The recent speculation about the death of Delicious and it subsequent rebirth, possibly under a new owner, has given rise to discussions about when use of Cloud Services in an institutional context are appropriate. For some, the answer may be never. I have heard people say “I’ll never use Google to index my Web site; I don’t have access to the source code, so I can’t fix it if it goes wrong“.

This is nowadays probably regarded as a fairly extreme position, but how should one go about answering the question: “When might use of a freely-available Cloud Service be appropriate for use in an institutional context?“.

For me the answer is simple: “Always!“.  Or, to refine this answer slightly: “Use of freely-available Cloud Services should always be considered for use in an institutional context“.

In part this is a response to the need to be seen to be making cost-effective use of tax-payers money.  Although this has always been the case, as we have seen from the Daily Telegraph’s recent FOI request the media is looking at how the public sector spends money – to be paying for the development and support of a service within the institution when there are free alternatives available could lead to questions as to the appropriateness of such decisions.

In addition to the need to be seen to be aware of the financial considerations, freely-available Cloud Services also provide an opportunity for evaluating options and identifying popular features and patterns of use. And following on from the question of the features provided by services there is also the question of the community which can have an important role to play in social sharing services. This is an area in which a service hosted within an institution may fail to gain a sufficient user base in order to gain the benefits of network effects.

Of course there will be a need to consider the sustainability of possibly options.  As I described in a post on Lessons From Delicious’s (Non)-Demise this issue seems to have been the top concern when Niall Sclater made his point: “@mweller @psychemedia delicious. i rest my case.“.

But considerations regarding the sustainability of services in nothing new in the IT industry.  In my time in the business I have worked on computers which are no longer made (IBM and ICL mainframes, VAX minicomputers, Apollo and Sun workshops, BBC and Commodore micros, etc.) and have seen software come and go.  Such concerns can affect large-scale enterprise software and business developments (Blackboard buying WebCT, Oracle buying Sun, etc.) . But now we are also seeing political and economic factors affecting the sustainability of institutional services.  We have already heard recently the news that “Welsh universities to merge” – and we can’t realistically expect that there won’t continue to be significant changes across the sector.  So although we might be able to speculate on the risks of commercial providers of such services (such as Yahoo’s failure to develop the Delicious service and, in retrospect, the failure to accept Microsoft offerto buy the search engine company Yahoo for $44.6bn (£22.4bn) in cash and shares“) we also need to be honest about institutional risks. So if we wish to speculate that, for example, Microsoft’s proposed purchase of Yahoo services could lead to cherry-picking and closing down services which compete with existing Microsoft services we might also speculate that proposals to install services in-house are being proposed in order to justify the continued existence of the unit providing the service.

There are risks in using Cloud Services; there are risks in purchasing commercial software and there are risks in deploying services in-house. This is nothing new.   What is different today is that we are no longer in a period of growth – so we will need to be prepared to understand and address the risks of institutional provision of services.  I do feel that “use of freely-available Cloud Services should always be considered for use in an institutional context” – and whilst this does not necessarily mean that such services should be deployed we will need to have exit strategies in place for in-house alternatives.