Development of Rubust Framework for Distributed Collaborateve High Performance Edge Computing

2023 Fiscal Year Final Research Report

• Project/Area Number: 20K11744
• Research Category: Grant-in-Aid for Scientific Research (C)
• Allocation Type: Multi-year Fund
• Section: 一般
• Review Section: Basic Section 60050:Software-related
• Research Institution: University of Fukui
• Principal Investigator: Mori Shin-ichiro 福井大学, 学術研究院工学系部門, 教授 (20243058)
• Project Period (FY): 2020-04-01 – 2024-03-31
• Keywords: 高性能計算 / 耐故障性 / エッジコンピューティング / 並列・分散処理 / 防災 / インタラクティブシミュレーション / サイバーフィジカルシステム / ヒューマンインザループ

Outline of Final Research Achievements

In this research, they have developed a distributed collaborative simulation framework to improve the accuracy of simulation without introducing unacceptable delay. The key component of this framework is a coordination mechanism that indirectly and hierarchically coordinates groups of autonomous servers that are geographically distributed. By making the coordination non-strict, or a weak, best-effort style, this framework obtains the robustness against communication failures so that it became free from system-wide failures that are an issue in large-scale coordinated systems. The framework also has automatic functions for the graceful degradation and the recovery at the occurrence of and recovery from failures. In their 3 days long continuous experimental study with three servers placed South America and Japan, they confirmed the functionality of the framework. They also confirmed the sustainability of their frame work by intentionally introducing the communication failure.

Free Research Field

計算機システム

Academic Significance and Societal Importance of the Research Achievements

様々な場所で大量のデータが発生するIoT環境において，データ集約が必要な現在の集約型のクラウド計算技術は，通信帯域の不足や応答時間の遅さが実時間応答性に関する性能限界を規定してしまう．そのため次世代のクラウド技術として，自律・分散・協調を実現する新しい高性能エッジコンピューティングアーキテクチャの開発が喫緊の課題となっている．
この課題に対して，本研究では自律動作可能な複数のサーバを階層的に連携集約し，間接的な協調連携を行わせることで，個々のサーバ単独では得られない質（計算精度や実時間応答性等）の高い計算サービスの提供と障害耐性を持った広域分散型の大規模連携シミュレーション環境を構築している．