Skip to content

Instantly share code, notes, and snippets.

What would you like to do?
2020-03-12 Eth Impl Call #35
- Created targets for March sprint
- Interop
- Starting auditing
- Dropping completely libp2p daemon
- Almost at v0.10.1 with the new BLS changes
98% work done
delayed due low-level bugs
- Progress on attestation aggregation work
not fully done but mergeable
- CI automatically launch a local network and ensure finalization
- Discovery: can interop with Geth discv5
cannot interop with Lighthouse yet
- Several improvement to libp2p stability
- Log analysis tool to track libp2p errors
- Fix a critical performance bug in libp2p
that lead to buffer of size 1 instead of 1MB
- Impact: we can handle 5x more traffic
and now we generate 5x more logs
v0.11: DOS hardening on the network
pyspecs, cleaning up
changes in state_transition
proposer_index added to block to ease using block without the state
No more changes in state transition expected
More changes expected for networking/multiclient testate
Next tests: fork choices
Proto: ZCLI/ZRNT update to 0.10
Rumor for gossip sub
Client updates
Research on () system
ENR in its own crate
Chat with sigp, and felix to improve Discv5 speed
ENR (?) query
Staker survey
[… missed]
2 new developers
Interop with lighthouse on EthDenver branch
0.1 version of slashing service live
—> report double votes
New SSZ, significantly faster
Uses a code generator that output a marshal function
Issue was Go absence of support of generics
Dynamic Substitutions of Subnets
Discv5 merged in
Can discover Lighthouse but cannot discover own nodes
Good progress on RPC
- interest in Pegasys to standardize across clients
- strong interest with Infura
Perf: remerkleable:
4x speedup on Block processing
4x slowdown in Epoch processing
Next perf target: sync / database slowdown
Fast verification of aggregate signature:
Focus on stability
Discv5: interop with Geth and Lighthouse
Updates to latest spec release
Audit will result will be made public soon
- trying to sync with other testnet
- Stability of code
Nethermind Tomasz (sound issue)
- 10.1 spec!
Audit results. Will be released ASAP
Phase 1: major PR to be merged soon
TXRX - Joseph
Writeup on safety of Eth1 follow distance
Cross-shard transaction
Polynomial commitments for state roots (longer term proposal)
as a replacement to binary merkle tree
Allows very short witnesses for large (?)
Drop the bottleneck for stateless clients by 90%
Requires digest using polynomial commitment:
Relies on Math invented 2 weeks ago
For starter:
- block bodies
- or receipts
- Eth1.X, looking into Eth2 research that can be reused for Eth1.x
See notes from last week call from Ben
Proto: EthCC working with Dmitry (Status) and Felix
March: 0.11, multi client experimentation
April: Multiclient testnets
in-situ interop event, not anytime soon
V: Phase 1 implementation?
—> party interested but not started.
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment