The UK Cabinet Office thinks that cloud computing can deliver cost savings to public sector IT expenditure; but insourcing outsourcing is throwing up some dilemmas.
Strange things seem to be happening around the UK government's grand G-cloud project to condense IT costs and build a competitive infrastructure for the future. The omens seem sound – the promised savings would contribute nicely to the government's spending cuts over the next few years, and the scheme dovetails neatly with Prime Minister David Cameron's avowed intention to plant deep sustainable roots within the UK economy.
Yet in other ways the fizz seems to have gone out of the project, following a series of setbacks which threaten to delay and water-down G-cloud's ambits. The change of government may be part of the problem, as the incoming coalition seems not as committed to the cloud concept as its Labour predecessor. Then came the spending review aimed at achieving 40 per cent cuts, but this would seem to favour G-cloud, given its 'promise to cut public sector IT costs by some 30 per cent. One problem was that the G-cloud plan as initially formulated did not yield jam soon enough for the Treasury – the full 30 per cent savings would not come home until 2020 – far too late to do much for the immediate budgetary crisis, or to assist the coalition at the next general election.
This led the Treasury to urge the Cabinet Office, which runs the project, to come up with a way of bringing the savings forward, leading to a new option to maximise cost reductions within the next three years up until 2014, while minimising the investment needed during that period.
The Cabinet Office reckoned that this seemingly short-term and politically motivated option would yield extra savings of '70m in the coming financial year, rising to '399m for 2012/13, and £430m for financial year ending April 2014. These are relatively modest amounts and raise the question whether it is worth compromising the strategy to achieve them.
Drumming up support for G-cloud
In any case it looks like the moment may be passing, since the government has yet to come out firmly behind G-cloud. This may be partly because of issues relating to competition, following concerns raised by the European Union that G-cloud may contravene EU procurement law.
This highlighted a central dilemma of G-cloud, which is how to gain the utmost economy of scale without allowing a few big suppliers to become dominant and in a position to dictate terms.
There is a contradiction bedevilling G-cloud's ambits. The more suppliers involved the less economy of scale there is; but if there are only a few suppliers, the pressure for competitive pricing is reduced. The G-cloud strategy attempted to square this by cutting duplicated procurement processes, but formulating the contracts in such a way that there was competition for individual components of the cloud. The impact of the EU's procurement law may be to revise this, although there will still be scope for savings by sharing infrastructure between multiple public sector departments.
Indeed, this has already happened on a smaller scale between some neighbouring local authorities, which have become aware of the potential for avoiding duplicated procurement processes and sharing infrastructure. This has been almost forced upon them by the government's 28 per cent reduction in its contribution to local government funding. In London, for example, Hammersmith & Fulham, Westminster City Council and the Borough of Kensington and Chelsea have embarked on a merger of services in general, including adult social care and environmental services, as well as IT. Significantly, such local authorities are ardent advocates of G-cloud, because they believe this would be more efficient and achieve greater savings than ad hoc bilateral or trilateral arrangements.
Leeds City Council is perhaps a typical case, with about 90 per cent of its IT services hosted in-house. Its chief ICT officer Dylan Roberts, chair of strategy and policy at the Society of Information Technology Management (SOCITM), has stated his belief that G-cloud, if properly implemented, would bring this down to under 50 per cent, with the majority of services delivered from the cloud. Roberts has been a consistent cheerleader for G-cloud and has been urging other local authorities to rally behind it, arguing that local government can play a key role in shaping its course. This is partly because, unlike central government, a number of local authorities have experience of sharing IT facilities or services.
Since last autumn enthusiasm has been dampened by the realisation that successful implementation of G-cloud will depend on several issues that have nothing directly to do with IT. The technological components needed to assemble G-cloud are mostly available now, which is precisely why hopes were running so high. Yet apart from the EU competition issue, another stumbling block has emerged in the shape of data privacy.
There have been rumblings about the impact of data privacy on public cloud services for some time, but it was only recently that the issue was related specifically to G-cloud by one of the government's own major contractors, Lockheed Martin. In both the UK and much of Europe, data privacy laws prevent the movement of personal data outside of the body responsible for it, as was pointed out by Melvin Greer, chief strategist of Lockheed Martin, which is already one of the largest providers of cloud services to the US government. At IBM's recent Pulse 2011 infrastructure conference in Las Vegas, Greer indicated that prevention of data movement between jurisdictions is in total opposition to the whole cloud concept. It would particularly inhibit government initiatives such as G-cloud, which can only fulfil their potential if data privacy legislation catches up with IT strategies and possibilities.
As Greer noted, security concerns over cloud can largely be fixed by technical developments; but the fundamental flaw of linking data privacy to islands of IT can only be fixed by politicians.
The Cabinet Office declined to comment directly on these points, but a spokesperson did acknowledge some of the concerns, admitting that there was a difficult balance to be struck between the needs of an open marketplace and the desire to drive volume discounts through aggregation of demand. The spokesperson also agreed that in attempting to share not just infrastructure but also citizen-facing applications and data between departments, G-cloud would run into uncharted areas where the public sector, at present, had limited expertise.
The upshot is that G-cloud has become enmeshed in non-technical issues after getting off to a flying start. This may well have contributed to the departure of the high-profile former government CIO, John Suffolk, at the end of 2010. Suffolk remained a strong advocate of G-cloud, insisting at the Software & Information Industry Association On Demand conference held in London October last year that it would yield savings of £1.2bn a year by 2014/15 and £4bn a year by 2019/20, through elimination of duplicated back-end processes such as identity and authentication. This could immediately allow six data centres to be closed down, Suffolk argued, although some of the proposed consolidation looks as if it could fall foul of the data protection issue.
Suffolk has been careful not to criticise the government explicitly since his departure, but he has expressed frustration at the coalition's loss of momentum over G-cloud. Suffolk's replacement is Joe Harley, formerly CIO at the Department of Work and Pensions. Significantly, he has been given a new deputy with the title 'director of ICT Futures', who will focus on the evolving G-cloud strategy. This is an admission of the need for further revision of the strategy before full blown deployment can begin. It may also signal a shift in direction away from a monolithic single platform to a more hybrid approach combining multiple environments flexibly across a common network. This may circumvent both privacy and security concerns while realising at least a large proportion of the savings.
While it is possible that G-cloud's objectives may shrink in pure IT terms, they may actually expand in their vision. Rather than just being concerned with cutting IT costs, it may be that G-cloud becomes positioned as a way of enabling government departments to increase all-round efficiency in the hope of mitigating the pain of the cuts. There is even the ambition, articulated by Suffolk before he departed his post as government CIO, of turning G-cloud into a public service offering shared IT facilities to SMEs – even individuals. In this way, G-cloud could become both a revenue source and stimulus for the UK economy.
While such sentiments may sound wishful at present while G-cloud appears stuck in something of a rut, they also could also provide the impetus to liberate the project, which is sure to be realised eventually in some form. The question is how much resemblance the eventual G-cloud will bear to the original conception. *