IT and the Business Cycle
Putting the house in order
There is a regular business cycle, which lasts for about 9 years. The cycle is characterised by a period of growth, then strong growth and then recession. Unfortunately, the cycle isn't exact and it isn't dependable, or else you could make money out of it, by gambling on it. Sometimes it lasts 7 years, sometimes 10 or 11.
In the later stages of the last business cycle some odd things were happening. Growth in the US economy was much higher than anyone expected and unemployment much lower. Productivity statistics from the US Government suggested that from 1996 productivity was improving at an average 2.2 percent per year, which was a dramatic improvement on the 1 percent average for the previous 25 years.
The point at which productivity turned up coincided with the point at which the Internet started to become visible. That may have been a coincidence, but there was another strange quirk in the figures. Investment by US companies since 1990 has been static in every area except in IT, where it rose dramatically by a factor of 14 over the decade. And to cap it all, the bulk of the productivity improvement in the US economy was confined to the IT industry itself.
In theory, IT should be counter-cyclical and it usually is. The benefit that IT is supposed to deliver is automation. It either cuts costs and/or improves productivity, accordingly. The figures from the US suggest that it was doing the latter in the glorious 90s and particularly in the later years. In the current part of the business cycle it will be cost cutting that matters. The heady days of optimism are over and the era of cautious IT investment has arrived. So what are the information technologies that will do well in this era?
Here's one thought. Consider the anomaly of Moore's Law. This suggests that CPU power will double every 18 months and it has done just that for over 30 years. Actually Moore's law doesn't just apply to CPUs, but also memory, disk, buses and just about every aspect of a computer or network. One would think then, that it would have brought down the cost of computing as a matter of course. However, it didn't have that effect until recently, because most of the accelerated capability was delivered to the PC where its contribution to productivity was minimal.
Now however, we are beginning to see it pay dividends within the network, with NAS (network attached storage) and SAN (Storage Area Network) technology. These technologies are about sharing storage capacity and they cut costs by allowing data to be managed from the centre. They also improve the level of resource sharing. In networks that have grown wild by incremental change, they can deliver very genuine - and ongoing - cost reductions.
The same can be said of large servers. In the dot com era, many companies built up server farms by a simple add-another-box policy for intranets, web servers, email servers and the rest. All of this is ripe for consolidation onto large servers and the cost saving discipline of central management.
Whether or not these specific technology options have any relevance, in most IT sites there will be some economies that can be made by considering the general idea of network consolidation. In the boom years, very little was invested in the shepherding of IT resources, so many networks are now over-resourced - or if you like, under-utilised. If the business cycle prioritises cost cutting, as it is now doing, then it makes sense for IT to make its first contribution by putting its own house in order.
Sponsored: Benefits from the lessons learned in HPC