Skip to content
RemitBridge

Research

RemitBench

An open testbed for comparing blockchain scaling setups against an ordinary bank transfer, using synthetic traffic that behaves the way real remittance traffic does. Currently in development.

Still being built

RemitBench is not finished yet

The testbed is under active development and has not produced results we are willing to publish. What is on this page is the design: the question we are asking, the setups we intend to compare, and the six tests we will run against each of them. The comparison table below is assembled from published specifications, not from our own measurements.

When there are real numbers, they will appear here with the code and the runs behind them.

A row of desktop workstations in a computer lab

The question we keep coming back to is which setup — sharding, sidechains, or off-chain channels — actually moves more transfers for less money, and whether any of them can talk to the bank rails that already exist. That second part gets skipped a lot. A system that is fast but cannot hand off to SWIFT or RTGS is not much use to a family whose bank is the only place they can pick up cash.

Everything here will run on synthetic traffic. We generate the transfers ourselves, so no real family’s transfer history is involved anywhere in this.

How the setups compare

Drawn from each system\u2019s published specifications, for comparing designs against each other. These are not our measurements and should not be quoted as results.

SetupThroughputLatencyFinalityCost rangeTalks to banks via
Sharding (Layer-1)Splits the work across more lanes2,500 – 10,000+2.5 sProbabilistic (2–3 blocks)$0.001 – $0.05API gateway + bridge contract
Sidechains (Layer-2)Own rules, own currency handling1,500 – 5,0001.2 sMainnet checkpoint (15 min)$0.005 – $0.02Custom relay + vault contracts
Off-chain payment channelsCheapest and fastest, if the money is already parked50,000+ (off-chain)< 100 msInstant (state update)< $0.001HTLC + ISO 20022 intermediary
SWIFT / normal bank transferEvery bank already accepts it100 – 500 (batch)1 – 5 business daysSettlement account credit$15 – $45 (fixed + FX)Native ISO 20022 / MT messaging

The six tests

The six scenarios we plan to put every setup through. Each one is a different kind of bad day, because a setup that only wins on the easy test is not really winning. None of these have been run yet.

A normal week

Regular family transfers, somewhere between $50 and $300, going out across four busy corridors. This is the boring case, and it is the one that matters most.

Target throughput
500 TPS
Average transfer
$180
Latency threshold
< 3.0 s
Failure tolerance
0.01%