The traditional means of long distance data recovery is either by tape or array-based (synchronous) replication, but each approach has its limitations.
Tape is certainly an economical approach, but has a number of problems: creating a tape backup disrupts production operations; shipping tapes to a remote site takes time and exposes data to risk of loss; and tape backups are typically done once a day, so recovered data can be as much as a day out of date.
Array-based replication is typically limited to a single vendor replicating between like-for-like storage, so companies usually can't use lower-cost storage at their recovery sites.
Synchronous replication is the proven method for array-based solutions and supports recovery of real-time data at the remote site, but distances are limited to about 50 miles, and replication requires very high network bandwidth. A synchronous system copies an update to the remote site and has to wait for confirmation that the update was written to disk before the production application can continue processing. As distances increase, waiting for this turnaround for each update can affect application performance. Thus, cost and limitations make array-based approaches applicable for only the most mission-critical applications.
Topio Inc., of Santa Clara, Calif., thinks it has found a solution.
"Topio solves this problem by lowering the cost of recovery while still delivering the immediate recovery these applications need," says Topio vice president Chris Hyrne.
The Topio Data Protection Suite (TDPS) uses asynchronous replication to solve the distance problem. Updates to production data are continuously sent over the existing IP network as they happen. Software at the remote site processes the data stream to ensure updates are applied so they match production site data. This can be accomplished no matter what distance is involved.
"A continuous 'trickle feed' eliminates the disruption of creating tape backups and ensures updates are copied to the remote site as they happen," says Hyrne. "Efficient use of the customer's existing IP network saves significant network costs and leverages existing IT skills."
Maintaining a current, consistent data image at the remote site means recovery to a current state can occur immediately and applications can be back on line in minutes. TDPS also allows the replication of data between any vendor or type of storage as a way to reduce costs and avoid vendor lock-in.
Topio provides TDPS solutions tailored to disaster recovery, backup consolidation and data migration. In the backup consolidation area, for instance, TDPS continuously copies data from multiple remote sites to a central backup facility where point in time copies are made on disk and then transferred to tape.
This eliminates the need for backup software and hardware at the remote sites. And since remote site staff is not often trained in IT, backup consolidation enables skilled staff with the proper controls at the central site to ensure compliance.
What is different about TDPS compared to competing products in the asynchronous replication line? According to Hyrne, Topio has three main differentiators: the ability to achieve current recovered data as opposed to data from earlier points in time; guaranteeing that replicated data will not be corrupted at the recovery site; maintaining consistency in enterprise scale environments; and platform independence.
"Other solutions are proprietary (Veritas), only support Windows environments (NSI) or only support SAN-attached storage (replication appliances)," says Hyrne. "Topio is platform and storage agnostic, so you can replicate and recover your data no matter where it resides, using your existing infrastructure."
For now, the product can't be classified as continuous data protection (CDP), but this will come, says Hyrne.
"We will continue to improve the granularity of our point in time copies to provide rollback to any point in time," he says. "Our customers have told us that their adoption of this capability is well into the future, whereas their most pressing problem is near real time disaster recovery."
TDPS supports both server-based and fabric-based replication. It runs on Solaris, Windows AIX, HP UX, VMWare and Linux Platfoms. It functions on SAN, DAS, internal disk, iSCSI and solid-state storage. Pricing ranges from $25,000 to $50,000 per TB, depending on functionality.
"TDPS is used by enterprises to solve recovery issues in-house and by recovery service providers like IBM global services, who use TDPS to provide out-sourced business continuity services," says Hyrne.
Article courtesy of Enterprise IT Planet