Monday, 25 February 2008

“I think there is a world market for maybe five computers”

Quote attributed to Thomas J Watson President of IBM in 1943 (It is reckoned to be one of more famous ‘misquotes’ though. For more information see Click here)

I been writing about SaaS or Utility computing or Cloud computing or whatever the latest name goes by, since 1995. Indeed even readers who only got into the Holway’s HotView habit in the last year must be pretty familiar with the theme by now.

Yet another article on the subject Microsoft predicts the rise of the datacentre in the FT on 25th Feb, caught my eye. This time it is Microsoft saying that it will build a giant datacentre of incredible power. We already know that Amazon, Google, Intel and IBM are in various stages of planning/building such datacentres.

It all reminded me of one of the quotes that is often used to demonstrate the fallacy of making predictions. In 1943, Thomas J Watson, President of IBM, supposedly said “I think there is a world market for maybe five computers”. Ha, ha, ha – how we all laughed. Even though the prediction was actually true for about ten years after it was made! Almost every forecaster would settle for a ten year limit on the testing of their forecasts!. Of course, by the 1980s and the advent of the PC, such a statement looked plain daft.

But now you could have a really serious debate about whether, in say the next ten years, such a prediction might indeed be true. If Cloud computing really did take off we might, in theory, need just one mega computer/datacentre. Clearly competition and security issues would mean that one monopoly would not be allowed. So five seems a reasonable number!

You could then go on to muse whose will they be? Of course, that’s why the current Microsoft/Google tussle is so important. Microsoft has almost everything to lose if Watson’s prediction eventually comes true and they are not a player – having chosen instead to protect their desktop software cash cow.

Unfortunately I think it might well be Countries who control and/or “Sovereign Wealth Funds” who own these datacentres. Maybe the big global battlefield of the next 25 years will be fought on this issue.

The issue reminds me of my mixed emotions over the NHS medical records database. The logic for this is overwhelming but I am scared stiff that 800,000 NHS staff can access my medical records. Indeed, given our poodle-like relationship with the US, I could see them accessing it too. So after the usual 2 hour queue at US immigration in 2012 you could get “Why did you not declare that you went to your doctor with a drug problem in 1975?”.

That’s why the logic of Cloud computing, the ultimate Holway Martini Moment and the achievement my MobiTop, excites and scares me all at the same time.

1 comment:

Paul Wallis said...


As you say it is difficult to make predictions that remain accurate for a reasonable length of time, especially with technology changing so rapidly.

That said, I think we are some way from businesses putting mission-critical applications on “The Cloud”.

One of the problems at the moment is economics. The late Jim Gray of Microsoft analysed “On Demand” computing a few years ago, and he pointed out that it is only economical for very CPU intensive operations.

Although telecom prices have fallen and bandwidth has increased, processing power has increased much more rapidly, which means that CPUs in the Cloud will run inefficiently at the moment except for applications like, for example, image rendering.

On my blog I’ve discussed this and some of the other issues surrounding The Cloud. I’ve tried to place it in historical context, looking at the Cloud's forerunners and the problems they encountered before being adopted.

You can find the article here.

You may also be interested reading my thoughts about how “IT exists for one reason”.

Your comments or feedback are very welcome.