Technology executives are increasingly tinkering with cloud-first strategies that aim to starve data center investment, create counsels for projects and ultimately allow large companies to move as fast as smaller rivals.
Those trends were surfaced last week on at a panel at the Amazon Web Services re:Invent conference. What follows is a series of takeaways from CIOs working toward a cloud-first strategy:
Bharat Shyam, CIO for State of Washington:
Disaster recovery is a good use case to begin a cloud strategy. Many enterprises and institutions simply can't afford disaster recovery infrastructure. The cloud changes that equation.
Tackle no brainer projects first. Shyam began using Amazon Web Services to handle "bursty" Web traffic. "Our traffic increases by a factor of 10 when it snows and the site grounds to a halt. We were serving up stale traffic data," said Shyam (top right).
The more you go cloud computing the more your IT department will focus on application. "There are a lot of people maintaining servers and they have to be comfortable letting go," he said.
Cloud billing may not line up with your accounting requirements. Shyam had a billing beef with AWS. "Amazon isn't set up to get micro about allocations. The problem is that's the way governments allocate funds," Shyam explained.
Troy Otillio, cloud strategist at Intuit:
Outages can be a catalyst for a cloud strategy. Eighteen months ago, Intuit suffered through outages---led by a data center power failure. Intuit began using AWS following those outages and now has a cloud first strategy.
Keep your competition in mind when going cloud. Intuit's largest competitive set is a series of startups aiming for individual products. "Our biggest challenge is that there are hundreds of startups swarming around our point solutions," said Otillio.
Don't discount agility benefits. Otillio noted that time to market, happier product development managers and speed of innovation are overlooked benefits of a cloud strategy.
Cloud approaches require automation. The more cloud you have the more IT processes have to change. "The new model is to go completely automated," said Otillio. "We invested in our automation platform. You can't do manual."
Darren Person, CTO of Elsevier:
A cloud first strategy requires projects to spurn the urge to build your own and hug physical hardware. "Projects need to have a real good reason why the cloud isn't the right fit," said Person (check this).
Form a cloud council. Elsevier has a council that reviews cloud migrations, procurement processes, applications to be ported and security. "To have collective buy-in, you need everyone's input on moving forward," he said.
Cloud projects snowball and open up possibilities for new services that weren't pondered originally.
Disaster recovery is an obvious use case to get started with cloud computing.
Sean Perry, CIO, Robert Half International
Cost assessments are critical. Companies need to get visibility to on-premise costs to effectively make cloud choices. Track time of implementation, cascading labor costs to maintain and production requirements. "In doing assessments you don't have to be perfect, just close," said Perry (bottom right).
The biggest mistake companies can make is moving processes and applications you have now and just migrate it to the cloud. In other words, old processes just don't fit with a cloud approach. An attitude change has to happen, said Perry.
Starve the data center if cloud-first is a strategy. "We're actively not making our data center more attractive," said Perry. "It will run the way it is." Perry added that he's turning away vendors because he's in maintenance mode---not upgrade happy---when it comes to his data center.
People need change management too. Perry said that the people part of going cloud first is critical. Workers will have to be retooled and develop new skills. "Don't forget the people part of the equation. There's change management with the people too," said Perry.