What are the challenges of bootstrapping a PaaS?

PaaS abstracts infrastructure to make it more developer-friendly.

For one, there are many many competitors.

Also, PaaS requires a lot of trust.

You are giving the keys to your kingdom to somebody else. Not only are you sharing your code, you are also relying on the PaaS to safely guard your users' data.

Are you thinking of building a PaaS?

0 Likes
codexsw Author

@gabriel4649 I am. Around Serverless Background Tasks.

In a few words: Register your event types and subscribe your API endpoints to them. We take care of Security, Resiliency, Retries, etc.

Our solution provides a good developer experience via SDK (NodeJS/TS, Python) and several wrapper libraries for popular web frameworks (NextJS, ExpressJS, Flask, Django) and UI and CLI tools.

Building MVP, launching EOM.

0 Likes

@joshmanders is uniquely position to answer this.

0 Likes

It's can be difficult to reduce the scope of PaaS to be able to create an MVP. You often have to create most of the infrastructure and have high reliability before there is any value you can offer customers.

A lot of PaaS products are B2B, which may take a lot more effort for sales and the sales cycles can be a lot longer… sometimes even annual for large companies.

They can also be traps for software engineers who really enjoy building and refining architectures.

0 Likes
Andras Bacsai

I'm bootstrapping a hassle-free self-hostable and open-source "PaaS" but not yet provide a cloud version.

The MVP is hard to define because I would like to include a lot of stuff in it - but that's the case for almost every side-project I have 😂 Really include only the minimum features. It doesn't matter if you only provide 1% of your mission in the MVP.

The best thing you can do - fits almost every project IMO - is to do small iterations over-and-over again.

Mine looks like this:

  1. Simple UI, only host Node/Static sites/apps.
  2. Provide a hard self-hostable option; lots of manual work is needed. (oops 😄)
  3. Release beta - got minimal attention on Github/Twitter. 🎉
  4. Polish self-hosting part to be only one script and fully automatic. 🤖
  5. Better UI.
  6. Add a Database hosting option. (I'm here currently.)
  7. Release MVP / first non-beta version.

(There is no high-reliable option atm, as I'm focusing on indie hackers, hobbyists, where it's not the most important thing to start hacking.)

Imho, self-hosting and open-sourcing are the 2 things that differentiate mine from competitors. Not sure if I'm right, but I enjoy doing it and getting some positive feedback - so I'm happy if I'm incorrect 🤷‍♂️.

0 Likes

Please sign in to leave a comment.