OK, the MoD computer revision plan.
It aims to provide 150,000 terminals across 2,000 military sites — including on Royal Navy ships as well as for ground troops in war zones.
150,000 seats. OK. Sounds about right. Somewhere between a terminal for every one and every two of the Armed Forces (roughly, back of the envelope stylee).
And:
And the estimated cost has now risen from £6.2billion to £7.1billion.
Eh? I\’ve not dropped or added a zero here have I? They\’re really saying that a computer system is going to cost £47,000 per seat? Yes, I know, they need to be secure and all but how bloody much does a Faraday Cage cost?
A lot of the cost will be in training service personnel to use the system.
Still seems about five times too high, just the same. The cost is about a penny on the income tax, just for this.
Can they no just buy them Blackberrys?
To have any meaning whatsoever this analysis must say what the computers are actually doing. A machine to browse the web you can do for about 200-500 quid on that scale of purchase. However if this includes, say, computers requiring a purpose designed and built real time operating system that has an extremely high level of assurance attached to it that you might want to use for, say, controlling a missile with a nuclear warhead, you’re going to pay a lot for that. Skimping on that cost is probably the wrong option. A faraday cage hasn’t helped you, me and everyone else in the world if it malfunctions you see.
Talking about “computers” or “terminals” is like talking about “metals.” If what you need is platinum it’s no good whinging that this is costing us more than tin. It is worthwile to check you’re not overpaying for platinum.
I have no clue about the sense or otherwise of the MoD’s purchasing and it would be unusual if they haven’t paid far to much for a rubbish solution but this is pure prejudice. Anyone bringing them to account needs to make sure they’re able to count effectively.
There is basically nobody in government, in the senior public service, in senior management or board of directors in a listed (non IT) company who has a clue about computing systems. You’d think that would make people who do have a clue a valuable commodity who would get promoted to such positions so they can actually keep an eye on the consultants, woudln’t you? It’s a very clear and very real market failure. But well anyone who understands what a computer is, how it works and has the experience to be able to design and build large scale systems with computers is obviously a bloody geek to be made fun of, belittled and generally kicked in the balls at every available opportunity. Pretty much the entirety of popular culture reinforces this. Can the cost of this market failure be measured do you think? Anyway we all accept massive cost overruns and rubbish product in every public sector and large corporate IT project because those bloody geeks are rubbish, couldn’t possibly be the stirling chaps calling the shots and reviewing the process who have so little idea what they’re looking at bareing any kind of responsibility now could it?
If I were to estimate the cost of this market failure, I think I’d just start adding up somp portion of the market caps of the major IT consulting firms. IBM, Accenture I think might well qualify for 100% on that measure from what I’ve seen. Every single project I’ve seen at least an order of magnitude more than it could have cost if they’d hired people with ability.
The extra costs won’t be Faraday cages, but they will be buying specialist equipent that doesn’t radiate secure and insecure data.
The data rooms will be hardened and the comms between the terminals and data centre will all be ultra secure and with at least one full diverse route per terminal and probably 2. (that means a completely seperate path, not just fibre, and therefore doubles/triples transmission costs which will be high anyway).
Lets assume they pay £20k per annum per link, not unreasonable as there will be some very long links, and there are 2 links per site and it is a 10 year plan that is £800m. Plus installation fees it is easy to see £1bn in transmission alone.
So it is easy to see how costs for these types of projects can quickly rack up but without detail I can’t even attempt to get any closer.
The core difficulty is not with “market failure” or even with the inability of the gov’t. personnel to properly apprehend their particular needs (and therefore their reasonable costs) but with the very structure of bureaucracy itself and of the particular mind-set of those drawn to and qualifying for such employment.
It is a “classic” type of problem for which no one has ever been able to propose a viable solution.
About the only generally applicable “policy” (on the part of society) is to recognize the problem and, in that recognition, bureaucratize as few areas of life as possible. Unfortunately, for government (including especially the defense function), no other viable alternative (to bureaucracy) kas ever been found or even suggested.
In government, one can have either bureaucracy or despotism. At core, that’s a principal reason to keep government as small as possible and especially to keep it out of as many areas as can be achieved. But, since defense is certainly one of the most legitimate and indispensible functions (of gov’t.), there’s not much chance of basic improvement.
Since it won’t work, there’s not much point discussing the cost.
“150,000 seats. OK. Sounds about right. Somewhere between a terminal for every one and every two of the Armed Forces (roughly, back of the envelope stylee)”
Note that DII will be being used by MOD civil servants as well.
@gene berman
There is a market for skilled labour and a the level of seniour management, boards of directors and seniour public servants it pays boatloads. Despite having and paying the money this market avoids the talent in this particular, extremely important area. So the talent exists. Not hiring the talent costs them a fortune in cost overruns and rippoffs. Hiring the talent would save this money. They are spending more than enough money in the skilled labour market to hire talent.
Yet no talent is hired.
There are many purchasers. All listed companies have a board of directors. All hire senior management and obviously there are many government departments.
There are many sellers who have studied the relevant field and proven themselves successful in designing and building large scale computer systems.
Many buyers, many sellers and yet the wrong thing is bought. Consistently.
How is this not a market failure?
If you dig up an old microeconomic textbook you see one of the assumptions of perfect competition is perfect information.
Nobody doing the hiring has sufficient clue to be able to diffentiate effectively between a salesman and an engineer. Indeed they tend to differentiate by hiring the salesman and abusing the engineer.
Think about all the places large scale computer systems are crucial in large companies and in government. Doesn’t it just frighten poo from your bottom that nobody in charge of these developments knows anything? It is this market failure that Accenture, for one, exploit with crackerjack salesman convincing fools to part with their shareholders’ money.
Bright kids straight out of university really can do better than this and charge a lot less but they’ll never ever ever get hired.
It’s a market failure, this is obvious. How we can fix it is not.
The appropriate comparison is large constructions projects. Budget blowouts and such in these, as dramatic as the papers like to paint them are small beer compared to the IT world. When was the last time you heard of a shopping centre or a stadium that failed utterly? That someone produced a rival building of the same size, shape and purpose for an order of magintude lower cost? These things are everyday occurances in IT.
The Sun article is based on the House of Commons Public Accounts Committee report, which I’ve just read. http://www.publications.parliament.uk/pa/cm/cmpubacc.htm The PAC were working on the basis of the National Audit Office report (from July). http://www.nao.org.uk/publications/0708/the_defence_information_infras.aspx I haven’t got to that yet, but I will just say that the project progress is measured in “installed terminals” but the project includes all the infrastructure behind that, including cabling, networking, servers, and some software, so the cost isn’t obviously excessive.
Look for a detailed review on Anomaly UK tonight or tomorrow.
http://anomalyuk.blogspot.com/2009/01/defence-information-infrastructure.html