You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
{{ message }}
This repository has been archived by the owner on Aug 2, 2022. It is now read-only.
To achieve true BFT guarantees requires a double confirmation from producers. The discussion with Vitalik in issue #2718 points out how our current improvement is incomplete.
LIB does not advance until 2/3+1 producers implicitly agree on the LIB
No new information is required in block headers, as each block has an implied LIB already which we will now consider the the Proposed LIB (PLIB) which only becomes LIB after 2/3+1 agree on the PLIB.
Side effect is to increase latency on producer schedule changes.
BFT messages for real-time confirmation of blocks (outside of block headers) will require the double-confirmation same as all other BFT protocols. This will increase low-latency proofs to 30 to 37 signatures plus 21 bytes of signaling information.
The text was updated successfully, but these errors were encountered:
To achieve true BFT guarantees requires a double confirmation from producers. The discussion with Vitalik in issue #2718 points out how our current improvement is incomplete.
No new information is required in block headers, as each block has an implied LIB already which we will now consider the the Proposed LIB (PLIB) which only becomes LIB after 2/3+1 agree on the PLIB.
Side effect is to increase latency on producer schedule changes.
BFT messages for real-time confirmation of blocks (outside of block headers) will require the double-confirmation same as all other BFT protocols. This will increase low-latency proofs to 30 to 37 signatures plus 21 bytes of signaling information.
The text was updated successfully, but these errors were encountered: