- More attitude from customers.
- On-call 24/7.
- Scale can be intimidating.
- Revenue generation part of the job description.
And third, "you now have to think about how to maximize your P&L impact by increasing margin and lowering cost," says Casusol. Specifically, "you need to think about scale, about deploying processes and tools that deal with volume," he explains. "You become a business within a business."
For those reasons, Casusol says, it may be difficult for traditional IT employees to make the transition to working for a cloud service provider. "You need to adopt a customer service mentality," he says. "Traditional IT is not looked at as customer-friendly, especially when they're enforcing policies. Going from being [a controlling influence] to being customer-friendly requires a cultural change."
Pat O'Day, Bluelock's CTO, echoes Casusol's last sentiment, as did other IT professionals. He spent 10 years as a Web services manager at a hospital, and then a little over a year in pharmaceutical manufacturer Eli Lilly's Internet group, running the security infrastructure. For him, traditional IT wasn't always fun. "I'll be completely honest: Sometimes we felt like we could almost hear the grumbles of people cursing IT," he says. "Is there more gratitude on the service-provider side? Yes."
High stakes
If there's more gratitude, there are also higher stakes and more demanding customers. Users choose the service provider they want to patronize, O'Day cautions, and they're free to choose to take their business elsewhere -- which will have an immediate impact on the cloud provider's bottom line.
"With a service provider, people get enraged if the network goes down for an hour," says O'Day. "You have to give 110% service. The way you behave -- how quickly you call people back, how quickly you address their issues -- affects the health of that customer relationship. There's a stress factor to consider."
Beyond that, service providers work on a much larger scale than most traditional data centers -- they may have as many as 2,000 machines in one cluster. Working on a platform of that scale can be intimidating, O'Day says. You don't want to screw up 2,000 machines in one shot, he warns.
But that scale can also be invigorating. Ken Owens spent the early part of his career doing various kinds of network architecture for financial services and telecommunications companies. Now he's technical vice president of security and virtualization technologies for Savvis, the cloud services arm of Internet and telecommunications provider CenturyLink.
