In moving to a Platform as a Service engagement, it's best to look as far into the future as possible.
Stephanie Mann and Jack Vaughan over at SearchSOA spoke with Geva Perry, former general manager of the cloud computing business unit at GigaSpaces Technologies, and just put together a brief (a brief brief) on questions that need to be asked before signing on with a Platform as a Service vendor:
- What languages are supported?
- How much scalability can be achieved?
- How easy it will be to move applications off the PaaS in the future?
These are critical areas that need to be addressed in any cloud agreement, and the third point -- ease of migration when it's time to switch -- is probably the most critical question of all.
All the work that's been done in service-oriented architecture over the past decade was intended to free up enterprises from any type of vendor or technology lock-in. Unfortunately, cloud has the potential to take things backward in that regard.
A few months back, I discussed this point with best-selling tech author Thomas Erl, who points out that companies need to look at service-oriented approaches -- and what has been learned over the past decade -- before taking on a potentially entangling and silo-creating cloud engagements. “With cloud environments, its kind of a new level of lock in. You can have your application. It can be standards-compliant for certain interoperability functions. But the actual hosting of the application, the actual requirements for that application to exist in a cloud environment, to connect to the virtualized resources and whatever administration tools the cloud providers may give you to configure and maintain the application, will be, for the most part, controlled by the cloud provider.”
A lot of these integration and governance issues have already been worked out within SOA efforts in recent years, Thomas says. For companies contemplating cloud, "the SOA community has already done the work for them," says Thomas. "Everything's there and documented now, in terms of models, patterns, and principals and best practices."