A new client who is going through a core systems upgrade was asked from the vendor about their Disaster Recovery (DR) Strategy and if they had a cold site, warm site, or hot site and what the difference was between them.
Simply put a cold site is the presence a facility to work out following a disaster of but is not equipped with computer equipment or even copies of data. A cold site may take weeks or months to become fully operational following a disaster. A cold site requires extensive administrator work including acquiring hardware and software. Often supplementing the cold site is a person who occasionally takes backup copies such as tapes offsite or stores them in a fireproof safe. Many times there are no network connections at a cold site and circuits or internet connections must be ordered and installed. A cold site is typically the least expensive to implement.
A hot site is full replication of the original datacenter(s) with pre-configured servers and network equipment. Typically a hot site is configured in such a way that data is replicated from the production servers to the DR site in real time or near-real time. A hot site can be configured to failover with no or limited administrator intervention and can be performed in seconds or up to 1 day. A hot site is usually the most expensive option to implement and maintain, as all IT investments and changes must be performed at the DR site.
A warm site lies somewhere between hot and cold. Many warm sites have servers and network equipment ready to go, but at reduced capacities and data backups lagging either hours or a couple of days behind. A warm site requires some administrator intervention and can usually be activated between 1 hour and 2 days. A warm site typically has network and internet connections pre-installed but often at reduced capacity.
A popular configuration for hot and warm sites utilize server virtualization and SAN-to-SAN replication technology. All virtual servers are replicated on a pre-determined basis to the DR site. Technology such as VMware Site Recovery Manager (SRM) or Cisco Local Area Mobility ensures that when the time comes to failover to the DR site that endpoints can connect to the DR site and access servers and data.
Alternative to an on-premises DR site lies the Cloud. The Cloud can be utilized in many ways, including running primary servers out of the cloud or replicating a production datacenter to the cloud in real-time or through backups. In the latter, image-based backups of servers are replicated to the cloud on an hourly or daily schedule and in the event of a DR, a network connection is made to the cloud across the Internet and data accessed there.
Which choice is appropriate for an organization depends on the organization’s Risk Analysis and Business Continuity Plans (BCP). It’s recommended that all organizations review their BCP’s regularly and update according to business requirements. RSM has consultants that specialize in BCP planning and designing and implementing DR plans. For more information on RSM’s offerings please check out our website. You can also contact RSM’s technology consulting professionals at 800.274.3978 or email us.