:warning: DIFFERENT CCS RULES ARE IN PLACE FOR THIS PROPOSAL! PLEASE READ THE FOLLOWING! :warning:

As a trial, this CCS proposal is going to operate on slightly different rules
given the unprecedented scope and duration of this proposal. For this proposal
ONLY, refunds will be issued in the event that the funding is not satisfactory
or the milestones are not completed. This differs from the standard of excess or
unused funds going to the general fund.

To qualify for a refund, the donator must send their tx ID, amount, and return
XMR address to [email protected] (PGP fingerprint:
FE6D D72A 19CD C5FC 6CB9  1696 BA18 1389 4EDD 58B9, full PGP key at
github.com/monero-project/monero/blob/master/utils/gpg_keys/luigi1111.asc) NO
LATER than ONE WEEK after their donation is made. Any remaining unclaimed funds
(in the event that the proposal is not completed) will be sent to the general
fund as usual. If refunds are to be issued, the funds will be returned via the
provided XMR address.

In summary, the funds can be either:

Unclaimed, leading to the general fund receiving them in the case of a failed
proposal.

Claimed within one week of the donation, leading to a refund in the case of a
failed proposal.

Note: The hope is that the refunds will not be needed, and the proposal will get
funded and completed. In the event of proposal completion, refunds will NOT be
issued. It is only if the proposal is not completed or funded to satisfaction,
and ONLY for this proposal.

Monero Atomic Swap implementation funding

Previous CCS: Monero Atomic Swaps research funding

Hi everyone,

Three months ago, I posted a CCS for continuing my research on Monero Atomic Swaps. That research is now complete and the results can be found here. The resulting protocol is implementable today; no more missing crypto! So much so that a PoC was implemented in no time; thank you, kayabaNerve and PlasmaPower! Thus I am reaching out to propose getting a team to work on implementing this protocol, with the end goal of creating a production-ready client/daemon for swapping Bitcoin and Monero. Our design enables to seamlessly extend support for more cryptocurrencies to swap with Monero. It would be very exciting to build that.

You can find the whitepaper that describes the full protocol here.

A ready-to-use implementation requires a lot of engineering work. Here, my colleagues and I attempt to break down the project into manageable parts, describing the dependencies that have to be fulfilled, and the general roadmap of the project.

Motivation

Trustless technologies are now emerging, creating the option of refusing to accept counter-party risk. You can make trades with your enemy, as they can't cheat on you. If you don't have to trust, you don't have to know who they are, either.

It is very unlikely that Monero will get banned by all centralized exchanges, but by having an open source atomic swap implementation, such banning mechanism is inefective, as Monero would still be available to anyone who could acquire Bitcoin, which is ubiquitous, and swap the coins online anonymously, trustlessly, with a random peer. Monero will be more robust than ever.

Bitcoin is traceable. This is used to recognize dirty coins, but also for untargeted surveillance and censorship. Bitcoiners, in need of strong privacy, might recognize the utility of a trustless path with low resistance to convert their bitcoin into monero, and become Monero users.

However, with power comes responsibility, atomic swaps enable users to exchange coins directly with each other. At the same time, if transacted value is significant, honest users MUST carry out their due diligence regarding the origin of the counterparty funds and possibly other anti-money laundering countermeasures, in order to comply with regulations. Trustlessness and no counter-party risk are narrowly defined terms of the atomic-swap literature, that ignores the context whereby the technology is deployed. Bitcoins accumulate dirt in their lifetimes, so swap your monero responsibly, because trustlessly receiving tainted bitcoins is a real counterparty-risk. The counterparties of a swap generate private and blockchain notarized cryptographic proofs of their private agreement, but the court of your jurisdiction might not like that explanation so much.

The crypto-ecosystem is rapidly moving towards interoperability. Atomic swaps unleash interoperability between Monero and other blockchains. Whether a user needs to open a lighting channel from the monero-bitcoin swap or wants to fund an arbitrary bitcoin contract, the swap protocol exposes the interop socket.

This project will also, as a beneficial side-effect, extend the Monero ecosystem in Rust. Multiple libraries are needed to support the full protocol. Most of them are related to cryptography, for example the "Discrete logarithm equality across groups" algorithm described in the MRL-0010 technical note by Sarang Noether (originally proposed by Andrew Poelstra), or directly at the Monero protocol level in the Monero Rust Library.

Our motivation to build this software is to empower individuals and businesses, who want to or need to exchange within a strong security and privacy context using P2P, trustless technologies.

This project has the potential of increasing Monero's liquidity and enabling Monero to get into the hands of more people.

We deem it critical to build this in a manner that fully aligns with the interests of the community. Thus we're reaching out to raise community money, to build this with the community, for the community, enabling the community to preserve its own interests.

What are we building?

We aim to build a collection of programs—similar to programs you are familiar with, such as the Monero daemon, wallet CLI, or wallet GUI—that have limited functionality individually but as a collection, serve the functions an end-user requires. One can launch these swap programs to exchange coins with a counterparty. We call those programs: swap clients (CLI or GUI), the swap daemon (like the Monero daemon), and chain-syncers (connected to full nodes). In the default configuration, this will mean opening the swap client and letting it launch and manage all other programs involved.

For example, if you, as an end-user, want to acquire monero and have bitcoin, you'll launch a swap client that connects to a swap daemon, and connects to a counterparty that has monero and is looking to trade them for bitcoin at an agreed upon price. The swap client will give you an address where to move your bitcoin and, at the end of the swap, the swap client will display the monero key-pair to import into your wallet. You now own monero. If at some point the swap is canceled for any reason, your bitcoin will be refunded at the address you chose, making this exchange trustless.

Connecting to a counterparty will require knowledge of their daemon's address, and the amounts traded (i.e. the price and quantity). Creating a platform such as a DEX, allowing people to find each other and "auto" connect with the correct arguments or negotiate the price, is out-of-scope for this project. Industry standards for such interfaces are yet to emerge.

Overview

R&D Institution: Cryp GmbH

Funding: Monero CCS

Duration: 7 months

Job completion: by Q2 2021

Contributors:

• h4sh3d
• kayabaNerve
• lederstrumpf
• the charlatan
• zkao

Licenses: The license for the code will be decided based on community feedback. Our current preference is LGPL-3.0. The specification will be released under CC-BY-4.0.

Expiration date: Funding will remain open until 31.12.2020. If materially underfunded until 31.12.2020, we'll either (1) agree with the community to deliver a subset of the deliverables and collect the funds, or (2) discuss how to re-allocate the funds with the community.

Architecture

The core project will be built in Rust. Rust's good coverage of cryptographic libraries and blockchain protocols, type safety, and language design makes it a very good candidate for such applications (and the prototype is also written in Rust, for the same reasons).

Here we present an overview of the project's architecture. More details of the individual components will be described in a forthcoming section under Deliverables.

The figure represents the general architecture of the swap components and their interactions.

The following table summarizes different aspects of each component.

Client/daemon segregation rationale

The rationale behind segregating the client and the daemon is not for security reasons at the moment (the client signs the transactions received from the daemon blindly, implying full trust), but for the flexibility and extensibility added.

Other clients can be created: mobile applications (that also run the daemon in background), heavy or light desktop GUIs, or even scripted/automated backends (e.g. in a business environment).

Future extensibility

The atomic swap protocol is just the first instantiation of a more generic interface to other systems—we aim to build this construction abstractly enough to allow clean extension ¹ to future protocols.

Deliverables

Below is a complete list of our deliverables.

Specifications

• Specification of swap-lib: Specify the interface and the requirements for adding support for a new chain, for one or both templates (Bitcoin-like and Monero-like).
• Specification of swap-daemon: Specify messages passed between swap-daemon and: swap-client and swap-daemon. These include protocol messages exchanged between swap participants, but also specify the medium of communication of swap-daemon and those components.
• Specification of chain-syncer: Specify the functionality and interface chain-syncer has to expose in order to permit the swap-daemon to carry out swaps. Specify the type of jobs a chain-syncer has to implement in order to support executing both templates.

Libraries and Components

• swap-lib: includes stateless libraries that implement the core protocol, without runtime, disk, nor network implementation. Knows how to create and verify all the transactions involved in the protocol: it understands and handles the crypto verification, including adaptor signatures and DLEQ proofs across groups, and contains two templates for the pair of exchanging chains (Bitcoin-like and Monero-like). The goal of swap-lib is to facilitate integration of the base protocol logic of all pairs of chains that implement the two templates, such that adding a new pair, e.g., Litecoin/Monero, only requires implementing Litecoin for the Bitcoin-like template and an ltc-chain-syncer (see below).
  • btc-swap-lib: an implementation for Bitcoin-like template exchanging bitcoin for monero.
  • xmr-swap-lib: an implementation for Monero-like template exchanging monero for bitcoin.

• swap-daemon: implements a daemon, based on swap-lib, uses chain-syncer as interface to the blockchain world, has the full picture of the state of the cross-chain-swap, as it's aware of the events on both chains and of exchange counterparty protocol messages, it fully understands the protocol, and contains the state machine to execute its respective role in the protocol.

• swap-client: used by the end-user to enter into the protocol, has access to secret keys, uses the swap-daemon to execute the protocol, and signs transactions when needed. swap-client trusts the daemon completely to execute the protocol on its behalf and to exchange protocol messages with the swap counterparty.
  • swap-cli: end-user CLI client that binds to the daemon for executing swaps and reporting the state of an ongoing swap.
  • swap-gui: minimal end-user GUI client that binds to the daemon for executing swaps and reporting the state of an ongoing swap.
• chain-syncer: connects and synchronizes the protocol universe to the blockchain universe by following its client's commands (swap-daemon). chain-syncer knows the transactions of interest based on what its client subscribes to and informs the client in case one of its transactions gets reorged away from the main chain. chain-syncer must guarantee to be online during the entire execution of the protocol, and carry out actions on behalf of its clients. It has the ability to play a job and respond with events.
  • btc-chain-syncer: a chain-syncer connected to a Bitcoin full node, it takes jobs such as listening for transaction confirmation or event-driven transaction broadcast.
  • xmr-chain-syncer: a chain-syncer connected to a Monero full node, it takes jobs such as listening for transaction confirmation.

• xgroup-dleq-lib: a cryptographic library implementing the MRL-0010 technical note. This library must support at least secp256k1 and ed25519 curves. secp256kfun will be used at first to speed-up the development and will later be replaced by a fork of libsecp256k1 and rust-secp256k1. dalek-cryptography will be used for ed25519 cryptography.

• ecdsa-adaptor-sig: a cryptographic library implementing ECDSA One-time VES over secp256k1. We are looking forward to how "Add ecdsa_adaptor module" evolves and wait on this to add support in rust-secp256k1.

Disclaimer: those cryptographic libraries will require review for being considered as safe-to-use in production.

We're currently in discussions with potential academic collaborators to extend the formal coverage of the protocol and its publication. We're also in the process of publishing a preprint of the swap paper on the Cryptology ePrint Archive.

The code will be released mainly under the monero-rs and swap-rs Github organisations.

Dependencies

We describe here a list of cryptographic and protocol dependencies likely needed for achieving a complete implementation (some dependencies are readily available, others will have to be extended or newly developed):

• Monero library for block parsing and for transaction parsing, manipulation, and verification logic
• Bitcoin library for block parsing and for transaction parsing, signing, manipulation and verification logic
• RPC and ZMQ libraries for listening to bitcoin and monero nodes; allow reacting and broadcasting transactions depending on the protocol execution
• Discrete logarithm equality across secp256k1 and ed25519 groups library, as described in MRL-0010 technical note
• ECDSA One-time VES over secp256k1 library as described in One-Time Verifiably Encrypted Signatures A.K.A. Adaptor Signatures, for signing some parts of the bitcoin transactions
• ECDSA signing over secp256k1 library, for signing bitcoin transactions

Currently available Monero dependencies:

• monero (https://github.com/monero-project/monero)
• monero-rs (https://github.com/monero-rs/monero-rs)
• monero-rpc-rs (https://github.com/monero-ecosystem/monero-rpc-rs)

Currently available Bitcoin dependencies:

• bitcoin-core (https://github.com/bitcoin/bitcoin)
• rust-bitcoin (https://github.com/rust-bitcoin/rust-bitcoin)
• rust-bitcoincore-rpc (https://github.com/rust-bitcoin/rust-bitcoincore-rpc)
• libsecp256k1 (https://github.com/bitcoin-core/secp256k1)
• rust-secp256k1 (https://github.com/rust-bitcoin/rust-secp256k1)
• secp256kfun (https://github.com/LLFourn/secp256kfun); for ECDSA One-time VES over secp256k1 signing implementation and prototyping DLEQ proofs

Currently available general dependencies:

• rust-zmq (https://github.com/erickt/rust-zmq)
• rust-libp2p (https://github.com/libp2p/rust-libp2p); as an option for daemon peer-to-peer communication
• dalek-cryptography (https://github.com/dalek-cryptography); for ed25519 arithmetic/curve operations

Milestones

We split the principal milestones (1, 2, and 3) into unordered sub-milestones, each releasing a fraction of the total funds for their principal milestone.

The goal is to share the advance of each individual sub-milestones in biweekly progress reports.

Milestone 1 (specs): 20% [6 weeks]

• Technical specifications: a list of specifications that covers all aspects of the protocol, resembling specifications such as BOLT for Lightning network or Cryptonote.

The specifications will demarcate which functionality falls in scope of the implementation in Milestone 2, and which functionality will be postponed to Milestone 3.

A. Specification of swap-daemon

Detailed description and specification of the swap-daemon, internal and external.

A.1 User-facing [35% of M1] [2 weeks]

• Specify swap-daemon's outer API layer of user interaction (swap-clients & counterparty swap-daemon) first to guide design of inner dependencies (swap-lib & chain-syncer) to match desired UX. This includes protocol messages exchanged between swap participants, that is: interactions with the counterparty swap-daemon.
• Specify the API that swap-cli and swap-gui consume from swap-daemon.
• Specify the networking stack between swap-daemon and: swap-daemon counterparty and swap-clients.

A.2 Service internals [16.25% of M1] [1 week]

Specify links between swap-daemon and the other deliverables it requires to facilitate a swap, but that are not user-facing, i.e. swap-lib and chain-syncers:

• Specify the subset of swap-lib's API strictly required for swap-daemon's operation
• Specify API and network stack for swap-daemon's required calls to chain-syncers

B. Specification of swap-lib (core protocol)

Detailed description and specification of all the libraries that, in conjunction, form swap-lib, including stateless transaction construction libraries, crypto-libraries and their wrappers, and state-machine libraries for executing the protocol.

B.1 External specification of swap-lib [16.25% of M1] [1 week]

• Specify swap-lib's public API to be consumed by swap-daemon only. Preliminarily, covers InitTx(), VrfyTx(), Vrfy(), and EncVrfy() from the whitepaper.
• Specify swap-lib's public API to be consumed by swap-clients only. Preliminarily, covers Sign(), EncSign(), DecSig(), RecKey(), and Rec() from the whitepaper.
• Specify swap-lib's public API to be consumed by both swap-daemon and swap-clients.

B.2 Internal specification of swap-lib [16.25% of M1] [1 week]

• Specify internal function calls of swap-lib.
• Specify a concrete implementation to support a chain, including all cryptographic primitives that must be supported.

C. Specification of chain-syncer [16.25% of M1] [1 week]

• Specify the functionality and interface chain-syncer has to expose in order to permit the swap-daemon to carry out swaps. Describe what jobs are, and what jobs must be supported.
• Specify the networking stack between swap-daemon and chain-syncers.

A core goal of the spec-writing process is to ensure that the deliverables are compatible and functional in the sum of their parts. This process in Milestone 1 thus aims to identify limitations of the architecture presented in this proposal, which can impact the structure of Milestones 2 and 3. In case changes are consequently required, we will propose them to the community at the completion of Milestone 1, and we will ensure that the same final functionality as set out in this proposal will be delivered.

Milestone 2 (implementation of core architecture): 45% [14 weeks]

• Minimal functionality: at completion-time of milestone 2, all components for executing atomic swaps successfully are implemented using our libraries as proposed in the presented architecture.
• BETA STATE: the software released is considered to be beta software and not ready for deployment.

This milestone achieves the initial implementation of all the components (without GUI client) interacting with each other. It implements only the minimally required functionalities specified in the specifications (Milestone 1).

A. Cryptographic libraries [7.5% of M2] [1 week]

For milestone 2 we intend to use Lloyd's experimental library for ECSA adaptor signatures (ecdsa-adaptor-sig) and DLEQ proofs (xgroup-dleq-lib).

We postpone the integration of a production-grade ECDSA adaptor signatures to the end of the project, namely milestone 3, as it is highly complex, and this gives tooling more time to mature. In case Bitcoin incorporates Schnorr signatures soon (BIP 340), ECDSA adaptor signatures may also not be needed and would then be replaced with Schnorr adaptor signatures.

At completion, we provide generic cryptographic libraries that expose a high level API, comparable to how rust-secp256k1 exposes a high level API for ECDSA signatures over secp256k1.

A.1 xgroup-dleq-lib

Experimental discrete-log equality across groups library using secp256kfun. At completion, this library enables zk-proving discrete log equality across groups secp256k1 and ed25519 for arbitrary private keys.

A.2 ecdsa-adaptor-sig

Experimental ecdsa-adaptor signature library using secp256kfun. At completion, this library enables producing and verifying ecdsa-adaptor signatures, and extracting secrets from clear and encrypted signatures.

B. swap-lib [25% of M2] [4 weeks]

swap-lib is the core for adding support for a new chain, either via the Bitcoin-like template or the Monero-like template. At completion, this library is an interface for enforcing the requirements of swap-daemon and swap-client for a selected chain, and the requirements for each chain-syncer type.

B.1 xmr-swap-lib

A concrete implementation of swap-lib that allows the creation of a swap-daemon and a swap-client for exchanging monero for bitcoin.

B.2 btc-swap-lib

A concrete implementation of swap-lib that allows the creation of a swap-daemon and a swap-client for exchanging bitcoin for monero.

C. swap-clients [12.5% of M2] [2 weeks]

At completion of Milestone 2, we provide only the swap-cli client.

C.1 swap-cli

A CLI client utilizing swap-daemon's API for executing swaps. At completion, a minimal CLI client for running Bitcoin and Monero swaps will be delivered.

• swap-cli can initialize a swap-daemon with parameters for a swap (swap-pair, swap-direction, counterparty daemon address, swap-amount, exchange-rate).
• swap-cli can sign transactions the swap-daemon provides.
• swap-cli sends signed transactions back to the swap-daemon.

D. swap-daemon [30% of M2] [4 weeks]

At completion, swap-daemon enables swap-client to successfully swap bitcoin for monero.

It coordinates all aspects of the swap by communicating with swap-client, chain-syncers and the counterparty's swap-daemon. It exposes an RPC server for client interaction.

At this stage:

• swap-daemon understands the semantics of all the public keys, unsigned, signed and partially-signed transactions, hash pre-images, and encrypted signatures involved in the protocol.
• has the full picture of the state of the cross-chain-swap, as it's aware of the events on both chains; each chain-syncer only has a partial view. swap-daemon knows the off-chain events as well.
• protocol implemented in 2 varieties: (1) one for executing btc-sender/xmr-receiver protocol (btc->xmr) and (2) another for xmr-sender/btc-receiver protocol (xmr->btc). In conjunction these 2 protocols cover the full protocol.
• at run-time, each swap-daemon instantiates a state-machine of only one of the varieties: either xmr->btc or btc->xmr.
• at run-time, the daemon can start in a master or slave mode. In master mode, swap-daemon accepts entering connections, in slave mode swap-daemon tries to connect to a counterparty daemon.
• reveals a high level API that lets swap-clients run the protocol.

E. chain-syncers [25% of M2] [3 weeks]

At completion, chain-syncer exposes enough functionality to permit the swap-daemon to carry out swaps between Monero and Bitcoin. It does not publish transactions on behalf of its client yet (e.g. do X once Y), just listens to transactions of interest and blocks, process them and push events to the swap-daemon (i.e. chain-syncer's client). It also exposes functionality for clients to publish transactions to the blockchain, if needed.

All chain-syncers implement the base functionality:

• interface to swap-daemon
• read transactions (arriving in the mempool, or with a number of confirmations)
• read new blocks
• detect block re-orgs

E.1 btc-chain-syncer

A bitcoin concrete implementation of the chain-syncer. In addition to the base functionality, btc-chain-syncer:

• connects to bitcoin full-node
• broadcasts transactions

E.2 xmr-chain-syncer

A monero concrete implementation of the chain-syncer. In addition to the base functionality, xmr-chain-syncer:

• connects to monero full-node

Milestone 3 (minimum viable product): 35% [12 weeks]

• Full integration of individual components: the entire functionality as defined in the specs is now implemented.
• MVP STATE: the software released is considered as stable for a minimum viable product.

At this stage, we replace/finish some components of Milestone 2 with MVP grade components, meaning all features specified in Milestone 1 are implemented.

A. Cryptographic libraries

At completion of this task, cryptographic libraries should benefit from more reviewed and more stable codebases.

A.1 xgroup-dleq-lib [25% of M3] [3 weeks]

We upgrade xgroup-dleq-lib to use rust-secp256k1 (fork or mainstream) and dalek. The goal is to implement in C in libsecp256k1 the high level API needed for cross group discrete logarithm equality proofs with secp256k1 curve and update rust-secp256k1 bindings (fork or mainstream).

A.2 ecdsa-adaptor-sig [25% of M3] [3 weeks]

We upgrade ecdsa-adaptor-sig to use rust-secp256k1 (fork or mainstream), based of the C implementation in libsecp256k1 currently in progress here.

B. chain-syncer [15% of M3] [2 week]

We upgrade the chain-syncers to enable pattern such as: do X if Y. This allows other daemon requirements to be supported, i.e. a daemon can be written such that going offline for a small amount of time is not problematic, allowing e.g. mobile daemons.

• swap-daemon can queue tasks such as execute refund path at block height X with a chain-syncer.

C. swap-clients

We improve swap-cli and release a new web-based GUI client, swap-gui.

C.1 swap-cli [10% of M3] [1 week]

swap-cli exposes additional functionality and better UI/UX:

• during a swap, swap-daemon relays to swap-cli what actions are currently available to swap-cli, and future actions that will become available after n blocks.
• swap-daemon pushes updates to swap-cli, in lieu of the swap-cli pulling from the swap-daemon.
• swap-cli can schedule actions to be performed by chain-syncers, as via the upgrade specified in M3.B.

C.2 swap-gui [15% of M3] [2 weeks]

A minimal web-based GUI client, swap-gui. Cryptographic requirements will be fulfilled with WASM or might be embedded as native libraries on an Electron webapp.

• swap-gui covers the functionality of swap-cli as implemented at the end of M2.
• web-based GUI that interacts with swap-daemon API.

D. swap-daemon [10% of M3] [1 weeks]

Implements all features defined in the specification for swap-daemon.

Cost

We ask for 2727 XMR from CCS.

For further information on our crowdfunding, please visit https://cryp.ee/crowdfund or write to us at [email protected] (PGP fingerprint A873 CF81 AAFE C016 EC1A  14BB D889 86F7 A3F7 37C4²).

Feedback

Due to the size of this project and the funding required for it, questions, comments, and feedback regarding this request are very welcome.

Communication channel

Besides the MR comments on GitLab and the Monero IRC channels such as #monero-community and #monero-research-lab, we opened the #monero-swap IRC channel for crowdfunding questions, project updates, and work management during the development time-frame.

We commit to make ourselves highly available for the community of researchers and developers, such as participating in MRL meetings, sharing progress, incorporating/giving feedback from/to the community in general, and responding to general inquiries, if they fall under our expertise.

-----BEGIN PGP PUBLIC KEY BLOCK-----
Version: OpenPGP.js v4.10.8
Comment: https://openpgpjs.org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=oDUJ
-----END PGP PUBLIC KEY BLOCK-----