Synchronous Replication
StorTrends offers the ultimate in flexibility and performance when it comes to disaster
recovery and business continuance. Remote Synchronous Replication is implemented on
the target stack. This imposes no burden on the application servers and is transparent to
the SAN initiators. Setting up a complete mirror-recovery site requires just a few clicks
of the mouse and in seconds the entire production server storage configuration is
duplicated on the recovery site storage server. Thereafter, initial resynchronization is
automatically started. StorTrends keeps track of written blocks at a very fine granularity,
and replicates just those blocks instead of the entire volume. By its very nature, most
Synchronous Replication setups protect organizations against physical site disasters.
Logical data errors (like viruses and accidental erasures) are not handled by synchronous
replication.
Due to inline nature of Synchronous Replication, latency plays a very big factor in its
viability. Traditional Copy-On-Write (COW) snapshots introduce sizable latencies to
almost rule it out as a workable solution. StorTrends, with its Near-CDP snapshots and
Redirect-On-Write architecture (with negligible latencies), allows creating closely spaced
snapshots. These low latency snapshots can be scheduled independently at the production
and recovery sites. This feature allows StorTrends to offer unmatched flexibility and
protection against both physical disasters and logical data corruptions.
StorTrends Advantage:
•
Affordable to setup
•
Fine – grained synchronization
•
Bi-directional replication up to 64 volumes
•
Only necessary data is replicated across power failures and link outages.
•
Low-latency snapshots in production and recovery sites
•
Tools to validate replication integrity
•
Link split/join control
•
Support for Failover and Fail-back
•
Extensive monitoring and manageability
Appendix E : Replication Overview 241