Don't just focus on technological challenges when considering moving internally delivered IT services to the public cloud. Instead, start with the business case--if it doesn't add up, no amount of technological wizardry will make cloud the right choice. If building that analysis is tough, think back to how you talked to business leaders about moving to virtualized hardware, which was a fairly significant technological and operational leap. From there, the move to infrastructure-as-a-service is, in some ways, a much smaller change. You can run virtual machines on a provider's infrastructure the same as you would internally, but you've now outsourced hardware- and facility-related headaches.
But moving to the public cloud also comes with risk. For one, your service provider may not offer the same networking and storage features that you had in your own data center. We've seen companies lose a significant amount of operational control around continuity and disaster recovery, and when cloud outages occur, internal IT can do very little about it even as employees or customers scream.
Then there's the financial picture. Data center costs are well understood, and CIOs can be reasonably confident when estimating budgets for new projects. But companies that haven't used cloud providers in the past may be dazzled by visions of low-cost, pennies-per-hour charges for virtual machines and then suffer sticker shock as usage ramps up or new services get layered on.
Security and compliance concerns also persist, as we saw in our InformationWeek 2012 Cloud Security and Risk Survey. Will our data be safe in the cloud? Who can get access to it? Where, in both a physical and logical sense, does my data actually reside?
It's no wonder that some CIOs approach migration to the public cloud with trepidation.
On the other hand, Amazon CTO Werner Vogels isn't completely off base to classify CIOs who flatly refuse to venture into the public cloud as "dinosaurs." While self-serving, Vogels' criticism isn't unwarranted because the public cloud has significant benefits compared with in-house virtualization for many use cases, perhaps most compellingly around dynamic workloads and disaster recovery. Thus, though they do need to proceed with caution, CIOs must at least consider the public cloud as an application platform. Start by asking these nine questions.
What's our business reason for moving to the public cloud?
The purpose of IT is to serve and support the business, so any significant technology migration plan should start with a clear definition of why the company as a whole will be better off as a result of moving a given service or application to the public cloud. Some of the most common justifications are increased scalability or high availability, improved disaster recovery or business agility, and infrastructure and staff cost savings.
For example, a consumer-facing retail site that experiences predictable bursts of activity--say, back-to-school or holiday shopping--might benefit from a provider's ability to scale resources on demand, rather than building the internal capacity to handle peak loads that occur a few times per year.
Don't move an application to the cloud just because you want to experiment or because someone has decided that you should have something "on the cloud." We've seen more than a few IT decisions heavily influenced by factors that are not supported by business justifications.
Download the InformationWeek September special issue on the cloud