System replication is an IT strategy for business continuity, high availability and disaster recovery. It involves the sharing of identical information or tasks (replicas) between redundant resources to improve reliability, fault tolerance and system accessibility. Data replication is the same data stored on multiple storage devices. Computational replication is the same computing task executed many times. A computational task may be replicated on separate devices, or executed repeatedly on a single device. System replication may be active or passive. Active replication processes the same request in every replica. Passive replication processes a single request on a single replica, then transfers its resultant state to the other replicas. Unlike backup, which saves a copy of data unchanged for a long period of time, replicas are updated frequently and quickly lose any historical state.
For database high availability, system replication usually involves setting up a standby secondary system that is configured as an exact copy of the active primary system. Each service instance of the primary system communicates with a counterpart in the secondary system. Multitier system replication adds a third system attached to the secondary system, forming a replication chain of three systems. The secondary system may be located near the primary system to serve as a rapid failover solution for planned downtime, or to handle storage corruption and other local faults. For disaster recovery, large enterprises may install the secondary system in a remote site. Both local and remote secondary systems can be chained together in a multitier replication chain.
System replication is recommended by SAP to ensure the high availability and disaster recovery of SAP HANA databases. It is available in every SAP HANA installation. SAP HANA provides the ability to replicate a SAP HANA system within the same or over two data centers. SUSE Linux Enterprise Server for SAP Applications includes support for SAP HANA System Replication to help automate high availability and the failover process.