Common Options for Disaster Recovery

Disaster recovery (DR) is based on three standard DR sites.

In this article, lets take a look at the differences in hot site vs. warm and cold sites in disaster recovery.

Hot site 

In a hot site approach, the organization duplicates its entire environment as the basis of its DR strategy — an approach which, as you’d expect, costs a lot in terms of investment and upkeep. Even with data duplication, keeping hot site servers and other components in sync is time consuming. A typical hot site consists of servers, storage systems, and network infrastructure that together comprise a logical duplication of the main processing site. Servers and other components are maintained and kept at the same release and patch level as their primary counterparts. Data at the primary site is usually replicated over a WAN link to the hot site. Failover may be automatic or manual, depending on business requirements and available resources. Organizations can run their sites in “active‐active” or “active‐ passive” mode. In active‐active mode, applications at primary and recovery sites are live all the time, and data is replicated bi‐directionally so that all databases are in sync. In active‐ passive mode, one site acts as primary, and data is replicated to the passive standby sites.

Warm site 

With a warm site approach, the organization essentially takes the middle road between the expensive hot site and the empty cold site. Perhaps there are servers in the warm site, but they might not be current. It takes a lot longer (typically a few days or more) to recover an application to a warm site than a hot site, but it’s also a lot less expensive.

Cold site 

Effectively a non‐plan, the cold site approach proposes that, after a disaster occurs, the organization sends backup media to an empty facility, in hopes that the new computers they purchase arrive in time and can support their applications and data. This is a desperate effort guaranteed to take days if not weeks. I don’t want to give you the impression that cold sites are bad for this reason. Based on an organization’s recoverability needs, some applications may appropriately be recovered to cold sites. Another reason that organizations opt for cold sites is that they are effectively betting that a disaster is not going to occur, and thus investment is unnecessary.