1.2.7 Failover

Failover is a concept that can be used to evaluate how fast can a site be recovered from any planned or unplanned outage, which has a closed relationship with the High Availability site. Recovery can be achieved by automation management tools or manually, which is up to which High Availability strategy you are applying in your site.

Generally, the measurement to failover can be classified by two aspects:

The system view is much narrower than the customers' view, so we should only concentrate on one specific tier. That is, if it is a database outage, from the system's point of view, how long can a database successfully restore from disaster should be a key measurement for failover from the system's point of view.

This is much more complex than evaluating failover time from the system's view, since WebSphere Commerce builds above a complex infrastructure, where the application server, Web server, database, and WebSphere Commerce should work together to deliver sufficient utilities. So that the information that the users (customers) can see should be the data that have been processed by the Web server, application server, and database. If there is an outage in one tier, it should impact tiers in the entire environment, while the major impact should be to customers. From the beginning of system failure untill the functionality becomes workable again to customers is the measurement with which to evaluate the failover capability from the customers' point of view.
xxxx