Government policy-makers should level the playing field for cloud services
Cloud services policies are being developed and iterated in all jurisdictions. Having reviewed a number of draft policies in recent months, we believe policy-makers need to work harder to create a level playing field for cloud services adoption, mindful of the potential for a type I procurement error (buying a “bad” cloud service) or a type II procurement error (buying a “bad” in-house, shared, or outsourced service, when a cloud service would have been better).
Policies tend to be biased toward avoiding type I errors, so even in policies that are well-intended the playing field is tilted away from cloud services. Type II errors can, however, create worse outcomes, arising from missed or delayed opportunities for productivity improvement and innovation in policy and service delivery.
Innovation-minded governments need to level the playing field
The logic of government cloud computing policy usually starts with the implicit assumption that cloud services are risky, ill-defined, and unproven compared to other ways of sourcing ICT, such as in-house IT, internal shared services, and outsourced/managed services.
Cloud services adoption in government faces a steady headwind caused by procedural, organizational, and cultural inertia. The following five statements may help policy-makers to question their assumptions and strengthen government cloud services policy.
1. There is no cloud
As far as government agencies are concerned, there is no such thing as “the cloud.” The term may be useful in the consumer market as a trendy synonym for the Internet, but it has no place in government ICT policy thinking. Policy should refer to a “cloud service” – a tangible service delivered on a professional basis by a trustworthy external service provider.
2. Cloud computing and cloud services are different things
Cloud computing is the suite of technology innovations, including scalable infrastructure, virtualization, automation, self-service provisioning portals, and multi-tenant architectures, that is used by a service provider to build and deliver a cloud service. But there is much more to providing a mature enterprise-grade cloud service than buying and installing cloud computing technology.
A cloud service is an established bundle of processes, people, organisation, and technology, which has been assembled and refined to deliver a well-defined and trustworthy shared service to many customers. Trustworthy, sustainable, cloud services require scale and are expensive and difficult to build and operate.
We often use the phrase “cloudy is as cloudy does” to make the point that a cloud service should be judged by what it is today. In contrast, a plan to buy and install cloud computing technologies – for example to build a so-called “private cloud” – is simply a promise that a service may exist in the future, if sufficient funding is available and implementation of the required processes, people, organisation, and technology is successful and sustained over time.
3. “Private” versus “public” is the wrong way to view cloud services
Most policies start with a mechanistic taxonomy of cloud services as either “public,” “private,” “hybrid,” or “community.” This taxonomy is rapidly becoming an anachronism because the categories are indistinct and become more blurred as the service offerings mature.
Worse still, it is often implied or explicitly stated that “public” cloud services are relatively unsafe for government, while “private” or “community” cloud services are safer. The track record of ICT project and shared service implementation failures in most jurisdictions proves that this very statement is a triumph of hope over experience.
One of the biggest benefits that cloud services can offer government is the reduction of implementation risk. The correct way to assess cloud services is not as either “public” or “private,” but rather whether or not they already exist – and are functional, reliable, trustworthy, and affordable. In short, does the service meet the agency’s business requirements?
4. The best way to understand cloud services is to showcase them in action
Policy documents often tend to be long on theoretical positions written by those with little or no hands-on experience of cloud services and short on real examples. This needs to be reversed. The most impactful way to explain cloud services is to show tangible examples of their use and to discuss the experiences of early adopters in the public sector.
Ovum’s case studies reveal that cloud services are better, faster, less costly, and (overall) less risky than more traditional sourcing approaches. The first task for anyone charged with developing a cloud services policy should be to research and publish 10 to 20 examples of early cloud services adoption experiences within the relevant jurisdiction. When it comes to understanding the practical benefit/risk tradeoffs, seeing is believing.
5. Type II procurement errors are just as bad as type I errors, if not worse
Overall, we need to counterbalance the bias inherent in most cloud services policies toward avoiding a type I error. A type II error can be just as bad, if not worse.
The risks of procuring a “bad” cloud service are relatively low and can be contained by well-established risk-management mechanisms and shorter contract terms. However, the risks of “bad” in-house, shared, and outsourced services are well understood; they are often substantial, long term, and difficult to mitigate.
The risk of a type II error in cloud service procurement is a deferred or missed opportunity for productivity improvement and innovation in policy and service delivery.
Potential transformational benefits should be fairly stated
A government cloud services policy should aim to fairly state the potential transformational benefits of cloud services and create a level playing field for agency adoption, mindful of the practical and long-term costs to government of both type I and type II procurement errors.
- » Amazon wins injunction to temporarily halt Microsoft JEDI contract award – reports
- » Hybrid cloud environments: A guide to the best applications, workloads, and strategies
- » Don’t forget to budget for business objectives to gain digital transformation ROI
- » A guide to computational storage: Boosting performance for SSD storage arrays
- » How SaaS, AI and machine learning are boosting sports broadcasting: A guide