Read the paper
Authenticating IoT Data on Ethereum after Pectra and Fusaka
on Zenodo
Imagine you run a postal service. You can stamp every letter individually (expensive, maximally trustworthy), or you can bundle a thousand letters into one sealed bag and stamp only the bag (cheap, but you trust the person who packed the bag). The question is: given the latest postal rates, which stamping strategy actually pencils out for a fleet of ten thousand devices? That's the core mechanism of this paper. Sie et al. take four cryptographic signature schemes — secp256k1 ECDSA, P-256 ECDSA, BLS12-381 aggregate signatures, and an identity-based signature (IBS) with batch verification — and measure their real gas costs on Ethereum under three successive rule sets: pre-Pectra, post-Pectra (May 2025), and post-Fusaka (expected late 2025). They organize IoT data anchoring into three trust levels: (1) a trusted aggregator posts a Merkle root (cheapest, but you trust the aggregator), (2) device signatures are verified off-chain and a Merkle proof is anchored, (3) every device signature is verified on-chain (most expensive, least trust). The headline result is that batching plus Merkle anchoring with secp256k1 ECDSA gets you to 8,963 gas per record — roughly a sixfold reduction from naive single-record posting — and nothing else comes close. The Pectra and Fusaka upgrades added precompiles for BLS12-381 pairing operations and P-256 verification (EIP-2537, EIP-7212), repriced calldata and MODEXP, and capped transaction gas at 2^24. The intuition might be that these new precompiles would make pairing-based schemes competitive. They don't. BLS aggregate signatures cost 3.9× more per record than secp256k1 ECDSA, dominated by the hash-to-curve operation which still runs in EVM bytecode rather than as a precompile. P-256 lands at 1.6× and the identity-based design at 2.7×. The upgrades mainly made these designs *feasible* (they were impossible before due to gas limits), but they didn't flip the cost ranking. Capacity constraints matter as much as per-record cost. A single Ethereum transaction can carry at most 1,754 ECDSA-authenticated records but only 411 BLS-aggregated ones, constrained by the 2^24 gas cap. For agricultural supply chains with thousands of sensors, this means BLS batches require 4× more transactions to process the same data volume. On the device side, signing on a Raspberry Pi Zero 2 W takes 0.29 ms for ECDSA versus 5.7 ms for the pairing-based scheme — a 20× penalty that matters when you're running on battery-powered edge hardware. The paper's trust-level framing is the most useful contribution for practitioners. Level 1 (Merkle root via trusted aggregator) is cheapest but requires you to trust someone to faithfully include all records. Level 3 (every signature verified on-chain) is the most trustless but prohibitively expensive for large fleets. Level 2 — off-chain signature verification with on-chain Merkle anchoring — is the practical sweet spot, and the paper provides concrete gas numbers for each combination of signature scheme and trust level. This is the kind of decision matrix that engineers building IoT-to-blockchain pipelines actually need. The integrity story is solid for an engineering benchmark paper. All four schemes are implemented in Solidity, gas is measured against the actual EVM opcode pricing schedules (not simulated), and the benchmark harness plus data are released on Zenodo. The limitation is that this is a cost study, not a security analysis — the paper does not attempt to prove anything about the schemes themselves, only to measure their on-chain cost under real fee schedules. The identity-based scheme's key escrow risk (the key generation center can recover any device's private key) is noted but not deeply analyzed. The obvious experiment not run: a live deployment on Ethereum mainnet with real IoT devices transmitting real sensor data under actual network congestion. The paper measures gas, which is deterministic given the opcode schedule, but actual cost in USD depends on gas price, which fluctuates wildly. A weeklong mainnet trial with, say, 500 devices would have turned this from an engineering benchmark into a deployment study. The honest read is budget and scope: this is a Taiwanese academic group benchmarking cryptographic costs, not a startup burning ETH on mainnet.
The analogy
You run a postal service and need to authenticate thousands of letters. You can stamp each one (expensive, trustless), bundle them into sealed bags and stamp the bag (cheap, trust the packer), or use fancy new holographic seals the post office just started accepting. The question isn't which seal is fanciest — it's which costs less per letter at today's postage rates. The new seals are finally accepted, but they're still more expensive per stamp.
Postal stamping
Choosing between individual letter stamps, sealed-bag stamps, or new holographic seals based on per-letter postage cost
IoT gas costing
Choosing between ECDSA, BLS, or IBS signatures based on per-record gas cost under Ethereum's updated opcode pricing
Honest assessment
A well-executed engineering benchmark that answers a narrow but practical question: what does each signature scheme actually cost on post-upgrade Ethereum? The answer — ECDSA still wins by a wide margin — is useful precisely because it's boring. This is a reference table for builders, not a research breakthrough. Worth reading if you're designing an IoT-to-blockchain pipeline and need to pick a signature scheme; skip if you're looking for new cryptography.
What Is CLAIMS?
60
/ 100
BedrockNews applies a structured analytical framework called CLAIMS to academic papers and technical research releases.
How CLAIMS worksClaim
2/5
Ladder
3/5
Architecture
4/5
Integrity
3/5
Milestone
3/5
Successor
3/5
What is this paper actually committing to?
Useful but incremental: first gas benchmark of four signature schemes under Pectra/Fusaka rules. An engineering contribution, not a cryptographic one. The 'never done before' is narrow — these specific fee schedules are new.
The field fight
Do Ethereum's new cryptographic precompiles (BLS12-381, P-256) change the practical economics of on-chain data authentication?
Side A
New precompiles enable pairing-based designs that offer richer trust properties (aggregation, identity-based keys) at competitive cost
Held by: EIP-2537 proponents, BLS aggregation advocates
Side B
secp256k1 ECDSA remains cheapest by a wide margin; precompiles made exotic schemes feasible but not competitive on cost
Held by: This paper, pragmatic IoT-on-chain builders
Milestone tracker
Unlocks: BLS aggregate signatures become cost-competitive, enabling trustless aggregation without per-device on-chain verification
Current
8,963
gas/record
secp256k1 ECDSA batched
Next
<5,000
gas/record
BLS with hash-to-curve precompile
2+ years; requires a new EIP for hash-to-G1/G2 precompile, no proposal yet
Verified Facts (6)
Batching and Merkle anchoring cut the cost of ECDSA authentication sixfold, to 8,963 gas per record
P-256 costs 1.6 times as much, the identity-based design 2.7 times and BLS aggregation 3.9 times
One transaction carries at most 1,754 ECDSA-authenticated records but only 411 BLS-aggregated ones
On a Raspberry Pi Zero 2 W, pairing-based signing takes 5.7 ms against 0.29 ms for ECDSA
Ethereum's Pectra and Fusaka upgrades added precompiles for BLS12-381 and P-256, repriced calldata and MODEXP, and capped transaction gas at 2^24
The benchmark harness is released for reproducibility (DOI: 10.5281/zenodo.24328819)
Practical Implications (5)
If you're building IoT data anchoring on Ethereum, use secp256k1 ECDSA with Merkle batching — 8,963 gas/record is the floor today and nothing else comes close
BLS aggregate signatures are now technically feasible post-Pectra but cost 3.9× more due to hash-to-curve overhead; don't adopt them for cost savings alone
A single Ethereum transaction fits at most 1,754 ECDSA records vs 411 BLS — plan your batch sizing around this constraint
On resource-constrained IoT hardware (RPi Zero 2 W class), ECDSA signing is 20× faster than pairing-based signing (0.29 ms vs 5.7 ms) — matters for battery budgets
The identity-based scheme introduces key escrow at the KGC — only use it if your trust model already accepts a centralized key authority
More latest news
Read all