Implement multi-hop routing of consensus messages #472
Conversation
… but is not yet integrated.
- reqwest doesn't support local sockets, so I've hacked around it by creating an in-process proxy
…a new consensus. This was causing premature commitment when the old consensus was a single producer
- Reorganize to use unittest - Add control for node logging
- Wrong index caused incorrect behavior of is_sole_producer and could also lead to dereferencing a null pointer - set_producers can now enter the default implementation in more cases, because we call it with the actual producers of each block
- View change now includes a base block - Fork restrictions are now based on the view change rather than the best prepared block - Ensure that view changes are synchronized network-wide - Propagage and use joint producer view changes, including when the new producers are not BFT - Fix missing connection state change to ready when we have no blocks
…the first error in each loop
…r end of the connection.
| - Y has been committed by 2f+1 producers, because that is the requirement for it to be considered irreversible | ||
| - let Z be the earliest block that is an ancestor of X (inclusive of X), conflicts with Y, is better than Y, and is prepared by 2f+1 producers. Such a block is guaranteed to exist because X itself must be prepared by 2f+1 producers in order to be committed by any honest producer. | ||
|
|
||
| The existence of Z implies that at least f+1 producers have violated the protocol |
There was a problem hiding this comment.
Doesn't the existence of X and Y already imply that at least f+1 producers have violated the protocol?
There was a problem hiding this comment.
It does. This list is the breakdown of exactly how they violated the protocol.
| { | ||
| if (route.second.metric == RouteMetric::infinite) | ||
| { | ||
| return true; |
There was a problem hiding this comment.
Shouldn't infinite metric return !feasible?
There was a problem hiding this comment.
https://datatracker.ietf.org/doc/html/rfc8966#feasibility-condition
"Note further that retractions (updates with infinite metric) are always feasible, since they cannot possibly cause a routing loop"
There was a problem hiding this comment.
Based on the name, I expected this to test that the routing would follow the shortest path, but the tests really seem only to verify that multi-hop routing works (and will only route messages to producers).
There was a problem hiding this comment.
I'm not exactly sure how I would test that. The exact path followed is kind of an implementation detail.
James-Mart
added
the
Node infra
No description provided.