Performance Analysis Of Failover Cluster For System Recovery

Abstract

The need for up-to-date information systems encourages web server service providers to build systems with high availability. One who needs it is the Online Shop Business and large-scale companies. The technique that can be used is implementing server clustering. Cluster Server is a technology that combines several resources that work together so that it looks as though a single system. Cluster technique that can be used to provide a system with high availability that is failover cluster. In this study, the failover cluster system was built using the VMware provided Vmware's FmFL failover feature. Failover cluster built consists of three physical servers with Vmware ESXi operating system. One server is used as data traffic monitoring and Shared Storage while two servers are used as nodes of the Failover Cluster which will be used to run the Guest OS. Each node in the Failover Cluster will be tested using a System Failure Scenario created based on potential interruptions in the company. Parameters measured in this study are Availability and Network Utilization. The Availability trial results performed using Vmware features called Vmware Fault Tolerance can generate a 94.44% Availability Value for 7 System Failure Scenarios. The Network Utilization value of 13,041.54% was obtained when the failover occurred between nodes in failure scenarios performed in a row.

Keywords: Failover Cluster, Availability, and Utilization.