One of the claimed benefits for Cloud Computing is the notion that it heralds a new era of enterprise flexibility. No longer, say cloud campaigners, will project lead times be constrained by the logistics of hardware provision, rack-space or the need for more air-conditioning.

Well, we can all dream – but the CIO is, despite rumours circulating in the Finance Department, merely human and, like the rest of us, must battle against the normal nightmares of under-resourced over-expectations.

Where previously CIO's and their development teams adjusted their facilities and systems to take advantage of distributed computing and vanishingly small storage costs, now they can rethink the larger issue of enterprise flexibility or agility – tailoring resources to match fluctuating demand.

These new freedoms run counter to years of effort spent trying to stabilise demand and iron out fluctuations. Where aggregation of computing resources was just too difficult to manage, distributed computing made life more modular. But now, it seems, Cloud Computing offers both the economic efficiency of aggregation and the granularity needed by fast changing enterprises.

In the public sector there's an even stronger reason to congregate in the Cloud when, for example, the system requirements in the neighbouring parish are almost exactly identical – so why build it twice?

In the popular imagination agility and government IT might seem unlikely to be found on the same page but, according to the UK Government's new deputy CIO, perceived gaps between private enterprise performance and public sector services for the electorate should be consigned to history.

“The Citizen-Consumer expects”, says Bill McCluggage, “services in the public sector to be as good and as responsive to changing needs as they experience elsewhere – and the reality is that quite often our large government systems are actually out-performing private enterprise. One of the government's key reasons for adopting Cloud Computing is to serve the very large fluctuations in public service demand.”

The flexibility of Cloud Computing - dealing with load fluctuations and zapping lead times for hardware provision - prompts questions about planning priorities; which systems should stay grounded in-house and which should be floated in the Cloud?

There is an argument that Cloud Computing is best used for mundane low-value tasks that would otherwise clog the superior servers of your IT department's finest fellows. The 'clear away the dregs' approach is fuelled by the nature of some fairly sketchy cloud computing services with limited platform options, network linkage constraints and the distances that data must travel to and fro. So the CIO's project planning view of candidates for the Cloud is very much dependent on just exactly what sort of Cloud he has in mind.

Enthusiasm (and wishful thinking) is endemic in the IT industry. How quickly we forget why now we have distributed desktop computing - and how easily, in our supposedly broadband world, we overlook the glaring inconsistencies; the patchy coverage/quality of access infrastructures that are, for example, supposed to sustain 'working from home'.

It has taken a few decades to overcome the simplistic and low-value automation of paper processes. In the migration to Cloud Computing there is surely little point in simply replicating the constrained (or unimaginative) legacy. For corporate Cloud Computing the acid test of any new ability to respond to changing business needs can be tested by looking at the impact on enterprise strategies – not relegated to the marginal re-jigging of business as normal.

CIO's contemplating 'what might go and what might stay?' might do well to start out by considering the real roadblocks to systems and operational development – not just the convenient accounting shifts from Capex to Opex but the potential for delivering significantly higher value to their enterprise.

Careful analysis might well show that something as precious and closely held as complex systems development might well deliver faster results if conducted in the Cloud – whilst the systems set up to handle lower level tasks are already optimised and certainly not broken enough to need fixing by further upheaval.

This focuses attention on just exactly what sort of Cloud Computing capability is actually required. For Clouds to be taken seriously they (and their connectivity) need to offer a richer resource appropriately scaled. Only then can they be considered as a 'heavy lifting' resource to add real strength to the CIO's ability to respond to enterprise needs.

The CIO's responsibility is not to knock the optimists but to understand that the Cloud claims of 'response-ability' need to be tested and proven in the context of future requirements.