Serverless — What’s all the fuss about?

For years the approach for any application and platform was to set up physical servers inside our data centres, and when more capacity was needed, we just dispatched our engineers with more boxes to install and deploy. The issue with that was the amount of time as well as downtime required to get the system up and running. We then enabled virtualisation, allowing our engineers to set up new instances from the comfort of their desk minimising the downtime.

Many companies took that approach when moving onto the cloud. And at the very beginning that was the correct approach, however, the cloud continued to evolve and have switched the approach to serverless design. Now, what does it mean for you? First of all the serverless word isn’t the most correct here, but it’s been universally adopted. It doesn’t mean that there aren’t any servers running; it just means that you don’t have the burden of management or maintenance of those servers. Your team no longer need to worry, plan, manage and operate the server infrastructure required to run their application or platform. More importantly, the serverless approach removes the need to worry about plugins, patching or even load balancing.

This approach allows your team to concentrate on the code and features freeing them up to be creative in places where value gets generated instead of being bogged down with delays and compatibility issues. Critically this approach allows you to increase flexibility while limiting the amount of time required to go to market.

So to capture critical benefits of going serverless design:

  • Cost optimisation — you only pay for the time when your code runs
  • No infrastructure to manage — it’s not that there isn’t any, it’s just someone else makes sure it’s operational and available for you
  • Automated scaling — in case of the spike of request the application will scale up to meet the demand

Now one thing to stress, “Serverless” isn’t one fit for all. There are still many use cases where taking the traditional route of dedicated instances might be required due to performance, security or availability. Here at Deployflow, we had the privilege of helping many customers making the best decision for their platform. Why don’t you let us help you?



Published on April 30, 2019