Why do you need a Serverless strategy

I recently was asked what is the use case for leaning towards Serverless when the same can be done with microservices. It is a great question and points a critical decision that most of us face today.

Lets go over the terminology and then we will talk about the progression.

Serverless and Microservices are orthogonal. Microservice is a design construct while serverless is an operational construct.

You can build an app with microservices design with serverless implementation. Lets go over the journey on the what one needs to move to the next level and what outcomes to expect.

We started from monoliths, they are easier to manage and maintain. Then as the product grew successful they became legacy. Always remember, companies have a legacy problem because their cutting edge monolith (idea) was very popular and needed to be scaled now. Focus on business and solving a problem versus spending too much time on focusing potential problems.

Monolith

Microservices

Serverless

Running a world class operations is not easy and you should avoid solving it unless you really have to.

Great breakdown of the decision matrix by @dwmkerr on how to decide what to pick.

TL;DR

Serverless is the future, ensure that you have a strategy for the future. If you can get out of the business of managing servers ( EC2, K8S agents) then I would highly recommend going down that route. Kubernetes is worth the overhead if you are at that scale and/or managing a data center and require a seamless transition OR have a requirement to run OS (scale, security, custom implementation, stateful or legacy apps)

Where is the industry headed

References

1. https://www.dwmkerr.com/the-death-of-microservice-madness-in-2018/
2. Comparison of FaaS: https://winderresearch.com/a-comparison-of-serverless-frameworks-for-kubernetes-openfaas-openwhisk-fission-kubeless-and-more/
3. https://www.suse.com/media/report/rightscale_2018_state_of_the_cloud_report.pdf