Replicating your critical data far away is widely held as a good thing. Now, at least two companies are also promoting the idea of time-stamping the replicated data. That way, in the event of disaster, not only can you run your business off a replicated server, but you can roll back to a stable file system image.
Proponents include XOsoft, which pares its Data Rewinder with its WANSync software, and Kashya, which combines wide-area replication capabilities with "small aperture snapshots," taken at very frequent intervals.
PowerDsine, an Israeli engineering firm, for one, uses XOsoft's WANSync and Data Rewinder software packages to replicate Exchange and Oracle servers from Israel to Long Island. The company also evaluated replication software from Veritas and NSI, but Data Rewinder tipped the scales in XOsofts' favor. "In a normal system, corrupted files, they're replicated as well," explains Asaf Silberstein, PowerDsine vice president of operations.
Kashya followed in XOsoft's footsteps last month when the KBX4000. Based on an appliance, it compares favorably to host-based solutions like Veritas' Volume Replicator, says Amar Rao, Kashya senior vice president of marketing, as those "need to be installed on each host you want replicated."