Updating and making it easier to run benchmark measurements for ODoH

[sudheesh001]

• Updates the ease with which benchmarks can be run
• Makes corrections to the --out benchmark output
• Adds scripts to download, and prepare dataset (Tranco) as an example

Signed-off-by: Sudheesh Singanamalla sudheesh@cs.washington.edu

[sudheesh001]

@chris-wood @tanyav2 What do you think?

[chris-wood]

Is there a reason why this needs to be in this repo, as opposed to some other script or program that uses the client directly?

[sudheesh001]

I think it'd be useful to have this directly with this client codebase as a reproducible artifact for the paper. A reviewer or someone wanting to test the benchmark portion of the client should be readily able to test it out of the box. So the documentation updates and getting the dataset and preparing it correctly might be valuable.

[chris-wood]

I think it'd be useful to have this directly with this client codebase as a reproducible artifact for the paper.

I understand, but that doesn't mean it needs to be in this repo, right? One could, for example, put the script to fetch data and drive the client in a separate repo, and that repo could just clone this one in the process and build the client.

[sudheesh001]

That is a possibility.

[chris-wood]

Let's do that then. I'm not sure piling on things here is best. Closing!

[sudheesh001]

Sure, I'll open a separate PR for fixing the --out issue