Fog Computing
      by Wyatt WalterCloud computing and virtualization has been all the hype over the last few years in the technology industry. The idea of being able to start up hundreds of servers (or even one or two, for that matter) nearly instantly and having them ready to start handling loads is life-changing for systems administrators around the world. Having to trust a cloud computing provider with critical data and systems is also life-changing for systems administrators. I’m not the only one who thinks so.
Just like any new technology, there’s those who are ready to jump aboard and never look back and there are those who are more cautious and not wanting to jump only to have nothing to catch them. Both have valid points. On one hand, you don’t want to miss out on an opportunity to leverage new technology to gain an advantage over your competitors. On the other hand, there’s a reason we call it the ‘bleeding edge’ of technology. It’s definitely not for the faint of heart.
Moving infrastructure out into a cloud environment offers tremendous flexibility, but it also comes at a cost since machines have to run all the time. Amazon’s EC2 instances (which are among the cheaper ones I’ve seen) cost about $.10 an hour.. for a small instance. Let’s say, though that I have a fairly resource-intensive application. I’m going to need an extra large instance to make my application work which costs $.80 an hour, still not bad. That means I need to run my instance 24×7. To provide software redundancy for high availability, now I need to have two of those instances. Suddenly, my $.80 an hour turns into $1152 per month. This cost comes before bandwidth going into and out from your instance. Now, I know that there’s other considerations such as networking equipment, rack space, etc. but I’m pretty sure a good lease on some hardware will come under that cost.
Naturally, a good compromise has become common in IT circles. Given the cost of running the instances 24×7 and trust concerns, a lot of shops have been utilizing cloud services as a hybrid alongside their existing infrastructure. Creating new instances during peak traffic times to handle loads or using it for running batch jobs is where the real cost savings comes into play. If you can use 25 servers for 5 hours a day rather than spending money on 25 physical (or space for 25 virtual servers) in your environment and spend (25 machines x 5 hours x $.8) $100 a day, you’ve just saved yourself a ton of money and now you don’t have to deal with the headaches of managing the extra under-utilized hardware.
Employees at Amazon have called this hybrid implementation of cloud computing ‘Cloudbursting’, but I’d like to throw out my own term: ‘fog computing’. There, I said it. I’ve been watching others play on the term cloud computing in their blogs for months and wanted to join in on the action. I couldn’t resist. We can have a cloud but keep it close to the ground, thus the fog. This trend makes perfect sense, and you can see real cost savings right up front rather than having to buy additional servers to accommodate additional temporary or cyclical loads. There’s definite value in the cloud, but a nice balance can be struck by using a hybrid approach and keeping your cloud ‘close to the ground’.
Tags: amazon, cloud computing
Filed under Tech Trends :
Comments (2) :
Sep 12th, 2008
-
http://www.fogcomputing.co.uk Peter Moss
-
Greg Pierson