An increasingly cloudy outlook

You're going to hear a lot more about cloud computing, but the marketing people seem to be determined to - ahem - fog up what cloud computing is and what it means for businesses large and small. More importantly, cloud computing is going to have an impact upon a wide range of telecom-businesses over the next five years.

Part of the problem is that cloud computing has some of its marketing muddling roots in the original concept of grid computing. Grid computing realizes that a lot of computers large and small have idle cycles, so if you have a bunch of desktop and servers sitting around late at night doing nothing more than warming the building, you can use software to harness all that excess compute power to work on big complex problems, like looking for new drug combinations, searching for ET or cracking codes. Very high geek, very boring stuff.

Meanwhile, we had the Internet - the "grid" writ large. Folks like Amazon are building large server farms for ecommerce and getting them powered up in the six weeks leading to December 25, which will be followed by a significant drop-off in capacity needed. In engineering diagrams, the Internet is a big "cloud" and you don't know - or want to know - what's behind the curtain.  Instead, you just want the services to work when you need them.

Amazon (and to be fair, others) figured out if you had to build web services to scale up and down, you might as well do it in a relatively generic fashion, so you can put whatever you need up and have the ability to scale capacity to peaks and demands.

Tada! Cloud computing!

Put more simply in <ahem> my definition, cloud computing provides a platform/service for hosted applications that can easily scale up or down. Hosting on steroids, basically. Annnd, you're able to sell the services to others. If you're operating a solution in-house, it's grid computing in my book, but you'll likely call it cloud computing because you want to use the latest buzz words.

Why is this important for telecom? Three reasons:

1) Big companies running hosted services now will likely offer cloud computing services in the future. They already have the data centers and the infrastructure, all they need to do is to start off with a row full of racks and "scale" upward. They can squeeze more money out of existing hardware and infrastructure while (potentially) being able to drop prices for more attractive SMBs.

Expect AT&T and Verizon to offer cloud computing services. Maybe IBM and other services establishments.

2) Companies running hosted applications (i.e. they rent servers in a machine room from the guys in 1) and throw their apps on top) will want to look at cloud computing services if they have peaks and valleys in the demand for their applications; in other words, either seasonal events (Christmas) or major activities (elections, Super Bowl, Final 4) that result in heavy demand for their services - but not all the time. Companies currently running stock applications with a fairly predicable and steady rate of growth will stick with what they know. 

Color your typical CDN as looking at cloud computing

3) Start-up companies and other new ventures will tend to drift to cloud computing, because it allows them to buy what resources they need when they need it, rather than having to plunk down either large capital expenditures to build their own servers or buy a hosted-server-at-a-time step function without a way to scale up and down more easily.

We're starting to see telecom apps appear on clouds already. Sylantro has its apps running on Amazon's Web services and Microsoft's Tellme division operates its own in-house cloud.  Hosted UC and call center service will most likely have a cloud engine driving them.

It's not going to happen overnight, but the clouds are coming.

- Doug