Tuesday, April 13, 2010

The Cloud

I have been thinking about "The Cloud" for some time now, and trying to understand what is new about it.

When I started in computing in 1980 I worked for a company called Hoskyns, which was a leader in what was called "facilities management".  This meant that they would run your computers (mainframes) for you - so that you could concentrate on the core business of the company.  Sometimes it worked well for both parties and sometimes not.  It tended to work well (as I remember) because Hoskyns could take over the entire computing for the company and were often providing much of the software anyhow.

In the 1990s there was much interest in outsourcing which was driven by the idea that businesses should concentrate on their core business and all other services should be handled by specialists brought in to provide that service.  This was especially true of cleaning and catering, but many other services were outsourced, including computing.  Whereas cleaning and catering has worked (in the sense that companies have stuck with it on the whole - though often to the detriment of the working condition of those actually doing the work) the picture in IT has been much more mixed with some notable disasters.  There are many reasons for this, but I would suggest that amongst them are:

  • The lack of flexibility that most outsourcing contracts offer so that changes can become prohibitively expensive.
  • Parts of the computing are effectively providing core services to the company so that close control is needed.
  • Lack of understanding of all the issues by the company that is contracting the service (the outsourcing provider is a specialist doing it all the time, the company outsourcing its services does it once and does not fully understand the issues.
  • Loss of control of things like when upgrades happen causing additional costs.

No doubt there are others, but there have been expensive fiascoes - especially in the public sector.

Now, these tended to involve moving control of existing services to a third party.  However, with the cloud we seem to be wanting to change our systems at the same time.  This worries me hugely - the costs of implementing a new system are large.  The staff development costs are usually the largest single cost (though often hidden in higher education as the loss of staff productivity through time out for training, lower efficiency while systems are learnt as they are expected to be absorbed).

So I think that the cloud needs to be approached with caution, with the following being areas where it may make sense:

  • Where new systems are being adopted anyhow
  • Systems that do not fully belong within the institution.  I am thinking here especially of VLEs and e-portfolios where continued access from graduates of the institution makes sense.

In short I think that there are some opportunities that the could offers, but it is nothing like as great as many of its advocates seem to suggest.