Be careful of the cloud providers!

I was drawn into a fascinating discussion around “price shocks” by the cloud service providers (like Google/Amazon) who have a stranglehold over the market for enterprise cloud computing. One of the back links went as far as 2011 to The Register:

On Wednesday, as it announced that App Engine will officially lose its “preview” tag in the second half of September, Google also rolled out a tool that lets existing users calculate how much they’ll pay under the service’s new pricing setup. Predictably, many developers are rather peeved at what they’re seeing. Google is essentially raising prices in an effort to turn the service into a sustainable business.

But this to-move-or-not-to-move conundrum has led some to question the long-term health of the service, which includes some unique APIs and uses a data model that may not easily translate to other platforms. “What has always been the biggest concern about App Engine? Lock-in. You’re at the mercy of Google,” writes one user.

This trend has accelerated. Proprietary lock ins and knee capping are eventually going to happen.

Why as a radiation oncologist am I writing about the enterprise cloud computing and app engines? Well, as the technology enters the hospitals, we need to be more conversant around what’s happening in the cloud business (Azure/Amazon and Google). They have built their businesses based on “certain criminalistic tendencies”- roughshod over what’s the greater common good for the end users and maximising the profits for their owners. It is fine as long as it is not the health data. It crushes the smaller businesses that have refused to adapt or are forced to pay a higher price for “listing”. That’s “rent seeking”.

I think the way forward for all users is to understand that freebies are over. It takes real infrastructure to maintain the services and a fair distributed “subscription” would actually lower the prices as the incentive to push through price shocks will be minimised. Besides, it is essential to create health developing options with data portability. I’ll keep these lessons in mind.