HP Preps Pricey Disaster Setup

Readies three-site disaster recovery offering with a starting price tag of (gulp) $6M

August 1, 2003

4 Min Read
Network Computing logo

Finding new ways to disaster-proof critical business data seems to have become one of the summers hottest tickets. The latest vendor delivering an offering in this area is Hewlett-Packard Co. (NYSE: HPQ), which is readying a very high-end, three-site remote replication offering that's not for the faint of heart: The starting price is a cool $6 million.

The computer giant is gearing up to launch a high-end disaster recovery solution next month, promising that it will safeguard customers’ data even through the most unthinkable of disasters. The HP StorageWorks Multi-Site Disaster Tolerant Solution includes HP’s XP disk arrays, as well as its Metrocluster and Continental Cluster software for servers, providing synchronous replication between two proximal sites and asynchronous replication to a third remote site.

Like most synchronous remote replication solutions, HP’s new offering ensures that the second site -- which cannot be more than 100 kilometers from the primary site -- always has an exact replica of all of the first site’s data. The data is sent to the second site via dense wavelength-division multiplexing (DWDM) equipment; HP says it supports any standard DWDM equipment. In the case of a local disaster at the primary site, such as a fire, the second site can automatically pick up operations.

But what happens if an entire region is hit by disaster and both sites are wiped out? That’s where the third, very remote location comes in, HP says. Depending on the importance of the application, companies can choose to asynchronously mirror their data to the third site, or send point-in-time copies to the site at regular intervals, like once every 15 minutes or every hour. For sending data asynchronously across long distances, HP offers SAN extension equipment from CNT (Nasdaq: CMNT).

If some colossal catastrophe were to wipe out both primary sites, customers could quickly get their operations up and running again from the third site, HP claims. While they would lose any data that had yet to be sent, or that was in transit, they would still have almost all of their data intact.Of course, a setup like this doesn’t come cheap. According to James Wilson, HP’s worldwide XP disk array product manager, a customer will need at least three XP disk arrays, as well as the Metrocluster and Continental Cluster software and the software to tie the whole solution together. All in all, he says, the total package -- including DWDM equipment, SAN extension hardware, Fibre Channel switches, storage, and servers -- will cost between $6 million and $20 million.

"This is a non-trivial solution," he says. [Ed. note: We think he means "super-expensive solution."] "If the customer is going to go to this kind of effort, they have to get back into real production, not pretend production... They want to implement a real production-capable solution at each of the three sites."

Wilson says that HP completed the software to pull this offering together in May, and that it has a number of potential customers evaluating it. Initially, he says, HP customers currently using the company’s two-site replication solutions will probably sign up to add a third site. "I think it will build gradually," he says. "This is not an inexpensive solution. Enterprise-class decisions are usually really well thought out... It’s going to take a little while for these things to roll out."

HP isn’t the only storage vendor going after the very high-end disaster recovery space. Yesterday, EMC Corp. (NYSE: EMC) announced several software enhancements for its Symmetrix Remote Data Facility (SRDF), offering asynchronous replication over thousands of miles for disaster recovery (see EMC Debuts DMX, Part Deux). And in May, IBM Corp. (NYSE: IBM) announced that it would be adding multi-hop replication capabilities to its Enterprise Storage Server (a.k.a. Shark) (see IBM Pushes Shark's Copy Buttons).

"Fundamentally, you’re going to see everybody try to fill that void in their offerings," says Dave Purdy, EMC’s director of business continuity, insisting that EMC has been offering a mix of synchronous and asynchronous replication for four years now.According to a recent Gartner Inc. report, EMC had 46 percent of the worldwide data replication market last year, while HP only had 5.3 percent (see our report on Data Protection).

Wilson, however, is quick to point out that EMC only offers one piece of the puzzle. "They just have the storage piece," he says. "They don’t have all of the other glue that wraps around this solution, like the clustering expertise, the server expertise... That isn’t there."

— Eugénie Larson, Senior Editor, Byte and Switch

SUBSCRIBE TO OUR NEWSLETTER
Stay informed! Sign up to get expert advice and insight delivered direct to your inbox
More Insights