The Lure of (Really) Remote Infrastructure Management

"Where on earth are you?" is one of the last questions IFC's Piatt says he asks of a potential infrastructure provider. He wants to assess their domain knowledge, their experience managing the technologies IFC has chosen and the work they've done for similar organizations. "Electrons need no visas," says Piatt whose full portfolio of infrastructure support is done by a combination of onshore, offshore and nearshore resources, both in-house and outsourced. "Anyone can do this work from anywhere," Piatt says.

Even before Piatt arrived at IFC last year, the company was on the bleeding edge of remote infrastructure management adoption. The finance arm of the World Bank had signed a deal with Satyam to set up an offshore network operations centre in 2003. It was successful enough that "we went through all of our services and catalogued what else could be done from Chennai", explains Zafar Azhar, IFC's information officer, who manages the company's offshore relationships. "The only thing we won't let them touch is our production servers."

Piatt and Azhar aren't alone. One reason offshore infrastructure management makes perfect sense to some CIOs is that the technology for managing and monitoring IT operations is relatively standard. Furthermore, whether answering a help desk call or monitoring a server, processes like ITIL for managing services and Cobit for governance give service providers "a common language", says Gartner's Potter.

When Bill Maguire signed on as CIO for Virgin America in 2006, he knew he wanted to offshore infrastructure management. Given that the airline was in start-up mode, he didn't expect to have a big staff. Rather than fight to increase IT headcount, he opened up the bidding to US and Indian vendors. "You have to consider the big guys who are good at that stuff," says Maguire, who had past experience with CSC, IBM and EDS. Ultimately, he hired Cybernet-SlashSupport (CSS), which, although it is based in San Jose, California, manages Virgin America's production applications, infrastructure and help desk from Chennai. IBM could have done the work offshore, too. But Big Blue cost three to four times more and, says Maguire, wanted a bigger piece of the airline's infrastructure pie. Then there was the issue of leverage. "We don't have a huge operation," says Maguire. "We'd be like a flea on an elephant's butt [with a big vendor]." Maguire also thought CSS was better able to support Virgin America's largely open-source shop. For others, a legacy provider proves the better choice for offshore support.

Nortel's Bandrowczak has a NOC in India, run, not surprisingly, by Nortel Services. The company outsources other infrastructure support, including system management to HP and database administration to CSC, which provides the services from points around the globe. "Absolutely I'd consider a Wipro or Infosys or Tata. They're all great partners," he says. "It's just that the system integrators have a leg up. They know their own equipment. And we've got a lot of HP equipment." Like most companies that outsource infrastructure management to an offshore provider, Virgin America still owns all its IT assets. Recent US acquisitions aside, most India-centric providers remain selective about buying their customers' equipment or hiring their staff.

That's fine with Piatt. In a traditional "asset-heavy" deal, unexpected changes in technology require renegotiation. "If someone else owns all those assets, each time you want to make a change to the network topology or server configuration, they have to reprice that contract. It's a big administrative discussion rather than a technology decision," says Piatt. With technology cycles shortening, having to bring in the lawyers every time you want to make a change "may make it difficult to nimbly respond to market opportunities", says Jim Meadows, partner in the global technology and outsourcing practice at law firm Hunton & Williams. Say you're a midsize company with Intel servers and you decide you want to move to Linux. "If you signed away your assets for seven years, you're out of luck," says Everest's Tisnovsky. "That deal has a technology road map and a refresh schedule that the vendor has amortized over seven years."

With an asset-light remote infrastructure management deal, the customer retains control. "IBM has built up very rigid operational procedures and processes over the years," says Maguire. "If I need to make a change in two hours, and I'm dealing with the big boys, it would be difficult given that overhead." Not that dynamic IT operations don't create challenges for CIOs who send infrastructure support offshore. It can be hard to keep the distant team up to speed. "As you're making changes, you have to train the remote people. Training someone 12 hours away takes much more effort," says Maguire, who ends up doing a lot of teleconferences and Webinars with the offshore team. Sometimes a key CSS employee has to hop a plane from Chennai to California to learn new systems and go back to train the rest of the team.

After enduring some hiccups with Satyam early on, IFC had to bring key members of the offshore team to Washington for six months to be trained on its systems. IFC's offshore manager Azhar also learned that offshore support personnel tend to be less experienced. Some skills - like Citrix expertise - are hard to find abroad, "but all in all", Azhar says, "we've still been able to pull this off at a lower cost", IFC has even started to give the Satyam team special products like designing network infrastructure.