グループ通信における意味的なロールバック復旧

概要

論文の詳細を見る
In distributed applications like teleconferences, a group of multiple autonomous objects are required to be cooperated to achieve some objectives. In the group communication, messages sent in the group have to be causally delivered to the objects. If some object o is faulty and rolled back to the checkpoint, objects which have received messages causally following the messages sent by o have to be rolled back in order to keep the system state consistent. It is discussed how multiple objects take the consistent checkpoints in. Leong and Agrawal discuss the concept of significant messages. On receipt of a message m, m is referred to as significant iff the state of an object is changed. If o is rolled back, only objects which have received significant messages sent by o are rolled back. Thus, the number of objects to be rolled back can be reduced. In the distributed computation, objects receive request messages and send back the response messages. The request may change the state of the object, and the response may carry the data. The paper does not consider the request and response messages. In this paper, we would like to define the influential messages based on significant causal precedence relation, and discuss how to reduce the number of objects to be rolled back. In section 2, we first present the system model. In section 3, we present briefly the traditional rollback methods. In section 4, we discuss the influential messages and the recovery protocol.
一般社団法人情報処理学会の論文
1996-03-06

グループ通信における意味的なロールバック復旧

スポンサーリンク

概要

著者

関連論文

スポンサーリンク