If the title of this post makes you go cross-eyed, don’t worry. All will become clear. Let’s explain. Active/active IT configurations consist of computer servers that are connected in a network and that share a common database. The ‘active/active’ part refers to the capability to handle server failure. First, if one server fails, it does not affect the other servers. Second, users on a server that fails are then rapidly switched to another server that works. The database that the servers use is also replicated so that there is always one copy available. Now for the other two acronyms: HA stands for high availability; DR (of course) for disaster recovery. It is DR that is more affected in this case.
The basic idea behind active/active configurations is to accept failure – and to then fix it fast. Fast means quickly enough for users not to notice any significant outage as they are switched from a failed server to a working one. In fact, users may think that an individual transaction has failed, and that they then have to resubmit that transaction. This can happen in normal circumstances in any case. The result is high availability or disaster tolerance. When servers are located in different geographical places, disaster tolerance goes up. However, so does processing time for distributed copies of the common database, or for users obliged to use remote server because their own local server failed.
That still leaves you with disaster recovery to be done on the failed node. The advantage is that some of the pressure is off, because all users should still be correctly served via the remaining working servers. The active/active configuration also allows you to eliminate even planned downtime, because any server that needs to be stopped does not have to affect users (they are all switched elsewhere). In terms of disaster recovery metrics, RTO (recovery time objective) in active/active systems is effectively zero. RPO (recovery point objective) may be zero or a few seconds depending on how the database replication is set up. Smart active/active organization can also cost less than ‘brute force’ duplication of every server in your IT infrastructure.