Spring in Practice

Willie Wheeler's Spring blog

Thoughts on Cloud Portability

| Comments

A colleague sent me a NY Times story on cloud portability today.

There’s lots of interest on both the consumer and provider side in cloud portability. Consumers want to avoid vendor lock-in and providers often want to lock customers in. Among the lock-in approaches are the following:

  • Technology-based lock-in: Offer compelling technology that other providers don’t offer, or technology that’s hard to migrate away from.
  • Data-based lock-in: Make it hard for customers to move data from one provider to another.

Those are tried and true approaches, but historically they have been overcome. For example, providers copy each others’ technology and create migration tools. Industry defines standards to commoditize individual offerings.

Where cloud portability gets interesting is on the data side. There are two types:

  • Data the customer owns
  • Data the provider owns, and offers up as a feature of the platform

Provider-owned data is inherently unportable. Or at least it sure seems that way to me.

To the extent that customers depend on provider-proprietary data, there is a real lock-in danger. I think that’s one of the major reasons that lots of companies are excited about the idea of providing cloud services. Build a platform with compelling data that nobody else has, and you have a powerful lock-in mechanism.

Any organization contemplating moving applications to the cloud needs to consider how it wants to deal with this sort of lock-in risk.