Improve self-repair summaries download #385
Labels
core team
Assigned to the core team
feature
New feature request
P2P
Involve P2P networking
self repair
Involve SelfRepair mechanism
Is your feature request related to a problem?
In development environment this case is really present, but since you are adding some network latency around 150-200ms, the time to get the beacon summaries for all subset for a given date, become really long as minutes.
Hence we should propose a way to optimize the requests to reduce the latency
Describe the solution you'd like
The self-repair could be optimized in few ways:
Batch of the beacon summaries
For now, for each subset, we are doing multi requests to the same node, leading to latency (nb subsets(=256) * latency).
We should aggregate trough the election to know which nodes are responsible for which subset and then aggregate through a
GetBeaconSummaries
messageStream responses
Because of the aggregation of summaries, network latencies and timeouts can occurred if the number of summaries and the transaction summaries inside are significant, as it's a single P2P call.
So we can stream transaction summaries like we are streaming transaction in the
GetTransactionChain
request, by doing multiple callsAdditional context
No response
The text was updated successfully, but these errors were encountered: