Information Technology Reference
In-Depth Information
configuration and recovery from backup media to be completed
before operations resume. Warm sites often employ older equipment
removed from service during the previous tech refresh cycle, reduc-
ing procurement costs but offering limited capacity in comparison
to the hot site.
•
Cold site
—The least expensive alternative, with the longest recovery
time requirement (except for a completely new site, which requires
starting from nothing). Cold sites are locations with basic utilities
such as power and water. They may not have equipment, networking,
software, or other requirements for data center operational recov-
ery. Recovery using a cold site will generally require procurement of
hardware, software, or utilities before recovery becomes possible.
Alternate Equipment
Recovery following a disaster that results in loss of equipment will gener-
ally result in new types or versions of equipment supporting existing func-
tions. Plan testing scenarios should include strategies for recovery using
alternative equipment, particularly when using warm or cold backup sites
that may have equipment with far less capacity than existing primary data
center equipment.
As mentioned earlier in this topic, server virtualization and remote
desktop access can improve portability between hardware hosting alter-
natives, provided the hypervisor is compatible with that on the original
host. Figure 13.1 illustrates some of the ways in which virtualized servers
might be migrated to facilitate recovery operational requirements. Due to
the hardware-independent nature of virtual machines, restoration of the
virtualized server's files and appropriate network settings will return func-
tionality with the level of resources available on the new host computer.
Alternate Communications
Following hurricanes Katrina and Rita, telephone service was lost through-
out the area. Due to the extended power outage, cellular phone service
became unavailable in some areas and intermittent in others. Remote per-
sonnel in many cases could not access organizational Web servers because
those servers were located in data centers within the affected area. One
