LibraryInteroperability2018Design paperCorpus record
Atomic Cross-Chain Swaps
Atomic swaps. Maurice Herlihy.
A swap across ledgers where either every transfer completes or none does, using hash locks and time locks. There is no custodian in the paper. A party can still walk away and leave the other waiting.
Herlihy's atomic swaps lock assets on several chains under the same hash and release them with one preimage, or return them after a timeout. There is no custodian. There is also no guarantee the other party will finish.
- Evidence
- Primary paper
- Re-measured
- No
- Assumptions
- 4
- Records linked
- 7
Code
Source, repository, and what is actually hosted
- arXiv:1801.09515https://arxiv.org/abs/1801.09515
- The source URL is not a GitHub repository. No code remote is invented.
- Full text is not mirrored. Public availability is not a licence to copy the file onto this site.
- Add to a reading list
01 Claim ledger
What the paper is allowed to say
Each row is a sentence already in the study. The status is the same on every row: a model claim, not a live measurement.
Claim 01 · Paper model
One secret, many chains
Whoever knows the preimage can claim. Publishing the preimage to take the asset on one chain reveals it to the counterparty, who takes the asset on the other. That coupling is the atomicity.
Claim 02 · Paper model
Timeouts are mandatory
If the secret is never published, every escrow must be able to refund. A swap with no timeout is a donation. The order of timeouts matters, because the party who moves second needs time to react.
Claim 03 · Paper model
Not every graph works
The paper's result is about which sets of pairwise swaps can be made atomic. A product that draws an arbitrary mesh of trades and calls it Herlihy has not checked the condition.
Claim 04 · Paper model
Price is out of scope
The protocol completes a pre-agreed exchange or reverts. It does not discover a fair price, route liquidity, or insure anyone against moving the market.
02 Three cuts
Observation, model, falsifier
A desk does not stop at the summary. Each claim is cut three ways, using only this study's own assumptions and checks. Nothing here is a new figure.
01 One secret, many chains
Observation
What the study says
Whoever knows the preimage can claim. Publishing the preimage to take the asset on one chain reveals it to the counterparty, who takes the asset on the other. That coupling is the atomicity.
Model
What has to hold
Timeouts are long enough for an honest party to observe the preimage and claim, and ordered so the leader cannot claim at the last moment and leave the other stranded.
02 Timeouts are mandatory
Observation
What the study says
If the secret is never published, every escrow must be able to refund. A swap with no timeout is a donation. The order of timeouts matters, because the party who moves second needs time to react.
Model
What has to hold
Every chain in the swap can hash the same way and can enforce a timeout refund.
Falsifier
What would retire it
What are the timeouts, and who must react last?
03 Not every graph works
Observation
What the study says
The paper's result is about which sets of pairwise swaps can be made atomic. A product that draws an arbitrary mesh of trades and calls it Herlihy has not checked the condition.
Model
What has to hold
Every chain in the swap can hash the same way and can enforce a timeout refund.
Falsifier
What would retire it
Does the swap graph match a pattern the protocol actually supports?
04 Price is out of scope
Observation
What the study says
The protocol completes a pre-agreed exchange or reverts. It does not discover a fair price, route liquidity, or insure anyone against moving the market.
Model
What has to hold
Contracts cannot be upgraded mid-swap by an admin the paper does not include.
Falsifier
What would retire it
Can anyone other than the preimage holder release the funds?
03 Sequence
One action, as an operating tape
- 01Parties agree the amounts and a hash of a secret one of them chose.
- 02Each locks funds in a contract that pays the counterparty if shown the preimage before a deadline, and refunds otherwise.
- 03Deadlines are staggered so the party who must react still has time after the secret appears.
- 04The secret-holder publishes the preimage on the chain where they want to claim.
- 05The other party sees it and claims on their chain. If the secret never appears, refunds fire.
04 Load-bearing
The argument, and where a pitch drops it
Atomic is not fair in the informal sense
The cut
A party can lock, wait, and refund, having used the other side's time and information. The paper stops them taking the asset without paying. It does not stop them wasting the counterparty's timeout. Calling that trustless requires saying what was actually guaranteed.
Why it carries weight
If this cut is skipped, the paper's name is being used without the mechanism that makes the name mean anything.
Where it is dropped
A timeout that is too short fails honest users. A timeout that is too long freezes funds. That choice is operational.
A committee that can complete the swap early is a custodian
The cut
The moment a third set of keys can release the escrow, the construction has left the paper and become a bridge. That may be a product choice. It is not an atomic swap.
Why it carries weight
If this cut is skipped, the paper's name is being used without the mechanism that makes the name mean anything.
Where it is dropped
This is not a bridge, not wrapped bitcoin, and not a liquidity network.
05 Register
What has to be true
Cryptography · Not re-measured
Every chain in the swap can hash the same way and can enforce a timeout refund.
Adversary · Not re-measured
Timeouts are long enough for an honest party to observe the preimage and claim, and ordered so the leader cannot claim at the last moment and leave the other stranded.
Cryptography · Not re-measured
The hash is preimage-resistant. A guessable secret breaks the lock.
Model · Not re-measured
Contracts cannot be upgraded mid-swap by an admin the paper does not include.
06 Divergence
What happened after the paper
Lightning uses the same hash-lock idea inside channels rather than as a one-shot exchange of on-chain assets. Bridge products usually replaced the preimage with a committee. Herlihy's 2018 paper is the citation for custodian-free atomicity and for the graph condition. It is not a citation for wrapped assets.
A later client, parameter set, or reward formula is a different object. Cite this paper for the mechanism. Cite a primary release for the network. This desk has not re-run the proof.
07 Pre-mortem
What to check before you use the idea
- 0 of 4 marked on this browser. A mark is a reading note, not a pass, a rating, or a recommendation.
08 Anatomy
The paper, in the order a builder needs
The problem it names
Two chains cannot read each other's state. Herlihy asks when a group of parties can still exchange assets so that no coalition gets paid without paying.
What the design proposes
- A secret preimage. Publishing it on one chain lets the counterparty claim on the other.
- Timeouts so a locked asset returns if the secret never appears.
- A graph of who swaps with whom. Not every directed graph has an atomic protocol under the paper's rules.
How the mechanism is specified
- The hash has to be the same primitive on every chain involved. A chain that cannot escrow and hash does not fit.
- Atomicity is about completion, not about price. The paper does not discover a fair exchange rate.
- The leader of the protocol learns the secret first in some positions. The timeout has to cover that advantage.
What this page does not treat as proven
09 Lexicon
Terms, opened into the record
- Timelock
- A deadline after which the original holder can take the funds back.
10 Repository
Every linked record on this page
Underlined words open a page that already exists: a concept, a protocol profile, a failure record, or another paper. If a word is not underlined, this desk does not have a record for it.
This is Blockchain Lab's reading of a public design paper. It is not the paper, not a copy of it, and not an offer of tokens, equity, custody or a partnership. Later network behaviour can diverge from the text. Nothing here is investment, legal or technical advice.
Research status: Design paper. Last reviewed: 1 October 2026. This is a reading of a public paper, not investment, legal or security advice.