Feasibility of the discv5 feasibility study

For a lack of a better name, this is what I shall call this post. One of the issues I was working at vac was creating a feasibility study for discv5. The full issue can be seen here. A breakdown issue of how I personally approached this was also created which can be viewed here.

Now that I am a few days into researching discv5, a several things come up which affect the feasibility of the study itself and how best to conduct it.

So, here is what we know:

  • discv5, according to Felix Lange will “hopefully” be finalized by August.
  • The nim implementation is missing probably one of the most important components for the discv5 study which is topic advertisement.
  • There is a go pr that contains the initial implementation of discv5 that is closer to spec than that of nim. See #20750.

With the information we have we are in the perfect position to make a decision. These are the options I see:

  1. Instead of using nim for the feasibility study, use the go implementation.
  2. Take the time to develop the nim implementation AND then do the feasibility study.
  3. Delay the feasibility study until we are closer to the date at which discv5 finalizes.

I would currently be leaning toward 1.

It is true that the nim implementation is missing the whole topic advertisement part. Next to that, the “basic” implementation also still has several issues and is not fully compliant to specification even.

So running a simulation now with this implementation would probably make you hit several issues that need to be resolved first. Next to the fact that you can’t test with topic advertisement, which is a very important part for the study.

So yeah, opt for 1. or 2.
3. Better not, as I think this study might actually give good feedback to possible adjustments of the protocol. We are probably only one of the few that actually want to test this against resource restricted device usage?

1 Like