Study on Intelligent Network System for Dynamic Optimization Control in Distributed Computing Environment

2017 Fiscal Year Final Research Report

• Project/Area Number: 15K00169
• Research Category: Grant-in-Aid for Scientific Research (C)
• Allocation Type: Multi-year Fund
• Section: 一般
• Research Field: High performance computing
• Research Institution: Kyoto Institute of Technology
• Principal Investigator: Nunome Atsushi 京都工芸繊維大学, 情報工学・人間科学系, 助教 (60335320)
• Research Collaborator: ICHII Sekai ; SHIMANO Shingo ; SHOJI Yuki ; TASHIRO Saki ; DEJIMA Takashi ; HAYASHI Shohei ; HIRATA Hiroaki ; FUJISAWA Kohei ; YAMASAKI Hiroki ; YOKOI Yuta
• Project Period (FY): 2015-04-01 – 2018-03-31
• Keywords: 計算機システム / ハイパフォーマンスコンピューティング / 分散処理 / 動的負荷分散制御

Outline of Final Research Achievements

In a massively distributed computing environment, it is important to tune the system from both aspects of software and hardware to bring out potential performance to the maximum. In this study, we have proposed a scheme to perform dynamic optimization control for the heterogeneous environments where various types of node are mixed. It also considers differences in the effective performance which is occurred at run-time. Formerly, a high overhead is regarded as a problem because the information needed for such control must be collected at run-time. In our scheme, the overhead accompanied with the optimization control will be suppressed by appending the system management information onto ordinary packets by the network device. We have established one of a fundamental technology which is capable of applying even if the management information enlarges for more complex dynamic optimization control in the near future.

Free Research Field

情報工学