Thomas Bittman

It seems like some hard lines have been drawn in the market over cloud computing versus on-premises computing. On the one hand, the proponents of cloud computing are promoting a massive shift of software development toward cloud platforms, designing for multitenancy and massive scale. No more software packages – just buy software as a service. How can you possibly compete on price with service providers leveraging huge economies of scale?

On the other hand, those defending traditional IT are pointing out the many and glaring flaws in today’s cloud computing services, and lining up behind the private cloud computing bulwark. IT vendors interested in maintaining the enterprise business they enjoy today are also co-opting the private cloud computing concept as only a slight modification of what enterprises have been doing all along.

This cloud computing argument is a lot like belly buttons. You’re either an innie or an outie. The unfortunate thing is this has led to some serious investment in navel-gazing.

Perhaps a dose of pragmatism would help. Or lint removal?

The cloud computing services needed to deliver the majority of IT services needed by customers do not yet exist. There are limited SaaS offerings today, service-level requirements can’t always be met, glaring security holes exist, regulatory compliance requirements haven’t caught up with technological capability, cloud providers tend to be proprietary and monolithic – just another opportunity for lock-in.

On the other hand, private cloud computing services cannot have the economies of scale that many large providers will enjoy. The complexity and speed of technology change will be hard for any internal IT organization to handle, especially smaller ones. The investment needed to build a private cloud service may be immense, and the resulting architecture could be a dead-end.

Gartner believes that there are quite a few services available today from cloud computing providers that are ready, and cloud computing will gradually fill more and more computing service needs. Where opportunities exist for a business model, cloud providers will fill those gaps – even when the number of potential customers for a service range in the hundreds, rather than the millions. Brokers and interoperability standards will emerge. SLA and security guarantees (for a price) will evolve. And don’t forget the completely new services that will emerge because of cloud computing and scale. It’s coming.

But it may take many years for some services. So while it is evolving, private cloud computing may make sense. It’s a question of ROI, not religion (or belly button architecture). So how you build your private cloud services matters – it’s important to build it with relatively rapid return on investment, in a way that eases a migration to external cloud computing at some point in the future. That point could be very far away. Or a hybrid model (both private and public) for some services might make the most sense in the future.

And there will be some services where cloud computing won’t ever make sense – perhaps there just won’t be a business model for a service provider, or perhaps it is simply too customized for a specific business – a real differentiator. Or perhaps it changes often, and doesn’t easily fit a concept of standard and relatively static interfaces. I think this kind of service will be a minority, but it will exist.

So hopefully we can raise our eyes from our navels, and make intelligent business decisions that will embrace the cloud computing concept as a part of our toolbox, and build a rational evolution to take advantage of cloud where and when it makes sense. And cents.

4 Comments »

Category: Cloud Future of Infrastructure     Tags: cloud computing, Future of Infrastructure, private cloud

Share

Tweet