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
Using a similar mechanism like the one to negotiate validator set changes, DynamicHoneyBadger could also allow changing parameters like the batch size (max_future_epochs — that doesn't have to be the same among different nodes anyway), etc.
Also, if we don't find a good way to allow new nodes to join in any epoch (see #129), we should allow the user to restart HoneyBadger for no reason, so that new observers can join.
Implementation
We'd probably add a new Change variant BatchSize (and maybe RestartHoneyBadger). It would allow validators to vote for batch size changes in the same way they vote for adding or removing validators. Once such a change has a majority, no key generation needs to be initialized. We'd just restart HoneyBadger with the new parameter.
The text was updated successfully, but these errors were encountered:
Just starting with this issue. I'm relatively new on Rust and also in blockchain :).
First readed "The Honey Badger of BFT Protocols" and refreshed Rust a bit.
Then I added BatchSize(usize) and RestartHoneyBadger() to enum Change and fixed compile errors.
Next will be tests and implementation.
Thank you for taking on this issue!
Yes, ideally batch size, threshold encryption (see #209) and subset optimization (see #243) should be dynamically configurable, and allowing a Honey Badger restart without a parameter change might also be a good idea.
If you have questions, don't hesitate to ask us (here or on Gitter).
The threshold encryption settings are already dynamic; other settings could easily be added to that. But I'd say it's not a high priority at the moment.
Using a similar mechanism like the one to negotiate validator set changes,
DynamicHoneyBadger
could also allow changing parameters like the batch size (— that doesn't have to be the same among different nodes anyway), etc.max_future_epochs
Also, if we don't find a good way to allow new nodes to join in any epoch (see #129), we should allow the user to restart
HoneyBadger
for no reason, so that new observers can join.Implementation
We'd probably add a new
Change
variantBatchSize
(and maybeRestartHoneyBadger
). It would allow validators to vote for batch size changes in the same way they vote for adding or removing validators. Once such a change has a majority, no key generation needs to be initialized. We'd just restartHoneyBadger
with the new parameter.The text was updated successfully, but these errors were encountered: