Design of the notification system for failure detectors. (8th June 2009)

• Record Type: Journal Article
• Authors: Hayashibara, Naohiro; Takizawa, Makoto
• Is Part Of: International journal of high performance computing and networking. Volume 6:Number 1(2009)
• Page Start: 25
• Page End: 34
• Publication Date: 2009-06-08
• Subjects: failure detection -- accrual failure detectors -- overlay network -- topology -- self-organising algorithms -- distributed systems -- monitoring networks -- simulation; High performance computing -- Periodicals; Computer networks -- Periodicals; High performance computing; Periodicals; 004.05
• Languages: English
• ISSNs: 1740-0562

Abstract

It is widely recognised that distributed systems would greatly benefit from the availability of a generic failure detection service. In this paper, we highlighted the issue on the construction of the monitoring network of failure detectors. We proposed an algorithm to construct and manage the monitoring network that each failure detector is monitored by some failure detectors. Notification of failures is propagated along the network. Especially it can involve various types of failure detectors from simple timeout-based failure detectors to accrual failure detectors, and help to spread information on suspected processes/nodes. In addition, we have made a simulation of the proposed algorithm for constructing the monitoring network. It shows that the algorithm is scalable for increasing the number of failure detectors.