Anonymous - 2023-10-16

Originally posted by: marta-lokhova

To clarify here, the functionality we'll need here is specifically for testing performance of the transaction subsystem. Previously, this was achieved by taking a historical snapshot of a ledger, scaling it by some factor, then replaying scaled history on top, and profiling the results (note that this is different from performance testing of SCP/overlay that we do with supercluster). With Soroban, we'll need to rethink how we test perf, since we likely won't have much meaningful historical data for some time.