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
Want to build out a network in which each peer --> peer relationship can be modeled by a fixed "connectivity" that will then be taken into account in the delay function.
For example:
Validator A might have a 0.9 connectivity rating against Validator B, but just a 0.5 connectivity rating against Validator C.
When A sends a new message to the network, delay(A, B) will take into account the high connectivity between A and B and take a relatively short time (can still be random on some distribution based off that 0.9). While delay(A, C) will see the lower connectivity (0.5) and will more likely take longer to propagate to C
Proposed Implementation
Create new network delay function called peer_delay
function maintains a closure with a connectivity matric. connectivity[sender][receiver] can then be used by delay to figure out how long a message should take to send from sender to receiver
connectivity can be static or can be a 3d array also indexed by round (to allow connectivity to evolve over time).
The text was updated successfully, but these errors were encountered:
Issue
Want to build out a network in which each peer --> peer relationship can be modeled by a fixed "connectivity" that will then be taken into account in the
delay
function.For example:
delay(A, C)
will see the lower connectivity (0.5) and will more likely take longer to propagate to CProposed Implementation
peer_delay
connectivity
matric.connectivity[sender][receiver]
can then be used bydelay
to figure out how long a message should take to send fromsender
toreceiver
connectivity
can be static or can be a 3d array also indexed by round (to allow connectivity to evolve over time).The text was updated successfully, but these errors were encountered: