Going the Distance with Disaster Recovery

Surviving a disaster — natural or man-made — is never something an enterprise can predict in terms of what might happen in the event of a strong hurricane, widespread power outage, fire or explosion. Yet during the disaster, the company should have a viable way to use a remote disaster recovery site to stay online or to quickly reconnect. The disaster recovery (DR) site should be neither too far nor too close to the damaged zone, but just right.

“When it comes to disaster recovery, concerns now include the increasing number of natural disasters like fires, hurricanes, floods and earthquakes,” says Fred Moore, an analyst with Horison Information Strategies. “CIO’s are also concerned about catastrophic power loss, hackers, employee sabotage and terrorist attacks.”

Opinions vary as to what is truly needed for disaster recovery. Mike Karp, an analyst with Enterprise Management Associates, suggests that clustering, SAN mirroring [define] and replication [define] are some of the more important technologies. The KBX5000 Data Protection Platform by Kashya Inc., of San Jose, Calif., for example, addresses the problem of data center availability.

“The KBX5000 delivers enterprise-class data protection so that a business — across the enterprise — can ramp up quickly during disaster recovery,” says Rick Walsworth, Kashya’s vice president of marketing.

The KBX5000 is an appliance-based system that provides continuous data protection (CDP) [define] and remote replication. The network-based architecture handles stringent data requirements and can replicate from Fibre Channel SANs to IP infrastructures without any distance limitation across the IP environment.

This factor of distance is a vital one in the DR arena. As last year’s hurricane season pointed out dramatically, it’s no good having one data center in New Orleans and another in Biloxi or Houston. In the former, the same weather event could simultaneously take out both data centers. In the case of the latter, two different hurricanes one month apart could serve up a one two punch that took out the backup data center shortly after the main center was destroyed.

The key, then, is to place a remote disaster recovery site far enough from the main office, but not too far away from where it’s needed the most. According to Karp, it is best to find a happy medium where events at one location are unlikely to be duplicated in the other location.

The KBX500 makes it possible to have remote data centers operating at a considerable distance without the usual distance limitations that challenge replication systems. You simply place one Kashya box in one location and another in the mirrored data centers and you are in business. These systems can sustain up to 2 GB/S (16 Gbps) of total throughput across a clustered environment.

User-defined policies in the KBX5000 map directly to specific application requirements. They integrate with Oracle, Microsoft SQL, Microsoft Exchange and other applications for time and event-based recovery. Advanced data protection capabilities include local and remote continuous data protection, remote replicator, bandwidth reduction, compression and WAN optimization.

“The total system costs begin at around $50,000 and for disaster recovery application an enterprise typically deploys a system at each location,” says Walsworth. “The KBX5000 is a complete data protection solution with user defined policies that map directly to specific application requirements and high levels of integration with key enterprise database and applications.”

Who does Kashya compete against? In most cases, Kashya fights against array- and host-based solutions that provide replication and snapshot-based technologies. Walsworth adds that the key differentiators for Kashya are the capability to provide superior replication and CDP capabilities across heterogeneous storage and server infrastructures at lower cost points.

Further Apart With double hurricanes or multistate power outages taking out whole regions, companies are now starting to pay more attention to geographical and other factors that could cause two or more data centers to be down at the same time. For those with the budget to do so, it makes sense to mirror data as far apart as possible. And tools like the Kashya KBX5000 can help facilitate such an architecture. A disaster recovery center placed hundreds or even thousands of miles distant from the main center could well turn out to be a small price to pay in the event of a serious disaster.

Some analysts say 30 miles might be far enough. Others, however, suggest that in certain circumstances, the huge potential losses involved might make it feasible to place centers as far apart as possible.

“The other side of the continent might be necessary to play it safe,” says Karp.

Article originally appeared on Enterprise IT Planet.com.