I sat in a meeting last year as everyone round the table was asked to not use DevOps as the answer. As a company, we were not using it well so what other options were available to us.
For many, that was a bit of a shock, but I really understood where that person was coming from: writing individual scripts for huge cloud deployments were not working for us.
I took that away and finally started working on augmented delivery. The basic idea is that any deliverable can be defined as a model. That model should enable you to define the pertinent information for any product and then generate all the human readable information you need.
Does that sound interesting?
Top comments (1)
I guess it depends on how you define "DevOps", a term that's frequently misused (especially in large organisations!). Going to ignore that political discussion for now :)
I do like the idea of a model-driven delivery system - declare what you need (bad example: "standard 3-tier application platform with SQL storage, scalable from 1-1000TPS, observable though dashboards and able to support separately deployable containerised services") then the tooling generates your delivery assets, such as cloud templates, deployment pipelines and gives you cost controls and purchase options. When you're happy it spits out the documents for auditors too. Kind of a meta-model above today's declarative cloud templating tools?