Right now buffer nodes submit tickets upfront, so they pay gas even when the top N all submit and no replacement is needed.
Proposed behavior:
Only allow buffer nodes to submit after a fault is detected. For example: a fixed window (e.g. 10 minutes) for the top N to submit; when the committee is finalized, if the threshold is not met, open a short buffer window (e.g. 5 minutes) where other nodes can submit tickets to fill the remaining slots. That way buffer nodes only spend gas when they’re actually needed.
Requires contract and node logic changes; to be implemented later.
Related: PR #1224
Right now buffer nodes submit tickets upfront, so they pay gas even when the top N all submit and no replacement is needed.
Proposed behavior:
Only allow buffer nodes to submit after a fault is detected. For example: a fixed window (e.g. 10 minutes) for the top N to submit; when the committee is finalized, if the threshold is not met, open a short buffer window (e.g. 5 minutes) where other nodes can submit tickets to fill the remaining slots. That way buffer nodes only spend gas when they’re actually needed.
Requires contract and node logic changes; to be implemented later.
Related: PR #1224