106.txt

A Proposed Fault Tolerance Model for Cloud
System Based on the Distributed Shared
Memory
Mohamed A. Elsharkawey, Hosam E. Refaat
Dept. of Information System Faculty of Computers and Informatics
Suez Canal University, Egypt
melshrkawey1964@yahoo.com, hosam.refaat@ci.suez.edu.eg
Abstract
Nowadays, the cloud computing appears as a magic solution for a lot of companies and
scientific problems to have extendable resources with a suitable budget. The increasing of the
utilization of cloud computing in different fields creates challenge in the reliability and
availability of the system. In this paper, a middle layer is proposed to increase the reliability.
This layer places between the application layer and the virtualization layer in cloud system. It
aims to handles the failure of the nodes. This layer monitors the system resources and
periodically updates the status of all the virtual machine functionality. Based on these
statuses, when the failure is occurred, the appropriate virtual machine will be selected to
handle object failure. Moreover, the failure is handled in two cases, firstly, when the failure
occurred in one of the running virtual machine, secondly, when the failure is occurred at the
virtual machine management.