System Recovery—Breaking through the Dissimilar Hardware Restore Challenge
Even better, the preceding steps do not have to be journaled and they can be identically replicated every time, with no special training required (see figure 4).
New Option for Meeting Strict Recovery Time Objectives (RTO) and Upping Disaster Tolerance Existing technology solutions provide many types of failover for the most critical systems, but rapid system recovery solutions are scarce, and the expense of high-end failover technologies cannot be justified in most cases. Restore Anyware technology offers a new option for meeting stringent recovery-time objectives that do not require immediate failover. This is a much-needed recovery solution for systems that do not warrant the high cost of clustering or mirror sites - but that must nonetheless be recoverable in minutes, not hours or days. One factor in determining the appropriate solution is disaster tolerance.
Defining disaster tolerance For many organizations, recovery-time objectives are short and do not allow sufficient time to order new computers and wait for them to arrive. To shorten recovery time, a system's disaster tolerance (its ability to survive a disaster, most often from multiple points of failure-perhaps even the loss of an entire data center or facility and all its functions) must be increased. How can a server be made disaster tolerant? The answer depends on the desired degree of tolerance to multiple failures, which in turn has financial ramifications, because the most fault-resilient systems are also the most expensive. Each level of protection has its own requirements and associated costs and benefits. Figure 5 shows a common mirroring scenario.