Cloud as Code != DevOps

Flashback to two years ago. A customer indicated a DevOps interest. We polished our decks, we polished ourselves and showed them a crisp collateral by all standards - less than 10 slides, specific, excellent vocals and great body language to go with it. Once the pitch was over - customer, rather politely, clarified that they are expecting us to propose an organization restructure aligned to DevOps.

A foot in mouth moment, that is what happens with cliched terms such as DevOps. Thrown around too often, they lose their meaning and essence and can translate into a lot of things for different people depending on context. Cloud as Code sounds more around Public Cloud.

Cloud as Code is the real differentiator that Public Cloud brings about when contrasted with private cloud offerings, or SaaS products.

Resources, irrespective of whether infrastructure, applications, or integrations, can be created or modified on the go and discarded later. The use case could be a virtual data center to be provisioned, an environment for a specific business purpose, or even a single virtual machine for the next MVP. Such flexibility is possible at within few seconds to few days' time.

Cloud as Code is not an out of box offering.

Automation via Cloud as Code is possible only if Enterprises have a setup in place, which, at the get go, means -

• A coherent Public Cloud strategy - think Azure DevOps versus AWS CodeDeploy
• Hyperscaler compliant Landing Zone(s)
• Mature coding practices i.e. uniform environment principles such as minimum three level environments, e.g. Dev, Test, Prod
• Cloud CoE to enable roles needed for reviews and releases
• Cloud SIAM, needed down the road for a multi-Cloud strategy

Once this is in place, the following steps begin -

• The most common use cases need to be established to derive the maximum payoff from effort that will take to implement this automation via DevOps (1 week).
• After the use case is established, and implemented via Cloud as Code (1-2 weeks), it can be cascaded to other stack layers (Infrastructure --> Applications or vice versa) to enable end to end automation (1-2 weeks).
• Once automated, journey can move to Optimization phase - think code reuse and decoupled architecture

This is not a huge effort, and the gain is mammoth- not cost but efficiency via faster release cycles -> higher change velocity, which ultimately accelerates the speed of business for your enterprise.

So the next time you read about DevOps and its benefits, nudge yourself to think bigger and ask yourself if you could do Cloud as Code instead?