Virtualized Systems as a Basis for Redundancy

**RSS Robot** · 08-23-2012, 11:40 PM

Foundation for Business Continuity, Disaster Recovery, and Managed Failover

Many people seem to know they want or need to virtualize their systems, but far less seem to have a good grasp of why they should do so. Virtualization holds a great deal of promise for redundancy to support business continuity and similar needs; the big question is how much capability comes “out of the box” and how
much other work is required to have a robust solution?
This document will discuss several levels of infrastructure redundancy with a view to accomplishing maximum possible uptime based on various risk factors, and suggest several possible architectures that correspond to different levels of risk mitigation. These will range from internal factors within a given component’s design to large-scale infrastructure designs to mitigate catastrophic failure of a site. Several of these factors are not
unique to a virtualized environment and will be identified. While several different parts of the overall infrastructure will be addressed, areas such as power and Internet Service Provider (ISP) redundancy require a far more extensive investigation and this paper will only touch on the topics needing to be covered.
Types of Redundancy within Individual Components (Server and Storage)

Within a given host computer, it is recommended that you have at least two forms of redundancy, assuming there is no local storage. Local storage, that is disks or RAID arrays installed within a single server, are not available to other servers in case of system failure and are therefore not suitable. In addition, storage that is considered highly available requires some specific design features for redundancy.

Power Supply Redundancy allows you to continue full operation in the event of a single power supply component failure, along with the ability to hot-swap a replacement power supply for the failed unit without interrupting production. Within Kodak, any of the Premium Workflow servers come with hot-swappable redundant power supplies by default. If selecting host computers from another vendor, ensure that the systems contains redundant power supplies. This also applies to any server or storage component in the system.