フォールトトレラント因果順序保存グループ通信

概要

論文の詳細を見る
In distributed systems executing such applications as teleconferences and telemedicines, a group of multiple application processes communicate with each other. The application requires the processes in the group to receive reliably the messages in the causal order. There are kinds of distributed systems. The first one is composed of two types of processes, i. e. clients and servers. There is another kind of distributed systems i. e. processes which autonomously compute and communicate with other processes. In these systems, a group of multiple autonomous processes are cooperated to achieve some objects. Here, it is required to achieve the intra-group communication where the processes communicate with each other in the group. The processes in the distributed system may suffer from kinds of faults. An approach towards making the system fault-tolerant is to replicate the processes in the system. In this paper, in order to support the fault-tolerant group communication, each process is replicated into a collection of multiple replicas, which is named a cluster. Our protocol supports the inter-cluster communication among the replicas in the clusters in order to tolerate the Byzantine faults of processes in the group. In section 2, we discuss the replication schemes. In section 3, we present the properties of the intrar-group communication. In section 4, we discuss the inter-cluster communication in the group In section 5, we discuss how to support the causally ordered and fault-tolerant delivery of messages in the group.
一般社団法人情報処理学会の論文
1996-03-06

フォールトトレラント因果順序保存グループ通信

スポンサーリンク

概要

著者

関連論文

スポンサーリンク