Generating Distributed Programs from High Level Specification Based on Combinatorial Topology

2020 Fiscal Year Final Research Report

• Project/Area Number: 16K00016
• Research Category: Grant-in-Aid for Scientific Research (C)
• Allocation Type: Multi-year Fund
• Section: 一般
• Research Field: Theory of informatics
• Research Institution: Kyoto University
• Principal Investigator: Nishimura Susumu 京都大学, 理学研究科, 准教授 (10283681)
• Project Period (FY): 2016-04-01 – 2021-03-31
• Keywords: 並行分散プログラム / 位相幾何的手法 / プログラム導出

Outline of Final Research Achievements

We have developed an algorithm that derives an executable program from a given specification of a distribute system, which is defined by a mapping from a simplicial complex (i.e., a combinatorially-constructed geometric object of higher dimension) to another simplicial complex. Inspired by topological perspective, we have also redesigned the immediate snapshot, the core distributed protocol that characterizes the standard model of distributed computing, by multiple steps of simpler distributed computation. A program optimization method has been proposed for this topologically refined snapshot algorithm. Furthermore, we have studied the topological properties of distributed computing models other than the standard model and we have also worked out a novel proof for the impossibility of certain distributed tasks, making use of epistemic logic, i.e., a formal logic of knowledge held by the processes in a distributed system.

Free Research Field

計算機科学

Academic Significance and Societal Importance of the Research Achievements

並行分散システムは現代ITインフラの重要な構成要素であるが、その設計と運用は実効順序の組み合わせに起因する非決定性と耐故障性への要求のため、非常に困難である。並行分散システムの性質を理解し、プログラムを正しく定義しこれを実行するための研究は、システムをより容易にかつ正しく設計・運用するための基礎となるものであり、より一層のIT社会の発展に資するものである。