Over the past three years centralized cryptocurrency exchanges have experienced a real boom, with most of these structures developed and deployed in a relatively short period of time. Given the high complexity of the assets they have to deal with (crypto tokens versus securities), as well as the short testing periods and lack of established history, it is not surprising that most of these exchanges are vulnerable to attacks. In fact, most of the leading cryptocurrency exchanges in the same three years, became victims of such attacks, which almost always led to huge losses: Bithumb, MtGox, ShapeShift, Bitfinex, Yapizon, Cryptsy
– The list can be continued if you want.
In addition to its vulnerability, the centralization principle behind such exchanges goes against the very spirit of decentralized blockchain. These reasons led investors and developers to turn their attention to decentralized exchanges or DEX (DEcentralized eXchange).
The three-year period has brought many of these projects-most of them still in development, but several working platforms have already gained a reputation.
The vast majority of decentralized exchanges are concentrated in the Etherium ecosystem – these will be discussed below (ether and ERC20 tokens). As for support for other cryptocurrencies, the main hopes are pinned on atomic swaps, although so far it is only an experimental development.
EtherDelta, OasisDEX
and 0x, which launched in August, are already up and running and will be the foundation for other exchanges.. Bancor is in development, and the AirSwap
and OmegaOne
are still preparing to hold an ICO.
Interestingly, up to 40,000 transactions a day are now passing through EtherDelta, which is about 14% of all daily transactions on the Etherium blockchain. You could say that this is the most popular Etherium app at the moment.
Pros and cons of centralized exchanges
A cryptocurrency exchange must perform three basic functions:
-
Liquidity – quickly find a partner to make an exchange;
-
Efficient value determination, leaving as few opportunities for arbitrage transactions as possible;
-
Quick settlements.
A centralized exchange is able to speed up both matching and settlement-the order book and settlement execution are on the exchange’s servers, and the transactions can be done much faster than on blockchain. The effectiveness of value determination is largely determined by exchange volumes.
The flip side of these advantages are security problems (centralization) and the presence of a vulnerable intermediary (the threat of both fraud and any action by the state).
How can a decentralized exchange solve these problems?
Security – DEX users control their funds on their own, right up to the moment of blockchain settlement (although the speed of this transaction is much lower). Other problems are a bit more complicated, and DEX projects offer different ways to solve them.
Liquidity, or finding a partner quickly: Finding a counter-order can be done through an order book (some degree of centralization, and delays between order placement and settlement can lead to hanging orders); having your own reserves (maintaining your own balance); or using a direct P2P transaction (difficulty finding a partner).
-
Order book: can be offchain or onchain. Keeping an off-blockchain order book (EtherDelta, 0x) is easier if you don’t consider the risks of increased centralization. Exchanges are solving this problem by rejecting the automatic matching of counter orders (rejecting the intermediary) – the party responding to the order must sign a specific counter-order, which only then goes to the smart contract. Such a scheme has a negative impact on liquidity. 0x goes a bit further here – Relayers, which anyone can become, can create multiple order books, versus a single centralized book in EtherDelta;
-
On the other hand, a blockchain-based order book (OasisDEX), while ideal for DAapp, many users aren’t happy about having to pay a gas fee for every transaction-cancel order, price change, deposit/withdrawal, etc.. д.
-
Reserves: Orders can be executed quickly if the exchange has its own reserves (Kyber, Bancor Protocol), but this requires it to maintain its own balance sheet. Kyber offers a solution in which a third party would be able to create a pool of tokens backed by other participants or centralized exchanges.
-
P2P: Direct price agreement between participants (AirSwap, 0x) via private messages – however, it’s a long process. In addition, picking up a counter-order requires a third party (centralization), and determining an agreed-upon price can be time-consuming.
-
Reference Price: In transactions, most decentralized exchanges have to rely on order books/reserve prices to determine the reference price. However, as long as the volume of trading and the number of participants is insufficient, price determination may not be optimal, and the exchange will have to rely on external sources or centralized exchanges. In turn, P2P exchanges will always depend on a third party to determine the reference price.
The table shows the comparative characteristics of decentralized exchanges:
Who will win?
The smart contracts described above are much safer than centralized exchanges, but security will probably not be the deciding factor. Users would prefer speed, no market manipulation, and low transaction costs.
-
Speed: Order-based exchanges on blockchain (OasisDEX) and P2P (AirSwap) will remain slow and will not increase in speed with volume. The speed of exchanges with the order book outside the blockchain (EtherDelta, 0x), although it will be low, but will grow with the growth of volumes.
Note: the situation may change once the scaling of Etherium is resolved.
Transactions by exchanges with their own reserves (Kyber, and possibly Bancor) could be almost instantaneous, but only assuming that Kyber and peers manage to build and maintain adequate liquid reserves. -
No middleman: No middleman in OasisDEX. Exchanges with centralized offchain order books (EtherDelta, 0x) are vulnerable to DoS attack by order book managers; orders “hanging” in the book are a target for arbitrage transactions. In AirSwap, there is a need to trust the service that helps find counter orders (Indexer) and the API oracle that provides the cost of. In Kyber, the network operator decides which reserve providers to allow and which not. Determining the price of the token by the provider is also a problem: an error can lead to rapid depletion of the reserve. It was the methodology of price determination that was the subject of harsh criticism of the Bancor protocol by Emin Gun Sirer.
-
Cost: Blockchain settlements require gas commissions. As for other costs, the winner will only be determined by competition: Relayers in 0x-based exchanges, as well as liquidity providers in Kyber, intend to charge users fees, the amount of which will vary.
Conclusion
Many decentralized exchanges are expected to appear in the very near future. About a dozen teams have already announced the creation of exchanges based on the protocol 0x. The first such project to reach the stage of open beta testing was RadarRelay, and operator OasisDEX MakerDAO has already announced a partnership with 0x for further development. On October 10 will ICO AirSwap, but so far the undoubted leader remains EtherDelta, incidentally, developed without any third-party funding.