Active/Active Systems: Theory and Practice

This course describes how active/active systems comprising two or more servers can provide uptimes measured in centuries. An active/active system is a network of independent, geographically-distributed processing nodes cooperating in a common application. Each node has access to an up-to-date copy of the application database. Should a node or a database copy fail, all that needs to be done is to switch over that nodes users to a surviving node. Recovery is in sub seconds to seconds. The underlying concept of active/active systems is let it fail, but fix it fast. If users never notice an outage, then in effect, an outage hasnt occurred.