Bok Consulting Pty Ltd

The Web3 Foundation organised a Decentralised Exchange (DEX) workshop in Berlin, inviting most of the major DEXes in the Ethereum space.

Web3 Foundation Decentralized Exchange Workshop : Jan 22 – 24

The Web3 Foundation is organizing the top teams working on decentralized exchanges to facilitate a 1-day workshop. Our goal is to come away with a list of prioritized issues which affect a number of projects. From there, we aim to align resources and find pragmatic solutions to the issues outlined. This could be in the form of bounty or grant programs, funding specific research and development initiatives, etc.

Participants

• Melonport
  • Reto Trinkler
  • Gautham Nekkanti
• 0x
  • Fabio Berger
  • Leonid Logvinov
• Kyber
  • TN
  • Yaron Velner
• Airswap
  • Michael Oved
  • Don
• Radar Relay
  • Alan Curtis
  • Caleb Tebbe
  • Brad Stachurski
• Gnosis
  • Stefan George
  • Martin Köppelman
  • Christiane Ernst
  • Alex Hermann
• Augur
  • Tom Kysar
  • Perry
• OmiseGo
  • Pawel Peregud
  • David Knott
• CryptoDerivatives.Market, GazeCoin & OAX
  • Bok Khoo (bokkypoobah)
• Web3
  • Matt Di Ferrante (Dexy)
  • Peter Czaban
  • Ashley Tyson
  • Maria Paula Fernandez
• Independent
  • Alexey Akhunov (ledgerwatch)
• Polychain
  • Chase Lochmiller
  • Albert Ni
• Herdius
  • Balazs Deme
• Vlad Zamfir

My Notes

Currently trying to categories the ~13 projects into the different types of DEXes:
1. Order Book - On or Off Chain
2. Provably Fair Matching Y/N
3. Order placement
4. Throughput
5. Latency
6. Cost
7. Privacy
8. Legal
9. Target users
10. Assets
11. Worst case failure
12. Dev stage


0x Protocol
- Off chain order book, on chain protocol
- Use ERC20 allowance of tokens to the 0x contracts (vs EtherDelta where you deposit tokens in the ED wallet)
- Relayers job to watch the orders and prune orders
- Also a cancel, but that can be frontrun
- Like the EtherDelta smart contract, but more an API
- Diff to EtherDelta
** ED has a deposit contract, and have to withdraw
** 0x - user always have the funds in the wallet
- 0x token for governance, fee in 0x
- Radar relay has a head start, most volume going through them
1. Off chain matching, onchain settlement
2. Up to the relayer
3. ?
4. Settlement limited by Ethereum blockchain. Every trade is settled individually, cannot group ATM
5. Blocktime
6. Gas cost + relayers fee
7. None, once settled
8. They have a law firm. 0x is an open source protocol. Relayers have to decide how to comply.
9. Crypto holders + Dapps
10. ERC20. Will be adding EIP721 non-fungible token support
11.
12. Working on V2 currently. 3 relayers on mainnet.



Radar Relay, based on the 0x protocol:
1. Off chain
2. Open orders - no matching
3. Off chain order book, on chain settlement


Gnosis DX
- DEX based on Ethereum. ERC20/ERC20
- Dutch auction exchange for any kind of token pairing
- No cancelled orders
1. On chain
2. Yes
3. No front-running
4. Ethereum, slow
5. Slow
6. Gas + Fees (distributed to the ecosystem), half paid in Gnosis token
7. Everything is onchain
8. Legal - decentralised
9. No high frequency traders
10. ERC20 + wrapped ETH
11. No liquidity
12. Code complete


Airswap
- No price discovery, local auction you have access to
- Order information - price and size
- Request sent to cparty (counterparty), cparty sends back signed message, request sender can then execute
- Network can be spammed
- P2P model, can do reputation system
- Can Wrapped ETH be made part of the ETH protocol? Vlad - Don't want ERC20 ETH on protocol level
1. Peer to peer design
2. N/A
3. Orders passed directly P2P. Trade request communicated directly between peers
4. OK. Limited by Ethereum network
5. OK.
6. Gas
7. Trading P2P (no price discovery). As private as P2P implementation
8. Legal
9. Mid-sized / HFT
10. ERC20 (will move into BTC if they can work out some bonding mechanism, wrapped BTC, BTCRelay works only one direction)
11. People can be ripped off - no price discovery
12. Went on Mainnet last week. 20 people. Private testing.


OmiseGo (OMG)
- Discussion of Plasma, using Merkle trees. Tree of chains, each chain secured by parent
- Multiple children, potential multiple parent anchor points
- Exchange to be built after Plasma developed
- Tendermint is a temporary solution while Plasma is being built
- Tokens emitted by corporate clients (loyalty points) of OMG, not ERC20
1. On chain, later to Plasma
2. (unknown, new area for OMG)
3.
4. Tendermint PoS chain. OMG token used for staking.
5. 1s
5.
6. Will have gas like Ethereum
7. No in Tendermin, Yes in Plasma
8.
9. Corporate clients of OMG, ERC20 holders
10. Corporate tokens
11.
12. Pre-alpha


Augur
- Changed quite a bit in the last 6 months
- Prediction market platform
- Outcome token is swapped for ETH
1. On chain
2. Fair
3. ETH
4. ETH
5. ETH
6. Gas
7. No - ETH
8. Meh
9. Traders, betters, people who want to create markets
10. ETH, outcome tokens
11. Lots, in the reporting system. ETH.
12. In 3rd stage of audits. 2 to 3 weeks of preparation, then live.


DEXY
- State channel, Tendermint
1. & 2. Tendermint
3. Off chain P2P
4. 100 - 1K
5. 1 second
6. Low + taker fees
7. No privacy in the open version, oblivious tx transfer
8. Don't give a fuck. Buy a boat and live on international waters.
9. All users
10. ERC20 + EIP721
11. ETH
12. Pre release
13. Use cases - ?


CryptoDerivatives.Market / GazeCoin / OAX
- CryptoDerivatives.Market - GNT, individual contracts
- OAX - Asset Gateway, collateralised ERC20 tokens
- State Channels next project
GazeCoin details below:
1. On chain
2. Yes
3. ?
4. ETH
5. ETH
6. ETH
7. None
8. Worried
9. VR users
10. ERC20/721
11. ETH
12. Planning
13. Biggest challenge: gas cost


Kyber
- ShapeShift on the blockchain
- B2B platform
1. On chain
2. N/A
3. small issue
4. ETH
5. 1 block
6. Fees, spread, 200k
7. No
8. Complex, Singapore based, accommodate Singapore regulation. Decentralised exchange loosely defined in Singapore law. Security/Utility should not have any consequence. Looking at setting up entities in Belarus. Guidance, not licence. Money changer.
9. B2B, Dapps
10. ERC20
11. Lose gas
12. Mainnet alpha next month


Herdius
- Want to be cheaper than Ethereum
- Sidechain
- 2 private keys generated for wallets
- Threshold signature to trigger movement
1. On chain. People who hold at least 0.2% stake run the chain
2. Yes
3. 45 txs/sec
5. 15 seconds
6. < $1 per trade
7. No
8. Issue going forward (3 or 4 months of legal work before approaching German regulators, then they took 6 weeks of back & forth)
9. Regulated crypto traders
10. All
11. Sidechain goes down. Everything stolen / locked up
12. Prototype
13. Regulation and KYC


Vlad - LLZK - MISC (Matching in a state channel)
1. Order book in State Channel. Matching engine maintains and gossips order book
2. Trusted matching, not provably fair
3. a. Deposit funds into channel; b. Place orders
4. State channel throughput
5. Very low latency
6. State channel cost, 0.15%/0.25% maker/taker fees
7. ZK proof when placing order. Users hold their account state and produce the ZKP on the right to place/cancel orders and withdraw
8. Legal for Vlad to specify
9. ETH
10. ERC20
11. Unfair matching, vanishing/reverting state
12. Not planned
13. Writing


Melonport
- Asset management on the blockchain
- Building Melonchain
- Token contract duplicated on 2nd chain with an additional `mint(...)` function, like Parity bridge


DEX Issues - Breakout Groups
a. Front Run Prevention [5 people]
- Distributed key generation
- Commit and reveal
b. Ongoing vs Batch [3]
c. Scaling [Loss]
- Polkadot, Cosmos, Plasma, DPoS
d. Legal [9]
- Web3 Foundation may try to set up something where we can share legal advice
e. Privacy [5]
- zkSNARKs (lose ability to know message sender), too much overhead
- Ring signatures (Passive - like Monero, active)
- Set accumulator
f. Liquidity [4]
- No funds
- Legally does not
- Investment risk
- To approach market makers
- Partnership with CEX
- Intradex settlement

For each issue:
- Define
- Map out
- Criteria
- Action item


Legal Breakout Group
- Coala have a working group on ICOs
- Functional equivalences - conversations with regulators
- Web3 Foundation - token sales, KYC and AML
- US - Perkins Coie (Marko)
- Gnosis
** Gibralta - DLT regulation
- Working with regulators, working with law firms
- DLT licence in Gibraltar like BitLicence in NY, but more business friendly
- Gnosis could engage with regulators in Gibralta
- Gnosis - blacklist US from UI
- Airswap is a HK company. Are these securities that are trading on the exchange
- Paradex only listing non-security tokens (US jurisdication)
- Airswap - manual verification of tokens
- Kyber - manual shortlist of tokens