How to get the most out of the Hyper-scalers
Justin Campbell CTO@junkshon
I know I’m sounding like a stuck record here, but cloud migration should never be viewed as a ‘lift and shift’ exercise. You’re not just moving your IT workloads from your data centre into someone else’s, and if you view it that way you’re missing out on a huge opportunity to modernise and future-proof your IT for years to come. Not only that, but if all you are trying to do is take advantage of the economics of cloud, then there’s a decent chance you won’t even achieve that. And that’s because moving outdated, inefficient systems into a third-party data centre will just port over the same legacy issues you had already, and you might well find – once you’ve ‘lifted and shifted’ it all, that you’re spending as much as – if not more – than you were before.
Cloud migration represents a brilliant opportunity to extract value from the billions that Microsoft, AWS and Google are investing in their cloud platforms to help you transform your IT in a way that could be game-changing for your business. They are constantly innovating and bringing to market new services to make your journey easier and better. With the best will in the world, there is no way you’ll have the expertise or agility to out-innovate them. So, your best bet is to shift your mindset to one where you just look to ensure you get absolutely maximum value from these behemoths, whichever one you choose. And focus your own resources on your core competence.
At Junkshon, we’re actually a pretty good example of exactly the point I’m trying to make. Our development focus is on our Aspire platform and its ability to automate the process of building a migration strategy based on the IT estate data we ingest. Working with the cloud providers, we’re able to leave the infrastructure heavy-lifting to them; in other words, we don’t need to commit our own resources to building infrastructure components. We also use their database / PaaS capabilities so we’re effectively outsourcing the scaling and high availability element as well.
As I said, the key here is that the hyperscalers allow you to focus on what matters most to your business; develop application code that’s directly relevant to your specific area of competitive advantage and leave them to do the rest.
To a large extent, the decision you have to make here is one of ‘build or buy’. You might start out with the intention to use the hyperscalers purely for IaaS and do all your own application build. But you’ve got to ask yourself if this is the best use of your resources, and whether or not, in the end, it will be cheaper doing it this way.
The ‘buy’ option will almost certainly give you more agility, cost less, and insert you in a stream of constant innovation from your chosen hyperscaler which will give you more and more gains over time. As an example, you’ll see a performance improvement every time the hyperscaler introduces a new chipset or new software innovation, an example of this is the emergence of serverless technology to only pay for the time your actual code is running.
Remember, the hyperscalers really do want your workloads, and they are competing against each other to make their platforms, and their programmes of innovation, as attractive as possible. It’s a buyer’s market. For example, each service provider will provide technical resources to help you build a landing zone; the first critical step on the path to your migration. They will help you move away from legacy, proprietary database technology and adopt more modern approaches.
Having said all this, it definitely helps to work with a partner who can help you navigate the hyperscaler landscape, from choosing which one (or ones) to work with, through to building a multi-year roadmap of progressive migration and modernisation that tracks the hyperscalers’ own innovation plans.
Obviously, there are lots of companies out there offering services around the hyperscaler ecosystem, so how do you find the right partner?
Experience – at Junkshon, we’re basically practitioners. Yes, we have tools to automate complex processes and help compress timescales, but the DNA of our organisation is that we are experts in migration and optimisation. Not only can we match your use cases and workload characteristics with the most appropriate hyperscaler, but we’re actively looking for proof that things will actually work in practice. We’re also building up a huge reference library of application workload patterns to help you make informed application decisions.
In-depth knowledge of Hyperscalers’ programmes – each hyperscaler has its own workflow migration programme, and you need to be able to tap into those in order to leverage their commercial and technical experience. At Junkshon, we’re constantly optimising our pricing engines and working in lock-step with the hyperscalers, so we know these programmes inside out and can make sure you benefit from them.
Ongoing optimisation – as I’ve said previously, we build long-term migration plans where we help you progressively optimise performance over time. As part of that, we do a quarterly review and health check, where all your data is refreshed.
My final piece of advice, here, is that even with Junkshon at your side, we’d recommend you have a cloud champion in your business; ideally, someone who has experience of working with the hyperscalers.
We can guide you, but we’re not truly inside your business. A cloud champion will also help get the hyperscalers invested in your programme and, not surprisingly, that can make a huge difference.