Cloud is easy. Cloud is hard.

Written By:
Published:
Content Copyright © 2022 Bloor. All Rights Reserved.
Also posted on: Bloor blogs

Cloud is easy. Cloud is hard. banner

Somewhere in the dim and distant past (probably round about 2010) an alliance of cloud native enthusiasts and technology vendors spun a wonderful story about the Cloud and how businesses could get more, for less, faster by adopting it. To be fair, if the world was all Cloud-Native, they’d probably be right. But the world isn’t like that.

As an IT analyst, I sometimes feel as if I exist in two parallel worlds…one where the inhabitants are all IT specialists, and most are Cloud natives, and the other where the inhabitants are mostly businesspeople and a few IT old timers who never got sucked into this new world. Allegedly, both worlds speak English as their main language, but to businesspeople the words coming out of the mouths of IT folk more resemble French… you recognise some of the words, but you don’t really know what they mean.

Fed on the stories of faster development of business applications to meet rapidly changing business needs, not having to buy IT servers and retiring costly in-house data centres, companies having been busily mandating cloud first, or even cloud only development policies. Add in the apparent imperative for companies to digitally transform, and you can see that the Cloud might look like the answer to a drowning person’s prayers. The reality is somewhat different. Moving legacy systems into the public cloud is complicated. The reasons why these might fail, or, at best, be disappointing in terms of both performance and cost, are many. Often, they lead to the system being repatriated to in-house facilities. This can get to be a very detailed and technical debate, but type Dilbert Kubernetes into Google and you will get a humorous, scarily accurate picture of the situation in a 3-picture strip cartoon.

Even in a cloud native world, complexity and complications abound. Open source, DevOps, micro-services, containerisation and other developments have contributed hugely to the success of the Cloud. But the pace of application development and deployment has left IT operations teams paddling like fury to keep up. In a world where what was once handled by the vendor inside the mainframe, has now been broken out into its component parts and physically distributed all over the network. Despite the promise of DevOps collaboration, in a Cloud world it always seems that IT Ops is left to plug the holes that appear.

Let me use a cricketing analogy to illustrate the point. Early cloud developers were like opening batsmen in the 1960s going out with only pads, fairly small gloves and a box, to face the awesome sight of the West Indian duo, Wes Hall and Charlie Griffiths running in and bowling 100 MPH bouncers at you. After people had been hit, and hurt, often enough, protection got better. Body pads were added. Gloves became bigger and more padded. Eventually helmets replaced caps. Then health and safety kicked in so that faceguards were added to helmets. Arguably, cricket is now more exciting, and safer, than ever.

I could go on, but hopefully you get my point. Containerisation was great, but orchestrating them was a pain. Enter Kubernetes, which was great at orchestrating containers, but fiendishly complex to configure. Enter Terraform to automate the configuration tasks. DevOps drove the development of CI/CD pipelines (Continuous Integration/Continuous Deployment). Yet the Ops part of DevOps, seemed to have been left out. Enter Gitops to provide operations with a single version of the truth and the ability to participate fully in the CI/CD pipeline. Replicate these sorts of issues and rapid fixes across the challenges of turning “I need a worldwide, web based, on-line order system for my consumers, by tomorrow”, into binary machine code and then deploy it so it is secure, resilient and performant is not a simple task, however the Cloud may make it appear on the surface.

Business and IT are desperately in need of finding a common language to communicate in and explain that there are speed vs cost vs risk implications to agree on to ensure that this wonderful Cloud world can be delivered without the bouncers hurting too much.

Post a public comment?