Skip to the content

Follow us @UKAuthority

We should view cloud as an operating model

09/06/20

Industry Voice

Cloud is at its best when it can rapidly adapt to change, writes Andrew Puddephatt, head of public sector at Nutanix

Public sector adoption of the cloud has gone through many phases.

From early experimentation, rising enthusiasm for wholesale migration developed across the sector – reflected by the Government’s ‘cloud first’ approach. But this has now matured into a change of tack towards a growing consensus in favour of ‘hybrid cloud’, a combination of public and private clouds, the latter of which can be built on-premise.

Preferences are likely to change again over time, leading to a growing appreciation that cloud should not be seen as a journey with a long term destination, or a choice of place for systems and data to reside, but as an operating model that can be amended to meet new demands and open up opportunities from new technologies and services.

This journey formed the subject of a recent UKA Live debate in which I took part with Andy Jones, lead infrastructure engineer at the Department for Work and Pensions, Andy Grogan, head of enterprise technology, projects and innovation at the Orbis Partnership of local authorities, Mark Adams-Wright, director of Socitm Advisory and Helen Olsen, publisher of UKAuthority.

It produced a series of insights that I believe reinforce the case to view cloud as an operating model.

Unpicking complexity

One is that experience is beginning to count, with valuable stories on organisations’ use of the cloud, including the teething pains of early adopters, unexpectedly complicated costs, and people beginning to understand which applications are suited to the cloud – and those that are not.

Much of this responds to organisational priorities and has fed the increasing use of the term ‘cloud appropriate’. This involves looking at each legacy application and digital asset, how they link together and how they can run within public and private clouds. Along with this is the need to avoid the major hazard of creating new infrastructure silos that do not work smoothly with each other.

Other considerations come into play: data governance, with the need to ensure the data is stored and managed in line with regulatory requirements; and cost governance, ensuring that the spending on the service is in line with expectations. The latter reflects unpleasant surprises that have prompted a few organisations to repatriate services back on-premise.

This makes it necessary to spend time on getting the architecture of the infrastructure correct, with the aim of ensuring that the components of the public and private clouds will interoperate with each other, and that you have a realistic view of what it will provide technologically and financially.

There are no short cuts to this, and there is a need to understand the investments that have already been made, the technical demands of switching services and whether the decisions are commercially justifiable. An organisation has to have a firm grasp of these factors before moving to any cloud environment.

Architecture forms the foundation

Tools are available to make this possible: the capacity to provide APIs between different services; software defined operations that can manage the volumes and set priorities for data traffic; and the hyperconvergence of computing, storage and networking into a single system to reduce the complexity of their management. There are also tools for monitoring performance, capacity and any failures in a data centre. These can help in the efforts to get the architecture right and provide an infrastructure environment fit for the future.

Other factors are at work. The emergence of cloud, particularly the as-service-models, have placed an emphasis on opex spending, when many organisations find it easier to make the case for capex, and those working on long cycles are likely to tilt towards on-premise models.

But recently more flexible offerings have emerged around the private cloud, with features such as software on subscription and hardware on demand. This is going to make the option more attractive to some, and there is scope to work with suppliers on new ways to present the billing to match procurement need.

Another crucial factor is that it is a mistake to assume the cloud provider takes full responsibility for security. Any organisation using a service has to acknowledge that it is responsible for the data it holds and the infrastructure platform it uses, and has to ensure that appropriate security applications are in place and that its people follow best practice.

There also has to be serious thinking about business continuity and disaster recovery, building them into the cloud architecture as a step towards organisational resilience.

Not where but how

All this creates a complex picture that will vary depending on the legacy technology, service demands and financial priorities of individual bodies; but the main message about cloud is that it should not matter where, but 'how'. It comes down to how you are architecting environments to deliver better services.

If you create an architecture that makes it possible to move applications from on-premise to cloud providers, you can take advantage of new services and avoid vendor lock-in and the creation of new silos.  Also, a hyperconverged infrastructure is an enabler for the environment because it provides a level of orchestration and automation, and it significantly reduces the footprint of the data centre, which produces benefits in terms of costs, energy usage and carbon neutrality.

The discussion was framed around the question of whether cloud should be seen as an operating model or a journey to a destination, and the consensus leaned strongly towards the former. The view was that aiming for a final destination creates the risk of an approach that falls short of interoperability and a capacity to find the right design of components.

We look at it as an operating model as – provided it has cloud characteristics, with agility and flexibility – location does not really matter to the consumer of the service.

Click here to watch the full UKA Live programme 'Cloud: an operating model or a location?'

Andrew will be presenting at UKAuthority Digital Health & Social Care at the end of June, outlining the way in which the NHS and local government have harnessed cloud to enable staff to continue working during the pandmic - register here to join this virtual event

For more information or to discuss the themes from the live webcast with Nutanix visit https://www.nutanix.com/uk/public-sector or contact Andrew Puddephatt and the team below:

powered by Typeform

Register: Library & Alerts

Keep informed - Get the latest news about the use of technology, digital & data for the public good in your inbox from UKAuthority.