'Cloud Computing' holds dangers as well as promise

Way back in the 1960's, when computers were still great big monstrosities that occupied immense rooms, people wanted to move from 'batch' to 'online' or 'interactive' computing. There were no personal computers at the time, and very few computer networks. What developed was a computing technique where data and programs were stored and executed on a central mainframe computer, with 'dumb terminals' (basically video screens with keyboards, but with no local processing or storage facilities) connected via a network to the central system. I programmed on such systems for several years, including developing online 'real-time' computer systems using software such as CICS and programming languages such as COBOL. (Boy, have I just dated myself . . . )

With the advent of personal computers in the late 1970's and early 1980's, users wanted more local control over their computer systems. Gradually a 'distributed processing' model of software evolved, where data would be drawn from a central database or repository, but processed locally by programs running on a personal computer. Many office networks function like this at present; for example, a secretary may call up a document from the network file server, but edit it on her own computer using a local word processing program, then save the amended document back to the network file server. Furthermore, data that was the property of an individual user would often be stored locally, without ever putting it on a central server (except for purposes of backup).

The Internet accelerated the distributed processing model, by making information and consumer applications available to millions of people. However, there was no central storage for users' data - it had to be stored locally. Also, it was expensive (at first) to buy software programs to perform all the functions users wanted on their own computers, such as word processing, home accounting software, and so on. Increasingly, companies began to try to provide such programs on the Internet, so that users could run them over the network but still store their data locally.

The latest development has been so-called 'cloud computing': storing all one's programs and data on a central server, and accessing it over the Internet whenever needed. Google's Gmail is such an application; all your e-mails, incoming and outgoing, are stored on Google's servers. Unless you tell your e-mail software to download local copies, you won't have any - they'll all reside on central storage facilities. This is quite convenient when it works . . . but it doesn't always work, as a number of companies and their customers recently found out to their cost. If the cloud servers go down, or Internet access is disrupted, suddenly you can't get to your data or your software. You're stuck.

Now Apple has launched a 'iCloud' iTunes service, where you can keep all your music online, accessing it on any device that can connect to the Internet, but not keeping it locally. I call this just another recipe for disaster. What happens if the Internet goes down for a prolonged period? There are several ways this could happen, from natural disasters, through accidental damage or cyber-criminal activity, to cyber-warfare from potential opponents. I certainly don't want to lose all my accumulated music, let alone my data and programs! I'll be keeping them all firmly on my local computer, thank you very much - with multiple backups, too!

I understand that 'cloud computing' is very convenient, and it's a great thing from the perspective of companies like Amazon, Apple and Google who make billions of dollars in revenue from it. From the perspective of individual users like you and I, though, there are also some significant disadvantages, and some very real dangers. I plan to avoid it, as far as possible.

Peter