Fork detection and recovery

[mikera]

In some extreme situations, network forks may occur, e.g.:

• Network partitions
• Sufficiently powerful attacks by adversaries (e.g. >33% stake attacks)

Peers should have the capability to detect and resolve forks. Strategies are likely to be configurable by peer operators and could include:

1. Give operator decision to abandon minority fork and rejoin majority fork
2. Automatic fork rollback and recovery (if permissible and possible, e.g. no clients affected)