Providing Hardware Redundancy for Highly Available Services in Virtualized Environments

Abstract

High-Availability requires hardware and software redundancy. Virtualization is a technique - among others - for improving the utilization of hardware resources by making virtual (rather than actual) versions of hardware, operating system, etc. and collocating them on the same hardware. In virtualized environments virtual machines (VMs) are used for the deployment of the software entities. When VMs hosting redundant software entities providing and protecting some service are collocated on the same physical node, the hardware redundancy is lost and the failure of this physical node certainly leads to service outage. To achieve high availability, we need to avoid such single points of failure even in the presence of VM migration. This paper tackles this issue in the context of a standardized middleware for service high-availability.