing routing method.

CiNii Books

researchmap

Language：Japanese   Publisher：The Institute of Electronics, Information and Communication Engineers  

On general content delivery networks, replica allocation is effective in reducing propagation delay when users get contents. On the other hand, the storage capacity limitations do not allow servers to store all contents. Then, a conventional method that considers popularities of contents has been proposed. However, the method performance is insufficient because of not considering file sizes. In this paper, we propose a novel content replication algorithm that considers the popularities per file size to reduce propagation delay. And, we show this method provides better perfonnance than the conventional method in any simulation condition.

CiNii Books

researchmap

Language：Japanese   Publisher：The Institute of Electronics, Information and Communication Engineers  

Information retrieval systems usually provide only one metadata selected from multiple different metadata corresponding to particular content (one content to one metadata model). However, an expectated value of browsing cost for successful retrieval increases when the selected metadata is not consistent with what the user thinks is interesting. On the other hand, as integrated metasearch services do, if multiple different metadata corresponding to single content are provided (one content to many metadata model), it is highly possible that we can reduce the expected cost of successful retrieval. Then, in this study, we defined three models of user's retrieval behavior to reveal the expected value of cost for successful retrieval under the one content to many metadata model. In this paper, we studied the derivation of evaluation formula to decide the optimal metadata ranking that archieves minimum expected cost for successful retrieval.

CiNii Books

researchmap

Language：Japanese   Publisher：The Institute of Electronics, Information and Communication Engineers  

Flow convergence on particular nodes causes various problems such as degradation of QoS. We previously proposed a new routing method to reduce flow convergence in order to reduce the maximum number of relay flows on the node that relays the most flows (referred to as "the maximum number of relay flows"). However, we only showed the characteristics of network delay on a network in which all nodes had an equal processing rate μ and it is preferable to show the characteristics on a network in which the nodes have different processing rates. Consequently, we define the degree of flow convergence and propose a routing algorithm to reduce convergence on a node determined to have the maximum degree of convergence (referred to as "the maximum convergence degree"). Finally, we show that the maximum convergence degree and network delay can be reduced by using our proposed method on a network in which the nodes have different processing rates.

CiNii Books

researchmap

Language：Japanese   Publisher：The Institute of Electronics, Information and Communication Engineers  

Networks including high-functional nodes (HNs) that have packet recovery processes such as retransmission are currently being studied in order to improve live streaming QoS. To utilize packet recovery processes of HN effectively, Novel Link Weight Method has been proposed as a path configuration method for HN networks. However, this method may sometimes configure inappropriate paths because this link weight does not take into account interaction of delay among all links on a path. Thus, we propose a new path configuration method that considers delay (including packet recovery delay) on each link in a path. In our method, we define the probability distribution of delay and use convolution of the probability distribution. We also define the transmission success rate, that is, the probability of completing receiving packets within allowable delay. Moreover, we propose an algorithm that configures paths having maximum transmission success rate. In a simulation, we showed that our method is more suitable for path configuration for HN networks than conventional methods.

CiNii Books

researchmap

Language：Japanese   Publisher：The Institute of Electronics, Information and Communication Engineers  

CiNii Books

researchmap

Language：Japanese   Publisher：The Institute of Electronics, Information and Communication Engineers  

Reduction of average service latency for file delivery is one of the crucial problems of file distribution systems in widespread use, such as CDN (Contents Delivery Network), Web Proxy System, or P2P file sharing system (Peer-to-Peer file delivery system). As one of the solutions, in our previous works, we already developed Hop-by-Hop local optimal file delivery method, which performed nearly at the overall theoretical efficiency limit of the system under the limited case, where the system had adequate cache capacity and cache overflow never occurred. However, practical system has finite cache capacity, and cache algorithm which pushes unnecessary files out of cache when cache overflow occurs, should be introduced into the system. In this case, average service latency depends on interaction between file delivery method and cache algorithm, because the files which are pushed out of cache by cache algorithm should be resended when users require the files again, and the service latency of the resends are produced by file delivery method. Therefore, in this paper, we introduce a cache algorithm suitable for Hop-by-Hop local optimal file delivery method under the case where overflow occurs, and evaluate the performance of it. The results show that our algorithm performs effectively reduction of average service latency compared to FIFO (First-In First-Out) and LFU (Least Frequently Used).

CiNii Books

researchmap

Language：Japanese   Publisher：The Institute of Electronics, Information and Communication Engineers  

Media broadcasting services using Internet Protocol (IP) multicast are emerging. On the other hand, IEEE 802.11 wireless LAN (WLAN) has been widely deployed. Because of such a background, the opportunities to receive multicast streaming over WLAN are expected to be more often. However, due to the restrictions on the standard, the transmission rate and the reliability of multicast over WLAN are both lower than those of unicast. We proposed a transparent multicast / unicast translation method which translates a multicast packet to multiple unicast packets so that we can utilize the higher transmission rates and retransmission function that are only available for unicast. However, we showed the proposed method efficiency only by theoretical analysis. Therefore, we evaluated the method by experiments and showed that the proposed method improves the efficiency and reliability of multicast over WLAN if the number of receivers is limited.

CiNii Books

researchmap

Language：Japanese   Publisher：The Institute of Electronics, Information and Communication Engineers  

Flow convergence on particular node causes various problems such as degradation of QoS. Then, we proposed a new routing method for reducing the maximum number of relay flows on the node which relays maximum flows (called "the maximum number of relay flows"). However, we showed the characteristic of maximum number of relay flows by simulation only on random network topology. Then, we examine the effectiveness of our proposed method and the characteristic of the network delay by Queueing Theory on scale free network topology. Finally, we show the maximum number of relay flows and the network delay can be reduced by our proposed method and our proposed method is more effective on scale free network topology.

CiNii Books

researchmap

Language：Japanese   Publisher：The Institute of Electronics, Information and Communication Engineers  

When high-functional nodes with retransmission processing are placed throughout a network, other routing methods considering the retransmission are needed instead of present methods. Multicast tree reconfiguration method is proposed as one of those method, but it is only for reconfiguration of a multicast tree which has a bottleneck link in bad condition. On the other hand, novel link weight method, which is a general routing method using Dijkstra's algorithm and a new link weight that integrates both the link delay and the packet loss, is proposed, and it is shown that this method is suitable to construct paths consisting of high-functional nodes. In this paper we focus on applying novel link weight method to constructing a multicast tree consisting of high-functional nodes, and this method can be also applicable to reconfiguration of a multicast tree to avoid a bottleneck link. Therefore we compare the performance of above two methods through the simulation. The simulation result shows that novel link weight method is suprior to tree reconfiguration method at the point of lower loss rate in application layer, that is, novel link weight method is more suitable to reconstruct a multicast tree.

CiNii Books

researchmap

Language：Japanese   Publisher：The Institute of Electronics, Information and Communication Engineers  

CiNii Books

researchmap

Language：Japanese   Publisher：The Institute of Electronics, Information and Communication Engineers  

CiNii Books

researchmap

Language：Japanese   Publisher：The Institute of Electronics, Information and Communication Engineers  

CiNii Books

researchmap

Language：Japanese   Publisher：The Institute of Electronics, Information and Communication Engineers  

CiNii Books

researchmap

Language：Japanese   Publisher：The Institute of Electronics, Information and Communication Engineers  

CiNii Books

researchmap

Language：Japanese   Publisher：The Institute of Electronics, Information and Communication Engineers  

CiNii Books

researchmap

Language：Japanese   Publisher：The Institute of Electronics, Information and Communication Engineers  

It is important to use an ALM (application layer multicast) tree with broad bandwidth for quality of streaming applications. However, when constructing an ALM tree without information of underlay topology, we cannot know a link congestion. Therefore if unexpected link congestion occurs, the performance of the ALM transmission may become worse. On the other hands, no polynomial-time algorithms for maximum-bandwidth ALM tree construction with underlay topology information exist. In our previous work, we assume that the underlay network has a tree network topology, and we have shown a polynomial-time algorithm that constructs a maximum-bandwidth ALM tree, and we have proved that we can obtain an optimal ALM tree. However, we have not evaluated our algorithm by simulation yet. In this paper, we evaluate our algorithm and a conventional algorithm. As a result, our algorithm can achieve a broader bandwidth ALM tree with low computational cost especially on a large size network.

CiNii Books

researchmap

Language：Japanese   Publisher：The Institute of Electronics, Information and Communication Engineers  

Stream flows Dynamic Class Assignment (DCA) has been proposed for reducing negative interactions between multimedia streaming application traffic and the other conventional data application traffic. To obtain traffic information on the bottleneck node, DCA requires all Core Node to add new functions for advertising information. However, additional functions of all Core Node increase loads of Core Nodes, and increase costs when DCA is applied to large-scale real networks. Therefore, in this paper, we introduce a new threshold that we termed Virtual Bandwidth, and propose an expansion of DCA method by using the Virtual Bandwidth. This expansion method can be operated by the cooperation of Edge Node only, and does not require additional functions to Core Node. We investigated an effective updating method of the Virtual Bandwidth by computer simulation. Then, we evaluated the effectiveness of our method by comparison with conventional DCA method.

CiNii Books

researchmap

Language：Japanese   Publisher：The Institute of Electronics, Information and Communication Engineers  

Flow convergence on particular node causes various problems such as degradation of QoS. In this paper, we proposed a new routing method for reducing the maximum number of relay flows on the node which relays maximum flows (called "the maximum number of relay flows"). Full search is required for calculating optimal algorithm of this problem. Then, we propose an approximate algorithm to calculate an near optimal solution by feasible computational costs. Moreover, when call loss is permitted on this method, the maximum number of relay flows can be reduced again. However, generally, there is trade-off relation between call loss rate and the maximum number of relay flows. Then, we show numerical relation between the call loss rate and the number of maximum relay flows. Finally, we show the effectiveness of our proposed method by simulation on variety network toporogy and traffic situations.

CiNii Books

researchmap

Language：Japanese   Publisher：The Institute of Electronics, Information and Communication Engineers  

Recently, multimedia applications such as video and audio are popular. Then, various kinds of heterogeneous stream flows occupy the most part of bandwidth of the network. Therefore, Call Admission Control (CAC) is required to maintain QoS (quality of services) that judges whether a flow should be accommodated or not. Well, an optimum CAC is proposed by using semi-Marcov decision process. This method optimizes judgement whether a newly arrived flow is accommodated or not by using a number of accommodating flows of each class. However, this CAC is complex and is not feasible. Moreover, this CAC should be derived whenever traffic conditions are changed. Then, this CAC lacks general versatility. Therefore, in this paper, we proposed an optimal admission control by using a machine learning method "SVM" which is supervised by optimum CACs. Performance evaluations show that our proposed CAC can realize almost the same total call-blocking rate as optimum CACs.

CiNii Books

researchmap

Language：Japanese   Publisher：The Institute of Electronics, Information and Communication Engineers  

Recently, multimedia applications such as video and audio come into wide use. When heterogeneous traffic is accommodated in a network, Call Admission Control (CAC) will be required for maintaining quality of services. Because the values of various applications are not always proportional to the bandwidth, therefore, conventional CACs, which aim to increase resource utilization under the condition that users' satisfaction is proportional to their own bandwidth, have difficulties in maximizing user satisfaction. As stated above, we proposed a novel CAC strategy for maximizing total accommodated flows based on a new philosophy that heterogeneous flows should be treated equally in networks. And we presented a theoretical analysis of the model and numerical analysis that show the effectiveness of our CAC. To obtain an optimal threshold that makes total call-blocking rate minimise, we need a full search from all candidates of threshold. In this paper, we developed an approximation formula to obtain suboptimal threshold based on theoretical characteristic analysis, and presented that the approximation formula achieves a good performance.

CiNii Books

researchmap

Language：Japanese   Publisher：The Institute of Electronics, Information and Communication Engineers  

CiNii Books

researchmap

Language：Japanese   Publisher：The Institute of Electronics, Information and Communication Engineers  

CiNii Books

researchmap

Language：Japanese   Publisher：Information Processing Society of Japan (IPSJ)  

CiNii Books

researchmap

Language：Japanese   Publisher：The Institute of Image Information and Television Engineers  

Recently, multimedia applications such as video and audio come into wide use. When heterogeneous traffic is accommodated in a network, Call Admission Control(CAC) will be required for maintaining quality of services. Because the values of various applications are not always proportional to the bandwidth, therefore, conventional CACs, which aim to increase resource utilization under the condition that users' satisfaction is proportional to their own bandwidth, have difficulties in maximizing user satisfaction. As stated above, we proposed a novel CAC strategy for maximizing total accommodated flows based on a new philosophy that heterogeneous flows should be treated equally in networks. And we presented a theoretical analysis of the model and numerical analyses that show the effectiveness of our CAC. In this paper, we conducted a qualitative characteristic analysis of a change of an optimal threshold to realize minimum total call-blocking rate, when both traffic intensity of narrowband and broadband flow change. As a result in this characteristic analysis, we elucidated that total call blocking-rate correlate with optimal threshold under some given traffic parameter conditions. Then, we intrduce some guiding principles for developing a practical threshold configuring method.

CiNii Books

researchmap

Language：Japanese   Publisher：Information Processing Society of Japan (IPSJ)  

Recently, multimedia applications such as video and audio come into wide use. When heterogeneous traffic is accommodated in a network, Call Admission Control(CAC) will be required for maintaining quality of services. Because the values of various applications are not always proportional to the bandwidth, therefore, conventional CACs, which aim to increase resource utilization under the condition that users' satisfaction is proportional to their own bandwidth, have difficulties in maximizing user satisfaction. As stated above, we proposed a novel CAC strategy for maximizing total accommodated flows based on a new philosophy that heterogeneous flows should be treated equally in networks. And we presented a theoretical analysis of the model and numerical analyses that show the effectiveness of our CAC. In this paper, we conducted a qualitative characteristic analysis of a change of an optimal threshold to realize minimum total call-blocking rate, when both traffic intensity of narrowband and broadband flow change. As a result in this characteristic analysis, we elucidated that total call blocking-rate correlate with optimal threshold under some given traffic parameter conditions. Then, we intrduce some guiding principles for developing a practical threshold configuring method.

CiNii Books

researchmap

Publisher：社団法人電子情報通信学会  

音声や動画像などのマルチメディアアプリケーションの普及に伴い,帯域の異なる通信を同一網に収容する際,通信品質維持のため,受付制御(Call Admission Control)を行う必要がある.しかし,網の利用効率向上が目的の従来方式では,帯域が通信の価値をそのまま表すとはできないため,ユーザに大きな不満を生じさせる.そこで,これまでに全ユーザの網への収容により得られる満足の最大化を目標とした,新しい受付制御方式を提案し,理論解析により,最適制御パラメータを導出し,提案方式の有効性を示している.そこで本稿では,狭帯域広帯域両フローのトラヒック密度を様々に変化させた場合の,トータル呼損率が最小になる最適な閾値の変動について,定性的な特性解析を行った.閾値を変動させながらトータル呼損率特性を解析した結果,与えられたトラヒック条件によっては,最適な閾値とトータル呼損率の間に相関性があることを明らかにし,より実用的な閾値設定法を構築するための方針を示した.

researchmap

Language：Japanese  

DOI： 10.1587/transcom.E93.B.3564

CiNii Books

researchmap

Language：Japanese  

CiNii Books

researchmap

Language：Japanese   Publisher：The Institute of Electronics, Information and Communication Engineers  

CiNii Books

researchmap

Language：Japanese   Publisher：The Institute of Electronics, Information and Communication Engineers  

CiNii Books

researchmap

Language：Japanese   Publisher：The Institute of Electronics, Information and Communication Engineers  

CiNii Books

researchmap

Language：Japanese   Publisher：The Institute of Electronics, Information and Communication Engineers  

CiNii Books

researchmap

Language：Japanese   Publisher：The Institute of Electronics, Information and Communication Engineers  

Current service authorization systems have mostly assumed administrators are divided from ordinal users and the all the responsibility and cost for managing services are concentrated on the administrators. Moreover, as the number of users gets larger, the human relation between administrators and users get more scarce and users tend to behave more selfish and maliciously. In this paper, we propose a new architecture named Social Networking Architecture (SNA) in which users authorize each other when using services. In the SNA, a user who is authorized to use a service is able to be an inviter for the service. An inviter is allowed to invite his or her trusted people based on his or her own responsibility and share the resources and permissions that he or she has on the service with invitees. An invitee is also allowed to be an inviter and able to invite his or her trusted people. Since each invitation is done with the human trust between two people, a network based on the human trust chain is constructed for each service. Because of the properties, under the SNA, the management cost and risk for malicious users due to the increase of number of users does not overwhelm administrators, and therefore, it is possible to construct a scalable service network. Furthermore, since an invitation is performed based on the human relationship in the actual society, invitees are discouraged to behave selfish and cause their inviters trouble. Because of this feature, it is expected that the SNA automatically makes a secure service network.

CiNii Books

researchmap

Language：Japanese   Publisher：The Institute of Electronics, Information and Communication Engineers  

Media broadcasting services using Internet Protocol (IP) multicast are emerging. On the other hand, IEEE 802.11 wireless LAN (WLAN) has been widely deployed. Because of such a background, the opportunities to receive multicast streaming over WLAN are expected to be more often. However, due to the restrictions on the standard, the transmission rate and the reliability of multicast over WLAN are both lower than those of unicast. In this paper, we proposed a transparent multicast / unicast translation method which translates a multicast packet to multiple unicast packets so that we can utilize the higher transmission rates and retransmission function that are only available for unicast. Our theoretical analysis showed that the proposed method improves the efficiency and reliability of multicast over WLAN if the number of receivers is limited to a number which an access point is assumed to reasonably support.

CiNii Books

researchmap

Language：Japanese   Publisher：The Institute of Electronics, Information and Communication Engineers  

Recently, multimedia applications such as video and audio come into wide use. Because of these heterogeneous traffics occupying most of networks, Call Admission Control (CAC) will be required for maintaining quality of services. User satisfaction depends on the success of flow accommodation by the CAC. However, conventional CACs have not considered user satisfaction because the main purpose of conventional CACs is an improvement of resources utilization. Therefore in this paper, we propose a novel CAC for maximizing total accommodated flows with a new philosophy that heterogeneous traffics are fair in networks. Optimal parameters are derived under some traffic configurations by theoretical analysis. Moreover, numerical analyses show the effectiveness of our proposed CAC.

CiNii Books

researchmap

Language：Japanese   Publisher：The Institute of Electronics, Information and Communication Engineers  

In case of IP-based real-time media, the recovery from losses due to congestion in network, needs to be made within a short range of delay constraint, and it happens to be the key factor determining the expected QoE (Quality of Experience) for media. One basic mechanism available for error control is ARQ, that provides loss recovery by retransmission of lost packets. Our goal is to find an optimal mechanism for real-time media using this retransmission approach that works in a receiver-oriented way and is based on the prediction of losses. In the scheme we propose here, the receiver sends retransmission requests for potential lost packets considering the network conditions and delay budget. We focus on the time instant when receiver-end should request a retransmission. Given network conditions, we examine the impact of time instant on the achieved performance at Receiver-end. Theoretical analysis as well as simulations have been performed to quantify the gains possible using proposed approach. We finish this paper with directions for an optimal policy for practice.

CiNii Books

researchmap

Language：Japanese   Publisher：The Institute of Electronics, Information and Communication Engineers  

CiNii Books

researchmap

Language：Japanese   Publisher：The Institute of Electronics, Information and Communication Engineers  

CiNii Books

researchmap

Language：Japanese   Publisher：The Institute of Electronics, Information and Communication Engineers  

CiNii Books

researchmap

Language：Japanese   Publisher：The Institute of Electronics, Information and Communication Engineers  

CiNii Books

researchmap

Language：Japanese   Publisher：The Institute of Electronics, Information and Communication Engineers  

CiNii Books

researchmap

Language：Japanese   Publisher：The Institute of Electronics, Information and Communication Engineers  

When a IP Multicast network is constructed on switch-based network, a lot of broadcasts of the multicast packet are generated. For solving this problem, IGMP Snooping is proposed, and it is implemented on various switches. However, some switches have only insufficient IGMP Snooping implementation actually. This problem is also reported in RFC4541 etc. Therefore, Then, we proposes the FDAR(Flooding Detection And Reduce) mechanism for solving this insufficient IGMP Snooping problem, in this research. FDAR transmits dummy IGMP Queries, and dummy joins to the flooding Multicast group under the environment of insufficient IGMP Snooping implementation switch. The experimental evaluation shows the effectiveness of our proposed method that the load of a network decreases because of the effective restraint control of Multicast packet flooding by the proposal method.

CiNii Books

researchmap

Language：Japanese   Publisher：The Institute of Electronics, Information and Communication Engineers  

We have so far proposed a Dynamic Class Assignment method for stream flows (DCA) that aggregates multimedia streaming application traffic and the other usual application traffic, maintaining quality of the both. DCA can extend its scalability by integrating with our proposed admission control scheme called Tentative Accommodating and Congestion Confirming Strategy (TACCS). However, we found considerable increase of undesired flow rejection could be caused by the characteristic of TACCS if we integrated DCA with TACCS without modifying DCA's algorithm. Therefore, in this paper, we proposed a new DCA algorithm that is more suitable for integrating it with TACCS and verify its effectiveness through computer simulation.

CiNii Books

researchmap

Language：Japanese   Publisher：The Institute of Electronics, Information and Communication Engineers  

CiNii Books

researchmap

Language：Japanese   Publisher：The Institute of Electronics, Information and Communication Engineers  

Recently, many contents distribution systems such as CDN and P2P file sharing system are proposed. In these systems, file type contents require complete download when they are played. And, they have no value before download finishing. Therefore, a total downloading time of content is very important, and reduction of service periods is required in order to satisfy the users. From the viewpoint, we discuss a hop-by-hop file delivery system which is suitable for file type contents delivery because the sum of a service period is reduced by using the request condition. on ..Moreover, we proposed a file delivery scheduling algorithm on one link model and on the condition that the distribution of contents request frequency is unknown, as a basic study. Its algotithm is based on local optimal stratedy. Finally, we performed the characteristic analysis by the computer simulation. The results showed that our proposed algorithm has performed nearly the theoretical efficient limit of the hop-by-hop system when distribution of requests frequency to each contents has deviation.

CiNii Books

researchmap

Language：Japanese   Publisher：The Institute of Electronics, Information and Communication Engineers  

A search method under success rate management is proposed on network environment that nodes have lack of memory. It has already shown that its method is able to satisfy search requirement of users flexibly such as search success rate, search time, and search cost. However, search parameters setting up method have not yet been clear that satisfy users' requirement. Therefore, we propose an auto setting up method of the optimum search parameters which satisfy the search requirement of users such as search success rate and search time. For our theoretical analysis, we assume networks are tree structure, and the search procedure is terminated when search target is found. Under these conditions, we derive the auto setting up algorithm in view that new querying and searching the rest of the nodes is unnecessary after the search has succeeded. Finally, we evaluate the proposed method under various search requirements. The results show the effectiveness that proposed method satisfies the search requirement f users, and the expected value of traffic by proposed method is the smallest compared with the other parameter setting.

CiNii Books

researchmap

Language：Japanese   Publisher：The Institute of Electronics, Information and Communication Engineers  

In the design of a public communication network, it is general to make a constant quality filled when a constant value in which the traffic change is considered beforehand is set as a standard of the traffic and the traffic joins. Moreover, the quality depends on the change for the traffic for a short time. Therefore, we should forecast measuring the traffic in detail can observe the change for a short time to request the maximum traffic that satisfies the quality for the distribution of the traffic in the future. However, we should put large equipment and the cost on the accumulation of data necessary to provide execution and the distribution of a detailed traffic measurement. Then, we report on the result of evaluating the possibility and accuracy, etc. by proposing the technique for presuming the distribution of the change for a short time of every the second from the measurement data of about 15 minutes at the cycle that can be executed, and using open to the public traffic data.

CiNii Books

researchmap

Language：Japanese   Publisher：The Institute of Electronics, Information and Communication Engineers  

When high functional node with the retransmission processing exists, we proposed new link weight which integrate the delay, the loss rate, and the packets retransmission for delay reducing and loss recovery. And we showed that our proposed link weight is suitable for path construction. In this paper, we show that our proposed link weight is more suitable for path construction than other link weight(delay, loss rate, delay expection). And prposed link weight satisfies conditional expression is inadequacy. So by excluding the path which satisfies this expression, we show that path construction using proposed link weight is improved.

CiNii Books

researchmap

Language：Japanese   Publisher：The Institute of Electronics, Information and Communication Engineers  

CiNii Books

researchmap

Language：Japanese   Publisher：The Institute of Electronics, Information and Communication Engineers  

CiNii Books

researchmap

Language：Japanese   Publisher：The Institute of Electronics, Information and Communication Engineers  

CiNii Books

researchmap

Language：Japanese   Publisher：The Institute of Electronics, Information and Communication Engineers  

CiNii Books

researchmap

Publisher：FIT(電子情報通信学会・情報処理学会)推進委員会  

researchmap

Language：Japanese   Publisher：The Institute of Electronics, Information and Communication Engineers  

As multimedia streaming services become common, admission control is becoming an essential technique for Internet protocol (IP) networks. We previously proposed an admission control scheme we call "edge-based tentative accommodating and congestion confirming strategy (edge-based TACCS)". Edge-based TACCS is an extended version of TACCS and it works based on cooperation of high functional edge nodes while original TACCS needs to add some functions to core nodes. However, in edge-based TACCS, some flows, which should be rejected due to the original TACCS, may be wrongly accommodated (Rejection failure). Therefore, in this paper, we show the rejection failure does not become a great factor to degrade the quality of accommodated flows through the theoretical analysis, and we verify that Edge-based TACCS is more scalable and easier to deploy than the original.

CiNii Books

researchmap

Language：Japanese   Publisher：Information Processing Society of Japan (IPSJ)  

Internet-based content distribution systems are now providing new business opportunities. Given this background, this paper discusses the fingerprinting, which act as a psychological deterrent to illegal copying and distribution of copyrighted contents. As a fingerprint code, this paper assumes the use of TA code because it is superior in term of tolerance for collusion. We also present fingerprinting method to prevent leakage of content that has not yet had watermarks embedded. In proposed method, content are scrambled at the server and are descrambled at the receiver. However, some are left scrambled. The scrambled positions in the content constitute the fingerprint. Implementation of an application system is also described. We demonstrate that an Internet-based pay broadcasting system can be implemented efficiently using proposed fingerprinting to protect copyrighted contents.

CiNii Books

researchmap

Language：Japanese   Publisher：Information Processing Society of Japan (IPSJ)  

Internet-based content distribution systems are now providing new business opportunities. Given this background, this paper discusses the fingerprinting, which act as a psychological deterrent to illegal copying and distribution of copyrighted contents. As a fingerprint code, this paper assumes the use of TA code because it is superior in term of tolerance for collusion. We also present fingerprinting method to prevent leakage of content that has not yet had watermarks embedded. In proposed method, content are scrambled at the server and are descrambled at the receiver. However, some are left scrambled. The scrambled positions in the content constitute the fingerprint. Implementation of an application system is also described. We demonstrate that an Internet-based pay broadcasting system can be implemented efficiently using proposed fingerprinting to protect copyrighted contents.

CiNii Books

researchmap

Language：Japanese   Publisher：The Institute of Electronics, Information and Communication Engineers  

CiNii Books

researchmap

Language：Japanese   Publisher：The Institute of Electronics, Information and Communication Engineers  

CiNii Books

researchmap

Language：Japanese   Publisher：The Institute of Electronics, Information and Communication Engineers  

CiNii Books

researchmap

Language：Japanese   Publisher：The Institute of Electronics, Information and Communication Engineers  

CiNii Books

researchmap

Language：Japanese   Publisher：The Institute of Electronics, Information and Communication Engineers  

CiNii Books

researchmap

Language：Japanese   Publisher：The Institute of Electronics, Information and Communication Engineers  

CiNii Books

researchmap

Language：Japanese   Publisher：The Institute of Electronics, Information and Communication Engineers  

In recent years, many functionalities which support applications has been proposed to be added into routers to improve the performance of the applications. These functionalities improve the performances of the applications, but they also increase loads of routers. These heavy loads may cause reducing network performance. To solve this problem, flow processing number should be limited on routers. When the number of the candidate flows to be processed is more than the limitation, the router should choose some suitable flows. In this paper, we propose new flow choosing algorithms on high functional routers. The goal of these algorithms is that more number of flows satisfy their allowable application loss rate. We evaluated our proposed algorithms by computer simulations.

CiNii Books

researchmap

Language：Japanese   Publisher：The Institute of Electronics, Information and Communication Engineers  

The delay or loss rate have been usually used as a link weight. However, when high functional node with the retransmission processing exists, we should consider the delay reducing and loss recovering effectiveness of packets retransmission between nodes. Therefore, we focus that network environment, and we propose a new link cost which integrate the delay, the loss rate, and the packets retransmission for delay reducing and loss recovery. We show that our proposed link cost itself is suitable. Moreover, we show that it is suitable for path construction on that network environment.

CiNii Books

researchmap

Language：Japanese   Publisher：The Institute of Electronics, Information and Communication Engineers  

As multimedia streaming services become common, admission control is becoming an essential technique for Internet Protocol (IP) networks. We have proposed an admission control scheme called Tentative Accommodating and Congestion Confirming Strategy (TACCS). TACCS is based on the idea to accommodate incoming flows tentatively and confirm congestion after a certain period. In TACCS, ingress nodes of a domain make flow accommodating decisions based on information about packet loss events. We assumed the information is advertised from Congestion Detection Agents (CDAs) that are located in the domain, however, if a comparable performance to TACCS can be achieved without adding CDA functions to core nodes, it is possible to suppress costs for deploying TACCS and improve its scalability. Therefore, in this paper, we propose an enhanced TACCS called Edge-based TACCS that achieves admission control based on cooperation of high functional edge nodes but depends on CDAs in the core network. Through the simulation experiments, we showed Edge-based TACCS has a comparable performance to conventional TACCS that depends on CDAs in the core network.

CiNii Books

researchmap

Language：Japanese   Publisher：The Institute of Electronics, Information and Communication Engineers  

Generally, when users retrieve information on a network, retrieval time and success rate are important for every user. Conventionally, there were retrieval methods, such as the control system of traffic and success rate by the hop limits etc, sequential retrieval which retrieves information by repeating a single query message and flooding method. However, when information used for retrieval was not memorized by each node, and a margin is in retrieval time, conventional retrieval methods cannot achieve effective reduction of retrieval traffic. We have already proposed a new retrieval method which can control retrieval range by processing query packets at each node autonomously using two parameters, energy and diverge rate given to query packets. In this paper, we assume the retrieval which has a range in a retrieval time demand, in this search method, and propose an efficient reduction method of retrieval traffic using two or more query packets. Moreover, we evaluate the setting indicator of each parameter, and the performance of a proposed method. The results also showed that our proposed method has advantages over conventional retrieval techniques.

CiNii Books

researchmap

Language：English   Publisher：The Institute of Electronics, Information and Communication Engineers  

Content searching is one of the main problems in Peer-to-Peer CDNs, which are effective for load and traffic distribution. NISHA, a content search algorithm that uses distributed content search graphs, has been proposed as a solution to this problem. In this paper, we provide an overview of NISHA and compare the traffic characteristics of NISHA with existing search algorithms. We also explain the protocol used to update the content search graph used in NISHA for dynamic events, such as addition and deletion of content in CDNs, and describe the traffic generated in this process. We conclude with a discussion of limitations and various issues regarding the applicability of NISHA to real CDNs.

CiNii Books

researchmap

Language：Japanese  

CiNii Books

researchmap

Language：Japanese  

CiNii Books

researchmap

Publisher：社団法人電子情報通信学会  

ユーザの課金状態に応じて動画像の品質を制御したい場合や,HDDに録り貯めた多数の高画質動画像を比較的低画質の動画像にトランスコードして保存する場合などには,ビットレートが一定であることよりも,符号化した動画像の主観品質が一定であることが望ましい.一般的に,動画像の主観品質は時間方向の品質(動きの滑らかさ)と空間方向の品質(静止画像としての綺麗さ)の両方によって決まる.また,動画像全体の主観品質は,当該動画像における最も主観品質の低い箇所の影響を強く受けることが知られている.これに対し,著者らはこれまでに時間方向および空間方向の主観品質の制御法と予測法を提案してきた.本稿では,これらの成果を応用し,符号化時のビットレート,フレームレートを目的の品質と画像内容に応じて動的に変動させることにより,主観品質の時間的な変動を抑制することで,任意の動画像について目的の主観品質を効率よく実現する符号化法について提案する.

researchmap

Language：Japanese   Publisher：The Institute of Electronics, Information and Communication Engineers  

We proposed a rate control method for video coding to achieve a requested subjective quality. In general, a subjective quality of video is strongly affected by the part with the lowest quality. Therefore, in order to achieve the requested subjective quality efficiently, it is necessary to restrain the temporal variation of subjective quality. This approach requires variable bit rate, variable frame rate video coding, and estimation methods of subjective quality. We already proposed such method for temporal quality (motion smoothness) and spatial quality (quality of each frame). Applying these methods allows us to develop a simple method for achieving the requested subjective quality. Moreover, we implemented the proposed method on the general MPEG-4 encoder. Experimental results show that the proposed method satisfy the requested quality.

CiNii Books

researchmap

Language：Japanese   Publisher：The Institute of Electronics, Information and Communication Engineers  

In content based video retrieval, it is difficult to obtain a query (not keywords) from a user who only has an ambiguous memory about what he wants. On the other hand, we have already proposed the user interface by using relevance feedback that achieves automatic query generation for still image retrieval. In this paper, we proposed a video retrieval system by introducing motion feature and motion histogram matching to our conventional system. The proposed system presents nine automatically generated videos for a user as candidates of the query. Then, a user selects video which are similar to what he wants. After user's selection, the system creates new nine videos based on the video selected by user. This interaction process allows us to create the appropriate query easily. Experimental results show the effectiveness of our proposed system.

CiNii Books

researchmap

Publisher：社団法人電子情報通信学会  

あいまいな記憶からキーワードを用いず動画像検索を行う画像内容検索では,クエリ作成,及び,ユーザインタフェースが重要である.これに対し,我々は,静止画像検索において,候補画像の自動生成を行う適合性フィードバックによって,簡単な操作でのクエリ作成を可能とするユーザインタフェースを,提案してきた.本稿では,これを動画像検索に適用するために,動きの特徴量として縦横方向の動きの大きさをインタフェースに導入し,これからクエリを生成,また,動きベクトルのヒストグラム空間でのマッチング手法を用いた動画像検索システムを提案する.提案システムでは,所望の動画像に似ているものを提示される候補画像の中から選択することの繰り返しにより,属性情報が絞り込まれ,クエリを作成する.提案手法を実装した試作システムにより,動画像検索においても,選択式の簡単なインタフェースを持つシステムが有効であることを確認した.

researchmap

Language：Japanese   Publisher：社団法人電子情報通信学会  

CiNii Books

researchmap

Language：Japanese   Publisher：The Institute of Electronics, Information and Communication Engineers  

Delay jitter causes QoS degradation in delay sensitive real-time multimedia communications. We focused on delay jitter reduction by utilizing multipath, i.e., forwarding duplicate data over a set of paths, and proposed a path set configuration scheme that selects a suitable set of paths for this purpose. Our scheme configures a path set that satisfies a delay constraint, and that has small shared nodes and delay difference. In our previous work, we evaluated our scheme through simulations under random networks and cross-traffic with Poisson processes, and showed that our scheme could reduce delay jitter more effectively than a conventional scheme that chooses the shortest path as the primary path. In this paper, we evaluated a path set configured by our scheme under networks with a property of power-law connectivity, and under cross-traffic with a property of self-similarity. From the results, we demonstrated that our scheme is more effective in realistic network models.

CiNii Books

researchmap

Language：Japanese   Publisher：The Institute of Electronics, Information and Communication Engineers  

Delay jitter degrades the quality of delay-sensitive streaming services. We investigated the use of multipath-based delay jitter reduction method in which duplicate packets are sent over multiple paths in parallel, and proposed a nearly-equal-delay path-set configuration (NEED-PC) scheme. The NEED-PC scheme configures a pair of a maximally node-disjoint paths which have nearly equal path delays and satisfy a given delay constraint. In our previous work, we evaluated NEED-PC scheme through simulations under realistic network models, and showed that our scheme could reduce delay jitter more effectively than a conventional scheme that chooses the shortest path as the primary path. However, since sending duplicate packets of course increases the traffic load, the use of this method itself can be a cause of the delay jitter increase. In this paper, we therefore investigate the actual effect of this method taking into account the increased traffic load due to sending duplicate packets. We showed from the results that the method applying the NEED-PC scheme can improve the quality of some limited amount of high priority traffic and also estimated an appropriate limit for the amount of traffic applied this method.

CiNii Books

researchmap

Language：Japanese   Publisher：The Institute of Electronics, Information and Communication Engineers  

CiNii Books

researchmap

Language：English   Publisher：The Institute of Electronics, Information and Communication Engineers  

CiNii Books

researchmap

researchmap

Publisher：社団法人情報処理学会  

コンテンツ配信ネットワーク(CDN)における重要な問題の一つとしてコンテンツ検索の問題があげられる。しかし, 従来の分散型コンテンツ検索手法はコストの高いフラッディングに基づいているかコンテンツ配置アルゴリズムに制約をおいた上で効率的探索を可能とするものが多い。そこで, 本稿では近傍ノードのコンテンツによる検索木の分散構成に基づく効率的コンテンツ検索手法を提案する。提案手法の重要な特徴はコンテンツ配置アルゴリズムに依存しないという点である。本稿では提案手法を説明し, 従来のコンテンツ検索アルゴリズムと性能の比較を行うことによって提案手法は検索時に生じるトラフィックを少なく保ちながらもコンテンツ発見確率の向上に効果的であることを示す。

researchmap

Language：Japanese   Publisher：The Institute of Electronics, Information and Communication Engineers  

Proposed QMLS (QoS Multicast for Live Streaming) protocol has focused on an end-to-end allowable delay, and this improves the performance of live streaming multicast. We have proposed a path reconfiguration method for QMLS multicast tree to address the case that a bottleneck has developed at a certain link on the multicast tree. This method has been shown its efectiveness as compared to a conventional tree reconfiguration method based on shortest delay. However, it does not improve the QoS in only a few networks. Therefore, we propose a more efficient path reconfiguration method to solve the problem. In selecting a suitable path, we consider not only the delay and the number of QMLS routers on the path but also the distance betwen QMLS routers and retransmission delay. Finally, we evaluate the performance of our method through simulation. The simulation results show that our method provides a multicast tree that is tolerant to variations in traffic after path reconfiguration.

CiNii Books

researchmap

Language：English   Publisher：The Institute of Electronics, Information and Communication Engineers  

One of the most important issues in Content Distribution Networks (CDN) is the algorithm for searching the contents. The existing distributed content location systems for CDNs are either dependent upon inefficient flooding techniques or on methods imposing a restriction on content placement algorithms. In this paper, we propose an efficient content location algorithm for CDNs based on distributed construction of search tree using the contents of proximal nodes. One of the important characteristics of our algorithm is its independence on the content placement algorithm. We describe our algorithm, compare it with the existing content location algorithms and show its effectiveness in increasing the success rate of queries while maintaing the traffic generated very low.

CiNii Books

researchmap

Language：Japanese   Publisher：The Institute of Electronics, Information and Communication Engineers  

CiNii Books

researchmap

Language：Japanese   Publisher：The Institute of Electronics, Information and Communication Engineers  

CiNii Books

researchmap

Language：Japanese   Publisher：The Institute of Electronics, Information and Communication Engineers  

CiNii Books

researchmap

Language：Japanese   Publisher：The Institute of Electronics, Information and Communication Engineers  

CiNii Books

researchmap

Language：Japanese   Publisher：The Institute of Electronics, Information and Communication Engineers  

CiNii Books

researchmap

Language：Japanese   Publisher：The Institute of Electronics, Information and Communication Engineers  

CiNii Books

researchmap

Language：Japanese   Publisher：The Institute of Electronics, Information and Communication Engineers  

CiNii Books

researchmap

Language：Japanese   Publisher：The Institute of Electronics, Information and Communication Engineers  

Recently CDN(Content Delivery Network) has been proposed as a mechanism to delivery content to users. It is necessary to allocate contents according to popularity to avoid the concentration of load to a particular server's storage. In this paper we propose the optimum content allocation in the static model considering the reference cost from storage of center and propose the optimum estimation of query distribution in the dynamic model. Finally we propose the effective content allocation in the dynamic model.

CiNii Books

researchmap

Language：Japanese   Publisher：The Institute of Electronics, Information and Communication Engineers  

Recently CDN(Content Delivery Network) has been proposed as a mechanism to delivery content to users. It is necessary to allocate contents according to popularity to avoid the concentration of load to a particular server's storage. In this paper we propose the optimum content allocation in the static model considering the reference cost from storage of center and propose the optimum estimation of query distribution in the dynamic model. Finally we propose the effective content allocation in the dynamic model.

CiNii Books

researchmap

researchmap

Language：Japanese   Publisher：The Institute of Electronics, Information and Communication Engineers  

Delay jitter and packet loss cause QoS degradation in real-time multimedia communication applications. Previous studies have reported that sending redundant packets through multiple paths can reduce packet loss. However, these studies did not focus on delay jitter when establishing multiple paths. We propose a multiple-path routing method for reducing delay jitter. Our algorithm outputs a set of paths that satisfy a pre-specified delay constraint. and the paths in the set have a minimum of shared nodes and delay difference. Using computer simulations in which redundant packets were sent through the sets of paths, we demonstrated that our method can reduce delay jitter more effectively than conventional methods.

CiNii Books

researchmap

Language：Japanese   Publisher：The Institute of Electronics, Information and Communication Engineers  

Live streaming media. such as streaming video or audio, is delay sensitive, and its allowable delay between end to end is an important factor. Therefore, the proposed QMLS (QoS Multicast for Live Streaming) protocol has focused on an end-to-end allowable delay, and this improves the performance of live streaming multicast. QMLS routers (relay nodes). which are placed somewhere on the multicast tree, detect packet loss and request retransmission to a previous QMLS router. These functions reduce retransmission delays, and therefore, the number of packet losses caused by delays in excess of the allowable delay is reduced. However, when the traffic conditions of a certain link on the multicast tree are getting worse and a bottleneck has developed, even QMLS protocol cannot recover the QoS. Therefore, we propose an efficient path reconfiguration method for QMLS multicast tree to avoid this unsuitable link. In selecting a suitable path. it is important to take the characteristics of QMLS protocol into consideration. Therefore, we consider not only the delay on the path but also the number of QMLS routers for path selection. Our method provides a multicast tree that is tolerant to variations in traffic after path reconfiguration. Finally, we evaluate the performance of our method through simulation. The simulation results show the effectiveness of our proposed method as compared to a conventional tree reconfiguration method based on shortest delay.

CiNii Books

researchmap

Language：Japanese   Publisher：The Institute of Electronics, Information and Communication Engineers  

Live streaming media. such as streaming video or audio, is delay sensitive, and its allowable delay between end to end is an important factor. Therefore, the proposed QMLS (QoS Multicast for Live Streaming) protocol has focused on an end-to-end allowable delay, and this improves the performance of live streaming multicast. QMLS routers (relay nodes). which are placed somewhere on the multicast tree, detect packet loss and request retransmission to a previous QMLS router. These functions reduce retransmission delays, and therefore, the number of packet losses caused by delays in excess of the allowable delay is reduced. However, when the traffic conditions of a certain link on the multicast tree are getting worse and a bottleneck has developed, even QMLS protocol cannot recover the QoS. Therefore, we propose an efficient path reconfiguration method for QMLS multicast tree to avoid this unsuitable link. In selecting a suitable path. it is important to take the characteristics of QMLS protocol into consideration. Therefore, we consider not only the delay on the path but also the number of QMLS routers for path selection. Our method provides a multicast tree that is tolerant to variations in traffic after path reconfiguration. Finally, we evaluate the performance of our method through simulation. The simulation results show the effectiveness of our proposed method as compared to a conventional tree reconfiguration method based on shortest delay.

CiNii Books

researchmap

Language：Japanese   Publisher：The Institute of Electronics, Information and Communication Engineers  

Admission control is becoming an essential technique for Internet Protocol (IP) networks to provide full-fledged multimedia streaming services. In this paper, we propose a new concept of admission control called Tentative Accommodating and Congestion Confirming Strategy (TACCS), which is suitable for packet switching networks. TACCS is not based on the same idea as signaling based admission control techniques, but based on the new idea. i.e., to accommodate incoming flows tentatively and confirm congestion after a certain period. From the results of mathematical analysis based on the combination of M^X/M/c/c and M/G/1/K queuing systems, we show that TACCS makes it possible for a domain to control admission without a centralized management agent. We also provide a guideline for configuring the parameters that determine the behavior of TACCS.

CiNii Books

researchmap

Language：Japanese   Publisher：The Institute of Electronics, Information and Communication Engineers  

Delay jitter and packet loss cause QoS degradation in real-time multimedia communication applications. Previous studies have reported that sending redundant packets through multiple paths can reduce packet loss. However, these studies did not focus on delay jitter when establishing multiple paths. We propose a multiple-path routing method for reducing delay jitter. Our algorithm outputs a set of paths that satisfy a pre-specified delay constraint. and the paths in the set have a minimum of shared nodes and delay difference. Using computer simulations in which redundant packets were sent through the sets of paths, we demonstrated that our method can reduce delay jitter more effectively than conventional methods.

CiNii Books

researchmap

researchmap

Language：Japanese   Publisher：The Institute of Electronics, Information and Communication Engineers  

Admission control is becoming an essential technique for Internet Protocol (IP) networks to provide full-fledged multimedia streaming services. In this paper, we propose a new concept of admission control called Tentative Accommodating and Congestion Confirming Strategy (TACCS), which is suitable for packet switching networks. TACCS is not based on the same idea as signaling based admission control techniques, but based on the new idea. i.e., to accommodate incoming flows tentatively and confirm congestion after a certain period. From the results of mathematical analysis based on the combination of M^X/M/c/c and M/G/1/K queuing systems, we show that TACCS makes it possible for a domain to control admission without a centralized management agent. We also provide a guideline for configuring the parameters that determine the behavior of TACCS.

CiNii Books

researchmap

Language：English   Publisher：The Institute of Electronics, Information and Communication Engineers  

The proposed QoS Multicast for Live Streaming or QMLS protocol focusing on an allowable delay improves the performance of live streaming multicast. QMLS routers detect loss and request retransmission to reduce delays of retransmitted packets and therefore losses are decreased. However, the process in QMLS router increases packet delays. Consequently, activating all routers to perform QMLS protocol might not always improve the performance. This research proposes the adaptive method for QMLS protocol to select routers in the multicast tree to perform QMLS protocol according to loss rates in the multicast tree. Analysis and simulation results reveal that QMLS protocol using this adaptive method has better performance than activating all routers in the multicast tree.

CiNii Books

researchmap

Publisher：社団法人電子情報通信学会  

IPでの移動性を実現するモバイルIPに関してハンドオフ時の切断時間とそれに伴うパケットロスに関する様々な検討が行われているが,ストリーミングアプリケーション利用時にハンドオフ時の経路変更が引き起こす遅延変動により生じる移動端末の再生品質劣化についてはほとんど検討されていない.そこで本研究ではこの問題を解決する新しいハンドオフ手法を提案する.本提案方式では,ハンドオフごとに複数の経路を設定,保持し,ストリーミングの再生に適した遅延変動の少ない経路を移動端末がハンドオフ時に受信経路として選択することによりハンドオフ時のストリーミング再生品質劣化を減少することが可能となる.これにより再生ストリーミングの初期バッファ量を減少させることができ,再生遅延やバッファ量の減少が可能となる.計算機シミュレーションにより提案方式の有効性を示した.

researchmap

Publisher：社団法人電子情報通信学会  

IPでの移動性を実現するモバイルIPに関してハンドオフ時の切断時間とそれに伴うパケットロスに関する様々な検討が行われているが,ストリーミングアプリケーション利用時にハンドオフ時の経路変更が引き起こす遅延変動により生じる移動端末の再生品質劣化についてはほとんど検討されていない.そこで本研究ではこの問題を解決する新しいハンドオフ手法を提案する.本提案方式では,ハンドオフごとに複数の経路を設定,保持し,ストリーミングの再生に適した遅延変動の少ない経路を移動端末がハンドオフ時に受信経路として選択することによりハンドオフ時のストリーミング再生品質劣化を減少することが可能となる.これにより再生ストリーミングの初期バッファ量を減少させることができ,再生遅延やバッファ量の減少が可能となる.計算機シミュレーションにより提案方式の有効性を示した.

researchmap

Language：Japanese   Publisher：The Institute of Electronics, Information and Communication Engineers  

The proposed QMLS protocol focusing on allowable delay provides quality of service for live streaming by relay nodes placed along the multicast tree. The relay nodes detect a loss and retransmit the data. However, checking the sequence number for loss detection and changing the sequence number for avoiding duplicate nacks in relay node increase process delay. The relay nodes should be placed efficiently in view of application layer loss and packet delay. This paper describes adaptive loss recovery multicast protocol that each router decides to start and stop relay by autonomous decision based on packet loss rate. Simulation result comparing to the model which all routers are relay nodes show the effectiveness of this method.

CiNii Books

researchmap

Language：Japanese   Publisher：The Institute of Electronics, Information and Communication Engineers  

When information retrieving on a network, the network traffic and finding probability should be suitably controlled to find the required information. However, conventional retrieval methods do not have functions for simple control of network traffic and finding probability. In this paper, we propose a novel network retrieval method, which achieves flexible retrieval and meets user expectations. The method includes two new parameters, energy E and division probability P, in query messages. These parameters can be easily controlled by users and networks. When a node receives a query message, its node retrieves itself and processes query message autonomously using the parameters E and P. First, the user inserts the parameters E and P in the query message, which is then transmitted over the network. At each node that receives the query message, the message is divided by P, and E is also distributed into each divided qeury message. When the query message passes any links, its E is reduced by the energy of the respective link. When E is less than zero, the query message is eliminated by the node. In this paper, we describe an evaluation of our proposed method using computer simulation. Under the simulation conditions, energy was distributed equally when a query message was divided, and E was reduced by one when a query message passed any link, for basic examination. The results of the simulation showed that our proposed method achieved flexible retrieval from a viewpoint of both the network traffic and finding probability. The results also showed that our proposed method has advantages over conventional flooding retrieval techniques.

CiNii Books

researchmap

Language：Japanese   Publisher：The Institute of Electronics, Information and Communication Engineers  

In this paper, we term multimedia streaming application traffic "stream flows" and the other usual application traffic "non-stream flows". Many problems occur when both flows are aggregated on a shared link because different TCP and UDP behaviors cause negative interactions. One way to solve these problems is to isolate stream and non-stream flows to different classes. However, under conditions where traffic provisioning fails and the bandwidth allocation for each class is unsuitable for the traffic, the QoS will be worse than without isolation. It is considered to change bandwidth allocation ratio adaptively, but it is hard to implement on large scale networks and impossible if the ratio is fixed. Therefore, we have proposed a dynamic class assignment method that maintains the QoS and that has a higher scalability than dynamic bandwidth allocation schemes. It is workable on Diffserv AF PHB. The outline is as follows. Non-stream flows are classified into four classes, and statically assigned to four AF subclasses. Stream flows are dynamically assigned to four subclasses, taking conditions and interactions into consideration. In this paper, we show the algorithm for our dynamic class assignment method and provide results of simulation experiments done on a small-scale intra domain model. The results show that our method provides constant qualities with stream and non-stream flows, adapting to changing traffic.

CiNii Books

researchmap

Language：Japanese   Publisher：The Institute of Electronics, Information and Communication Engineers  

CiNii Books

researchmap

Language：Japanese   Publisher：The Institute of Image Electronics Engineers of Japan  

DOI： 10.11371/iieej.33.242

J-GLOBAL

researchmap

Language：Japanese   Publisher：The Institute of Electronics, Information and Communication Engineers  

CiNii Books

researchmap

Language：Japanese   Publisher：The Institute of Electronics, Information and Communication Engineers  

CiNii Books

researchmap

Language：Japanese   Publisher：The Institute of Electronics, Information and Communication Engineers  

CiNii Books

researchmap

Language：English  

CiNii Books

researchmap

Language：Japanese   Publisher：The Institute of Electronics, Information and Communication Engineers  

The more multimedia streaming application users increase, the more the Internet will be occupied by their traffic. Streaming applications often use UDP to transport their data because they need real-time data arrivals. Against this, other usual applications use TCP because they need reliable data arrivals. Where UDP and TCP flows are aggregated on a single link, the QoS of TCP flows is degraded becouse responsive TCP flows tend to give up bandwidth to unresponsive UDP flows. To maintain the QoS of TCP flows, it is studied to discard UDP flow packts with a higher precedence than those of TCP flows. MRED (Multi-level RED) schemes are generally used to implement this strategy. However, MRED schemes discard packets based on an average queue size calcuration and it is reported that this causes many problems. Therefore, we propose to discard UDP flow packets based on Real Queue Size. We obtained the results that showed this method could maintain the QoS of TCP flows by discarding less UDP flow packets than a general MRED scheme. In this paper, we describe the effectiveness of this method and show some guidelines for configuring parameters that determine the behavior of it.

CiNii Books

researchmap

Language：Japanese   Publisher：The Institute of Electronics, Information and Communication Engineers  

The Content Based Network which administrates contents by its meta datas and the contnet drift model which is the content delivery method on this network are proposed. However, it is hard to clear its content derivery characteristics by mathmatical analysis because every content drifts indepentely of each other. To cope whith this problem, we are to implement the evaluation system and evaluate it by. the comparison between the conventional caching and the content drifting from the viewpoints of the costs to download contents and the costs of content drifts. The target caching algorithms are LRU and LFU, and the target content drift algorithms are the ones which use LRU and LFU to determine the contents to move or delete respectively. Key words

CiNii Books

researchmap

Language：Japanese   Publisher：The Institute of Electronics, Information and Communication Engineers  

Satellite communication has delay allocation at the beginning of communication. For improving this delay-overhead, we proposed a method of 'Standby Bandwidth'. Standby Bandwidth is a bandwidth that any transmitting stations can use without waiting for an allocation notification. Therefore, a transmitting station can communicate with a receiving station by this bandwidth at once. But this bandwidth decrease in proportion to increase transmitting station, and the characteristic of this decrease hasn't been made clear. In this report, we make clear this characteristic using approximate analysis. First, we establish analysis of 'probability of loss' using 'M/M/S/S Queuing Theory'. Second, we compare the analysis results with simulation results, and inspect that this approximate analysis is valid. Lastly, we present the effective application range of the Standby Bandwidth method using a formula of loss probability.

CiNii Books

researchmap

Language：Japanese   Publisher：The Institute of Electronics, Information and Communication Engineers  

The more multimedia streaming application users increase, the more the Internet will be occupied by their traffic. Streaming applications often use UDP to transport their data because they need real-time data arrivals. Against this, other usual applications use TCP because they need reliable data arrivals. Where UDP and TCP flows are aggregated on a single link, the QoS of TCP flows is degraded becouse responsive TCP flows tend to give up bandwidth to unresponsive UDP flows. To maintain the QoS of TCP flows, it is studied to discard UDP flow packts with a higher precedence than those of TCP flows. MRED (Multi-level RED) schemes are generally used to implement this strategy. However, MRED schemes discard packets based on an average queue size calcuration and it is reported that this causes many problems. Therefore, we propose to discard UDP flow packets based on Real Queue Size. We obtained the results that showed this method could maintain the QoS of TCP flows by discarding less UDP flow packets than a general MRED scheme. In this paper, we describe the effectiveness of this method and show some guidelines for configuring parameters that determine the behavior of it.

CiNii Books

researchmap

Language：Japanese   Publisher：The Institute of Electronics, Information and Communication Engineers  

We already proposed a contents synchronous arrival method [1], [2] with multicast transmissions. The method is carried out to minimize the total SLA evaluation value of multicast groups. Therefore, the more the number of multicast groups is larger, the more the probability of contention is higher. It is a problem for the method to solve, if we adapt to the complicated SLA conditions. This paper describes a simpler and enables method to deliver the contents in a same time for a several SLAs. To carry out these objects, the encrypted contents data were delivered with multicasts for each SLA level. The keys which start time of delivering are adjusted for each client are used to synchronize the contents delivery time. The key is small enough to deliver in the same time.

CiNii Books

researchmap

Language：Japanese   Publisher：The Institute of Electronics, Information and Communication Engineers  

Satellite communication has delay allocation at the beginning of communication. For improving this delay-overhead, we proposed a method of 'Standby Bandwidth'. Standby Bandwidth is a bandwidth that any transmitting stations can use without waiting for an allocation notification. Therefore, a transmitting station can communicate with a receiving station by this bandwidth at once. But this bandwidth decrease in proportion to increase transmitting station, and the characteristic of this decrease hasn't been made clear. In this report, we make clear this characteristic using approximate analysis. First, we establish analysis of 'probability of loss' using 'M/M/S/S Queuing Theory'. Second, we compare the analysis results with simulation results, and inspect that this approximate analysis is valid. Lastly, we present the effective application range of the Standby Bandwidth method using a formula of loss probability.

CiNii Books

researchmap

Language：Japanese   Publisher：The Institute of Electronics, Information and Communication Engineers  

We already proposed a contents synchronous arrival method [1], [2] with multicast transmissions. The method is carried out to minimize the total SLA evaluation value of multicast groups. Therefore, the more the number of multicast groups is larger, the more the probability of contention is higher. It is a problem for the method to solve, if we adapt to the complicated SLA conditions. This paper describes a simpler and enables method to deliver the contents in a same time for a several SLAs. To carry out these objects, the encrypted contents data were delivered with multicasts for each SLA level. The keys which start time of delivering are adjusted for each client are used to synchronize the contents delivery time. The key is small enough to deliver in the same time.

CiNii Books

researchmap

Language：Japanese   Publisher：The Institute of Electronics, Information and Communication Engineers  

The Content Based Network which administrates contents by its meta datas and the contnet drift model which is the content delivery method on this network are proposed. However, it is hard to clear its content derivery characteristics by mathmatical analysis because every content drifts indepentely of each other. To cope whith this problem, we are to implement the evaluation system and evaluate it by. the comparison between the conventional caching and the content drifting from the viewpoints of the costs to download contents and the costs of content drifts. The target caching algorithms are LRU and LFU, and the target content drift algorithms are the ones which use LRU and LFU to determine the contents to move or delete respectively. Key words

CiNii Books

researchmap

Language：Japanese   Publisher：The Institute of Electronics, Information and Communication Engineers  

The Content Based Network (CBN) is a novel platform for content distribution, and a Content Drift Model is used to place contents on the CBN efficiently according to user's changing preferences. The conventional cache method considers the retrieval histories of the contents only at the cache, but the Content Drift Model puts meta-data which describes retrieval histories into the content itself. Thus, it is able to place contents more efficiently than the cache method can. In this paper, to study the basic attributes of the drift method, we add popularity to content as meta-data, and we describe a drift method that reduces the number of links the user has to go through when downloading content. Simulations show the effectiveness of the method.

CiNii Books

researchmap

Language：Japanese   Publisher：The Institute of Electronics, Information and Communication Engineers  

For related information retrieval on the Web which is formed by hyperlinks, restricting the search space to a specific topic can increase the precision of the search results. One of the easiest ways to do it, is to limit the search space to Web contents within a fixed link distance from the Web page specified by the user. However, the relationship between the link distance and the semantic distance of the Web contents is not proportional. Thus, using only a fixed link distance is insufficient to remove unnecessary search results. In this paper, we propose that the strength of the relationship between two linked Web contents, is represented by using the similarity of their contents, and each similarity reflects the link distances respectively. When searching, the search space is formed by gathering Web contents which are closely related according to similarities, when tracing both forward and backward links from the Web page specified by the user. To verify the effectiveness of our method, we applied it to the actual Web link space. The experimental result shows that the precision of the search rasults increases using the proposed method.

CiNii Books

researchmap

Language：English  

CiNii Books

researchmap

Language：English  

CiNii Books

researchmap

Language：Japanese   Publisher：The Institute of Electronics, Information and Communication Engineers  

Notwithstanding the attractiveness of Peer-to-Peer Content Distribution to people, their focus is mainly on file sharing efficiency. In this paper, a linked architecture with a content sharing layer and a metadata distribution layer, and a content distribution model, in which each peer plays a role as an economic peer, are proposed. And a vitalization scheme of content distribution is discussed under the aegis of a reward as each peer's incentive to cooperate.

CiNii Books

researchmap

Publisher：社団法人電子情報通信学会  

従来のPeer-to-Peer (P2P)アプリケーションでは,各ピアがコンテンツの宣伝情報を伝搬させる動機がないために,コンテンツの存在を能動的に認知させる手段がないという問題がある.そこで本稿は認知活動の見返りとして報酬を付与するP2Pコンテンツ流通モデルに対し,各ピアにおけるコストの前提条件等を明らかにし,宣伝情報伝播範囲,コンテンツ販売者,各ピアおよび系全体の利益特性を,数値的に解析することで, P2Pコミュニティを対象にしたコンテンツ流通の経済活動に有効性があることを示す.

researchmap

Language：Japanese   Publisher：The Institute of Electronics, Information and Communication Engineers  

Notwithstanding the attractiveness of Peer-to-Peer Content Distribution to people, their focus is mainly on file sharing efficiency. In this paper, a linked architecture with a content sharing layer and a metadata distribution layer, and a content distribution model, in which each peer plays a role as an economic peer, are proposed. And a vitalization scheme of content distribution is discussed under the aegis of a reward as each peer's incentive to cooperate.

CiNii Books

researchmap

Language：Japanese   Publisher：The Institute of Electronics, Information and Communication Engineers  

The conventional Peer-to-Peer(P2P) applications for a content distribution encounter a serious problem, which is no way to inform consumers of content since each peer has no incentive to propagate an advertisement. In this paper, we propose a P2P reward-based content distribution model taking peer's cost into consideration, in which peers obtain rewards according to informing activities. The number of informed peers and the properties of the net gain for a distributor, each peer and a whole system are numerically analyzed in the proposed model after setting out preconditions of peer's costs. In conclusion, efficiency of P2P distribution model is dusccussed on the basis of these analyses.

CiNii Books

researchmap

Language：Japanese   Publisher：The Institute of Electronics, Information and Communication Engineers  

Notwithstanding the attractiveness of Peer-to-Peer Content Distribution to people, their focus is mainly on file sharing efficiency. In this paper, a linked architecture with a content sharing layer and a metadata distribution layer, and a content distribution model, in which each peer plays a role as an economic peer, are proposed. And a vitalization scheme of content distribution is discussed under the aegis of a reward as each peer's incentive to cooperate.

CiNii Books

researchmap

Language：Japanese   Publisher：The Institute of Electronics, Information and Communication Engineers  

The conventional Peer-to-Peer(P2P) applications for a content distribution encounter a serious problem, which is no way to inform consumers of content since each peer has no incentive to propagate an advertisement. In this paper, we propose a P2P reward-based content distribution model taking peer's cost into consideration, in which peers obtain rewards according to informing activities. The number of informed peers and the properties of the net gain for a distributor, each peer and a whole system are numerically analyzed in the proposed model after setting out preconditions of peer's costs. In conclusion, efficiency of P2P distribution model is dusccussed on the basis of these analyses.

CiNii Books

researchmap

researchmap

Language：Japanese   Publisher：The Institute of Electronics, Information and Communication Engineers  

Wireless LAN has considered as a infrastructure to realize mobility service because of its fast transmission speed. Generally, Mobile IP realizes mobility function at IP layer. However, in wireless LAN environment, packet loss and transmission delay while mobilenode(MN) handoff between access routers(ARs) cause reduction of application quality. In this paper, we propose the reducing method not only packet loss and transmission delay but also jitter through AR establishes 3 routes at most between the Corresponding node(CN) and itself, and selects minimum jitter route.

CiNii Books

researchmap

Publisher：社団法人電子情報通信学会  

移動通信サービスを実現するために,その高速性から無線LANの利用が検討されている.IP層での移動性を実現するためには一般にMobile IPが利用されるが,無線LAN環境では,モバイルノードが基地局間をハンドオフする間に生じるパケットロスや伝送遅延は,ストリーミングなどのリアルタイムアプリケーションにとって致命的な品質劣化をもたらす.本研究では,アクセスルータはバッファを持ち,通信ノードとの間に最大3つの通信ルートを確立して,MNへパケットを送る際に最もジッタが少なくなるルートを選択することで,パケットロス,遅延を減少すると共に,ジッタの減少を目的とした新しいハンドオフ方式を提案する.

researchmap

Publisher：社団法人電子情報通信学会  

移動通信サービスを実現するために,その高速性から無線LANの利用が検討されている.IP層での移動性を実現するためには一般にMobile IPが利用されるが,無線LAN環境では,モバイルノードが基地局間をハンドオフする間に生じるパケットロスや伝送遅延は,ストリーミングなどのリアルタイムアプリケーションにとって致命的な品質劣化をもたらす.本研究では,アクセスルータはバッファを持ち,通信ノードとの間に最大3つの通信ルートを確立して,MNへパケットを送る際に最もジッタが少なくなるルートを選択することで,パケットロス,遅延を減少すると共に,ジッタの減少を目的とした新しいハンドオフ方式を提案する.

researchmap

Language：Japanese   Publisher：The Institute of Electronics, Information and Communication Engineers  

Hyperlinks have an important role on the Web. A high possibility exists that semantic relevance is between neighborhood content in the link space constituted by the content connected by the hyperlink. Therefore, more precision can be expected, not by searching a wide area using a general search engine, but by searching more specific areas by defining the topics or referring to link contexts, which is how they are linked to other content. To achieve this retrieval, we made a personal information retrieval system using cooperative servers (PIRCS), which is a search engine that considers the function share requested to the server and the client and link context as a whole. We expanded the cooperative Web architecture on the server side, we introduced fundamental management of retrieval indexes, an evaluation of similarity, a notice function for consistency maintenance of the index between the servers, etc. We designed the client side to have a link space browser where the connection of the content in the link space can be read and a personalization function. We introduced the expansion header of HTTP for communicating additional information between the servers and between the server and client. And, we introduced a search filter in a plug-in form for a similarity decision between the search key and the content that is the essential processing of the retrieval correspond to various Content-type and the purpose of retrieval.

CiNii Books

researchmap

researchmap

Language：Japanese   Publisher：The Institute of Electronics, Information and Communication Engineers  

CiNii Books

researchmap

Language：Japanese   Publisher：The Institute of Electronics, Information and Communication Engineers  

CiNii Books

researchmap

Language：Japanese   Publisher：The Institute of Electronics, Information and Communication Engineers  

CiNii Books

researchmap

Language：Japanese   Publisher：The Institute of Electronics, Information and Communication Engineers  

CiNii Books

researchmap

Language：Japanese   Publisher：The Institute of Electronics, Information and Communication Engineers  

CiNii Books

researchmap

Language：Japanese   Publisher：The Institute of Electronics, Information and Communication Engineers  

CiNii Books

researchmap

Language：Japanese   Publisher：The Institute of Electronics, Information and Communication Engineers  

Satellite communication technology has been looked to extending the terrestrial network to remote areas in recent years. It has been used for the purpose of providing a dedicated line from one to another or of broadcasting from a specified host. However the demand for a switching network such as the Internet, where a host communicates with one of a number of unspecified hosts, is increasing, and expectations are high that satellite technology will be adapted. There are some problems when a satellite network are used as switching network. In this paper, we focus on delay allocation at the beginning of communication. For improving this delay-overhead, we propose a method of 'standby bandwidth'. This method enables bandwidth allocation logically as independent from bandwidth reservation of a physical layer. First, we suggest the proposed method in a VSAT network, which is the most popular satellite network. Finally, we evaluate the proposed method in simulation and show that the proposed method performs reliably and effectively.

CiNii Books

researchmap

Language：Japanese   Publisher：The Institute of Electronics, Information and Communication Engineers  

Satellite communication technology has been looked to extending the terrestrial network to remote areas in recent years. It has been used for the purpose of providing a dedicated line from one to another or of broadcasting from a specified host. However the demand for a switching network such as the Internet, where a host communicates with one of a number of unspecified hosts, is increasing, and expectations are high that satellite technology will be adapted. There are some problems when a satellite network are used as switching network. In this paper, we focus on delay allocation at the beginning of communication. For improving this delay-overhead, we propose a method of 'standby bandwidth'. This method enables bandwidth allocation logically as independent from bandwidth reservation of a physical layer. First, we suggest the proposed method in a VSAT network, which is the most popular satellite network. Finally, we evaluate the proposed method in simulation and show that the proposed method performs reliably and effectively.

CiNii Books

researchmap

Publisher：社団法人電子情報通信学会  

衛星通信技術は,地上網を遠隔地へ拡大する手段として,回線交換による一対一の利用や,衛星放送のような単一ホストからのブロードキャスト通信の形態で利用されてきた.しかしながら,インターネットのようなひとつのホストが数多くの不特定のホストと通信を行なうパケット交換網に対する需要が近年高まってきており,衛星通信に対しても適用が望まれている.衛星通信をパケット交換網として利用する場合,いくつかの問題が存在することが知られている.本研究では特に,通信開始時の帯域割り当て遅延について焦点をあて,この遅延オーバーヘッドを軽減するための「予備帯域方式」を提案する.これは,衛星ネットワークのチャネル割り当てとは独立に,空き帯域を論理的に分配することにより実現される.本稿では,最も一般的な衛星ネットワークであるVSATネットワークを対象として予備帯域によりオーバーヘッド削減を実現する方式を提案し,計算機シミュレーションにより提案方式の動作

researchmap

Language：Japanese   Publisher：The Institute of Electronics, Information and Communication Engineers  

Satellite communication technology has been looked to extending the terrestrial network to remote areas in recent years. It has been used for the purpose of providing a dedicated line from one to another or of broadcasting from a specified host. However the demand for a switching network such as the Internet, where a host communicates with one of a number of unspecified hosts, is increasing, and expectations are high that satellite technology will be adapted. There are some problems when a satellite network are used as switching network. In this paper, we focus on delay allocation at the beginning of communication. For improving this delay-overhead, we propose a method of 'standby bandwidth'. This method enables bandwidth allocation logically as independent from bandwidth reservation of a physical layer. First, we suggest the proposed method in a VSAT network, which is the most popular satellite network. Finally, we evaluate the proposed method in simulation and show that the proposed method performs reliably and effectively.

CiNii Books

researchmap

Language：Japanese   Publisher：The Institute of Electronics, Information and Communication Engineers  

Where UDP stream flows and other usual application traffic are aggregated on a link, there are many problems because different TCP and UDP begaviors cause negative interactions. From the reason, we have proposed a method which reduces their nagative interactions. From the reason, we have proposed a method which reduces their negative interactions. In our method, we proposed that stream flow packets should be dropped earlier than other flow packets when congestion occurs. In this paper, we reveal how to decide parameter settings of RIO-C to implement the strategy of dropping stream flow packets early. First, we specify which parameters determine the effect ot the strategy. Second, we describe the effective parameter setting we obtained from various pattern sumulation experiments, and show that the setting is widely usable as the default value of our system.

CiNii Books

researchmap

Language：Japanese   Publisher：Information Processing Society of Japan (IPSJ)  

This paper presents a methodology to realize a pay-broadcasting system which can supply the contents to support a number of clients simultaneously on the Internet. Various user-authentication methods on IP multicasting are compared under the constraint that the system should be constructed without any specific hardware for user identification, and it also must be installed without any changes on the network infrastructure of the Internet. To support scalable number of users, the contents of the system should be supplied to the users as the same data-stream. To keep the security of the contents, the encrypted single data-stream should be decoded by different keys individually distributed to the authorized users. Broadcasting Encryption and Tracing Traitors Schemes proposed by Naor et al. are considered to be suitable to realize such system. As a prototype of the system, a music broadcasting system named "MusicCast/AS" was implemented with user-authentication on IP multicasting. Its server and clients are implemented as software running on common Linux systems and the system do not need any changes in the existing campus network as a testbed. The practicality of the methodology was verified by the experimentation to broadcast encrypted music data to the campus network. The scalablity of the users and security of the keys are also discussed on the implementation.

CiNii Books

researchmap

Language：Japanese  

CiNii Books

researchmap

Language：Japanese  

CiNii Books

researchmap

Language：Japanese   Publisher：The Institute of Electronics, Information and Communication Engineers  

CiNii Books

researchmap

Language：Japanese   Publisher：The Institute of Electronics, Information and Communication Engineers  

CiNii Books

researchmap

Language：Japanese   Publisher：The Institute of Electronics, Information and Communication Engineers  

CiNii Books

researchmap

Language：Japanese   Publisher：The Institute of Electronics, Information and Communication Engineers  

CiNii Books

researchmap

Language：Japanese   Publisher：The Institute of Electronics, Information and Communication Engineers  

CiNii Books

researchmap

Language：Japanese   Publisher：The Institute of Electronics, Information and Communication Engineers  

CiNii Books

researchmap

Language：Japanese   Publisher：The Institute of Electronics, Information and Communication Engineers  

CiNii Books

researchmap

Language：Japanese   Publisher：The Institute of Electronics, Information and Communication Engineers  

CiNii Books

researchmap

Publisher：社団法人電子情報通信学会  

ストり-ムアプリケーションの普及に伴い,それらが発生する通信(以下,ストリームフロー)がIPネットワーク上のトラヒックの多くを占める状況が今後想定される.UDPにより通信されるストリームフローと,それ以外の一般的なTCPアプリケーション(以下,非ストリームフロー)のトラヒックが集約された場合,TCPとUDPの振る舞いの違いから,双方の相互影響によるサービス品質の劣化が問題となるため,これを緩和する新しい方式が必要となる.非ストリームフローは,その上位のアプリケーションにより,要求するQoS,トラヒックの統計的性質等に違いがあるため,それらを分類することにより,ストリームフローとの相互影響についても差異を生むことができると考えられる.そこで,これらをいくつかのクラスに分類し,それぞれのクラスに別のキューを割り当てた上で,各クラスの非ストリームフローとストリームフローとの相互影響の大きさを明らかにする,次に,ストリームフローを,到着時の混雑状況と,相

researchmap

Language：Japanese   Publisher：The Institute of Electronics, Information and Communication Engineers  

The more multimedia streaming application users increase, the more such application traffic will occupy IP network traffic. In this paper, we term such streaming application traffic as "stream flows", and other usual application traffic as "non-stream flows". Most non-stream flows are TCP flows, and most stream flows are UDP flows. There are many problems when both flows are queued into the same single queue, because different TCP and UDP behaviors cause adverse interactions. In this paper, we first classify non-stream flows into several classes using differences in statistical characteristics and QoS requirements of upper layer applications, because each class has different interactions between stream flows and non-stream flows. We then verify the characteristics of each class. Based on this, we propose a new method of class assignment which dynamically assigns a proper class to a newly arrived stream flow, considering the conditions and characteristics of each class. The method we propose can reduce these negative interactions, and we discuss effective results we obtained through simulation experiments.

CiNii Books

researchmap

Language：Japanese   Publisher：The Institute of Electronics, Information and Communication Engineers  

Real-time applications, such as VoIP and Video Conference, have been widely introduced to the Internet. We propose Packet Merge technique to improve QoS degradation caused by many and small packets that sending by real-time application. In this paper, we shows the procedure which computes QoS from a traffic parameter and a network performance parameter. In addition, we evaluate improvement effect of QoS by proposed technique theoretically. Moreover, we discuss about necessary performance of packet processing to implementation Packet Merge mechanism.

CiNii Books

researchmap

Language：Japanese   Publisher：The Institute of Electronics, Information and Communication Engineers  

Real-time applications, such as VoIP and Video Conference, have been widely introduced to the Internet. We propose Packet Merge technique to improve QoS degradation caused by many and small packets that sending by real-time application. In this paper, we shows the procedure which computes QoS from a traffic parameter and a network performance parameter. In addition, we evaluate improvement effect of QoS by proposed technique theoretically. Moreover, we discuss about necessary performance of packet processing to implementation Packet Merge mechanism.

CiNii Books

researchmap

Publisher：映像情報メディア学会  

researchmap

Publisher：社団法人情報処理学会  

筆者等は広域ネットワーク上のMPEGビデオストリーミングにおける再生品質の公平性を指針とした帯域制御方式[1]を提案して、ネットワーク内にボトルネック回線が存在する場合においても、輻輳を発生させず、ボトルネック回線を共有するフローの再生品質の均一化を実現した。しかしこの方式では、ルータ間の協調性が乏しいため、空き帯域や動画像フローの特性の変動に対する応答性が悪く、再生品質の迅速な収束を実現する仕組みが求められた。本稿では、フロー状態通知を利用した分散協調型ストリームフローQoS制御方式を提案する。本方式では、ルータが回線やその上を流れるフローの状態・状況を監視し、サーバに対する個別な制御を行い、サーバが伝送経路上のルータヘのフロー状態を通知することにより、QoSの分散協調制御を実現する。各ルータは、接続された回線の輻輳を検出すると、その上を流れるフローの状態および再生品質に基づいてそれらの送出元サーバに対

researchmap

researchmap

Publisher：社団法人電子情報通信学会  

筆者等は広域ネットワーク上のMPEGビデオストリーミングにおける再生品質の公平性を指針とした帯域制御方式[1]を提案して、ネットワーク内にボトルネック回線が存在する場合においても、輻輳を発生させず、ボトルネック回線を共有するフローの再生品質の均一化を実現した。しかしこの方式では、ルータ間の協調性が乏しいため、空き帯域や動画像フローの特性の変動に対する応答性が悪く、再生品質の迅速な収束を実現する仕組みが求められた。本稿では、フロー状態通知を利用した分散協調型ストリームフローQoS制御方式を提案する。本方式では、ルータが回線やその上を流れるフローの状態・状況を監視し、サーバに対する個別な制御を行い、サーバが伝送経路上のルータヘのフロー状態を通知することにより、QoSの分散協調制御を実現する。各ルータは、接続された回線の輻輳を検出すると、その上を流れるフローの状態および再生品質に基づいてそれらの送出元サーバに対

researchmap

Language：Japanese   Publisher：The Institute of Electronics, Information and Communication Engineers  

The bandwidth management scheme [1] was proposed, to obtain fairness for MPEG video streaming QoS on wide area network. In this scheme, when bottlenecks exists in the network, the QoS of video streams that share the bottleneck can be maintained. However, the lack of cooperativeness among the routers make response to the change of available bandwidth and video characteristic slow. Therefore, the scheme that can make quick response to the network dynamics is required. This paper proposes the distributed cooperative control scheme of stream flow based on flow state, video characteristic and available bandwidth. In this method, the router plays an important role for distributed QoS control. It has the function of monitoring link bandwidth, flow state of video streams which share that link, and sending message to streaming server. Futhermore, server has the function to acknowledge the flow state to all of the routers on the path. When congestion in a link is detected, router sends appropriate acknowledgements to the streaming servers which share that link, based on the flow state, video characteristic, and available bandwidth. By simulations, it is shown that the proposed scheme can give quick response to the change of network resources. We also study the relation between the responsiveness and parameters.

CiNii Books

researchmap

Language：Japanese   Publisher：The Institute of Electronics, Information and Communication Engineers  

CiNii Books

researchmap

Language：Japanese   Publisher：The Institute of Image Information and Television Engineers  

In this paper, we propose an autonomous distributed QoS control method for MPEG video streaming in the wide area network. The capacity of the links and the characteristics of the video streams change dynamically. However, it is difficult to manage the condition of all the links and the streams in the network intensively. On this account, the routers which compose a network observe the conditions of the links and the streams locally, and control the transmission rate of the stream server in the individual judgment. Picture quality oriented fairness is realized because the transmission rate of the stream which has the highest PSNR in the bottleneck link is reduced. The computer simulation results show that this control concept is effective.

DOI： 10.11485/itetr.26.82.0_7

researchmap

Publisher：社団法人電子情報通信学会  

リアルタイムメディア伝送で再送による誤り訂正を行った場合パケット誤り率やノード間の遅延が大きいと、再送による遅延増加が大きくなる。許容遅延を越えたパケットは廃棄されるためリアルタイムメディアのアプリケーションのQoSに影響を及ぼし、品質が劣化する問題がある。本研究ではプロトコル中継を用いたリアルタイムメディア伝送方式におけるルート設計と効果を検討する。プロトコル中継による廃棄率低減の効果を検証するため、UDP/Iを用いた実験ネットワークにおいてエンドーエンド間の廃棄率を測定した。また、廃棄率の低減に効率的なパス決定アルゴリズムを提案した。

researchmap

Publisher：社団法人電子情報通信学会  

リアルタイムメディア伝送で再送による誤り訂正を行った場合パケット誤り率やノード間の遅延が大きいと、再送による遅延増加が大きくなる。許容遅延を越えたパケットは廃棄されるためリアルタイムメディアのアプリケーションのQoSに影響を及ぼし、品質が劣化する問題がある。本研究ではプロトコル中継を用いたリアルタイムメディア伝送方式におけるルート設計と効果を検討する。プロトコル中継による廃棄率低減の効果を検証するため、UDP/Iを用いた実験ネットワークにおいてエンドーエンド間の廃棄率を測定した。また、廃棄率の低減に効率的なパス決定アルゴリズムを提案した。

researchmap

Language：English   Publisher：The Institute of Electronics, Information and Communication Engineers  

In real-time media communication, large propagation delay and probability of packet loss associated with the large-scale network increases the end-end packet delay and discard ratio by an allowable delay. Although packets can be reached to the destination however its delays, exceeding an allowable delay will be discarded by application level and this affects to the QoS of real-time applications. In this paper at first, we examine the relations of end-end delay, end-end packet discard rate and packet error rate by setting up an experiment network at application layer over UDP/IP. And then we propose an algorithm for determining the effective path in a packet transmission network model with the present of a number of protocol relay nodes.

CiNii Books

researchmap

Language：English   Publisher：The Institute of Electronics, Information and Communication Engineers  

Live stream media is delay sensitive with small allowable delay. Current conventional multicast protocols do not provide reliability guarantee. Even there are some researches provide reliability for multicast but long propagation and high packet loss rate make them inefficient. Any lost packet should be recovered within its allowable delay or else it will be useless. This paper proposes a new multicast protocol based on protocol relay concept. This proposed protocol focusing on allowable delay provides faster data recovery mechanism, which is appropriate for live stream media. Relay nodes are placed along the multicast tree. Data recovery is done between relay nodes. New techniques that enable protocol relay to request retransmission immediately without waiting time like in the conventional technique are proposed. It also reduces the number of time-over packets sent to receivers. We analyzed the expected traffic and loss probability of this proposed protocol comparing with other conventional multicasts. The results show that proposed protocol has loss smaller than the conventional protocols.

CiNii Books

researchmap

Language：English   Publisher：The Institute of Electronics, Information and Communication Engineers  

Live stream media is delay sensitive with small allowable delay. Current conventional multicast protocols do not provide reliability guarantee. Even there are some researches provide reliability for multicast but long propagation and high packet loss rate make them inefficient. Any lost packet should be recovered within its allowable delay or else it will be useless. This paper proposes a new multicast protocol based on protocol relay concept. This proposed protocol focusing on allowable delay provides faster data recovery mechanism, which is appropriate for live stream media. Relay nodes are placed along the multicast tree. Data recovery is done between relay nodes. New techniques that enable protocol relay to request retransmission immediately without waiting time like in the conventional technique are proposed. It also reduces the number of time-over packets sent to receivers. We analyzed the expected traffic and loss probability of this proposed protocol comparing with other conventional multicasts. The results show that proposed protocol has loss smaller than the conventional protocols.

CiNii Books

researchmap

Language：Japanese   Publisher：The Institute of Electronics, Information and Communication Engineers  

This paper describes a study of a multicast communication method, which can synchronize to deliver within a SLA of a time deviation. First, the SLA is converted into a SLA penalty function. Convoluting the SLA penalty function and prediction errors of communication channel state individually generates penalty functions of clients. The clients are divided into some groups. The number of group is calculated by the maximum bandwidth of the delivery server. And, the timings of delivery are decided to minimize the penalties of each group. The estimation method of communication channel state, the penalty function and the method of timing decision are mentioned in detail.

CiNii Books

researchmap

Publisher：社団法人電子情報通信学会  

インターネットの普及,高性能化に伴い,リアルタイム通信を要求するVoIPやビデオ会議といったアプリケーションがインターネット上でも使用され始めている.しかし,リアルタイム系アプリケーションのストリームは小サイズのパケットの高レートのフローとなる傾向があるため,その経路のQoSの低下を招く可能性がある.本稿では,フローを構成するパケットのレート,サイズにより,経路のQoSに与える影響が異なることに着目し,そのレートの調整によりQoS性能を向上させるパケットマージ手法を提案し,実ネットワークでの予備実験から,適応領域を示す.

researchmap

Language：English   Publisher：The Institute of Electronics, Information and Communication Engineers  

Live stream media is delay sensitive with small allowable delay. Current conventional multicast protocols do not provide reliability guarantee. Even there are some researches provide reliability for multicast but long propagation and high packet loss rate make them inefficient. Any lost packet should be recovered within its allowable delay or else it will be useless. This paper proposes a new multicast protocol based on protocol relay concept. This proposed protocol focusing on allowable delay provides faster data recovery mechanism, which is appropriate for live stream media. Relay nodes are placed along the multicast tree. Data recovery is done between relay nodes. New techniques that enable protocol relay to request retransmission immediately without waiting time like in the conventional technique are proposed. It also reduces the number of time-over packets sent to receivers. We analyzed the expected traffic and loss probability of this proposed protocol comparing with other conventional multicasts. The results show that proposed protocol has loss smaller than the conventional protocols.

CiNii Books

researchmap

Language：Japanese   Publisher：The Institute of Electronics, Information and Communication Engineers  

This paper describes a study of a multicast communication method, which can synchronize to deliver within a SLA of a time deviation. First, the SLA is converted into a SLA penalty function. Convoluting the SLA penalty function and prediction errors of communication channel state individually generates penalty functions of clients. The clients are divided into some groups. The number of group is calculated by the maximum bandwidth of the delivery server. And, the timings of delivery are decided to minimize the penalties of each group. The estimation method of communication channel state, the penalty function and the method of timing decision are mentioned in detail.

CiNii Books

researchmap

Language：Japanese   Publisher：The Institute of Electronics, Information and Communication Engineers  

This paper describes a study of a multicast communication method, which can synchronize to deliver within a SLA of a time deviation. First, the SLA is converted into a SLA penalty function. Convoluting the SLA penalty function and prediction errors of communication channel state individually generates penalty functions of clients. The clients are divided into some groups. The number of group is calculated by the maximum bandwidth of the delivery server. And, the timings of delivery are decided to minimize the penalties of each group. The estimation method of communication channel state, the penalty function and the method of timing decision are mentioned in detail.

CiNii Books

researchmap

Language：Japanese   Publisher：The Institute of Electronics, Information and Communication Engineers  

Real-time applications, such as VoIP and Video conference, have been widely introduced to the Internet. Usually a realtime application stream consist of many and small size packets per unit of time. It may cause a QoS reduction on some routers. In this paper, we focused on the difference in router performances in case that packet transfer rates are changed, and we proposed new network control method "Packet Merge Technique" that QoS is improved by packet transfer rates adjusting. Moreover, the applicable domain is shown by preliminary experiment on a real network.

CiNii Books

researchmap

Publisher：社団法人電子情報通信学会  

インターネットの普及,高性能化に伴い,リアルタイム通信を要求するVoIPやビデオ会議といったアプリケーションがインターネット上でも使用され始めている.しかし,リアルタイム系アプリケーションのストリームは小サイズのパケットの高レートのフローとなる傾向があるため,その経路のQoSの低下を招く可能性がある.本稿では,フローを構成するパケットのレート,サイズにより,経路のQoSに与える影響が異なることに着目し,そのレートの調整によりQoS性能を向上させるパケットマージ手法を提案し,実ネットワークでの予備実験から,適応領域を示す.

researchmap

Publisher：社団法人電子情報通信学会  

近年,インターネットは商取引やコンテンツ流通のプラットフォームとして注目を集めている.従来,高機能なサービスはアプリケーションレイヤで実現していたが,我々はコンテンツの発見・検索・配信などのサービスをネットワーク側で提供する「コンテンツ指向ネットワーク」(CBN)を提案してきている.またCBNの特徴を生かして,コンテンツを位置情報と分離し,属性情報や流通履歴を利用してネットワーク内を自律的に移動するオペレーションを与えることにより,より効果的なコンテンツ配置を行う「コンテンツ漂流モデル」を提案している.しかし,このネットワーク上ではコンテンツは移動するため,従来のWWWのURLなどによる位置情報を用いたコンテンツの入手は困難である.本研究では,属性情報を利用した,ネットワーク上を移動するコンテンツを発見するための漂流コンテンツ追跡法を提案し,またその評価をした.その結果,提案する手法は従来の手法と比べ,ネットワークに負荷をかけずに

researchmap

Language：Japanese   Publisher：The Institute of Electronics, Information and Communication Engineers  

Recently, the Internet has attracted the attention as a platform for content distribution or e-commerce. We have proposed Content Based Network (CBN) as a platform which provides functionality, such as effective content distribution, e-commerce and advertisement, for users, content providers and network operators. By the feature of CBN, we have proposed Content Drift System, in which content is able to make copies of itself and move to nodes that have higher concentrations of user requests for this content. However, in this system, it is difficult to discover the content using the location information like URL, which is used in current WWW. In this research, we propose a method of content tracing to discover moving content in the network using its attribute data. We make a simulation to evaluate the proposed method. The results show that the proposed method used network resources more effectively than the conventional method.

CiNii Books

researchmap

Publisher：社団法人電子情報通信学会  

近年,インターネットは商取引やコンテンツ流通のプラットフォームとして注目を集めている.従来,高機能なサービスはアプリケーションレイヤで実現していたが,我々はコンテンツの発見・検索・配信などのサービスをネットワーク側で提供する「コンテンツ指向ネットワーク」(CBN)を提案してきている.またCBNの特徴を生かして,コンテンツを位置情報と分離し,属性情報や流通履歴を利用してネットワーク内を自律的に移動するオペレーションを与えることにより,より効果的なコンテンツ配置を行う「コンテンツ漂流モデル」を提案している.しかし,このネットワーク上ではコンテンツは移動するため,従来のWWWのURLなどによる位置情報を用いたコンテンツの入手は困難である.本研究では,属性情報を利用した,ネットワーク上を移動するコンテンツを発見するための漂流コンテンツ追跡法を提案し,またその評価をした.その結果,提案する手法は従来の手法と比べ,ネットワークに負荷をかけずに

researchmap