Cloud computing in the world of DevOps. Will IT help or add to the confusion?

Cloud computing in the world of DevOps. Will IT help or add to the confusion?

Welcome to the (confused) world of DevOps.  Some call it a movement, a cultural revolution. Some call it continuous development. Others re-brand everything ever done as DevOps. May the confusion prevail. It is an opportunity for innovation.

Distance between Dev and Ops

Starting from the organizational structures that built a wall of mistrust between Dev and Ops (also called Run by some) to deeply conservative cultures that dislike more frequent than monthly Ops deployments… all add distance between Dev and Ops in organizations.  I measure the distance in both time spent and miles travelled.  To reduce the distance one needs to cut both the time spent and the miles travelled. To me, DevOps is just an improved way of working.  A lot more collaborative and arguably more efficient way of doing what we have been doing for ages - as we managed our IT landscapes. What are the top 5 items that the so - called DevOps movement is trying to change for a manager? And can Cloud Computing help? 

Increase Velocity: Cut down time spent

Firstly, let us talk about time spent.  Most develop continuously, but test and deploy in monthly slots.  Add backlog and the picture accumulates shades of Red.  Literally, Red Tape.   We have seen cases where developed features are just dropped as they are no more relevant after 3 months.  Time is lost due to infra bottlenecks, differences in environments, issues with configuration control, improper prioritization & planning.

Reduce Distance: Cut down miles travelled

Second, let us talk about miles travelled.  A developed component from Dev team gets tossed over to testing team, then to a UAT check team, a security check team, to a pre-prod validation team, to a Go-noGo decision council and then, on a Go, to Ops team for deployment. These are miles travelled crossing organizational boundaries.  More miles get added as legacy structures, once considered relevant, continue to resist joint ownership and merged roles.  Merged environments are where developers build & test in production like environments. Merged roles should increase build quality while taking out time and cost spends and if possible the need for post event testing. 

Change Policy: Move to ‘Deploy When Ready’

Third is to adopt a policy of “deploy when ready.”  Move from ‘Deploy Per Release Plan and Down time window capacity’ to ‘Deploy When Ready.’ Dump all those emergency release meetings to evaluate if something should move to Ops on priority and ahead of upcoming monthly release slot.  Cut down wait times on completed Dev items. Deploy as often as needed.

So the needle shifts left to see how to get Dev components and their integrations ready faster, so we enable faster deployment into real use. The faster we deploy, the faster business takes it to end customers and market and the sooner we get real feedback.

Old policies are sacred cows no more.

Automate Process:  Are we anywhere close?

Automation is the key to accelerate all of the above. Are we anywhere close? Automate the entire Dev2Ops process or focus on Dev first and Ops later?  Tough  question.  Which vendor tools help automate faster and better?  Put money on old & proven war horses like the IBMs or move over to the new age Chefs and Puppets and Dockers?    

The answer is different for different contexts.  But this step is key.   Without automation of the DevOps train route, we can neither bring down the number of train stops and nor the risk of breakdown.

Bring Cloud Computing and sprinkle some Change Management. 

So what can Cloud Computing do? 

Firstly, let us start with the basic cloud promise.  Address Infra bottlenecks, clean up environment variance issues and even handle config control items that help roll back and roll forward.  All at low cost and on demand on the Cloud. 

Second, the tooling to run the show.  While most IT automation tools started life automating DevOps in the on-prem world, most offer solutions to both worlds.  And we already see that almost all Cloud Dev PaaS platforms offer inbuilt DevOps automation opportunities to tap. We need to find the right tools that run the show to seamlessly administer our DevOps train route covering both on-prem and cloud portions of it. 

Here are my best wishes.  Sign up and stay ambitious. Reduce the Distance between Dev and Ops in your organization. DevOps is worth the try and Cloud Computing is here to help.