If you’re not actively engaged with what benefits Cloud computing can bring to an organisation, the economic benefits are a good place to start. The pay-as-you-go model enables a shift away from expensive upfront infrastructure investments (and refreshes), making the business case argument for using Public Cloud almost irrefutable. As with any large-scale migration, it gives companies an opportunity to rethink, re-architect and rationalise their IT. For example: Security – where most companies have historically focused on perimeter security for their datacentres and data assets, Cloud adoption gives an opportunity to rethink security-in-depth with a much more granular approach.
Current Cloud climate
A wealth of off-the-shelf solutions and components are now available, and the selection is ever growing. Almost any functionality required can be purchased as a virtual machine image or as-a-service. Public Cloud solutions can now also cope with higher performance and scale requirements than ever before – for example Cloud instances to support huge in-memory databases with literally hundreds of CPU cores are now available. It is now also possible to provide high performance personal computing from the Cloud with virtual desktop and application streaming backed by powerful processors and GPUs (graphics processing units). As well as providing an ever-increasing computing power, the geographic spread of public Cloud facilities has continued to grow with major service providers increasing their footprint in different geographies and legal jurisdictions, allowing customers to meet regulatory demands for data location.
Predictions for the near-future
Over the coming 1-3 years, I predict an increased focus on Data Residency with regards to the potential impacts of Brexit, GDPR and geopolitical concerns. For example, while some commentators in the days after the UK Referendum of June 2016 predicted reduced Cloud investment and a slowdown in the stand-up of Cloud datacentres in the UK, it is in fact possible that there will be increased investment as Cloud consumers look to repatriate data to the UK as a mitigation against any data regulation risk issues thrown up by Brexit. Furthermore, a recent example of Data Sovereignty tension saw Microsoft take on the United States of America and the FBI, blocking their efforts to seize email data stored in Dublin. Microsoft won this historic case, increasing pressures on European-American relations after the dismissal of the Safe Harbour agreement.
I also think that the Open Banking initiatives will have an impact on the public Cloud industry. While big tech companies such as Amazon, Google, Facebook and others look to exploit Open APIs to position themselves as financial service intermediaries, there may be resistance to their Public Cloud infrastructures being used to host services owned by the underlying banking institutions. Firstly, the public and regulatory bodies may expect to see clear separation between financial service intermediaries and the provider banks in order to allay any concerns of misuse of customer data. Secondly, the Big Tech companies themselves may come to be regarded as “too-big-to-fail” organisations representing a structural vulnerability in the economy’s infrastructure. Therefore, Open Banking may provide some push to enhance competition in Public Cloud provision.
Is security a concern?
The realisation is growing that security in Public Clouds is in many cases superior to the existing security arrangements in private datacentres, therefore migration to Cloud can be an opportunity to enhance (even re-mediate) current security. The key is that careful design is still required with Cloud computing; and security must be, as it always was, a prime requirement in design and not an afterthought.
Is there a preference between: Hybrid Cloud, Multi-Cloud or Cloud-to-the-Edge strategies? Or do each have their merits?
In short, they each have their merits. A Private on-premises Cloud instantly becomes a Hybrid Cloud as soon as an organisation deploys anything in a Public Cloud. Due to virtualisation and the emergence of converged infrastructure technologies, Private on-premises Clouds are not at all unusual; the management and deployment techniques for Cloud are therefore already well established in some organisation prior to their making any move towards Public Cloud deployments.
The arrival of technologies that allow seamless management across Private and Public Clouds strengthens this – Microsoft Azure Stack is one example.
Multi-Cloud is an increasingly valid strategy for several reasons – to provide a further level of Disaster Recovery (e.g. using a second Cloud provider as back-up for the primary provider) and as a means of producing best cost efficiencies. For example, which Cloud provider or Cloud region is providing the best price on this day? Technologies such as Cloud Foundry are now capable of automating the provisioning of the same Cloud setup in different Cloud services, therefore effectively acting as a broker for Cloud services, giving the consumer the best options for cost and availability at any given time.
Cloud-to-the-edge represents something different and revolutionary, and is an aspect of a shift in technology in the IT industry as a whole. IT has traditionally been built to serve the needs of “the User” – a person sitting at a desk with a device that accepts input and displays output, where the Human Computer Interaction (HCI) occurs. Successive generations of IT – punch cards, teletypes, green screens & mainframes, Client-server, web and mobile have all been flavours of this traditional model.
However, with the emergence of IoT (Internet of Things) and the massive scale of data analytics that Cloud computing can provide, there is a shift towards pervasive computing, where the devices at the “edge” are themselves the users and consumers of the centrally held data and analysis of that data. In order to utilise that data in concert with the information they themselves are collecting, these devices require a great deal of autonomous processing power. Artificial intelligence, machine learning and real-time analytics will play a vital role in devices which may not themselves have any direct human interaction but will be directly interfacing between the real world and complex IT systems extending all the way back to Cloud datacentres. Human interaction will not necessarily be with the edge devices – but with the outcomes that such devices bring about. Self-driving cars, automated traffic control, environmental monitoring and control are just the more obvious examples of the applications that are emerging. All this might sound sci-fi and indeed quite frightening – it remains to be seen just how human-friendly such systems will be. Where Human computer interaction is required (as it will be in thousands of use-cases) we will see a significant, but not total, shift away from traditional screen & keyboard (or mobile equivalent) towards augmented reality and even virtual reality Human Computer Interaction.
Get stuck in, try building something you want in a Cloud, start with something non-essential where you can afford to get it wrong and fail fast – maybe a development, test or proof-of-concept environment, keep trying until you have something that works. Embed security from the very beginning. Start looking at workloads in your datacentre and see what could most easily be translated to an external Cloud.
Exception can help assess your current IT landscape to identify the opportunities for your business. Our experienced staff have a proven track record of providing strategic guidance in upgrading legacy systems, managing Cloud computing migrations and delivering quality managed services. Whether you’re starting your Cloud journey or planning your next strategy and would like some guidance, we’d love to hear from you.
Alan Blain – Infrastructure Architect
Back to articles