Legacy applications are probably the biggest hurdle to overcome when moving to public cloud. How do we get these applications to run in the cloud?
There are obviously many options but sometimes the answer is simple: you don’t.
This week I showed a favorite slide during a cloud workshop with some of our clients. It’s deliberately provocative. It says: “Go native, or don’t go”. This slide always gets some interesting comments… but what does it really mean? And who really has this point of view?
Almost every migration framework starts with the Big Cloud Journey. What does your current estate look like and what are the options? Re-host, re-platform, refactor or even a complete redesign?
We at Fujitsu have adopted a framework that embeds all of these strategies. In our framework we talk about standard, rationalized and dynamic, which are basically:
- Standard: don’t change the applications, lift and shift to the cloud as they are
- Rationalized: introduce cloud solutions like PaaS
- Dynamic: the phase where the real interesting stuff happens. Think containers. Serverless. CI/CD. Service Mesh. Micro-services.
Whilst all steps on this journey are valid, there is no single answer in terms of the right approach. Where you start often depends on how technically ‘cloud-mature’ you are today… and of course how much so you are aiming to be in future.
But with more and more customer projects, I am reaching the conclusion that this can be a long, sometimes slow, and for some an almost endless path on their cloud journey.
Why?Because the actual business-case for cloud is not well-defined.
What do enterprises want to achieve in the public cloud? In most cases it doesn’t start with “get my on-prem’ systems to public cloud’”. The business doesn’t ask the IT function to move systems around from platform to platform.
Most of the time, they don’t care where or how they are hosted… and it’s fair to ask “why should they?”
The cloud journey starts with a whole different type of question:
- “We want to do predictive maintenance and anomaly detection on our systems in the field”
- “We want to run smart data analytics to know ourselves and our customers better”
- “We want to make our in-store point-of-sale systems deliver greater value”
- “We want to leverage breakthrough technologies at scale to drive a great customer experience”
You can think of anything – but these are real business driven questions – often underpinned by anything that produces data. As an enterprise, we want to reach a higher level in our services to our customers and to enable that, we want to make better use of that data.
Indeed, data is the new gold.
That’s where public cloud really comes in. Data lakes, data analytics, machine learning, deep learning. That’s a whole different beast than ‘we want to move our systems into the cloud’. In other words: IaaS is not the killer app in public cloud. Data is.
So why start with a migration plan to lift and shift systems to the public cloud? Whilst this can lay great foundations for getting more out of existing IT, this takes time… and there might not be a real business need for that move to happen right now, unless a client really asks for it.
I’ve seen many examples where the business case for simply ‘lifting and shifting’ systems to public cloud in isolation is really bad and there’s no business added value.
We need to start with the ideal business outcomes – and using a cloud-native approach that leverages data as the vehicle for achieving these; not falling into the trap of doing the same as you have always done, with the actual datacenter location – on premise or cloud – being the only difference.
Don’t let legacy with all your complicated enterprise core systems hold you back.
Yes, there needs to be a place and a future plan for those… but the biggest ‘business value-add’ of public cloud is in utilizing cloud native environments for (new) business functionality.
You can start today with that.