As with all waves of innovation, as a technology matures, it is refined and optimised to maximise the benefit that it provides. In each case, players initially compete on a technological basis in terms of speed, feeds and technological wizardry, but the market eventually becomes more closely focused on customer value by becoming more closely aligned to the needs - either of individual customers or of particular community groups.
Vertical-industry expertise and talent becomes the ultimate differentiator as customers want to know that their technology suppliers are just as committed to their industry and its specific needs as the customer itself is. In effect technology wizardry becomes table stakes, while customer expertise trumps all. And we are now seeing this in the Cloud arena.
Multi-Cloud and Community Cloud
The National Institute of Standards and Technology (NIST) definition lists five essential characteristics of cloud computing: on-demand self-service, broad network access, resource pooling, rapid elasticity or expansion, and measured service. Private cloud has failed to deliver on the promise or economics of cloud, as it lacks many of these key characteristics. Likewise, hybrid cloud retains many of the flawed characteristics of private cloud. This is why multi-cloud which is rapidly becoming the de-facto standard for cloud architectures as organisations seek to optimise workloads while avoiding vendor lock-in.
Every organisation is different: coming from a different starting point with a different array of existing applications and a different set of requirements, often complicated by systems inherited either through merger or acquisitions or via shadow IT. As these organisations seek to enhance and integrate their systems to meet the challenge of digital transformation, most are finding that there is no single one-size-fits-all solution that can meet all their diverse requirements. A multi-cloud approach enables customers to find the best fit for individual workloads on a case-by-case basis.
In some sectors where collaboration is desirable, community clouds have emerged and become a key component of any multi-cloud strategy. In manufacturing, where supply chains are integrated between suppliers, numerous organisations can collaborate by sharing access to key datasets to optimise the end to end processes, enabling just-in-time manufacturing. Likewise, pharmaceutical, life sciences and clinical health systems are also able to use community clouds to collaborate in the same way. This is especially true in the UK where healthcare is nationalised, eliminating direct competition, and where collaboration is possible across secure networks like N3 and HSCN that protect patient data, while allowing organisations to leverage access to key data sets like Genomics England’s 100,000 Genomes Project for genomic analytics. Similarly, public services in most countries are seeking to collaborate using key logistical and geospatial data sets for transportation, or seeking to use shared services, shared data sets and integrated applications to transform the provision of other public services. In the USA, the main public cloud providers have set up dedicated regions as community clouds to allow US government agencies at the federal, state and local level, along with contractors and educational institutions to collaborate using sensitive workloads and data sets while meeting specific regulatory and compliance needs. Meanwhile in the UK, UKCloud has created a community cloud for public sector and healthcare that has attracted over 220 projects, capturing over a third of the G-Cloud IaaS workloads.
Such community clouds can spark a clustering effect, where, as more customers from a particular sector join, it attracts specialist application providers, both Software-as-a-Service (SaaS) providers and Independent Software Vendors (ISVs), which in turn then attract more customers in what becomes a virtuous circle
A multi-cloud approach allows you to have the best of both worlds: it combines the advanced functionality in areas like Artificial Intelligence and Machine Learning that you get with public cloud, with customer centricity around key workloads and data sets that you get with community clouds. Multi-cloud also allows customers to create rich heterogenous solutions that address a wider set of requirements than is possible using only cloud native technologies or any single cloud platform, while maximising choice and flexibility and minimising lock-in.
Proximity and Latency
The move to customer-centricity at a strategic level is simultaneously being matched by a similar move at a technological level. There is a movement away from big centralized clouds, to clouds that are closer to their data origins and help cut down on latency. This is taking two forms: fog computing, and intelligent edge computing.
Cloud computing revolutionised virtualisation and ushered in the digital era with universal access to almost limitless remote computing resources. Often customers were expected to access these resources over insecure internet connections, and to accept that their data could be stored and processed anywhere in the world, introducing a level of latency (delays in the time taken by systems to respond). Now intelligence is being built into a fog of industrial IoT applications, sensors and VR-powered devices that are providing local or offline capabilities. This can be used to eliminate latency and deliver the kind of seamless, real-time experience that modern users have come to expect.
Intelligent edge computing takes this a step further, utilising cloud appliances from companies like Microsoft and Oracle, to extend the reach of public cloud services by placing cloud computing capacity as close as possible either to the customer or to the data sets that are critical for the customer.
Latency can occur either between the users and the workload that they are accessing, or between different workloads and datasets that need to work together, but are often based on different technology platforms. In the first instance, the appliance can be located as close to the main user groups as possible in order to minimise latency. In the second instance, it is better to locate the appliance within a community cloud alongside as many of the key datasets, workloads and platforms that need to interoperate and if possible to provide connectivity to this community cloud via secure, high performance networks.
For example, an NHS trust in the UK may have a collection of legacy workloads that are Microsoft or Oracle based, along with a few newer cloud native applications. It might also have legacy systems that cannot be moved to the cloud, but that could be hosted in a secure facility and it might want to access cloud based applications offered by leading health providers (either SaaS or ISV) as well as core data sets like the 100,000 Genomes Project database. Ideally the trust would want as much of this as possible available in a single community cloud with close proximity between systems to minimise latency. The trust would also want to be able to access this heterogenous environment via HSCN and also to be able to connect onwards to peripheral workloads hosted elsewhere or even to public clouds for things like artificial intelligence. Fortunately for healthcare and the public sector this is all available today.
Again, the clustering effect works with key workloads and datasets. As more workloads and datasets are hosted in close proximity to minimise latency, this provides an ever more compelling argument for locating further workloads in the same vicinity.
In 2018 we expect cloud to come of age – not only will it be the defacto platform for most workloads, but by leveraging multi-cloud and community cloud, as well as fog and edge computing, it will become truly customer-centric.