Content Copyright © 2019 Bloor. All Rights Reserved.
My colleague, Dr Richard Sykes and I have been looking at how fast the underlying IT infrastructure layer of compute, storage and networking has been diversifying.
In order to illuminate what has been happening we need to step back in time to when Richard was the new CIO of ICI (Imperial Chemical Industries). He remembers that the ‘battle’ then was the central data centre (he inherited a large one at Runcorn) vs client server boxes ‘down the corridor’. His inherited computing systems manager, proud of his big central ‘machine’ at Runcorn (their on-premises facility), wanted Richard to ban client server boxes ‘down the corridor’ in the individual ICI businesses. Richard explored, took advice and his computer systems manager went off to run the IT at Sheffield Hallam University.
Soon after, we got virtualisation and migration to client server architectures on-premlses. Then followed the arrival of the hyperscale public cloud providers like AWS, Microsoft Azure and Google Cloud Platform (to name but 3) and the development of the hybrid concept, where enterprises started to ‘blend’ on-premises and public cloud.
Now, let’s step forward to 2025. The geography of ‘compute, store and network’ has become much more diverse. You have a highly distributed IoT-driven ‘edge’. You have companies, like Bristol’s Graphcore, offering very powerful AI-focused machines that in physical terms are back to the local ‘box down the corridor’ but with the capacity of AWS. And, courtesy of Big Software, boundaries of what is data storage, what is data processing, what is data networking are sufficiently blurred that a whole range of entrepreneurs now offer (this is 2025 don’t forget) a plethora of ‘compute’ services. Furthermore, key innovators such as Apstra demonstrate that you can ‘architect’ the overall whole in digital terms to be self-managing, self-correcting, and assuring compliant in real time.
Imagine a multi-dimensional ‘digital fabric’, the flexing, re-weaving and re-tailoring of which, in real time, is quite straight forward. ArchOps is then a serious conceptual development. It says define clearly the operations (Ops) you require to deliver your business objectives. After that, standard processes will then scope and deliver the digital architecture required (Arch) to make them happen.
So what is happening now that leads Richard and I to envision that 2025 scenario? The flexible, hybrid infrastructure components we are seeing emerge, once in place, clearly enable rapid adjustment in infrastructure operations. The practical reality of ArchOps is that, with an Arch that is fully software based and virtualised, we can rapidly ‘tweak’ Arch to rapidly ‘tweak’ Ops. The strength of the Apstra offer, for example, is that Ops can be monitored and managed in real time, sensing and correcting systems, including regulatory failures. But the implication is that Ops can also be adjusted simply and in real time for other purposes – such as changing business requirements – by adjusting Arch.
Software can now be presented and managed as a hierarchy of services, such as underlying & enabling infrastructural elements – IaaS – and enabling platforms – PaaS, exploiting the new architectures of microservices, containers, open APIs, Kubernetes. These services may be delivered through in-house, on-premises operations or sourced externally – SaaS – but the overall message of the Apstra modus operandi is that the whole (it is all software after all) can be managed and self-managed holistically, whether in house or externally.
If we read it correctly, the parallel developments in data management (including and especially exploitation of data lakes etc.) mean that, providing an enterprise’s data is thus broadly structured in this contemporary fashion. Therefore the ArchOps ability to enable ‘flex’ will apply as much to data exploitation as the infrastructure.
IT’s sales pitch to the Board then becomes:
‘You likely associate your IT Function and the CIO with a steady stream of proposals for big, significant investments (big ‘throws’) that take time to shape and implement – and can be disruptive in the process of implementation.’
‘In this new era, your enterprise’s technical architecture can be developed to a state where it is both highly business focused (delivering on key business performance objectives in real time) and robust (i.e. cybersecure) while also being flexible and rapidly responsive to changes in business demands. And the cost of each change is small (small ‘throws’), with minimal disruption, and rapidly delivered.’
To hear more about ArchOps come and join Richard and me in the CIOW Digital Board session on Wednesday 24th April at 1PM