By Neil Thurston, Chief Technologist at Logicalis UKI
Thanks to William Shakespeare for lending us his eternal question to adapt. Unfortunately, when it comes to cloud the answer isn’t quite as binary - so let’s discuss this in a bit more detail over the next few minutes.
When cloud services first emerged you could literally describe them as someone else’s data centre or next generation hosting – they provided pay-as-you-go infrastructure services that suited short-lived workloads such as marketing campaign sites and software development environments. The appeal was the on-demand availability of the pre-built resources that didn’t need IT involved to do designs, procure and deploy kit or maintain and operate it – which saved weeks-to-months of waiting - and no capital budget was required, it could be expensed.
If the cloud providers had stopped development there, then cloud wouldn’t be the force it is today. Instead, they poured $billions into development of general platform and software services to simplify operations, serverless and low-code services to simplify and democratise software development, data lake and analytics services to simplify and democratise data science, industry-specific platforms, and so on. All services being available to instantly consume, without any upfront costs or waiting time. This naturally brings new challenges with it, such as skills shortages to operate these new technologies and the need to control operational costs.
So should you put everything into cloud? No one disagrees that you could technically put any workload into cloud today – but that doesn’t always mean you should. Many organisations are restricted by regulations for some workloads and data, for which the simplest solution is to have them on-prem. Others simply haven’t transformed their financial operations to switch from capital expenditure (capex) for IT and the tradition of sweating of assets, to the land of IT operational expenditure (opex) that requires constant consumption-based outlay.
Some of the cloud providers, like Microsoft and Oracle, are major software players and commercially incentivise their install base to migrate to their cloud offerings. The reticence to take up such offers is typically down to seeing this as more stick than carrot but software-as-a-service (SaaS) plays usually trump technical debates and tend to make business sense.
At the same time, the infrastructure in the data centre has been cloudified. Some literally with hybrid cloud options from the cloud providers pushing their technologies on-prem. In the main it has been the transition of infrastructure to virtualised and software-defined technologies that can be automated to the same levels as cloud infrastructure services. Finance models have changed too with most solutions these days available via consumption-based opex models, as well as capex, to cater for both financial models.
Whatever your stance on cloud, the answer is that you will end up consuming some of its services, one way or another. Equally, it’s likely that you will continue to run services on-prem (I include 3rd party data centres under that banner) - just look at the “born in the cloud” big tech digital disruptors of the last decade who now run large data centres integrated with cloud services.
With so many cloud options on the table – public, private, hybrid, multi, poly, industry, etc – the right thing to do appears to be to let the right workload sit in the right cloud, where the commercials, regulations, performance, etc define what ‘right’ is per workload. The issue with this approach is that you will end up with islands of innovation across a wide landscape, requiring multiple skillsets to manage and secure it. But that’s actually OK if you’ve planned for it and invested in appropriate hybrid multi-cloud skills, tools and policies to embrace it.
So what’s my conclusion? I think cloud poses a quantum question with several outcomes – to be in cloud and not to be in cloud at the same time, depending on the workload. The point is that left to evolve organically, your cloud environment could quickly become overwhelming to manage and secure, as your teams to choose best-fit locations for each workload or function. It’s far better to create a cloud strategy that maps out the end game, then everyone can prepare for it, knowing some things will need to be transformed to align to it. A strategy also maps out how you will cater for the next new cloud thing, so that when it comes along you can embrace it within the boundaries of your defined policies, gaining competitive advantage. Even if you have already ventured into cloud without a strategy, it’s never too late to create one and adapt what you’ve already deployed to align to it.
If you need help making sense of cloud, maybe with creating a cloud strategy or executing it, then book a workshop with us today at Logicalis and let’s have a chat about it. Logicalis Executive-Level Workshops