Elliptic Curve Pairing Stealth Address Protocols

Proof of concept implementation of ECP-SAP paper. Benchmarks stealth address parsing speed against BaseSAP (DKSAP) across BN254 and BLS12-377 curves.

What this does

Stealth addresses let you receive crypto without linking transactions to your public identity. The receiver scans announcements to find payments meant for them. This project implements and benchmarks different scanning approaches:

• bn254/ - dual-key stealth addresses on BN254 (pairing-based, original paper approach)
• bn254-singlekey/ - single-key variant with view tags and experiment runner
• bn254-keychange/ - key-change variant
• bls12-377/ - same protocol on BLS12-377 curve
• plot/ - benchmark results and plotting script

Run it

git clone https://github.com/dzmbs/ecp-sap.git
cd ecp-sap

# run any variant
go run ./bn254
go run ./bn254-singlekey
go run ./bls12-377
go run ./bn254-keychange

You can change the number of public keys to scan through in config/config.go:

const RunNumber = 5000

Plot results

The plot/ directory has pre-computed experiment results for 5k-80k public keys. To visualize:

pip install matplotlib numpy
python plot/plot.py

Results

Scanning speed (ms) across different announcement counts:

Protocol	5k	10k	20k	40k	80k
DKSAP	2362	4694	9393	18797	37558
DKSAP + view tag	308	614	1225	2450	4885
ECPDKSAP + view tag	296	588	1182	2355	4685
ECPSKSAP + view tag	3411	6813	13624	27272	54435

The pairing-based dual-key approach (ECPDKSAP) matches DKSAP+view_tag performance while providing stronger privacy guarantees.

References

• Elliptic Curve Pairing Stealth Address Protocols
• BaseSAP (ERC-5564)
• Built with gnark-crypto