Novell clusters go to the next level

Any business that positively, absolutely has to be up and running 24 hours a day, seven days a week, 52 weeks a year, needs clustered servers.

I’ve been beating the drum for Novell Cluster Services ever since it was introduced shortly after the release of NetWare 5. Before that, Novell had, with good reason, concentrated on hardware failure as the major cause of server and service downtime which was combated by using System Fault Tolerant (SFT) NetWare to provide redundant hardware – and even redundant servers – to guard against failure of any single component.

But five years ago two things had become apparent: multiple component failure was almost as common as single component failure; and software failure was causing more server crashes than hardware failure. While we still needed to protect against hardware faults, the SFT technology couldn’t handle problems in software. Bad software, in fact, would crash both servers in an SFT pair, which meant you were spending money for a solution to your problem but the solution didn’t work.

Clustered servers, though, could overcome that problem. Where SFT was built on a shared memory model (both servers’ RAM contained identical code which was constantly synchronized), clustered servers work on a shared storage model. There’s a good paper on the theory and implementation of the original Cluster Services still available from Dell’s Web site which you should read if you’ve not been exposed to the service:

https://www1.us.dell.com/content/topics/global.aspx/power/en/ps4q99_novell?c=us&cs=555&l=en&s=biz

As the authors state, while SFT was all about high availability, “the goals of a cluster are to provide high availability, scalability, and manageability. From the client’s perspective, the application or resource should always be available to service its requests. Scalability implies the ability to increase the performance of the overall cluster by adding more resources (nodes) to the cluster. A single-system image simplifies administration and management of the cluster.”

Now Novell has raised the bar by introducing “clustered clusters.” Novell Business Continuity Clustering (NBCC) Version 1.0 automates the configuration and management of high-availability, geographically dispersed, clustered servers to greatly improve availability, scalability and manageability.

While the earlier Cluster Services package did all it promised in terms of failover, there were limitations on how far apart the servers in the cluster could be. NBCC overcomes that limitation by providing for up to four geographically dispersed clusters of servers to be tied together in a “supercluster” so that the failure of one site doesn’t have to bring an enterprise to its knees.

It should go without saying (it should, but I’d better anyway) that NBCC is fully eDirectory-integrated and controlled. It’s really the power of the directory that makes it all possible.

Now Novell Business Continuity Clustering isn’t for everyone. But if uptime is crucial, if you have multiple computing centers and if – most importantly – you’re running NetWare 6.5 (no Linux allowed just yet), then you should find out more at: https://www.novell.com/products/businesscontinuity/