Researcher Database

Sugiki Akiyoshi
Information Initiative Center Systems Design
Associate Professor

Researcher Profile and Settings

Affiliation

  • Information Initiative Center Systems Design

Job Title

  • Associate Professor

URL

J-Global ID

Research Areas

  • Informatics / Software

Academic & Professional Experience

  • 2015/04 - Today Hokkaido University Information Initiative Center Associate Professor
  • 2009/04 - 2011/10 University of Tsukuba
  • 2011/10 University of Tsukuba Faculty of Engineering, Information, and Systems
  • 2007/04 - 2009/03 科学技術振興機構 CREST 研究員

Association Memberships

  • IEEE   ACM   Information Processing Society of Japan   USENIX   Institute of Electrical and Electronics Engineers (IEEE)   Association for Computing Machinery   

Research Activities

Published Papers

  • Phyo Thandar Thant, Courtney Powell, Akiyoshi Sugiki, Masaharu Munetomo
    2016 JOINT 8TH INTERNATIONAL CONFERENCE ON SOFT COMPUTING AND INTELLIGENT SYSTEMS (SCIS) AND 17TH INTERNATIONAL SYMPOSIUM ON ADVANCED INTELLIGENT SYSTEMS (ISIS) 293 - 298 2016 [Refereed][Not invited]
     
    Hadoop configuration optimization is very challenging because of the complexity of its framework. And optimized Hadoop parameter configuration settings depend significantly on the performance of MapReduce applications in the cluster. Although much research has been conducted on Hadoop parameters configuration optimization, configuring its resource setting parameters to minimize the execution time of MapReduce jobs in clusters still needs a lot of continuing researches. Further, determining the type of machine instances that should be used to minimize the resource usage cost for executing applications in clusters is also difficult. This paper addresses these problems by optimizing the instance resource usage and execution time of MapReduce tasks using a multi-objective steady-state Non-dominated Sorting Genetic Algorithm II (ssNSGA-II) approach. In this approach, the instance resource usage cost of MapReduce tasks is calculated based on the cost of machine instance types and the number of machine instances in the Hadoop cluster. The optimized configuration is identified by selecting an optimal setting that satisfies two objective functions associated with instance resource usage and execution time minimization, from Pareto optimal front solutions. Although dynamic machine instance type is considered within the search process in our system, dynamic cluster size is out of consideration and intended to be carried out in our future. Experiments conducting using workloads from the HiBench benchmark on a high specification 6-node Hadoop cluster verify the efficacy of our proposed approach.
  • 杉木,章義, 加藤,和彦
    Computer Software 日本ソフトウェア科学会 2013/10 [Refereed][Not invited]
  • 学内無線LANシステムにおける利用統計データの分析とその課題
    杉木, 章義, 佐藤, 聡, 和田, 耕一
    IPSJ SIG Technical Reports 一般社団法人情報処理学会 2013 (7) 1 - 5 0919-6072 2013/09 [Not refereed][Not invited]
     
    教育や研究環境の充実を目的として,多くの大学で学内無線 LAN システムが導入されている.大規模データ解析の手法を学内無線 LAN の統計データに応用すれば,従来の手法では見つけることのできなかった潜在的な問題を発見し,また戦略的なネットワークの配備計画が可能になることから,利用者から見た学内ネットワークの品質を大幅に向上させることができると期待される.本論文では,学内無線 LAN システムに関する利用統計データを蓄積し,対話的に解析を行うためのシステムについて提案する.また,いくつかの解析事例について紹介し,その際に課題となる利用者のプライバシーやセキュリティに関する配慮などについても議論する.
  • Mobile Devices in the Wild (Campus)
    杉木, 章義, 佐藤, 聡, 和田, 耕一
    IPSJ SIG Notes 一般社団法人情報処理学会 2013 (21) 1 - 8 2013/07 [Not refereed][Not invited]
     
    近年,従来からのノート PC に限らず,スマートフォンやタブレットなど,多様な移動端末が利用されている.近い将来,一人のユーザが複数の端末を当たり前のように活用し,生活を行うといったことが指摘されているものの,実際の我々の身近な利用状況について報告されているものはまだ少ない.本研究では,大学のキャンパスで利用されている Wi-Fi 端末の利用状況について,約 2 ヶ月半にわたり,475 台のアクセスポイントを利用して観測を行った.その結果について,本稿で報告する.
  • Measures and Issues of Information Infrastructure for University of Tsukuba at Tohoku Region Pacific Coast Earthquake
    Akira,Sato, Akiyoshi,Sugiki, Hanxiong,Chen, Kazutaka,Furuse, Kazuki,Katagishi, Hisashi,Nakai, Tomohiro,Haraikawa, Atsushi,Maeda, Kouichi,Wada
    IPSJ Journal 一般社団法人情報処理学会 54 (3) 1038 - 1049 0387-5806 2013/03 [Not refereed][Not invited]
     
    University of Tsukuba, which has main campus at Tsukuba City in Ibaraki, was struck by Tohoku Region Pacific Coast Earthquake that occurred on March 11, 2011. Moreover, our university experienced the power failure with the university surrounding area. We selected minimum systems which can run in limited power capacity considering the dependancy between systems, and have operated them. These plan and action have reflected on several disaster plan concerning the information infrastructure. Using analysis of the network system logs before/after the earthquake, we report how the campus from outside after the earthquake that the network was used from inside/outside, and clarify the effectiveness of the measures of information infrastructure in order to continue the announcement.
  • Akiyoshi Sugiki
    19th IEEE International Conference on Networks, ICON 2013, Singapore, December 11-13, 2013 IEEE 1 - 6 2013 [Refereed][Not invited]
  • Maintaining Performance and Consistency in Virtual Resource Management
    Akiyoshi,Sugiki, Kazuhiko,Kato
    情報処理学会論文誌. コンピューティングシステム 一般社団法人情報処理学会 6 (1) 31 - 44 0387-5806 2013/01 [Not refereed][Not invited]
     
    In recent years, cloud computing has been received much attention. Resources in such clouds are managed in a large-scale data center which consists of thousands of machines, network switches, and storages. Virtualization technology at each layer also makes management ever difficult. In this paper, we present a caching technique for such virtual infrastructure. By taking advantage of reference locality, it achieves high efficiency as well as maintaining consistency. Our approach is based on classical caching techniques for distributed objects since resources in the target platform are represented as distributed objects. Consistency is also managed by timeouts and invalidation protocol following the given policy. In the experiments, we have shown that overhead and performance improvement for each operation and demonstrated its effectiveness in virtual machine management.
  • Elements and Composition of Software-defined Data Centers
    Akiyoshi, Sugiki, Kazuhiko, Kato
    4th ACM/IFIP/USENIX Int'l Conf. on Middleware (MIddleware 2012), Demos and Posters Track ACM 2  2012/12 [Refereed][Not invited]
  • Towards Robust Tuning of Hadoop by using Design of Experiments Approach
    Akiyoshi, Sugiki, Kazuhiko, Kato
    18th IEEE Pacific Rim Int'l Symp. on Dependable Computing (PRDC 2012), Fast Abstract 2  2012/11 [Refereed][Not invited]
  • A Scripting Environment for Iterative Tuning Process of Server Applications
    Takuya,Aikawa, Akiyoshi,Sugiki, Kazuhiko,Kato
    情報処理学会論文誌. コンピューティングシステム 一般社団法人情報処理学会 5 (5) 138 - 151 0387-5806 2012/10 [Refereed][Not invited]
     
    Although parameter tuning is critical for server performance, that tuning process is error-prone and time consuming. An administrator must repeat many iterations to find an optimal configuration and even at each step non-trivial tasks, including proper configuration of a server and launching benchmark clients, are required. In this paper, we present a scripting environment for efficiently describing such tuning process. We offer distributed objects as ways to manipulate a server and benchmark clients. By this way, tuning process can be described by scripting them in an integrated environment. We also provide automatic tuning as a library for further efficiency. In the experiments, we confirmed that tuning of an Apache web server under SPECweb2005 benchmark and a Hadoop cluster were successfully possible in our tuning environment.
  • Enhancing Cloud Availability Through Failure-Oblivious Computing
    Akiyoshi,Sugiki, Akihito,Okuhata, Kazuhiko,Kato
    情報処理学会論文誌. コンピューティングシステム 一般社団法人情報処理学会 5 (5) 103 - 117 0387-5806 2012/10 [Refereed][Not invited]
     
    In recent years, cloud computing has drawn much attention from both industry and academia. A cloud is typically hosted in a large-scale data center that must be well operated despite various failures that may cause severe damage. In this paper, we present a technique to increase availability by introducing a failure-oblivious concept into a software cloud platform. As in the original concept, we made the cloud platform continue its operation even when the presence of failures. To confirm the effectiveness of this approach, we have implemented that concept into our cloud platform, Kumoi, and evaluated it in terms of continued execution and influence from side-effects.
  • クラウド基盤ソフトウェアにおけるLineageを利用した障害対策手法の検討
    奥畑, 聡仁, 杉木, 章義, 加藤, 和彦
    IPSJ SIG Notes 一般社団法人情報処理学会 2012 (21) 1 - 8 2012/07 [Not refereed][Not invited]
     
    近年のクラウドコンピューティングの普及に伴い,データセンターの規模が拡大している.これらの多数の計算機資源を管理するデータセンターにおいて,日常的に発生している障害への対策は大きな課題である.本研究では,クラウド環境で典型的に行われるリソースプールから資源を取り出すという操作に着目し,その障害対策を Lineage と呼ばれるモデルを用いて導入する.障害が発生した資源の代わりをクラウド基盤ソフトウェアが自動的に用意し,透過的に障害に対応することを目指す.本研究では,この手法を当研究室で開発しているクラウド基盤ソフトウェアである Kumoi に導入し,いくつかのスクリプトを用いて実験を行い,障害発生時の Lineage の挙動を確認した.
  • A Method for Regional Switching of the Network Booting Environment using Difference of the Latency
    TAKADA,SHINGO, SATO,AKIRA, NAKAI,HISASHI, SUGIKI,AKIYOSHI, SHINJO,YASUSHI
    IEICE technical report. Internet Architecture 一般社団法人電子情報通信学会 111 (485) 65 - 70 2012/03 [Not refereed][Not invited]
     
    Improving the usage rate of the computers is important in organizations that have a lot of computers. The authors propose a method to switch the computer environment which is distributed via the network. The proposed system uses the DHCP's characteristic: first-come-first-served. We inject a DHCP server which can respond to the client faster than the original one. Only netboot clients that receive the injected server's DHCP message faster than the original server can boot the different computer environment. The authors constructed a prototype of the proposed method and made experiments. The results show: (1) the original server's response is slower than the injected one because of the delay of the DHCP relaying, (2) the relaying delay would be reduced if the DHCP requests are concentrated.
  • Shingo Takada, Akira Sato, Yasushi Shinjo, Hisashi Nakai, Akiyoshi Sugiki, Kozo Itano
    2012 THIRD INTERNATIONAL CONFERENCE ON NETWORKING AND COMPUTING (ICNC 2012) 201 - 207 2012 [Refereed][Not invited]
     
    Network-booting is widely adopted in universities that have to maintain many client computers. In conventional network-booting systems, the primary bottleneck is the disk image distribution servers and the network to these servers. To eliminate this bottleneck, peer-to-peer (P2P) methods must work. However, existing P2P methods, including BitTorrent, do not work well for network-booting because they are highly optimized for distributing an entire large file, while network-booting requires certain parts of a large file. In the present study, aiming to solve the problems described above, we describe a new network-booting system that uses a P2P method. In our P2P-based network-booting system, a client node receives disk blocks not only from central distribution servers but also from the other client nodes that already have the demanded disk blocks. To the best of our knowledge, our network-booting system is the first effective implementation of a network-booting system that uses a P2P method in a local area network. Unlike conventional P2P systems, the proposed network-booting system can deal with demanded parts of a large disk image. We performed experiments with 112 client nodes in real classrooms on a university campus. The results of the experiments show that our implementation scales well as the number of client nodes grows.
  • A High-level Scripting Environment for Server Parameter Tuning in the Clouds
    Takuya,Aikawa, Akiyoshi,Sugiki, Kazuhiko,Kato
    IPSJ SIG Notes 一般社団法人情報処理学会 2011 (11) 1 - 8 2011/07 [Not refereed][Not invited]
     
    After the rise of cloud computing, there is a sharp increase in demand for effective management in a data center. Whereas parameter tuning is a most difficult task that requires expertise and knowledge, it strongly affects server performance. In this paper, we present a scripting environment for parameter tuning in the cloud environment. The system offers distributed language objects each of which corresponds to a benchmark components such as an application or a client. This allows high-level scripting in an integrated environment, being unconscious of distribution and reducing many tasks. To support a variety of applications, we also provide a compiler which generates distributed objects from Interface definitions. Our experiments with Apache under SPECweb2005 and Hadoop, and the results show that parameter tuning is successfully possible in our scripting environment.
  • Akiyoshi Sugiki, Kazuhiko Kato
    Proceedings - 2011 4th IEEE International Conference on Utility and Cloud Computing, UCC 2011 306 - 311 2011 [Refereed][Not invited]
     
    Virtualization has changed the ways computation is done, especially in utility and cloud computing. While several virtualization cloud platforms have emerged to provide interfaces between users and data centers, traditional operating systems (OSes) have defined interfaces between applications and hardware. We argue that virtualization cloud platforms and OSes can share many problems, thus, many techniques developed for OSes can be applied to such platforms. We have implemented a prototype of an OS-like middleware that adopted a micro kernel design, had resource abstraction achieved by objects and functions, and offered a scripting environment as a programming interface. The experimental results demonstrated that it had the ability to achieve elasticity and high-availability VMs. We also carried out qualitative comparisons with other middleware to clarify our research position. © 2011 IEEE.
  • Koji Hasebe, Kei Yamatozaki, Akiyoshi Sugiki, Kazuhiko Kato
    2011 4th IFIP International Conference on New Technologies, Mobility and Security, NTMS 2011 - Proceedings 1 - 6 2011 [Refereed][Not invited]
     
    We present a self-stabilizing passive replication mechanism for Internet service platforms. The proposed system based on this mechanism aims to provide services despite simultaneous failure of the majority of nodes. To achieve this objective, we use the self-stabilizing consensus algorithm introduced by Dolev et al. [7], instead of a majority-based consensus algorithm such as Paxos. Although our system allows temporary illegal behavior, it eventually converges on the desired state without interruption of services. Thus, the proposed system is useful for providing services that require high availability, but not strict consistency, a prime example of which is an Internet bulletin board for communication during large-scale natural disasters. We implemented a prototype and evaluated it with experiments to demonstrate availability through various patterns of failures.
  • Self-Stabilizing Passive Replication for Internet Service Platforms
    Kei,Yamatozaki, Koji,Hasebe, Akiyoshi,Sugiki, Kazuhiko,Kato
    IPSJ SIG Notes 一般社団法人情報処理学会 2011 (11) 1 - 8 2011/01 [Not refereed][Not invited]
  • Akiyoshi Sugiki, Kazuhiko Kato
    Proceedings - 2011 4th IEEE International Conference on Utility and Cloud Computing, UCC 2011 306 - 311 2011 [Refereed][Not invited]
     
    Virtualization has changed the ways computation is done, especially in utility and cloud computing. While several virtualization cloud platforms have emerged to provide interfaces between users and data centers, traditional operating systems (OSes) have defined interfaces between applications and hardware. We argue that virtualization cloud platforms and OSes can share many problems, thus, many techniques developed for OSes can be applied to such platforms. We have implemented a prototype of an OS-like middleware that adopted a micro kernel design, had resource abstraction achieved by objects and functions, and offered a scripting environment as a programming interface. The experimental results demonstrated that it had the ability to achieve elasticity and high-availability VMs. We also carried out qualitative comparisons with other middleware to clarify our research position. © 2011 IEEE.
  • Munetoshi Ishikawa, Koji Hasebe, Akiyoshi Sugiki, Kazuhiko Kato
    Service Oriented Computing and Applications 4 (4) 245 - 260 1863-2386 2010/12 [Refereed][Not invited]
     
    Power-saving has become a central issue for well-configured SOC platforms. In particular, as a high percentage of the total energy is used by the storage systems, the cost effectiveness of data management is equally as important as reliability and availability. To address this issue, we propose the dynamic grid quorum as a method for reducing the power consumption of large-scale distributed storage systems. The basic principle of our approach is to skew the workload toward a small number of quorums. This can be realized using the following three techniques. First, our system allows reconfiguration by exchanging nodes without any data migration, so that high-capacity nodes can be reallocated to busier quorums. Second, for more effective skewing of the workload, we introduce the notion of dual allocation, which makes it possible to consider two distinct allocations in the same grid for write and read quorums. Finally, we present an optimization algorithm to find a pair of a strategy and an allocation of nodes, which minimizes power for a given system setting and its workload. We also demonstrate that the dynamic grid quorum saves, on average, 14-25% energy compared with static configurations, when the intensity of the total workload changes. © 2010 Springer-Verlag London Limited.
  • Akiyoshi Sugiki, Kazuhiko Kato, Yoshiaki Ishii, Hiroki Taniguchi, Nobuyuki Hirooka
    Proceedings of the International Conference on Parallel and Distributed Systems - ICPADS 322 - 329 1521-9097 2010 [Refereed][Not invited]
     
    We have designed and implemented a scripting environment called "Kumoi" for managing collective VMs in a large-scale data center. Kumoi is unlike other scripting environments because it exploits strong typing with type inference and high-level description. Kumoi introduces several advancements, including treating virtual machines as first-class objects and decoupling the scripting model and its execution for hiding as many details as possible. We implemented Kumoi as an embedded domain-specific language based on Scala along with distributed agents running on each physical machine. Evaluation using example scripts showed that an administrator can more concisely write the instructions for performing complex VM lifecycle management tasks. Use of this environment should improve management efficiency and agility. © 2010 IEEE.
  • Koji Hasebe, Tatsuya Niwa, Akiyoshi Sugiki, Kazuhiko Kato
    Proceedings - 2nd IEEE International Conference on Cloud Computing Technology and Science, CloudCom 2010 266 - 273 2010 [Refereed][Not invited]
     
    We present a power-saving method for large-scale distributed storage systems. The key idea is to use virtual nodes and migrate them dynamically so as to skew the workload towards a small number of disks while not overloading them. Our proposed method consists of two kinds of algorithms, one for gathering or spreading virtual nodes according to the daily variation of workloads so that the active disks are reduced to a minimum, the other for coping with the changes in the popularity of data over a longer period. For this dynamic migration, data stored in virtual nodes are managed by a distributed hash table. Furthermore, to improve the reliability as well as to reduce the migration cost, we also propose an extension of our method by introducing a replication mechanism. The performance of our method is measured both by simulation and a prototype implementation. From the experiments, we observed that our method skews the workload so that the average load for the active physical nodes as a function of the overall capacity is 67%. At the same time, we maintain a preferred response time by setting a suitable maximum workload for each physical node. © 2010 IEEE.
  • Akiyoshi Sugiki, Kazuhiko Kato, Yoshiaki Ishii, Hiroki Taniguchi, Nobuyuki Hirooka
    Proceedings of the International Conference on Parallel and Distributed Systems - ICPADS 322 - 329 1521-9097 2010 [Refereed][Not invited]
     
    We have designed and implemented a scripting environment called "Kumoi" for managing collective VMs in a large-scale data center. Kumoi is unlike other scripting environments because it exploits strong typing with type inference and high-level description. Kumoi introduces several advancements, including treating virtual machines as first-class objects and decoupling the scripting model and its execution for hiding as many details as possible. We implemented Kumoi as an embedded domain-specific language based on Scala along with distributed agents running on each physical machine. Evaluation using example scripts showed that an administrator can more concisely write the instructions for performing complex VM lifecycle management tasks. Use of this environment should improve management efficiency and agility. © 2010 IEEE.
  • A Cooperative Virtual Machine-based Service Replication Platform for WAN Environments
    Akiyoshi,Sugiki, Kei,Yamatozaki, Kazuhiko,Kato
    情報処理学会論文誌. コンピューティングシステム 一般社団法人情報処理学会 2 (1) 1 - 11 0387-5806 2009/03 [Refereed][Not invited]
     
    Recent Internet services have been developed at extremely limited time and resources. Although service availability is highly demanded, it is generally difficult to achieve high availability within short developing time. Remote replication is a potentially demanded solution to protect a service from whole site down, but it is hardly adopted from small service providers due to unacceptable costs. In this paper, we present a cooperative virtual machine-based replication platform for wide-area environments. We provide a cost-effective, highlyavailable solution via service state replication over peer-to-peer platform. Virtual machines increase reliability over such platform by isolating each other and capturing complete service state. We have implemented a prototype supporting two virtual machines, Xen and Scrap Book for User-Mode Linux. Our experimental results show that the prototype successfully replicated two services on three sites in Japan and recovered them in case of failures.
  • Munetoshi Ishikawa, Koji Hasebe, Akiyoshi Sugiki, Kazuhiko Kato
    IEEE International Conference on Service-Oriented Computing and Applications, SOCA 2009, 14-15 December 2009, Taipei, Taiwan IEEE Computer Society 1 - 8 2009 [Refereed][Not invited]
  • 耐障害性向上のための自己安定性に基づくインターネットサービス基盤の構築
    大和崎啓, 長谷部浩二, 杉木章義, 加藤和彦
    コンピュータソフトウェア 日本ソフトウェア科学会 2009/01 [Refereed][Not invited]
  • Akiyoshi Sugiki, Kenji Kono, Hideya Iwasaki
    SOFTWARE-PRACTICE & EXPERIENCE 38 (12) 1215 - 1240 0038-0644 2008/10 [Refereed][Not invited]
     
    Apache web servers are widely used as stand-alone servers or front-ends in multi-tiered web servers. Despite the wide availability of software, it is quite difficult for many administrators to properly configure their web servers. In particular, setting the performance-related parameters is an error-prone and time-consuming task because their values heavily depend on the server environment. In this paper, two mechanisms are described for automatically tuning two performance-related parameters of Apache web servers: KeepAliveTimeout and MaxClients. These mechanisms are easy to deploy because no modifications to the server or the operating system are required. Moreover, they are parameter specific. Although interference between KeepAliveTimeout and MaxClients is inevitable, the tuning mechanisms minimize the correlation by using almost completely independent metrics. Experimental results show that these mechanisms work well for two different workloads; the parameter values are close to optimal and can adapt to workload changes. Copyright (C) 2007 John Wiley & Sons, Ltd.
  • Akiyoshi Sugiki, Kenji Kono, Hideya Iwasaki
    SOFTWARE-PRACTICE & EXPERIENCE 38 (12) 1215 - 1240 0038-0644 2008/10 [Not refereed][Not invited]
     
    Apache web servers are widely used as stand-alone servers or front-ends in multi-tiered web servers. Despite the wide availability of software, it is quite difficult for many administrators to properly configure their web servers. In particular, setting the performance-related parameters is an error-prone and time-consuming task because their values heavily depend on the server environment. In this paper, two mechanisms are described for automatically tuning two performance-related parameters of Apache web servers: KeepAliveTimeout and MaxClients. These mechanisms are easy to deploy because no modifications to the server or the operating system are required. Moreover, they are parameter specific. Although interference between KeepAliveTimeout and MaxClients is inevitable, the tuning mechanisms minimize the correlation by using almost completely independent metrics. Experimental results show that these mechanisms work well for two different workloads; the parameter values are close to optimal and can adapt to workload changes. Copyright (C) 2007 John Wiley & Sons, Ltd.
  • Akiyoshi Sugiki, Kei Yamatozaki, Richard Potter, Kazuhiko Kato
    International Service Availability Symposium 2008 Lecture Notes in Computer Science (LNCS) 5017 129-141  2008/05 [Refereed][Not invited]
  • Finding Major Parameters of a Web Server with Design of Experiments
    Sugiki,Akiyoshi, Kato,Kazuhiko
    IPSJ SIG Notes 一般社団法人情報処理学会 2008 (35) 33 - 40 0919-6072 2008/04 [Not refereed][Not invited]
     
    The performance of web servers is crucial for today's continual business process. Therefore, the proper management of performance is a major issue in web servers. Current web servers have a large number of parameters, making it impractical to tune all of them manually. In this paper, we present a technique based on design of experiments for finding parameters that influence server performance the most. We call these parameters major parameters. Even with the classical two-level orthogonal design, it can successfully find major parameters with a small number of experiments. We applied this technique to Apache web server under three work-loads of SPECweb2005. Experimental results show that our technique can select 6-9 major parameters out of 24 parameters with 32 experiments.
  • Akiyoshi Sugiki, Kei Yamatozaki,Richard Potter, Kazuhiko Kato
    Service Availability, 5th International Service Availability Symposium, ISAS 2008, Tokyo, Japan, May 19-21, 2008, Proceedings Springer 129 - 141 2008 [Refereed][Not invited]
  • Reliable Passive Standby of Virtual Machines using a Generic Snapshot Mechanism
    Sugiki,Akiyoshi, Potter,Richard, Kato,Kazuhiko
    IPSJ SIG Notes 一般社団法人情報処理学会 2008 (9) 79 - 86 0919-6072 2008/01 [Not refereed][Not invited]
     
    The use of virtualization technology on commodity PCs has become popular recently. Strong encapsulation and migration functionalities of virtualization make it easy to achieve disaster-tolerant systems. This paper presents a cheap and service-independent state replication technique based on virtual machines. Our approach has two notable features. First, it is based on a well-known Paxos consensus algorithm and we use it for passive standby. Second, our mechanism reduces network transfer cost and recovery time by using incremental snapshots and their merge mechanism. We implmented our prototype in Java with SBUML virtual machines. Our expriments show that our mechanism can sufficiently recover from failures for two standard benchmark workloads.
  • 杉木章義, 河野健二, 岩崎英哉
    コンピュータソフトウェア 日本ソフトウェア科学会 24 (2) 68-78  2007/01 [Refereed][Not invited]
  • A Sugiki, K Kono, H Iwasaki
    ADVANCES IN COMPUTER SCIENCE - ASIAN 2005, PROCEEDINGS 3818 146 - 159 0302-9743 2005 [Refereed][Not invited]
     
    This paper presents a practical approach to automatically tuning the parameters of the Apache Web server. In particular, two significant parameters, KeepAliveTimeout and MaxClients, are dealt with. The notable features of our approach are twofold. First, it is easy to deploy because no modifications to Apache or the underlying operating system are required. Second, our approach is based on the detailed analysis on how each parameter affects the server's behavior. Experimental results demonstrate that our prototype works well on different workloads; it can discover almost optimal values and quickly adapt to workload changes.
  • Akiyoshi Sugiki, Kenji Kono, Hideya Iwasaki
    Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics) 3818 146 - 159 0302-9743 2005 [Refereed][Not invited]
     
    This paper presents a practical approach to automatically tuning the parameters of the Apache Web server. In particular, two significant parameters, KeepAliveTimeout and MaxClients, are dealt with. The notable features of our approach are twofold. First, it is easy to deploy because no modifications to Apache or the underlying operating system are required. Second, our approach is based on the detailed analysis on how each parameter affects the server's behavior. Experimental results demonstrate that our prototype works well on different workloads it can discover almost optimal values and quickly adapt to workload changes. © Springer-Verlag Berlin Heidelberg 2005.

MISC

  • 杉木章義, 岩井良成  情報処理学会研究報告(Web)  2017-  (OS-141)  Vol.2017‐OS‐141,No.17,1‐7 (WEB ONLY)  2017/07/19  [Not refereed][Not invited]
  • 岩井良成, 杉木章義, 棟朝雅晴  情報処理学会研究報告(Web)  2017-  (OS-140)  Vol.2017‐OS‐140,No.13,1‐6 (WEB ONLY)  2017/05/09  [Not refereed][Not invited]
  • 市居 遼平, 棟朝 雅晴, 杉木 章義  電気学会研究会資料. ITS  2016-  (1)  83  -88  2016/03/10  [Not refereed][Not invited]
  • 市居 遼平, 棟朝 雅晴, 杉木 章義  電子情報通信学会技術研究報告 = IEICE technical report : 信学技報  115-  (504)  83  -88  2016/03/10  [Not refereed][Not invited]
  • 市居遼平, 棟朝雅晴, 杉木章義  電子情報通信学会技術研究報告  115-  (504(ITS2015 84-99))  83‐88  2016/03/03  [Not refereed][Not invited]
  • 杉木章義, 佐藤聡  情報処理学会研究報告(Web)  2014-  (OS-130)  VOL.2014-OS-130,NO.19 (WEB ONLY)  2014/07/21  [Not refereed][Not invited]
  • 杉木章義, 佐藤聡  研究報告システムソフトウェアとオペレーティング・システム(OS)  2014-  (19)  1  -6  2014/07/21  [Not refereed][Not invited]
     
    本論文では,Intel MIC アーキテクチャや CPU/GPU 統合プロセッサに代表される異種混在型アーキテクチャ上でのネットワークパケット処理について検討する.元々は GPGPU 用に記述された Longest Prefix Match (LPM),Packet Classifier,Aho-Corasick の 3 つのアルゴリズムを OpenCL に移植し,Xeon Phi 上で予備評価を行った結果について報告する.また,汎用マルチコア CPU 上で処理を行った場合との性能比較を行い,今後のアプローチの有効性について検討する.
  • 杉木 章義, 佐藤 聡, 和田 耕一  電子情報通信学会技術研究報告 = IEICE technical report : 信学技報  113-  (443)  171  -177  2014/02/27  [Not refereed][Not invited]
     
    筑波大学では,Microsoft EESライセンスを2013年3月より締結し,導入に向けた作業を行った.EESライセンスは国内の数多くの大学でも既に導入されているが,本学での状況に合わせて運用をカスタマイズし,学内ダウンロードによるソフトウェア配布とKMSライセンス認証を原則とした運用方針を採用した.本研究では,MS EESライセンスの導入を行った2013年3月から12月にかけての9ケ月以上にわたるKMSライセンス認証ログを蓄積し,その分析から,学内でのMS EESライセンスの活用状況やその問題点について報告する.
  • 杉木 章義, 佐藤 聡, 和田 耕一  電子情報通信学会技術研究報告 = IEICE technical report : 信学技報  113-  (442)  171  -177  2014/02/27  [Not refereed][Not invited]
     
    筑波大学では,Microsoft EESライセンスを2013年3月より締結し,導入に向けた作業を行った.EESライセンスは国内の数多くの大学でも既に導入されているが,本学での状況に合わせて運用をカスタマイズし,学内ダウンロードによるソフトウェア配布とKMSライセンス認証を原則とした運用方針を採用した.本研究では,MS EESライセンスの導入を行った2013年3月から12月にかけての9ケ月以上にわたるKMSライセンス認証ログを蓄積し,その分析から,学内でのMS EESライセンスの活用状況やその問題点について報告する.
  • 杉木章義, 佐藤聡, 和田耕一  電子情報通信学会技術研究報告  113-  (442(SITE2013 58-74))  171  -177  2014/02/20  [Not refereed][Not invited]
  • 杉木章義, 佐藤聡, 和田耕一  研究報告インターネットと運用技術(IOT)  2014-  (30)  1  -7  2014/02/20  [Not refereed][Not invited]
     
    筑波大学では,Microsoft EES ライセンスを 2013 年 3 月より締結し,導入に向けた作業を行った.EES ライセンスは国内の数多くの大学でも既に導入されているが,本学での状況に合わせて運用をカスタマイズし,学内ダウンロードによるソフトウェア配布と KMS ライセンス認証を原則とした運用方針を採用した.本研究では,MS EES ライセンスの導入を行った 2013 年 3 月から 12 月にかけての 9 ヶ月以上にわたる KMS ライセンス認証ログを蓄積し,その分析から,学内での MS EES ライセンスの活用状況やその問題点について報告する.
  • 杉木章義, 佐藤聡, 和田耕一  情報処理学会研究報告(Web)  2013-  (IOT-23)  VOL.2013-IOT-23,NO.7 (WEB ONLY)  2013/09/20  [Not refereed][Not invited]
  • 杉木章義, 佐藤聡, 和田耕一  情報処理学会研究報告(Web)  2013-  (OS-126)  WEB ONLY VOL.2013-OS-126,NO.21  2013/07/24  [Not refereed][Not invited]
  • 杉木章義, 加藤和彦  情報処理学会論文誌トランザクション(CD-ROM)  2012-  (2)  ROMBUNNO.KONPYUTINGUSHISUTEMU,VOL.6,NO.1,31-44  2013/04/15  [Not refereed][Not invited]
  • 杉木章義, 奥畑聡仁, 加藤和彦  情報処理学会論文誌トランザクション(CD-ROM)  2012-  (2)  ROMBUNNO.KONPYUTINGUSHISUTEMU,VOL.5,NO.5,103-117  2013/04/15  [Not refereed][Not invited]
  • 相川拓也, 杉木章義, 加藤和彦  情報処理学会論文誌トランザクション(CD-ROM)  2012-  (2)  ROMBUNNO.KONPYUTINGUSHISUTEMU,VOL.5,NO.5,138-151  2013/04/15  [Not refereed][Not invited]
  • 佐藤聡, 杉木章義, CHEN Hanxiong, 古瀬一隆, 片岸一起, 中井央, 秡川友宏, 前田敦司, 和田耕一  情報処理学会論文誌ジャーナル(CD-ROM)  54-  (3)  1038  -1049  2013/03/15  [Not refereed][Not invited]
  • 杉木章義, 加藤和彦  コンピュータソフトウェア  30-  (4)  4.3-4.17 (J-STAGE)  2013  [Not refereed][Not invited]
  • 奥畑聡仁, 杉木章義, 加藤和彦  情報処理学会研究報告(CD-ROM)  2012-  (3)  ROMBUNNO.OS-122,NO.21  2012/10/15  [Not refereed][Not invited]
  • 杉木 章義, 奥畑 聡仁, 加藤 和彦  先進的計算基盤システムシンポジウム論文集  2012-  168  -175  2012/05/09  [Not refereed][Not invited]
  • 高田真吾, 佐藤聡, 中井央, 杉木章義, 新城靖  電子情報通信学会技術研究報告  111-  (484(SITE2011 32-55))  65  -70  2012/03/08  [Not refereed][Not invited]
  • 廣岡誠之, 石井嘉明, 矢野恭平, 杉木章義, 加藤和彦  情報処理学会全国大会講演論文集  74th-  (1)  1.3-1.4  2012/03/06  [Not refereed][Not invited]
  • 矢野恭平, 石井嘉明, 廣岡誠之, 杉木章義, 加藤和彦  情報処理学会全国大会講演論文集  74th-  (1)  1.5-1.6  2012/03/06  [Not refereed][Not invited]
  • 石井嘉明, 矢野恭平, 廣岡誠之, 杉木章義, 加藤和彦  情報処理学会全国大会講演論文集  74th-  (1)  1.1-1.2  2012/03/06  [Not refereed][Not invited]
  • 矢野恭平, 石井嘉明, 廣岡誠之, 杉木章義, 加藤和彦  全国大会講演論文集  2012-  (1)  5  -7  2012/03/06  [Not refereed][Not invited]
     
    2011年に発生した東日本大震災では,サーバへのアクセス集中などにより,地方自治体や公共交通機関において,情報伝達に不具合が発生した.これらの機関では,大規模災害時においても継続した情報サービスの提供が求められ,システムには高い可用性が必要となる.本研究では,アクセス負荷に対して自律的にスケールアウトするミラーサイトシステムをクラウドコンピューティング基盤Kumoiを機能拡張することにより開発した.本システムは,複数のデータセンターを用いた広域分散環境を活用し,データセンター間でデータを同期することで,データセンター障害にも対応している.
  • 廣岡誠之, 石井嘉明, 矢野恭平, 杉木章義, 加藤和彦  全国大会講演論文集  2012-  (1)  3  -5  2012/03/06  [Not refereed][Not invited]
     
    近年,クラウドコンピューティングによるサービスが利用されるようになり,インターネットを通して安定したサービスの提供が求めれられている.しかし,クラウドコンピューティングを支える技術はサーバ環境を基本としており,クライアント環境やネットワーク環境についてはあまり考慮されていない.本研究では,クライアント環境,ネットワーク環境,サーバ環境を統合し,ディペンダブルなクラウドコンピューティング環境を構築した.本環境の性能検証をおこなった結果,ネットワーク障害やサーバ障害に対して自動的に復旧し,クライアントから継続的にサービスが利用できることを確認した.
  • 石井嘉明, 矢野恭平, 廣岡誠之, 杉木章義, 加藤和彦  全国大会講演論文集  2012-  (1)  1  -3  2012/03/06  [Not refereed][Not invited]
     
    近年,クラウドコンピューティングの普及が進んでいる.クラウドコンピューティング環境では,運用効率を高めるために物理計算機を仮想化する場合が多い.仮想化技術に基づくHAやスケーラビリティといった技術は,サービスの継続性を確保する上で重要である.しかし,これらの技術はシステム全体を復旧するものではない.本研究では,クラウドコンピューティング基盤Kumoiを用い,物理・仮想計算機障害に対して,アプリケーションを含めたシステム全体を自動復旧するクラウドシステムを開発した.また,リソース負荷を予見化し,スケールアウトする仮想計算機を事前に作成,起動することで,シームレスなスケーラビリティも実現した.
  • 矢野恭平, 石井嘉明, 廣岡誠之, 杉木章義, 加藤和彦  第74回全国大会講演論文集  2012-  (1)  5  -6  2012/03/06  [Not refereed][Not invited]
     
    2011年に発生した東日本大震災では,サーバへのアクセス集中などにより,地方自治体や公共交通機関において,情報伝達に不具合が発生した.これらの機関では,大規模災害時においても継続した情報サービスの提供が求められ,システムには高い可用性が必要となる.本研究では,アクセス負荷に対して自律的にスケールアウトするミラーサイトシステムをクラウドコンピューティング基盤Kumoiを機能拡張することにより開発した.本システムは,複数のデータセンターを用いた広域分散環境を活用し,データセンター間でデータを同期することで,データセンター障害にも対応している.
  • 廣岡誠之, 石井嘉明, 矢野恭平, 杉木章義, 加藤和彦  第74回全国大会講演論文集  2012-  (1)  3  -4  2012/03/06  [Not refereed][Not invited]
     
    近年,クラウドコンピューティングによるサービスが利用されるようになり,インターネットを通して安定したサービスの提供が求めれられている.しかし,クラウドコンピューティングを支える技術はサーバ環境を基本としており,クライアント環境やネットワーク環境についてはあまり考慮されていない.本研究では,クライアント環境,ネットワーク環境,サーバ環境を統合し,ディペンダブルなクラウドコンピューティング環境を構築した.本環境の性能検証をおこなった結果,ネットワーク障害やサーバ障害に対して自動的に復旧し,クライアントから継続的にサービスが利用できることを確認した.
  • 石井嘉明, 矢野恭平, 廣岡誠之, 杉木章義, 加藤和彦  第74回全国大会講演論文集  2012-  (1)  1  -2  2012/03/06  [Not refereed][Not invited]
     
    近年,クラウドコンピューティングの普及が進んでいる.クラウドコンピューティング環境では,運用効率を高めるために物理計算機を仮想化する場合が多い.仮想化技術に基づくHAやスケーラビリティといった技術は,サービスの継続性を確保する上で重要である.しかし,これらの技術はシステム全体を復旧するものではない.本研究では,クラウドコンピューティング基盤Kumoiを用い,物理・仮想計算機障害に対して,アプリケーションを含めたシステム全体を自動復旧するクラウドシステムを開発した.また,リソース負荷を予見化し,スケールアウトする仮想計算機を事前に作成,起動することで,シームレスなスケーラビリティも実現した.
  • 石井嘉明, 矢野恭平, 廣岡誠之, 杉木章義, 加藤和彦  情報処理学会研究報告(CD-ROM)  2011-  (5)  ROMBUNNO.CLE-6,NO.4  2012/02/15  [Not refereed][Not invited]
  • 相川拓也, 杉木章義, 加藤和彦  日本ソフトウエア科学会大会講演論文集(CD-ROM)  29th-  ROMBUNNO.DP-1  2012  [Not refereed][Not invited]
  • 石井 嘉明, 矢野 恭平, 廣岡 誠之, 杉木 章義, 加藤 和彦  研究報告教育学習支援情報システム(CLE)  2011-  (4)  1  -7  2011/11/24  [Not refereed][Not invited]
  • 相川拓也, 杉木章義, 加藤和彦  情報処理学会研究報告(CD-ROM)  2011-  (2)  ROMBUNNO.OS-118,NO.11  2011/08/15  [Not refereed][Not invited]
  • 大和崎啓, 長谷部浩二, 杉木章義, 加藤和彦  情報処理学会研究報告(CD-ROM)  2010-  (5)  ROMBUNNO.OS-116,NO.11  2011/02/15  [Not refereed][Not invited]
  • 大和崎啓, 長谷部浩二, 杉木章義, 加藤和彦  コンピュータソフトウェア  28-  (1)  1.248-1.257 (J-STAGE)  2011  [Not refereed][Not invited]
  • 丹羽 達也, 長谷部 浩二, 杉木 章義  日本ソフトウェア科学会大会論文集  27-  1  -9  2010/09/13  [Not refereed][Not invited]
  • 丹羽達也, 長谷部浩二, 杉木章義, 加藤和彦  日本ソフトウエア科学会大会講演論文集(CD-ROM)  27th-  ROMBUNNO.5B-2  2010  [Not refereed][Not invited]
  • 杉木章義, 加藤和彦  情報処理学会シンポジウム論文集  2009-  (13)  75  -82  2009/11/19  [Not refereed][Not invited]
  • 杉木章義, 大和崎啓, 加藤和彦  情報処理学会論文誌トランザクション(CD-ROM)  2008-  (2)  VOL.2NO.1,1-11  2009/04/15  [Not refereed][Not invited]
  • 丹羽達也, 長谷部浩二, 杉木章義, 加藤和彦  日本ソフトウエア科学会大会講演論文集(CD-ROM)  26th-  ROMBUNNO.4B-3  2009  [Not refereed][Not invited]
  • 大和崎啓, 長谷部浩二, 杉木章義, 加藤和彦  日本ソフトウエア科学会大会講演論文集(CD-ROM)  26th-  ROMBUNNO.3B-3  2009  [Not refereed][Not invited]
  • 杉木章義, 大和崎啓, 加藤和彦  情報処理学会シンポジウム論文集  2008-  (12)  3  -12  2008/11/05  [Not refereed][Not invited]
  • 石川宗寿, 長谷部浩二, 杉木章義, 品川高廣, 加藤和彦  電子情報通信学会技術研究報告  108-  (248(DC2008 22-27))  17  -22  2008/10/13  [Not refereed][Not invited]
  • ISHIKAWA Munetoshi, SUGIKI Akiyoshi, SHINAGAWA Takahiro, KATO Kazuhiko  IPSJ SIG Notes  2008-  (35)  107  -113  2008/04/24  [Not refereed][Not invited]
     
    A growing number of Internet services are provided through geographically distributed servers such as edge servers or clusters. In such environments, a certain level of replica consistency must be preserved. A Quorum-based protocol is a kind of replication protocol which can dynamically change the number of replicas that clients write to or read from depending on various circumstances. By changing the number of replicas, each service can satisfy a specific requirements for service availability and latency. In this paper, we present a flexible, latency-aware optimization method for quorum-based protocol. This paper specifically deals with the details of our protocol and discusses basic assumptions and future directions.
  • YAMATOZAKI Kei, SUGIKI Akiyoshi, KATO Kazuhiko  IPSJ SIG Notes  2008-  (35)  1  -8  2008/04/23  [Not refereed][Not invited]
     
    Service availability is crucial for today's Internet services. Traditionally, replicating service state on multiple physical hosts is widely used for high availability. But, preparing multiple machines in geographically diverse environment is expensive, especially for small Internet services. In this paper, we present a virtual machine-based cooperative platform for service backup. The advantage of our platform is twofold: First, by enabling many Internet services to back up their services cooperatively, the platform provides high availability with minimum cost. Second, by using virtual machines, the platform encapsulates the complete states of the services. Furthermore services can share a physical host while maintaining a high level of isolation. We have implemented a prototype of our platform and evaluated it by experiments. The results of our experiments show that participating hosts can back up each others, and that services can recover from host failures.
  • 石川宗寿, 杉木章義, 品川高廣, 加藤和彦  情報処理学会研究報告  2008-  (35(OS-108))  107  -113  2008/04/16  [Not refereed][Not invited]
  • 大和崎啓, 杉木章義, 加藤和彦  情報処理学会研究報告  2008-  (35(OS-108))  1  -8  2008/04/16  [Not refereed][Not invited]
  • 杉木章義, 加藤和彦  情報処理学会研究報告  2008-  (35(OS-108))  33  -40  2008/04/16  [Not refereed][Not invited]
  • 池嶋俊, 品川高廣, 杉木章義, 加藤和彦  情報処理学会研究報告  2008-  (9(OS-107))  25  -32  2008/01/30  [Not refereed][Not invited]
  • 杉木章義, POTTER Richard, 加藤和彦  情報処理学会研究報告  2008-  (9(OS-107))  79  -86  2008/01/30  [Not refereed][Not invited]
  • IKEJIMA Syun, SHINAGAWA Takahiro, SUGIKI Akiyoshi, KATO Kazuhiko  IPSJ SIG Notes  2008-  (9)  25  -32  2008/01/30  [Not refereed][Not invited]
     
    The number and types of Internet services are consistantly growing. Due to unpredictable network outages, it is difficult to make these services highly available. Currently, few highly-available network services exist because the creation of these services require a high development cost. We propose making Internet services highly available through distributed objects. With our system, if a network outage occurs, our library will ensure the concurrency of the distributed objects. Thus, the service author does not need to be concerned about micro-managing objects.
  • 杉木章義, 河野健二, 岩崎英哉  コンピュータソフトウェア  24-  (2)  68  -78  2007/04/24  [Not refereed][Not invited]
  • Akiyoshi Sugiki, Kenji Kono, Hideya Iwasaki  Computer Software  24-  (2)  68  -78  2007  [Not refereed][Not invited]
     
    Manual parameter-tuning of Internet servers causes high administrative costs because it requires administrator's expertise and huge amounts of time. The keep-alive parameter, which is one of major parameters in web servers, may cause severe degradation if it is not set properly. In this paper, we present an automatic tuning technique of the keep-alive parameter. Our mechanism adjusts the parameter without administrator's intervention so as to maintain active connections between clients and a server while closing inactive connections by observing request-waiting intervals. We implemented a prototype for Apache web server. Experimental results show that our prototype automatically adjusted the parameter and successfully yielded the nearly optimal server performance on two different workloads. © 2007, Japan Society for Software Science and Technology. All rights reserved.
  • SUGIKI Akiyoshi, KONO Kenji  IPSJ SIG Notes  2006-  (44)  37  -44  2006/05/13  [Not refereed][Not invited]
     
    Automatic parameter-tuning of Internet servers is demanded to reduce high administrative costs by manual tuning. In this paper, we present a tuning mechanism for the maximum number of connections of a web server. Our mechanism is based on the detection of server latency deviation because most servers show unstable behavior when it is overloaded. Experimental results show that our prototype achieved high throughput while reducing server latency on two different workloads.
  • 杉木章義, 河野健二  情報処理学会研究報告  2006-  (44(OS-102))  37  -44  2006/05/12  [Not refereed][Not invited]
  • 杉木章義, 河野健二  情報処理学会研究報告  2005-  (48(OS-99))  29  -36  2005/05/25  [Not refereed][Not invited]
  • SUGIKI Akiyoshi, KONO Kenji  IPSJ SIG Notes  2005-  (48)  29  -36  2005/05/25  [Not refereed][Not invited]
     
    Tuning performance parameters of Internet servers is a hard and time-consuming task for administrators. This paper presents an automatic mechanism of tuning the keep-alive timeout, a performance parameter of Web servers. If it is set improperly, it may cause throughput degradation due to poor usage of connections between clients and a server. To yield higher efficiency of connections, our mechanism tries to keep active connections linked and cut off poorly used connections by adjusting the keep-alive timeout. We have implemented a prototype of this mechanism for the Apache Web server. Experimental results show that our mechanism works well on two different workloads.
  • 杉木章義, 河野健二, 岩崎英哉  日本ソフトウエア科学会大会講演論文集(CD-ROM)  22nd-  7C-2  2005  [Not refereed][Not invited]
  • 杉木 章義, 河野 健二, 岩崎 英哉  日本ソフトウェア科学会大会講演論文集  22-  (0)  618  -627  2005  [Not refereed][Not invited]
     
    インターネット・サーバの性能パラメータ調整は,多くの時間や経験を要する作業である.Web サーバのkeep-alive 時間は,適切に調整しない場合,TCP 接続の使用率を下げスループットを低下させる.本発表では,データを送受信しない接続を切断するように keep-alive 時間を自動的にかつ適切に調整する機構を提案する.本機構を Apache を対象として実装した結果,異なる2 つの負荷に対して適切にkeep-alive 時間を設定し,サーバの性能を向上させることを確認した.
  • 杉木章義, 河野健二, 岩崎英哉, 益田隆司  日本ソフトウエア科学会大会論文集  20th-  (Pt.1)  230  -234  2003/09/16  [Not refereed][Not invited]
  • 杉木章義, 河野健二, 岩崎英哉, 益田隆司  情報処理学会シンポジウム論文集  2003-  (8)  267  -274  2003/05/28  [Not refereed][Not invited]
  • Sugiki Akiyoshi, Kono Kenji, Iwasaki Hideya, Masuda Takashi  Conference Proceedings of Japan Society for Software Science and Technology  2003-  (0)  11  -11  2003  [Not refereed][Not invited]
     
    インターネットサーバでは,提供するサービスに対する高いディペンダビリティが求められている.そのため,可用性・信頼性・管理容易性等,お互いにトレードオフの関係にあるディペンダビリティの構成要因を,アプリケーションの特性や目的に応じて適切に調整する必要がある.既存のアプローチでは,特定のサービスに特化した方法でディペンダビリティを達成したきたが,汎用性に欠け,他のサービスにその方法が転用し難いという問題点があった.この問題点を解決するため,本発表では,サーバシステムをいくつかのステージに分割し,アプリケーションの特性に応じてそれぞれの段階をカスタマイズ可能とすることによりディペンダビリティの達成を可能とするようなツールキットを提案し,その実現を示す.

Awards & Honors

  • 2010/03 情報処理学会 山下記念研究賞
     
    受賞者: 杉木 章義広域分散環境のための仮想機械を利用したサービス協調複製基盤


Copyright © MEDIA FUSION Co.,Ltd. All rights reserved.