Eth2 Call 42 Agenda

[djrtwo]

Eth2 Call 42 Agenda

call 41

Meeting Date/Time: Thursday 2020/6/25 at 14:00 GMT

Meeting Duration 1.5 hours

Livestream Link

1. Testing and Release Updates
2. Testnets
   a. altona
   b. attack nets
3. Client Updates
4. Research Updates
5. Networking
   • a handful of network interop items to discuss
6. Spec discussion
7. Open Discussion/Closing Remarks

[protolambda]

Spec discussion agenda point:

Discuss spec changes & progress for direction towards more mainnet settings.

Think of:

• With snappy as default for RPC, maybe disable and/or remove the non-snappy variant. (Like input here for possible use-cases of the non-snappy)
• Yamux support was recently discussed on discord
• Noise support, and/or fallback to Secio (part of Prysm audit discussion too)

P.S.: looks like the date in the agenda description is wrong? (Same as the previous call agenda?)

[AgeManning]

I'm not able to make this call, however my thoughts for each of these are:

• There's little overhead in supporting both. The downside of keeping non-snappy is that there could be implementations that don't add snappy requiring extra bandwidth for syncing etc. If we keep it, clients may want to limit requests on that protocol. I can't think of any immediate use case of keeping it, other than ease of developers to jump in and hit clients for requests without the added complication and dependency of snappy. I don't have a strong opinion either was to whether we keep or not.
• We are having issues with yamux support currently. It seems its just us (Lighthouse) and Prysm that currently support it. My understanding is that is more sophisticated than mplex and potentially better for scaling (although have no benchmarks or tests to justify this).
• Our plan is to remove Secio support and only support noise. We temporarily allow the Secio support so we can still connect to prysm.

[benjaminion]

My quick notes