This analytical article is by Steve Hodgkinson, the director of Ovum’s Government practice in Australia and New Zealand. Prior to joining Ovum, Hodgkinson was the Deputy CIO and Director of eGovernment Strategy & Policy for the Victoria state government in Melbourne for five years. This post originally appeared on Ovum’s site and is replicated here with permission.
analysis We believe it is time for governments to elevate their policy thinking about cloud services to confront their ICT strategy conundrum. They must address increasing fiscal constraints and disappointment with existing approaches to boosting ICT productivity with an approach that enables cross-agency sharing.
Mature enterprise cloud services are “capitalist economy” shared services that work. Cloud services break the cycle of agency investment in dedicated ICT solutions that are difficult or impossible to share. In contrast, each procurement of cloud services incrementally develops the capacity of the vendor to offer the same service to other agencies. A policy position of “cloud services first” is a strategic commitment by government to the development of the next generation of shared services.
Let’s get realistic about government’s ICT strategy conundrum
Ovum believes it is time for government policy executives to start considering a more visionary stance on cloud services, which explicitly acknowledges the stark realities of their ICT strategy conundrum:
- First, fiscal constraints mean that few agencies and few jurisdictions can now afford to sustain dedicated agency-by-agency, on-premise, or outsourced ICT environments and application portfolios. These constraints have led to unsustainable underinvestment, which will only worsen in the future.
- Second, “kicking the unruly ICT teenager out of the house” by entering into long-term outsourcing arrangements has, in many cases, proven to be a high-cost, high-risk, and inflexible strategy, with disappointing results.
- Third, government-owned in-house shared services have also, too often, proved to be high-cost, high-risk, and inflexible, with disappointing results. Their closed “socialist economy” markets, complex governance, underinvestment, and customer dynamics inevitably lead to real or perceived underperformance, compared to more open-market-based service offerings.
- Fourth, mandated ICT strategies that attempt to consolidate, standardize, and rationalise ICT infrastructure and applications across agencies have also proved to be high-cost, high-risk, and inflexible, with disappointing results. The organizational and cultural forces of devolution and differentiation tend to be stronger than the forces of central control and coordination.
How can governments tackle this conundrum?
From a strategic perspective, governments need infrastructure and application services that can be used by many agencies on an “arms-length,” shared services basis. A service must be designed to support diverse needs by enabling configuration of standardized functionality, rather than the creation of one-off customized solutions for each agency. It should deliver productivity benefits from economies of scale across agencies, enabling both reduced costs and the pooling of investment in service modernization and innovation.
It should evolve iteratively, on a continuous basis, to meet the changing needs of the user base. It must be transparent in its operations and have clearly defined interfaces for integration with other services. Finally, it must be trustworthy, secure, and reliable, and should not unnecessarily commit the agency to inflexible costs.
Agencies should use such a service because they believe it to be better, faster, and less expensive. They should behave as intelligent consumers of a shared service would – with a willingness to make pragmatic and agile cost/benefit decisions about desired functionality in order to work within the “as is” constraints of the service offering.
This represents the hopes and dreams of any government for its shared ICT services strategy, but in most cases these hopes and dreams remain substantially unrealized. It is also, however, an accurate statement of the value proposition actually delivered today by a mature, enterprise-grade cloud service that is deployed by a forward-thinking agency.
What comes first, the demand or the supply?
The issue for many jurisdictions is that, outside North America, the cloud services market at a local level is still at an early stage of development. Most governments prefer onshore data hosting, so the supply of cloud services is still rather thin across the spread of local applications that agencies require. In this regard, from a policy perspective, governments get the local cloud industries that they deserve. If a government has failed to demand cloud services in the past, it has failed to stimulate local investment in cloud services that can be supplied to agencies today.
A policy position of “cloud services first” is a commitment to shared services for government
The way out of this conundrum is for each agency to ensure that as much of their ICT spending as possible is directed toward the development and operation of shared services that benefit the country’s public sector as a whole. They can do this by buying cloud services from vendors that are either based in their country or committed to the deployment of global cloud services in their country.
Cloud services break the cycle of agency investment in dedicated ICT solutions that are difficult or impossible to share. In contrast, each procurement by an agency incrementally develops the capacity of the vendor to offer the same service to other agencies. Each procurement of cloud services boosts the breadth and depth of the local cloud services market for the benefit of all agencies in all jurisdictions in the country.
A policy position of “cloud services first” is a strategic commitment by government to the development of the next generation of shared services. It signals a buying preference that will stimulate the development of appropriate cloud services offerings in that country. It also requires agencies to change the logic of the way they source ICT capabilities and develop the mindset and skills needed to become more intelligent and agile consumers of shared, cloud services.
Image credit: Fred Fokkelman, royalty free.
FTA “They can do this by buying cloud services from vendors that are either based in their country or committed to the deployment of global cloud services in their country.”
As soon as your data is hosted in another country your data falls under that nations laws regarding privacy, retention laws etc. This is why the current DSD advice is to avoid internationally hosted cloud solutions.
From this link – http://www.neowin.net/news/us-government-all-your-cloud-data-are-belong-to-us – “Basically, they’re (the US Government) saying that the instant someone uploads their data to the cloud, they lose any rights they have to that data. That’s pretty scary stuff, especially when you consider the level of cloud integration in modern operating systems. And if you think you’re safe using SkyDrive, iCloud, or Google Drive, think twice – the government is taking this approach to all cloud data, not just Megaupload.”
And that’s just America.
The inherent risk of hosting Government data in a cloud solution is just too great to ignore.
Having said that, the benifits of cloud services to multiple government agencies can’t be ignored. I’d suggest the best way forward would be for Government to establish its own private Government cloud service for use solely by Government departments. It should be owned, managed by Government and supported by Government Staff in Government data centres build to DSD standards of security and reliability.
Government data is (or can be) just too sensitive to trust private enterprise with at times. We most certainly should never allow government data to be housed in a location that falls under the legal control of other nations governments.
From a vendor’s perspective, It’s actually not a policy that is needed but a procurement vehicle to allow agencies to purchase these services without having to constantly run costly Request’s for Tender. As a vendor ,I see Government agencies/departments are willing to purchase cloud services, the issue they have is that they don’t have the money or the time to procure them. They are after services worth a few thousand to a few million dollars.
Simplifiying the procurement process through a cloud panel or the Australian Government’s Department of Finance’s Data Centre as a Service Multi Use List or DCaaS MUL takes the sting out of the RFT process. The only issue with the DCaaS MUL is the restrictions around the max cost and length of contract, which is $80,000 and 12 months respectivley. I’ve spoken to clients who want a service that is under the $80k, but for longer than 12 months, so they cannot use the MUL.
If the Australian Government had a cloud panel or Multi Use List above these restrictions, they could access the services when required without spending money and wasting time running RFT’s.
So again it’s not the policy needing change, it’s the procurement methods.
Hi Mike,
Not surprisingly, I like this comment! I haven’t seen anyone seeking a service that is less than $80k but over a year as yet. If you were able to provide details I would be keen to explore how we could manage this.
Cheers
John
I have to agree with Mike’s comments. Policy will only get you so far, what you need is something to enable government to buy the services they want, quickly and easily. Build it and they will come.
A good example is the UK governments Gcloud, which has around £18 million in contracts signed and I think its got no restrictions.
I have to agree with Mike’s comments. Policy will only get you so far, what you need is something to enable government to buy the services they want, quickly and easily. Build it and they will come.
A good example is the UK governments Gcloud, which has around 18 million pounds in contracts signed and I think its got no restrictions.
Re “From a strategic perspective, governments need infrastructure and application services that can be used by many agencies on an “arms-length,” shared services basis. A service must be designed to support diverse needs by enabling configuration of standardized functionality, rather than the creation of one-off customized solutions for each agency. It should deliver productivity benefits from economies of scale across agencies, enabling both reduced costs and the pooling of investment in service modernization and innovation.”
Achieving these goals require time and money. Unfortunately government, at the political level, works against this. Politicians usually want systems as soon as possible and as cheaply as possible. The result is point solutions, often with failings that take effort and money (not always available) to fix. The desire to achieve politically set goals overrides procurement driven preference for cloud or any other shared service.
The reality is that, if you want to change the outcome that results from these political decisions, you’ll need to address political behaviour, not public sector behaviour.
Good luck. Just don’t hold your breath.
The UK Government is experiencing all of the same issues that you are endeavouring to resolve – see for the recently published strategic plan http://publications.cabinetoffice.gov.uk/digital/strategy/
I am the UK MD of UNIT4 Business Software – and I am pleased to say that our ERP solution Agresso was recently selected as the system to underpin the UK Government’s first independent cloud based shared service – see http://www.computerweekly.com/news/2240178815/DfT-outsources-shared-services-centre-to-Arvato
I would be very happy to share our experiences if this is of further interest
If an agency struggles with with the commercial elements of a outsourcing then I suspect that they will also have problems with a cloud provider.
I agree with Mike that a better procurement vehicle would assist Government agencies to purchase cloud services. The 80k and 12 month threshold for the DCAAS Multi use list is not ideal. Secondly, a Multi Use List still requires agencies to negotiate a contract with the cloud provider that addresses the risks that are spelt out in DSD and AGIMO policy guidance.
I understand that there are model contracts negotiated as part of the MUL process but would be interested to know whether the thorny issues relating to data, security, privacy etc are addressed to the satisfaction of the policy agencies (eg AGIMO, DSD, Archives) or whether line agencies still need to sort out this stuff with the cloud provider.
Hi Michael
The DCaaS MUL has a head agreement that contains the contract. All suppliers must sign it as a condition of joining the list. The process after that is for a request for quote to be sent to the potential suppliers as selected by the agency from those that meet its requirements. The agency considers the responses and then signs a work order with the preferred supplier. No extra contract negotiation is required.
The head agreement was drafted after public, online consultation with stakeholders. It doesn’t vary from supplier to supplier.
Since mid-October last year, there have been 12 orders signed under the MUL for a total of some $500,000. Several more are in the pipeline.
Regards
John
Comments are closed.