Ensuring High Availability

I get miffed when I try to buy a book online from some bookstore and have to wait more than five seconds for the check-out page to appear, or if the site temporarily seems to go down. What’s wrong with these guys? Don’t they understand high availability? What, are they running their entire store on a single box? Don’t they know single point of failure?

Whatever applications are critical to the operation of your business, you need to use some form of clustering to make sure they never go down or become inaccessible to customers. Windows Server 2008 includes two enhancements in the area of high availability. First, server clusters (now called failover clusters) have been significantly improved to make them simple to set up and configure, easier to manage, more secure, and more stable. Improvements have been made in the way the cluster communicates with storage, which can increase performance for both storage area network (SAN) and direct attached storage (DAS). Failover clusters also offer new configuration options that can eliminate the quorum resource from being a single point of failure.

Network Load Balancing (NLB) has also been improved in Windows Server 2008 to include support for IPv6 and the NDIS 6.0 specification. And the WMI provider has been enhanced with new functionality to make NLB solutions more manageable.

Has this piqued your interest? Check out Chapter 9, “Clustering Enhancements,” and find out more.