Cloud - Misunderstood
I've been "in cloud" for the best part of ten years all in the Capital Markets industry, although only really got my hands dirty in the last six years and really opened up the bonnet/hood in the last four. Looking back through the history books, it looks as though 2017’ish was the point of inflection where those that knew, we’re helping those that needed, and adoption accelerated. However, fast forward to 2024, and it seems that cloud has not touched everyone in the same way or, created that sense of revelation that it did for me.
Yes, we’ve got IaaS, PaaS, SaaS, FaaS and the list goes on… but then as someone reminded me, we also have “just someone else’s computers”, albeit nicely wrapped up with a console. The strange juxtaposition is that you could be OnPremise in a data centre, with a few racks of Dell blade servers, decided to migrate it all to your hyperscalar of choice and find that your compute moved across the floor/hall in the same building! But seriously, we know it’s more than just infrastructure. It's economies of scale, it's a service wrap, it's reliability...
What’s become more pressing recently is that there is a growing “cloud first” movement with impending deadlines attached, eg exit the on-premise datacentre by end of 2025, or, ensure compliance with the European DORA regulation for January 2025. That’s quite the corner to be backed into, but if nothing else it helps a business innovate, but what are the choices for migration? There’s the classic Lift and Shift where you redeploy your on-premise compute onto virtual machines (IaaS), you could add a little twist to that and use some of the hyperscalar’s ‘services’ such as databases (eg, AWS RDS), you could refactor the application and go containerised, you could…
One thing is for certain, there’s no one size fits all. Over the years I’ve witnessed some very brave individuals and teams, where the mantra is largely “Can we build/fix it? Yes we can!” (prizes for those that recognise the origin of that phrase!), but six months later having failed, they’ve realised they hadn’t considered; how will the business function with elastic (volatile) billing, what’s the business continuity strategy and SLA, what’s the application performance and monitoring framework look like, orchestration, infrastructure as code, project risk (costs, timeline, failure), support, operations, blast radius, skills, scalability (vertical/horizontal) and whilst there’s many more, security.
I drew a few conclusions from the experiences; for those pushing ahead with the “we can build it”, pause and seek peer review, it’s not just about the application. Secondly, is that, if SaaS is an option there are some key benefits, with predictability being one, and, another is the likely reduced total cost of ownership
TCO = infrastructure + software + people
Above all, don't contemplate running someone else's code in cloud unless it's thoroughly supported to run in cloud. Lastly, an organisation must think in a cloud first way, that means; being iterative about development and deployment, incremental with business benefits, ability to fund short term/agile projects, a level of abstraction on service provisioning, etc. I would also go out on a limb and say.. that if you haven't moved beyond bare metal servers your journey is probably longer that others who have.
Feel free to reach out if you want to learn more about my experiences, or some of the anecdotal stories that I've alluded to (no names though!), or just want to learn/talk more about cloud. I'm passionate about all things Open Source, Linux, Microsoft, Kubernetes, containerisation, DevOps, etc, and.. enjoy a coffee!
Cloud is definitely misunderstood but if you want to get hands-on it's easier than ever with no financial cost (AWS, Azure, GCP). Lastly, want to implement your own cloud, go look at LocalStack. Enjoy!
