In this age of the Internet, businesses and government agencies face increasing pressure to keep their systems up and running at all hours.
Back in the day it was good enough for systems to run only during normal business hours, but nowadays that’s almost unheard of. Just about every system has a Web component, and people have come to expect the Web to be available 24/7. As such, many organizations rely upon high-availability (HA) technologies to meet this requirement.
HA is not a new concept, but is being used by an increasing number of organizations under a wide variety of technological umbrellas. The basic principle of HA is simple: provide a means to ensure access to and usage of computer systems in case of a crash. Allowable downtime is measured in minutes, if not seconds.
HA is usually provided by some type of duplicate or redundant system. In just about every HA implementation, the backup system detects failure in the primary system. Measures are then taken to make the backup system function as the primary system. Of course, the backup system must be prepared to take over, which requires such precautions as "real-time" backups of data.
Oftentimes the HA system is housed in the same physical location as the primary system, and sometimes even on the same hardware. HA is designed to protect against "normal" everyday system outages, such as hardware component failure or even data corruption.
HA shares several technological methods with disaster recovery (DR). These two concepts, however, are not one and the same.
A DR system is typically housed miles away from the main production system. For example, on Oahu many folks house their DR systems on Maui or the mainland. Many California organizations house their DR systems in Arizona or Nevada. DR systems are meant to provide backup in case of a catastrophe, such as an earthquake, tsunami or hurricane.
While HA recovery time is measured in minutes or seconds, DR recovery time is measured in hours, sometimes days. Simple factors, like the speed of light and the cost associated with achieving such speed, govern the abilities of DR systems to be updated. Further, HA systems often run on the same network and electrical infrastructure. DR systems rely upon a separate infrastructure.
Regardless of whether you are implementing an HA or DR solution, one common activity must take place to be successful: testing. Semiannual or even quarterly role swaps should be scheduled and executed. In the case of HA, we often see folks simply swapping their main and backup systems on a regular basis.
———
John Agsalud is an IT expert with more than 20 years of information technology experience. Reach him at johnagsalud@yahoo.com.
