Fidelity’s Gamble With Server Virtualisation Pays Off

Summary:

When the I.T. head of Fidelity Business Services wanted to virtualize his servers—at a time when a few were—two forces defied him: The recession and management’s resistance. he had no choice but to prove them wrong.

Highlights:

server virtualization has brought down the number of physical servers, to 660 from 1,000The average server utilization has now gone up to 15 percent from two to three percent.

Reader ROI:

How contrafactual thinking can helps an individual make the right choicesHow server virtualisation can boost cost effeciency

Three years ago, Amit Gupta, head-technology infrastructure services, Fidelity Business Services, believed in going with the flow.

That’s why when the IT industry followed a ‘one application per server’ principle, he followed suit. “Owing to our emphasis on business continuity, we were actually working on a principle of ‘four servers per application’,” says Gupta.  Thanks to this principle, the company was sitting on an island of servers.  It had now created a huge server farm with scattered and underutilized computing resources. With over 1,000 servers—and an abysmally low average utilization of two to three percent—the company’s infrastructure was rigid.

That wasn’t all.  With the recession looming large, Fidelity’s underutilized IT infrastructure was beginning to put pressure on its finances. It was obvious to Gupta that the existing infrastructure left much to be desired.

Breaking the Mold Abiding by industry standards wasn’t helping Fidelity. And if he had to bring some sense back to his infrastructure, Gupta needed to do something different.

Like going for—the untried and untested—server virtualization. But the problem was, back in 2008, the technology was just an idea. According to IDC, less than 20 percent of Indian CIOs were implementing it at that time.

That was not the only problem. In recessionary times, organizations shy away from spending on what they deem ‘nice-to-have’ technology. And for most of them IT falls in that category. Getting management to fund the project wasn’t going to be easy. 

Gupta knew all of this. He had been watching virtualization solutions for the past two-three years. He knew that if he consolidated the server platforms he would be able to harness the investment into existing computing resources. With that insight in mind, Gupta decided to take the plunge. “I was convinced that I should place my bets on virtualization,” he says.

Though, it might seem obvious now, it wasn’t that easy to see at a time when there were no instances of server virtualization. But Gupta was still able to take a calculated risk. And that’s because Gupta is, what experts say, a counterfactual thinker.  Counterfactual thinking involves imagining what might happen. People equipped with this trait have the ability to imagine alternatives to past or present events or circumstances. This is usually done by imagining what-if scenarios.

“Counterfactual thinking heightens an individual’s ability to make the right choices,” says Laura Kray, associate professor, Hass Business School, and an expert on management philosophy and values. But knowing what Gupta wanted to do was only half the battle. Without getting the business—and funding—on board the project was as good as dead.

We had shivers running down our spines during this phase. Even a miniscule deterioration in performance could trigger resistance from our business teams

Winning Them Over Gupta was putting forth the idea of a transformative technology and hence he would have to justify every single penny to management. And that doesn’t get any easier if he was banking on a technology he hadn’t seen work. As Gupta had anticipated, management did not appreciate the value proposition of his initiative. “It is easy to get funding for a business project but in this case, I had to spend a lot of time explaining how business would also accrue numerous benefits. I did not want them to feel as if a black box was being pushed to them,” he says.

But it was also easy to see why the management was against the project. This was a complex implementation as it would affect all business critical apps. Even a small glitch could lead to serious business disruption. “Any business disruption could dent the organization’s reputation. The stakes were very high,” says Gupta.

Fidelity’s business is regulated by many overseas financial regulators. In the event of any business disruption, they would be penalized by the regulatory bodies.  “One wrong move could have a direct impact on our business,” he says.

But Gupta wasn’t going to let that bog him down. He told management that he would migrate the low-risk platforms to the virtual environment in the first phase. He presented a detailed cost-benefit analysis to them and impressed upon them that within two or three years the project would fund for itself.

Coming up with these arguments, says Kray, is natural to counterfactual thinkers. “They are more motivated and analytical in organizational settings.”  That was proved when Gupta’s astute moves finally bore fruits. “By speaking the lingo of cost optimization during the slowdown, I had hit the nail on the head.”

We had shivers running down our spines during this phase. Even a miniscule deterioration in performance could trigger resistance from our business teams

Taking Risk Down It was time for Gupta to deliver and ensure that the project was worth the risk. He began by migrating the organization’s low risk applications. “We had shivers running down our spines during this phase. Even a miniscule deterioration in performance could trigger resistance from our business teams,” says Gupta.  But in case of a mishap, he was prepared to roll the system back.

 He didn’t have to. The first phase was delivered ahead of the original schedule with no cost overrun. This stoked Gupta’s confidence. “I felt rewarded. I felt as if I was near the final goalpost,” he says.

But Gupta still needed to make sure that he doesn’t alienate his stakeholders. Because for any first-of-its-kind project, change management—especially for the users—is always a party pooper. Gupta knew that if he forced the project on the users, it could backfire.  “So, I encouraged a participative culture. There was proactive communication on everything related to the project. People had the trust and confidence that we were including them not just for the good news but also for the not-so-good news,” he says.  Now that Gupta had put everyone’s apprehensions to rest, it was time to reap    the benefits.

Today, server virtualization has brought down the number of physical servers, to 660 from 1,000. This, says Gupta, has trimmed maintenance cost by 20 percent and power consumption by 30 percent. What’s more, the average server utilization has now gone up to 15 percent from two to three percent.

The project has also increased availability of space within the datacenter by 20 percent and reduced provisioning time by 60 percent. And these numbers are a testimony to acknowledging that sometimes trying a different tack is worth the effort.