The ICT strategy will create a common, secure, flexible infrastructure that is available to the public sector. To achieve this, the strategy sets out the vision for:
- The Public Sector Network: A single holistic telecommunications infrastructure that will deliver converged voice and data communications. The PSN will deliver at least £500m savings per annum by 2014 and facilitate delivery of services to any location, running over a core network that is secure, based on open standards, interoperable, energy efficient and competitive
- The Government Cloud (G-Cloud): A Government cloud infrastructure that enables public bodies to host their ICT systems from a secure, resilient and cost-effective service environment. Multiple services will be available from multiple suppliers which will make it quicker and cheaper for public sector bodies to switch suppliers if they face service or delivery issues. The G-Cloud is a key enabler of the £3.2bn savings outlined in the Operational Efficiency Programme as it provides the access point for ICT services, applications and assets
- Data Centres: In order to deliver large cross-Government economies of scale, meet environmental and sustainability targets and provide secure, resilient services, it is necessary to significantly rationalise the current base of data centres in use by the public sector. Aligned with development of the G-Cloud, is a programme of activity that will consolidate and reduce the number of data centres in use from the current many hundreds to 10-12. This will deliver highly resilient, secure data centres that reduce cooling and power consumption by up to 75% on current infrastructure. It will also reduce IT infrastructure costs by up to £300m per annum by 2020.
- Government Applications Store (G-AS): The Government Application Store (G-AS) will be an online portal that enables sharing and reuse of business applications, services and components across the public sector. Rather than create bespoke solutions each time a requirement is identified, reuse will become the norm, with anticipated savings of over £500m per annum by 2020.
- Shared Services: In recent years, the Shared Services culture has been building both within and between departments – providing Finance, HR and Procurement services. This approach has saved both money and headcount and over 80% of civil servants are now supported by a shared service solution. By 2020, Shared Services will be provided via the G-AS and G-Cloud to further exploit opportunities. Shared Services will deliver £4bn savings as outlined in the Operational Efficiency Programme.
- Desktop Services: All public sector bodies need to provide their staff with access to functions such as email, word processing, spreadsheets and internet browsing. Historically, each public sector organisation has separately specified, built and run their desktop service. The ICT strategy will provide the public sector with a set of common desktop designs and all suppliers will be required to deliver common designs and shared services at the lowest price available. These designs will conform to Information Assurance (IA) and Sustainability requirements and will deliver significant economies of scale. A £100pa saving in operating cost per public sector desktop would yield a saving of £400 million per annum.
Data Centers
So we will have a limited number of data centers ripe for targeting by whoever whatever their political agenda – not very secure.
Never mind the integration costs, relocation costs, network infrastructure to link to the center, staff relocation / redundancy / recruitment etc. Making the many into a few does not reduce costs – it may even increase them.
This is all built upon the assumption that everything that government does is a secret and needs protecting with highly expensive common infrastructures. If you reverse the assumption then the need for all this expensive bespoke plumbing is automatically drastically reduced,
If government is truly open, then the things that need protecting are national secrets and citizen private data. National secrets will continue to be protected in the usual way whilst it is difficult to imagine how our personal secrets are protected when they are shared amongst millions of public sector workers and their contractors.
There is no small irony in the fact that the thing we care about most (our data) is not mentioned once here. Give it back to us and let us share it when and if we want to.
Mind you, I’m not sure what all the over-compensated government CIOs will have to justify their jobs with if there are no £billion projects to run. Perhaps they can all retire to their rare breeds hobby farms.
All right, Cloud Computing does not mean just having a common data centre, as is implied here. It means applications and data accessed via internet protocols rather than held locally and it critically depends on those services, and I mean services as an IT professional would understand the term, being well-designed to support such a model. The ‘G-Cloud’ will have to support services that are constructed in this way, which means third-parties and in-house IT conforming to standards that support them. The trouble is that these standards are not uniform across the industry and so decisions will have to be made on which route to take on this. Savings are dependent on common services being used by different parts of the government, but there is not indication on how this is to be brought about. There are truly formidable practical issues with producing genuinely reusable services and then presenting them in a Cloud Computing model and they are not even alluded to here.
Big green data centres? Well, modern servers will deliver the green bit almost automatically given recent improvements in their efficiency, so no marks there. More importantly though the organisation of the data centre is critical if you really want to support a cloud computing model. Virtualisation is not mentioned and that is a key enabler of both efficiency and a cloud computing model.
Again, we have this obsession with the desktop and a limited number of approaches to providing it, when the standards should be couched in terms of services and protocols. Desktop provision is evolving very rapidly and supposedly smart deals with suppliers can very quickly become restrictive if not formed in recognision of this.
I can certainly see the benefits in reduced numbers of data centres so long as there is sufficient redundancy in place at each site.
One item of technology that the government should probably make us of is Virtualization so that they could move the processing of one data centre to another in a short space of time.
It would also be great to see adoption of technologies such as VDI to allow government users a more secure environment at a limited cost.
Personally so long as sectors such as education (which I work in) are still allowed to conduct their own activities without a government mandate as to whom you can use as suppliers I’ll be happy.
Common Infrastructure themes, good … I am off to the supermarket, RFID.
I do not see the reasoning behind leaking this report, when all it does is strengthen the stance of Labour who are detailing how they expect to make IT efficiencies, therefore public spending savings, as they should be.
The fact that most government departments are being overcharged for failed systems by suppliers is the key – they, historically, do not deliver – not the departments who set out their business requirements.
•The Public Sector Network:
Good stuff. Let’s get some economies of scale going on in the provision of networks and data links, certainly cut costs. A single all encompassing network between everything will be great for sharing all that personal data, as long as we can figure out how to secure it properly of course. And get people to use it, because we know how popular GSI has been!
•The Government Cloud (G-Cloud):
Will save us some money if someone can manage to figure out the real definitions of some of the buzzwords that they included in this report. It certainly doesn’t look like there’s anyone with the right ideas at the moment.
•Data Centres:
Depends on your definition of “data centre”. There’s a lot of people around there who seem to think a cupboard with a server in it is a data centre, and to be perfectly frank; you can’t really get rid of them.
•Government Applications Store (G-AS):
If; someone can define some usage of enterprise architecture principles, come up with proper design and blueprints for systems, figure out how to define them properly against what and how they deliver, and enforce departments delivering not only the information to make the standards but also keeping to those standards – this might work. Otherwise it’s going to cost much, much more than you’d ever save.
•Shared Services:
This is the sort of transactional stuff that I’m fully behind rationalising and rationalising until it’s down to the level where there’s one group providing this service pan-government as long as they can get it right and don’t make life harder for public sector staff. It’s not difficult to get this sort of thing right, but I’ve seen a lot of agencies and departments make it very difficult.
•Desktop Services:
Why do you need a common set of designs? It’s a desktop PC. As a basic design there is nothing that needs more than 1 design for it.
Where you do need intelligence is within the procurement and licensing of the environment, and the delivery and support of the environment. This is again good transactional stuff that could be provided pan-government to shave those £’s off the provision cost.
There’s nothing in here about service management, nor about 1st line/helpdesk/service desk provision. I’m assuming those are dumped into one of the outlined categories as an “oh well of course we’ll do that too” type of thing when really its another transactional service that can benefit from some good economies of scale.
Overall I’m entertained by this report. I’d put money on it being drawn up by someone wandering around to their local technical team, asking “what could we do to save money?” and then picking the most spinable ideas and throwing them into a document to be given to someone to add random figures and political babble in the hope of looking credible for all audiences.
Something I’ve not seen mentioned yet, which it is possible I may have glazed over, is anything at all regarding the costs of this equipment. I’m willing to stake a lot on the fact that instead of having 3rd party companies supply equipment via rental prices, and tech support, so that when it goes wrong you call up a number, get through to a call centre, and you get a case number, and then within 24/48 hours its fixed, you had in house techs who actually knew about computer software & hardware, would could fix the problem that day, in the office. Not only would you get an increase in efficiency, but it’d also mean you’d no longer have to rent equipment. The equipment could be brought and doing so could save so much money. You can get a work computer for about £300, as a one off payment, that’ll last approximately 3 years. As oppose to paying lots to rent one. Chances are, if a company is offering to do the job for the government, they could do it themselves with a bit of initiate and some clever thinking.
Dear lord, do you lot know what a strategy is? It is intended to define a direction, not individual specific technologies or what the helpdesk number will be!
@Andrew: Data centre consolidation is a well recognised as bringing large savings. Pretty much every large corporation has either done it or are doing it.
@James Preston: Thats the problem this strategy is trying to address – stoping all the individual departments picking their own suppliers!!
Government data centre can only be a good thing for improving efficiency and security. Currently too many systems are provided for the public sector on their own set of servers, with their own firewalls, routing, networking and load balancing which then end up using only a tiny fraction of the hardware capacity that has been lovingly (and expensively) installed.
As a hugely rough guess I think you could reduce the number of servers currently in use by 10 – 20 times (assuming virtualization is used of course)
The concept of G-AS is nice but hugely ambitious as Daniel C states. Cannot really see any savings being made here but can see innovation and cross department apps being created if it was done correctly.
However, I think the concept of an app store needs to be defined further to include the following two main components:
– Services – a set of web services that serve up the information from the different departments and enables interaction with assets of those departments. These would probably be provided by trusted IT suppliers (we have to have something to do in the brave new world!) done through the traditional procurement process. They should be based on open and published standards that can be enforced.
– Clients – these could be a bit ‘looser’ and make use of all of the different services provided. I imagine it would be these that would be opened up to many more suppliers driving innovation. They would be able to easily access cross departmental services. I do have some concerns that some of the requirements of govt departments are not shared and are required but only by a handful (tens) of users. At this scale I am not sure you would ever get anything other than niche suppliers providing them.
Feel free to flame me.
The whole approach seems unlikely to succeed in my view. The IT strategy deals largely with IT and attempts to describe an environment which allows all of Government to arrive at common standards of technology and architecture – different from today. What is missing is the stability and unified standards of requirements across the many strands of ‘government’ over many years.
In one sense , unless a government does nothing, it is constantly requiring changes to computerised processes. Each change costs considerable effort and in many cases, changes to the applications dealing with data.
Current ‘best value’ legislation requires competitive tendering by Government agencies purchasing IT. As different histories of IT infrastructure are established at different times in different agencies ( or authorities, or centres) the evolution paths over time which appear to offer the best value at the point of decision , promulgate the differences.
Some style of Big Bang standardisation always looks attractive , but the seeds of self destruction are built into the decision making processes imposed by government on its various agencies. To really standardise and gain economies of scale, one would need a process whereby only one version of any application used in government was in use at one time. That is a degree of central control that is probably unachievable. Even having common architecture across government has seemed to date impossible to achieve, let alone common data standards. Common applications ? Can’t see it, myself.
There is another facet to government IT which is well worth considering. Whilst Government is a huge employer and a frenetic regulator of all it surveys , it can only purchase technology that is viable in the commercial world on a planetary scale. And that can change independently of national Government strategies. ( I won’t even discuss the stupidity of national governments inventing their own IT technologies )
This implies a risk that no Government IT strategy which defines technology into the future is robust; and that any robustness decreases with time.
To reduce the cost of Government collecting and processing data, it is not enough to rely on IT strategies: it would have to significanly reduce the volume and complexity of its use of data over time. (I.e. simplify, standardise and control most of the jobs involved) This is of course an anathema to most central governments, which is why ultimately IT projects never deliver savings over time – and I would also venture the same applies to IT strategies.
Uh oh…
“…Multiple services will be available from multiple suppliers…”
For those not “on the inside” of the IT/services industry, this multi-supplier approach is generally the consensus on what has caused the NHS project to go totally off of the rails.
Multiple (competing!!!) suppliers will quickly collide with each other, creating a them-and-us blame culture. Execs will clash, people will try to save face (and shareholder’s money), communication will break down entirely after going through a painfully slow and bureaucratic phase where decisions take months, not hours to get made. Deadlines are missed, budgets get blown out of the water.
Its a recipe for disaster.
The app store sounds great, but from my experience of these sorts of applications they are NEVER “stand-alone” and often come with huge huge huge huge huge dependencies on databases or particular other bits and pieces that not every department has access to.
I appreciate that the “common” monoculture approach will eliminate a lot these worries, but if an app has a big dependency on legacy CRM systems for example, large databases, specialist installation & setup requirements etc etc etc then I cannot see how it could be made to work “anywhere” without frankly absurd amounts of work from the original developers and on-going support teams.
It would of course work for the most basic of applications that are very much “stand-alone” (think Open Office, thin-client software etc) and I do like the idea of it.
It will be a living nightmare to manage compatibility and common interfaces for apps in the store though – it only works for Apple et al because they have an incredible amount of control over their platform, and the iPhone apps are pretty much stand-alone (i.e. no complex systems integration aspects)
I can see a great deal of sense in a limited number of dispersed shared data centres. Currently every department seems to either run their own facility or outsource it to a third party. There has been a huge mindset shift in the data centre design arena over the last few years and the way we have historically designed and built data centres is incredibly inefficient.
I’m sure the co-lo providers will be already providing some ecconomies to some departments and the in house facilities will be in various states, with some of them ancient and inefficient. I’m equally sure that the co-lo providers that have modern facilities could go into an inefficient facility and put forward a proposal to save that department some money, but we must always remember that the provider is still in it to make a profit.
Don’t get me wrong I’m not saying that there is anything intrinsically wrong with a business making a profit and for a small to medium size facility that may well be the most cost effective way of doing things as it allows you to tap into the ecconomies of scale that a small facility might struggle to achieve.
The public sector must have a large number of small and medium size facilities by working collaboratively I can see a vision of the public sector effectively becoming it’s own co-lo provider. The public sector would then achieve the ecconomies of scale that are possible in large facilities and, without a profit margin driving up the costs, it should also be cheaper than going to a co-lo provider.
With regards to Andrew’s comments about security; The likes of SIS and GCHQ must already have facilities that are a potential target. I would imagine that they are currently afforded greater protection than your average data centre. My guess is that they would be in underground bunkers. Other government facilities are also currently a soft target. This does not negate the the validity of collaborative facilities it merely means that you have additional parameters to consider. Part of that is the consideration that is already obviously there; all the eggs would not be in one basket the proposal was 10-12. Now I don’t know how they came up with that figure but I would envisage resilience being part of the thinking.
Building resilient efficient data centres in underground bunkers for the likes of SIS and GCHQ would certainly pose some engineering challenges but just because something presents a few challenges along the way doesn’t mean you shouldn’t do it. All it would effectively mean is that 2 or 3 of the 10 – 12 would have additional requirements.
The majority of public sector data centres are probably less of a target than the banks and the LSE. These private sector organisations do their business continuity planning and have duplicate facilities the public sector is no different in that respect.
As a public sector data centre manager who is currently looking to improve the efficiency of existing facilities, I would love to be involved in the design, building, commissioning and ongoing management of a set of facilities like these.
For at least 4 or 5 years, the current CIO has blocked attempts to share code and experiences across the public sector. He and his team have blocked innovation and allowed the elite incumbent SIs to take any good ideas and then rebuild them themselves, rather than buying from creative UK SMEs. Now he suddenly wants to start sharing code in the way that he has previously blocked. Why the change of heart? Is it just to copy Obama? Or is he worried new political masters will wonder what he has been doing since his appointment?
Providing standardised & centralised core services for commodity activities (such as computing infrastructure) is sensible. But what is proposed is a further anti-competitive move, aimed at making life easy for the CIO by further limiting the number of suppliers. The UK is the only country killing off supply-side capability and giving its money to a shortlist of ever-bigger and more profitable overseas companies. This is deliberate CIO strategy, further refinforced by OGC procurements (which block SMEs) and the lucrative monopoly deals with the likes of Microsoft and Oracle (who themselves admit they run 80% plus margin businesses).
It’s part of the failure of recruiting CIOs from private sector bodies (such as banks) with no experience of public sector, or any understanding that there’s a big difference between buying goods and services for a company and buying them for govt – particularly when govt is responsible for 50-60% of all IT spend in the UK. Every action it takes as a monopoly purchaser has huge impacts in the market.
As there is currently no competitive utility computing market, the proposed approach of outsourcing to a select group of vendors would be unwise. There’s also no mention of the interplay required with open standards and open source (such as the fact that standards – such as EC2 & S3 – are emerging).
A well-designed strategy would not adopt the lazy approach of handing all the govt’s IT business to a handful of elite suppliers (as the current CIO wants to do), but would be to adopt these emerging standards, build using only open sourced technologies (such as Ubuntu Enterprise Cloud), architect the design for a mix of private and hybrid cloud services and build in-house (outsourcing data centre floor space, power provision etc).
But the real question remains, as others have said, what is the objective and how does any of this relate to improving public services?
One of the biggest problems with NHS IT is that it is seen as an extension of an administrative task with people working their way up from data entry roles. Policy has been to outsource to external suppliers to minimise risk. In this environment staff rarely have adequate training to define requirements meaning everything starts from a very weak set of blueprints. Agenda for change compound this by not allow IT staff to be paid at competitive market rates and budget or staff are needed to progress beyond band 6. Good technical people will always better off elsewhere in terms of the technologies they are working with, the training they are given and the calibre of people. I have seen one cycle over the last 5 years due to having kids. All others that came in in the last down turn went elsewhere once the market turns up. I am waiting to see what will happen this time.
http://apps.gov/
1) The Public Sector Network – Can’t do any of the other stuff without this.
2) The Government Cloud (G-Cloud) – Debatable value .. do the other stuff here first.
3) Data Centres: Why didn’t this happen 5 years ago ?.
4) Government Applications Store (G-AS): Good in theory, but you have to actually make the applications work with the portal. Better to stick with the existing architecture without embedding in the portal. Then migrate backend components to central data centres. Systems that don’t perform under the centralized infrastructure become an evolutionary dead end … which is what you want any way (see the goals in the intro section). Products such as Citrix can be used short term to provide legacy apps over the HMG network where they are not architecturally suited to deployment over a wide area network.
Shared Services: If you want to make shared services work and have departments want to use the services you have to make it easy. Forcing departments to migrate to G-As, Portal, G-Cloud etc won’t recognize as many savings as hosting the existing apps from consolidated data centres. You don’t want to migrate the bulk of the applications you want to kill them off.
5) Desktop Services: There is absolutely no reason why more than one Desktop product set should be used. Though you will probably see multiple versions of the product set used as that’s unavoidable, for instance bulk of the desktops have Office 2007, a few stragglers have Office XP some have later versions. Common desktop standards significantly reduces the cost of developing applications by reducing the testing burden and making outcomes/users experiences more predictable. Most companies have set images that they role desktops with patches to match generally improving the security of the computing environment.
There always has been a common infrastructure: it’s called “interoperability”, if the focus were on this rather than the technologies or products then a lot of other issues would disappear.
And by interoperability I mean open, unencumbered standards. Remove vendor or interest group influence by rigid adherence to W3C, IETF, ODF (a genuine multi-vendor unencumbered document standard) If any new standard for anything is proposed it must be accompanied by existing not proposed interoperability infrastructure, e.g., “plugfests” see
http://www.odfworkshop.nl/
then just let the market operate
I may be wrong but this is NOT how I would expect a collaborative document to be evolved. The idea I would have thought is to go more to a WIKI style of evolution of the document with moderators etc to actually evolve a document that may be better than the current one. What appears to be happening is that lots of people are making lots of comments some sensible, some stupid but I can see very little evidence of how the conservatives are then expecting to evolve this into a better document – are they going to ready all of the comments and then rewrite the doc or have they misunderstood the way in which crowd sourcing can work?
On the document itself I can see the logic of the data centers and with perhaps fifteen of them sourcing won’t be too hard ie IBM, Fujitsu, Siemens, Atos, Logica, CMG, Cap etc – should keep most of the encumbent suppliers at least partially satisfied. The network again seems logical and sensible, the apps store – yes but probably not in my working lifetime (I’ve around ten years left before packing in, I’ve worked in the large SI environment for most of my life – delivering overpriced services into government). I think one of the most obvious things in Government has been the proliferation of the SI’s players across ALL of the different government functions – WHY? I can understand that perhaps HMRC and DWP etc require complex multi-facted systems, but why has the culture of employing an SI traveled all the way down to government functions perhaps employing 2000 to 3000 people…they do not need to be purchasing systems from big SI companies…but its a bit of a **** measuring competition….we need to use X,Y or Z because we are complicated and special – no you’re not – you could do most of what you need with a smaller delivery partner and at half the cost – BUT of course you can’t because they can’t get onto any of the government frameworks which ensure that only large SI players and consulting firms continue to exploit and abuse their position of authority. Try asking John Suffolk how many small companies he meets a week compared to how often he’s buggering off to some event from the big boys. Sorry off track a bit there – but what I don’t see in the document is any real attempt to address the behavioural and culture changes required to make this strategy actually happen
This bit, is assembled fashionable buzzwords at best and not a piece of strategy. Cloud is in fashion, shared services is in fashion etc. etc.
I agree with the comments above that the suppliers and multiplicity of suppliers (the usual suspects, big consultancies EDS, GSI, Cap-Gemini etc. etc.) are part of the problem. The other part being that nay civil service makes simple things complex and cannot manage IT projects, I say this as an ex ‘expert en informatique’ at the European Commission, I’ve been there.
The emphasis should be as (Gerry Gavigan says) on open standards and interoperability. This was also the ‘true’ way for the NHS, rather than monolithic failure-prone techno-Stalinism. Following this, and following the examples shown in France and Germany, a big emphasis on open source software, operating systems and components too.