DevOps - How infrastructure teams fit in?

DevOps as we all know is a cultural shift to an agile development methodology which helps reduce time to market by automating various stages of development, testing and deployment to production. This automation is not one time, the cycle repeats every time when there’s a new release, bug fix, enhancement or feature addition. This approach is critical to organizations who provide customer facing products (applications/websites etc.) especially when it is generating the revenue. In a competitive world, following are few of the key asks which are driving organizations to adopt to DevOps:

-         How to keep the product up and running while continuously enhancing its functionalities?

-         How to keep it secure?

-         How to reduce manual errors?

-         How to reduce time to market?

-         How to make it cost effective?

-         How to scale it based on requirement?

While tools and techniques of DevOps complement the software development lifecycle to address the queries above, what role does infrastructure plays in it? Is it just about automating provisioning of servers or the application tiers or is it more than that? How infrastructure people can complement the development and operational teams to meet the end goal?

It will be easier for any developer to understand the DevOps concepts as he/she would be able to correlate the tools and the concepts of development more quickly than any infrastructure engineer. Traditionally these two teams have been working in silos without much interaction with each other. However, this is almost unavoidable in case of DevOps and therefore, important for infrastructure teams to move up the hierarchy to understand and adopt to these tools and techniques. While they don’t need to learn how to code or develop the end product, but they certainly need to understand how and where the product is dependent on infrastructure and how to meet those demands at same pace as development. In simple words, infrastructure should neither increase the time to market for the application nor its performance in production no matter what the demand is.

At a high level, development in DevOps follows the concepts of Continuous Integration & Continuous Deployment (CI/CD):

As part of the life-cycle above, infrastructure teams need to complement the development and run part of the product/application to achieve the end results. To achieve these, they need to focus on the following areas:

-         Infrastructure Deployment (Provision servers stack, storage, containers etc.)

-         Infrastructure Testing (Ensure security, connectivity etc.)

-         Platform Deployment (Deploy web, database platforms etc.)

-         Configuration Management (Deploy approved code to the platform)

-         Run Operations (Auto scale, maintain server state, patching etc.)

When put together alongside development, this becomes the DevOps framework:

For each of these stages, there are a number of tools available in the market which will help automate any specific stage or an end to end solution. The choice of hosting (AWS, Azure, and Private/Hybrid Cloud) will help you identify which is the right fit tool for your environment and many of the tools used for managing application development can also be used to manage the infrastructure life cycle e.g. Git, Artifactory, Ansible, New Relic, Splunk etc.

In summary, for Infrastructure teams to adopt to DevOps culture, Infrastructure as Code, Infrastructure Testing, Configuration Management and Monitoring are key areas where focus will be required and this is irrespective of whether you are using virtual machines, containers or server-less computing as a platform to host/support the applications.