Serverless means extendibility

22 Jan 2019 · Two minute read · on Gianluca's blog

News! Tech related notes are NOW published to ShippingBytes. See you there! I always felt this was not the right place for me to write consistently about tech and tools. So if you want to read more about that see you at the other side

I wrote an article about a GitHub Action I recently created to deploy my code to kubernetes. Very nice. Writing the action and the post, I realized what serverless is all about. I wrote it in the incipit of the article, but I think this topic deserves its dedicated post. Serverless is not yet for web applications. I know some of you will probably disagree but this is my blog, and that’s why I have one, to write whatever I like!

I used Lambda and API Gateway to distribute two pdf I wrote about “how to scale Docker”, it looks to me way more complicated than a go daemon. So I wrote them because I got the free tier and because I like to try new things. There are excellent applications written in that way for example acloud.guru but I am probably not ready for that! My bad

Anyway, I know what I am ready for: We should use serverless to offer extendibility for our as a service platform.

Good for us distributed system and hispters applications are all based on events, Kafka and so on. Plus now we have runC, buildkit and a lot of the building blocks useful to implement a solid serverless implementation.

It is not easy, at scale this is a complicated problem but we are in a better situation now, and it is a massive improvement from a product perspective:

  1. Using containers, we can offer total isolation, and we can take a very carefully and self defensive approach.
  2. An API already provides extendibility but, you still need to have your server and to run your application by yourself to enjoy them. With a serverless approach, it will be much easy for the customer to implement their workflow.
  3. You can ask your customer to share their implementation creating a vibrant and virtuous ecosystem.

You can use a subset of the event that you write in Kafka as a trigger for the function, VaultDB to store secrets that will be injected inside the service and so on.

There is a lot more, but I am excited! Is somebody doing something like that? If so, let me know @gianarb, I would like to chat!

Something weird with this website? Let me know.