What Should Be Out There, In The Cloud Perhaps ?
Posted by Brian Kelly (UK Web Focus) on 14 November 2008
One Sunday afternoon a few weeks ago I observed a brief Twitter discussion between Paul Miller and Paul Walk which I found interesting. Paul Miller began by tweeted his thoughts:
Pondering… ‘semantic web’ as ‘data cloud’? Cf COMPUTING Cloud metaphor
Paul Walk responded:
@PaulMiller not sure…. cloud works for processing cos we want it to be invisible commodity… want data to be more visible?
Paul Miller replied:
@paulwalk – but shouldn’t data be commoditized too? Or at least AVAILABLE for ad hoc use
and Paul Walk concluded:
@PaulMiller not commoditized – we *care* about data, it’s provenance, accuracy. I don’t want to have to care much where my cycles come from
This discussion got me thinking what should be in the cloud or, more generally, what aspects of IT can be provided outside the organisation? Some thoughts on the benefits of using a variety of outsourced services are given below:
CPU cycles: As Paul suggested, nobody cares whether the CPU cycles are provided by the computer in front of you, a server hosted within the organisation, a national service or a global service. And if those CPUs cycles are provided by an organisation which can minimise the heat losses to the environment and provide cost-effective and energy-effective delivery of the CPU cycles then this will ensure that the organisation exploiting the service is addressing its own green agenda.
Applications: It’s not just the CPU cycles which can be delivered across the network, application software need no longer by tied to the desktop PC or institutional server. We are seeing examples of this ranging from bookmark management tools such as del.icio.us through to word processing and other office applications such as Google Docs, Google Spreadsheets, Adobe Buzzword, Zoho, etc.
Data hosting: As Paul Miller has suggested the Semantic Web can perhaps be regarded as a data cloud. But if this is a vision for the future, remote storage of data is very much a part of today’s IT environment, ranging form the personal data management services provided by companies such as BT through to institutional use of services such as Amazon’s S3.
Software development: Moving on from the IT infrastructure itself, we can also outsource IT development work. We are familiar with this from JISC’s development activities in which software development is funded by project money to develop software which is intended for deployment across the community.
Data creation, input and management: I recently read a press release entitled “Amazon Mechanical Turk Launches New Web-Based Tools That Bring the Power of an On-Demand Workforce to Businesses Worldwide” which announced the launch of “a new set of web-based tools for Amazon Mechanical Turk that make it easy for businesses to use Mechanical Turk to outsource work to an on-demand, scalable workforce via a simple graphical interface – in just a few minutes and without writing any code“. So yes, data input, metadata creation and management, etc can now be more easily out-sourced. There’s now need to have large teams of data preparation staff in your organisation – although, of course, this has been the case for some time now.
Policies: If institutional or sectoral policies are too onerous to comply with, you could choose to outsource your services which are more flexible. Consider, for example, the terms and conditions which cover registration for the UK Government communities forum which I blogged about recently. If you feel these terms and conditions are too stringent you can also make use of an alternative environment for hosting discussions.
Now there will be many issues which need to be addressed if organisations wish to make greater use of the out-sourcing options which are now available (sustainability, reliability, security, legal and ethical issues, etc.). But is the future, I wonder, a world in which organisations focus on their own strengths and the services which only they can provide, with the chore activities being provided by others? After all, as Andy Powell reported in his live blog summary of a talk by Sam Peters, Google, “does anyone get a competetive advantage by running their own email system?” (posted at 09.57). But if this future does appear to have much to offer we will need to develop a framework to support institutions in making such policy decisions. Out-sourcing metadata management to an environment which is more flexible, responsive and provides benefits of scale sounds great – but not if the work is out-sourced to children working in IT sweatshops. We will need, I feel, an equivalent of Fair Trade which ensures that the benefits are provided by more effective use of technologies and better management and not by exploitation.