This article can also be found in the Premium Editorial Download "Storage magazine: Exploring the most innovative midrange systems."
Download it now to read this article plus other related content.
Just because a company isn't listed on the Fortune 500 doesn't mean it doesn't need remote replication--a cornerstone of enterprise-class disaster recovery (DR) practices. But many midmarket firms are discovering that vendors usually associated with remote replication simply aren't meeting their needs.
"Talking to the big guys wasn't very satisfying," says Matt Kesner, CTO at Fenwick & West LLP, a high-tech law firm in Silicon Valley. With offices in San Francisco and Mountain View, CA, the firm began talking last year to top-tier DR providers like SunGard and AT&T, as well as storage vendors like EMC about ways to replicate data outside the firm's earthquake-prone area.
Things never went very far, though. "I found [the big-name vendors] had very impressive presentations, but they couldn't tell me specifics like what my actual bandwidth needs would be or what kind of asynchronous lag we'd have," Kesner says. Dealing with them, he says, "it was impossible to predict our ongoing costs." At the same time, he was being quoted solutions "in the high six figures and even low seven figures" that were way out of reach.
"You could tell they were geared toward big enterprises," he recalls. "They'd say to me, 'Your storage person can do this part-time.' But we don't have a storage person and we probably never will."
Eventually, Fenwick & West was introduced to InMage Systems Inc., a startup founded by former Brocade Communications' co-founder and CTO
Fenwick & West has replicated approximately 6TB of data, which may grow to as much as 40TB. They pay less than $5,000 per month in bandwidth costs, and the appliance itself has a list price of less than $25,000, InMage says.
Another problem midmarket firms face is that the kind of DR they can afford simply isn't practical with a limited IT staff, says Don Himsl, director of IT at CNS Inc., a personal care product manufacturer based in Minneapolis.
Currently, CNS' DR plan "is fairly traditional. We do full backups every week and incrementals at night," says Himsl. In case of a disaster, CNS would restore its servers from tape at a SunGard facility. But while Himsl is confident about being able to restore the systems within an acceptable recovery time objective--24 hours or less for critical systems, and two to four days for second-tier servers--with an IT staff of only five, "it becomes a real nightmare even to execute a test plan," he says.
"We want to take our DR up a notch," says Himsl. Asynchronous replication software, cheap SATA disk arrays and the fact that all of CNS' servers boot from the SAN would help CNS to improve its RTO, plus allow them to test their DR plan on a regular basis.
Himsl has petitioned Compellent, his SAN array vendor, for asynchronous replication functionality, which the company says it hopes to bring to market this year to complement its existing synchronous replication software.
Himsl isn't the only Compellent customer requesting this functionality. At a recent gathering of Compellent's Customer Council (C3), the company identified remote replication as its customers' highest priority.
Other startups gunning for the enterprise remote replication market are also taking a keen interest in medium-sized businesses. "Initially, we thought the mid-market wouldn't be interested in remote replication, but it isn't working out that way," says Mehran Hadipour, VP of product marketing at Kashya, which makes the Kashya Data Protection Appliance. Originally deployed in large firms like Raytheon and the BBC, the box has also been adopted by smaller organizations like the San Francisco Chronicle. "We've started to think that maybe the midmarket is a good place to be," says Hadipour.
This was first published in March 2005