I was torn about the title of this entry. Should it be Sting’s “Heavy Clouds But No Rain” or should it be Peter Allen’s and Carole Bayer Sager’s “Everything Old Is New Again?” I settled on the former. But either would work, as we’ll see!
The New York Times published no less than three very well written pieces on Cloud Computing from April 14th to April 18th! So, everyone knows that something is up in the world of cloud computing. But is there? What exactly is going on? The Times articles point out that traditional business models are threatened, and that companies “have concerns” about moving to the cloud, while new innovating startups jump on it immediately.
All true. So, let’s get a grip on the cloud, shall we?
First of all, what we need to do is define what this cloud thing really is: It currently is three things: Infrastructure-as-a-service, Platform-as-a-service, and Software-as-a-service.
Infrastructure-as-a-service is designed to eliminate the need for you to have and maintain data centers.
Platform-as-a-service is designed to provide you with a “place” to develop and run your own applications.
Software-as-a-service is designed to provide you with access to applications that are developed and run on the vendor’s data centers.
Wait a minute! Didn’t we already have all this? Absolutely. We called it mainframes, and time-sharing bureaus (service bureaus). Back in the day before internet (yes, little Timmy, there was a world before Internet) computing was centralized in large data centers. Clients would use computing resources and pay accordingly. Similarly, there would be service bureaus that would develop and run applications for specific industries, for example accounting or fulfillment applications. You would rent the use of this application, send your data over, and get back your reports. Ahh… those heady days of magnetic tape and printouts.
Hence, Everything Old Is New Again. So, what’s the big deal?
The difference here is how we are getting back to where we started. In the beginning, shortly after the “Let there be light” computers were slow, expensive, and difficult to maintain. Computing had no choice but be centralized. Then, just as the oceans settled down, the computer evolved and shrunk. Before you knew it those silicon chips were crawling all over your desk. Computing became cheap and easily accessible. No need for centralized data centers and the like. Evolution killed them. And, just as you were feeling safe to go back in the water, the Internet connected all those chips. Having them crawl all over the place became… bothersome. How about we use the internet and send all the chips to the cloud and have them talk to us from there? Great thinking! Off with their clamshells.
Now, no one said that evolution will be easy. For one, you have the clients that rightfully are paranoid about this clouds availability, reliability, and security. They are right. Vendor claims have not always been shall we say “entirely accurate.” On one hand they’ll sell 99% uptime, and on the other there will be internal memos saying that they’ll staff to 90% uptime because its cheaper to pay the penalty than to actually deliver to the promised service level agreement. Yeah. I’d be concerned too!
Then you have the software vendors themselves. The major players will license you software for a nice seven-digit-number, and then milk you for 20% annual maintenance fees. Plus, a whole army of implementers and integrators are there to leach off this transaction. They help you install, customize, train, etc., etc. Do you for one moment think that these constituencies are jumping up-and-down to move you to some cloud for $19.95 per user? Think again. When they do, you can be sure that the number will be substantially larger and that the aforementioned armies will have a special place in the cloud ecosystem.
And I don’t even need to tell you how all these nice people that actually build and sell servers and data centers feel about the cloud, do I? Good.
Where does this leave you? In the near term, and if you can accept the associated risk, compromises, and vendor growing pains, then move as many non-mission-critical applications to the cloud as possible. It will save you money and reduce your headache. In the longer term, start planning for a wholesale migration to the appropriate combination of services for your business. You still have to accept compromises, but those may be good for you. Especially if they start saving you those seven digits we talked about earlier.
Now, that’s rain!
