Outage tolerance

The tolerance of an outage varies from business to business. Each business, as a best-practice, should analyze the impact of an outage to their mission critical business processes. The results of this analysis are used to formulate a restoration plan. This plan includes an order of priority to the restoration if more than one process is identified.

Outage tolerance

A crucial factor in determining your availability needs is to ask how tolerant your business, or a specific system in your business, is to the occurrence of an outage. For example, a restaurant that operates a Web site primarily to publish menu information will not lose much revenue because of an occasional server outage. On the other hand, any outage on a stock exchange server that records transactions would be catastrophic. Thus, using a lot of resources to ensure the availability of the restaurant's server is 99.99% would not be cost-effective, whereas it certainly would be for the stock exchange.

When discussing tolerance two concepts should be kept in mind: time to recovery, and point of recovery.

Time to recovery is the time required to bring a business process or system back online.

Point of recovery is the historical point at which the business process or system is restored. In database terms, a plan would weigh the benefits of a quick restore that loses some transactions versus a complete restore that loses no transactions but which takes longer to perform.