Cloudless

Does every cloud have a silver lining?

In recent years there has been a momentous shift towards the cloud. Whether it’s lifting and shifting workloads or going cloud native, the cloud has forever changed the IT landscape. AWS, Lambdas, EC2, auto-scaling groups, GCP, Azure, containers, serverless... cloudless? You will generally hear these terms bandied about in any IT conversation these days, all except for one - “cloudless”. 

It all started as a bit of a laugh when meeting with a prospective customer. We were talking about their future direction and we asked about their plans to push more workloads into the cloud. It was a foregone conclusion, or so we thought, that they would be moving more and more of their workloads out of their self-managed data centers and into the cloud. “We’ve already begun pulling out of the cloud,” they said to us. “What! You’re going cloudless?” we asked in return. It was odd to hear that in a world where everyone was heading to the cloud someone was doing the opposite. We all had a laugh, but we also started to delve into the “whys” with regards to their journey to “cloudless”. 

The customer found that they didn’t reap the benefits they had envisioned through moving to the cloud. The problems that they had with managing their environments carried over to the cloud, and in fact, the convenience of the cloud actually magnified their issues. In addition to the problems they already had, sprawling amounts of machine instances, blow-outs in cost and no noticeable improvement in efficiency compounded their concerns. Simply put, the cloud was adding a layer of complexity that they didn’t need. With the price of infrastructure falling and advances with automation throughout the stack, hosting their own cloud-like environments had actually become more economical than the cloud (and provided them an easier opportunity to address some of the fundamental management challenges they were facing). 

It’s also not easy (or cheap) to re-architect applications for a cloud world. Many of the applications we deal with today were never designed to adapt to the architectures on offer in the rapidly changing cloud landscape. Separating your backend databases into RDS instances, running your front end in an autoscaling group and augmenting this with a bunch of Lambda functions sounds cool - until one of them experiences an outage and you’re at the mercy of your cloud provider to restore service. The shift to the cloud doesn’t necessarily de-risk you; sometimes it can just shift who is blamed if something goes wrong.

And just like the cloud, some organisations aren’t ready for containers. Containers are the new frontier, just as virtualisation and cloud were when they emerged. It’s the shiny new toy that people have started playing with but not everyone knows for sure how this new technology fits into their world. Not everyone can be on the cutting edge. Generally the shift can be gradual (or sometimes glacial), allowing time for an organisation’s paradigms to catch up to innovation. We’re seeing container maturity increase, but it still remains a relatively new solution for many organizations that will grow over time versus overnight. 

These days, the cloudless conversation isn’t as foreign as that first time. It’s not to say that people are pulling everything out of the cloud; rather, there’s a realisation that there can be value in what’s already in place. The future of the IT landscape is already here with new cloud-native tools and applications surfacing daily and we are convinced the shift is real (but the hyped up time frames around how quickly things are all going cloud-native may not be realistic). Look around and maybe you’ll find that your fields of green and brown are actually really valuable under a clear blue sky, with not a cloud in sight...at least not for now.

(Thanks to some key contributors, David Alexander, Zac Law and David Callaghan)