|
The answer to your question is a simple one, actually. It depends. There are other technologies besides clustering that can give you high availability but it depends on what you are trying to accomplish, what types of failures you want to protect against, how high a level of availability you are trying to achieve and what your critical applications are.
For most applications, if you want to protect against a system crash, a hardware failure or some other system-wide failure, there is nothing better than failover and clustering. However, for some applications there are better technologies.
If you are putting up a Web site with a database back end for instance, the database should probably be clustered. But for the Web servers there are other technologies such as round robin DNS or redirectors like Hydra's Hydra Web, Cisco's Local Director or Veritas' Cluster Server Traffic Director. Both technologies are better than clustering at recovering from the failure of a single server, and at redirecting client requests to different hosts.
Your choice of technologies will also depend on the level of availability you need. If you are only concerned about protecting the data and not achieving high levels of uptime, then mirroring and backups are sufficient. If you are concerned about a site outage, then local clustering will not be sufficient, and you'll probably want to consider replication to a disaster recovery site. It has been my position for quite some time that high availability is what you need it to be; it's the level of availability that allows your critical systems to achieve the goals that you purchased them for in the first place. Only you know if those goals require backups, mirrored disks, clustering, or replication. And different systems have different levels of requirements.
I hope this is helpful. Please follow up if you have other questions.
Evan L. Marcus
Editor's note: Do you agree with this expert's response? If you have more to share, post it in our Administrator Central discussion forum.
|
