After World War II the Swiss were the dominant force in watchmaking. Their products were high quality, and certified to meet an accuracy of -4 to +6 seconds per day. But then technology advanced; electronics and the quartz crystal could now replace the mechanical movement in consumer watches. In 1969 Seiko released the first Astron quartz watch to consumers; this operated at 8,192Hz instead of the 2-5Hz of a mechanical watch and so was accurate to +-5 seconds per month.

The first quartz watches were expensive, but mechanical couldn’t match their accuracy. However, rapid developments quickly drove down the cost of a quartz watch leading to the situation where quartz was now both more accurate and cheaper. At first, the Swiss watch industry did not embrace this technology. Within a few years  it was decimated by poor general economic conditions and the flood of cheap quartz and digital watches from Japan, Hong Kong and America. 

The surviving Swiss watchmakers decided to create a new market for the high quality mechanical watch and redefined the Swiss watch as a luxury item. In reality, everybody knows that you can get a more accurate electronic watch for a few dollars but that isn’t the point of a luxury item. With a huge marketing spend used to sponsor sporting events or by placing their watches on the wrists of the right people, (eg James Bond for Omega) they ‘re-presented’ the Swiss mechanical watch. It was now a desirable luxury item where the high price was, perversely, part of the attraction.

If you want a cheap, low TCO, accurate and reliable watch, buy a quartz, while the mechanical wristwatch is now a positional good (one whose value is determined by its desirability to others). The Swiss have now returned to a dominant position in today’s watch industry.

How does this relate to the data center industry?

The way we do IT is changing; instead of compromising availability with complexity we now build fault tolerant software. Availability is not “no single point of failure” or clusters of dual everything or expensive SANs. With virtual servers, we don’t really care what tin we have where – they are also easy and low risk to move and migrate.

Many existing apps are moving out to specialist providers - Rackspace for email, Salesforce for CRM etc. while new categories of application are specialist provider only (Google maps, Uber etc.). Commercial models are changing; there is a less centralised budget for IT, less desire to own capital assets such as data center capacity and IT equipment while the IT equipment and DC capacity are now seen as a commodity. Meanwhile the rapid ascent of cloud isn’t helping. Cloud operators build cheap, commodity capacity and then they sell that commodity by the hour.

Parallels with the Swiss quartz crisis

An economic contraction (2008 onwards) has created price pressure while new technologies have made IT capacity cheaper, faster to deploy and more flexible (making it easier to move your services). DC and IT equipment have become less attractive as a capital asset.

We have some vendors and providers whose business models don’t fit the new market, whose assets are less than optimal and sell products that carry “Premium” pricing despite a lack of real value differentiation. The rapid commoditization driven by all the above factors will expose these “premium” products and services for the poor TCO they are.

In my next blog on this subject I will look at whether modular is the answer and how to begin matching capacity to requirements.