VMware’s R&D Lab: A Little Piece of Palo Alto in the Heart of Kendall Square

When Xconomy set up shop in Kendall Square, Cambridge, last year, we knew it was the innovation hub of New England, but we still didn’t realize just how deep the bench is here. Practically every week we learn about a new startup or an established technology firm with an office near ours. My most recent discovery was VMware (NYSE: VMW)—the virtualization company I usually describe as “a Palo Alto, CA, subsidiary of Hopkinton, MA-based EMC.” Turns out VMware has a 150-strong research and development laboratory right in the Cambridge Center complex, a stone’s throw from the Kendall subway station and just upstairs from Google’s new Cambridge spread.

VMware has been in the spotlight for months, thanks mainly to a spectacular August IPO and the seemingly unstoppable stock climb that followed. (Share prices peaked at almost $125 on Halloween, but have since returned to the much more earthly $40 to $60 range.) But here at Xconomy we’ve also cast a critical eye on the company’s long delay in issuing a patch for a critical software vulnerability affecting three of its workstation virtualization products. A fix was finally released on March 18, more than five months after Boston-based security firm Core Security notified VMware about the problem.

I’d spoken with several company officials for those stories by phone, but recently I got an invitation to meet some of the folks at VMware personally. So I headed over to the Cambridge facility and had breakfast with senior director of R&D Julia Austin, who is the lab’s site director, as well as director of product management Ben Matheson. They gave me the lowdown on the lab’s activities, which focus on advanced product development—“visionary work looking at new architectures and solutions not core to VMware today but where we think there is a business opportunity three to five years out,” in Austin’s words.

VMware, whose software helps companies save money on computer servers by allowing multiple operating systems to run on the same machines, opened a small R&D shop in Kendall Square shortly after EMC (NYSE: EMC) acquired the company in 2003. About three years ago, the operation moved to its current location, a 50,000-square-foot space on the tenth floor at Five Cambridge Center. (Soon the lab will swallow up an additional floor and double its capacity to 300 staff members, Austin says.)

Siting the lab was a no-brainer. “Being right here in Kendall Square has been a phenomenal opportunity for us,” says Austin. “Being across the street from MIT, and just down the street from Harvard and BU and Northeastern, and not too far from Columbia and Carnegie-Mellon, has been very helpful for us, for recruiting and growing our academic programs.” In addition to its core staff, the VMware lab hosts 25 summer interns from local universities, as well as visiting researchers such Richard West from Boston University’s computer science department as Larry Rudolph and Saman Amarasinghe, both from MIT’s Computer Science and Artificial Intelligence Laboratory.

While there are at least three advanced development projects underway at the Cambridge lab, Austin and Matheson could only speak publicly about one of them—a system for disaster recovery that’s currently exiting the development stage and becoming a real product.

In a post-9/11, post-Katrina world, “the ability to recover from a disaster—natural or man-made—is one of the biggest problems across businesses of all sizes,” says Matheson. “Say an earthquake hits Boston. [Which isn’t as implausible as it may sound—the city was badly damaged by a 6.0-to-6.5 temblor in 1755.] An IT administrator hits the proverbial big red button, and our site recovery manager is going to start all of the servers at your backup site in the right order, in a very fast, reliable way, so that you can recover really quickly.”

Virtualized systems actually lend themselves to being transplanted from an operations site to a backup site, since virtual machines look pretty much like any other file—the software running on them can be picked up and moved from one server to another with little or no downtime (assuming, of course, that the data the machines need to operate upon is also being regularly copied to the backup site; data replication is a separate technology that most big companies already pay for).

But orchestrating all of that in advance and making sure it happens seamlessly when the need hits is the tricky part, and that’s where the skills of VMware’s Cambridge engineers came in. “It’s easy to make things complex, but it’s often hard to make them simple,” says Matheson. VMware’s site recovery software—which is being beta-tested right now, and should be commercially available in the second half of this year—lets IT administrators specify in advance which applications are most critical and should be restarted first, so they don’t have to worry about prioritization during a crisis. “To be able to push one button and restart your entire data center in a remote location—that’s elegant and requires a lot of work behind the scenes,” says Matheson.

One of Austin’s responsibilities is to manage her engineering teams as efforts like the site recovery manager evolve from development projects into full-scale products; some engineers want to stay with the products, she says, while others want to transition back to R&D. But another big part of her job to make sure that everyone who comes to work in VMware’s Cambridge facility recognizes it as part of the larger company.

“When people come here or visit from the West Coast, we want them to feel like they are at VMware, in terms of the culture, the look and feel,” Austin says. That includes replicating traditions from the company’s Palo Alto headquarters, including a Wednesday company lunch, weekly socials and tech talks, video game tournaments, and free snacks and drinks (along with an on-site gym where employees can burn off the snacks and drinks). Frequent video conferences keep the Cambridge staff connected with their counterparts in California and Europe. “My job is to make sure that [the lab] works, that it’s effective, that people don’t feel like they’re stuck on an island,” Austin says.

But that hasn’t been a problem, given the number of IT-related seminars, lectures, club meetings, and collaboration opportunities available around Kendall Square. Because they’re all working with the same university computer-science departments, VMware employees frequently cross paths with peers from Google, Akamai, IBM, Microsoft, and other big software companies with Kendall Square offices, Austin says. And the lab’s urban location has been a big plus, she says. “Our employees really feel like they’re in a technology hub. And Kendall Square has evolved quite a bit in the last three to five years; it’s created a technology presence right in the center of the city as opposed to out in the suburbs. Our younger employees, especially, are thrilled to be able to live in the city and stay local and have us be at their back door.”

Speaking of back doors—I couldn’t leave VMware without asking Austin about VMware’s procedures for patching software vulnerabilities. Those aren’t the Cambridge lab’s bailiwick, so Austin couldn’t speak about the specific workstation software flaw I covered. She did acknowledge, though, that “we weren’t as fast as we could have been” in issuing a fix for the workstation vulnerability.

“We take security very seriously, and priority number one is sending out patches as soon as possible,” Austin says. She adds that VMware is investing in a new update management system that allows the company to simultaneously patch both the “hypervisor” software at the core of its virtualization technology and the guest operating systems, such as Windows Server 2003, that run inside virtualized machines. “It’s something we’re definitely focused on doing better.”

Wade Roush is the producer and host of the podcast Soonish and a contributing editor at Xconomy. Follow @soonishpodcast

Trending on Xconomy