Computing in the "cloud" or "cloud computing" is the world coming full circle. When IBM, DEC, Tandem and Cray started out, they built room sized computers that were so big and ate so much energy, they required their own cooling systems. The programmers all worked directly for the manufacturer making software custom for each customer. The costs were staggering in today's dollars, and somehow the productivity gains made it all worth it. The First Main Frame that ran Bank Of America's ATM system throughout California had less memory and horsepower than today's $299 netbooks.
When the PC arrived on the scene, big companies started moving to mini computers and the client-server systems were born. Most of us are familiar with the client-server model and don't even know it. If you have an email program that connects to a server to get your email you are using client server computing. The difference here is even if you are not connected, you can create the e-mail and wait to send it until the next time you connect. Your computer in this case is the client and you trade or update data with a server. The process of trading or updating data is called synchronizing or "syncing" for short.
Cloud computing using netbooks goes back to the old mainframe model. the difference here is the software companies like Oracle, PeopleSoft and Google are in charge, not the hardware people like IBM, DEC (gone), Tandem (gone) or Cray. With cloud computing, if you don't have an internet connection to access iGoogle or MSN, you can not even write the e-mail. All of the computing power and software is really on their "servers", not your little netbook. The "servers" at these companies are really clusters of PC's with software designed to make them work together nicely and share the horsepower or processing power. Mini clouds within a company for applications makes sense, one computer gets the upgrade, everyone has new software.
The challenge for cloud computing is still access to the "cloud". Internet access isn't everywhere, and where it is, the access isn't always free or fast. If you don't have service, you really can't work or communicate. If your provider doesn't have service in a specific area, it is another challenge in itself.
For instance, if you fly two different airlines with a stop for lunch at an airport not served by your current provider and cell service is slow or your tether software quits, you could easily spend $70 in connection fees in one day. I know because I have done it. I have tethered services from AT&T, and they quit working one day. Airline 1, $15.99, Airport day pass $9.99, Airline 2, $15.99, Airport 2 day pass $4.99, Hotel day pass $14.99, day pass at area near meeting, $12.99. Yes it happens. Services are getting better but they are not that good yet.
The happy medium can't be far off though. The next step in the client server and distributed computing model could be combined with cloud theory to create a distributed sync model. In distributed sync, your data would still exist on centralized servers and on your computer or other storage device. Some parts of the software would exist on your computer, others on theirs. When you are within range of a network your computer would become part of the grid, offering computing power and network hops to everyone in range when you don't need it.
The first part of the distributed sync model holding it back is standards. If I change a document on my iPhone and my laptop before my next sync, which one wins? Two significant changes need to occur for this to work, computers need more power and networks need to have more speed to allow very fast sync for changed items only. If I change one item in my blackberry, I need to go through the entire sync process. The same was true of my windows mobile phone as well. This is a limit of sync computing that is solved by the pure "cloud" solutions.
The second hold back is network access. Some carriers charge you twice for your phone and laptop, others let you tether. They don't have interoperability agreements so when I am out of range of my carriers signal like today, I am dead.
The final question is of course the data. When "they" can sift through all of your e-mail, word processing, spreadsheets and presentations, what are they learning? Facebook recently agreed they didn't "own" the data, but they also didn't say they wouldn't "use" it while it was on their servers either. Scott McNealy, a guy I consider a pioneer in the silicon valley and on the internet said some time ago that privacy was dead, give it up. Should we?
Cloud computing is a great idea for a world under constant change. The hardest part is keeping up. This morning we found our Google Client Center and our Skype Business Center interfaces were completely different. We lost a lot of productivity learning on the fly with customers on the phone. Cloud computing in some form is here to stay, the mainframe and terminal just look different.