Exchanges have been a weak spot in bitcoin going back to the Mt. Gox days. It is a decentralized currency, so why do so many people store their coins in centralized depositories?
There are a lot of reasons, chief among them are ease of use and accessibility. Most of the time it is just laziness. But day traders have a legitimate reason: they need near instantaneous access to their coins. Constantly moving coins from local wallets to exchanges for the purpose of trading would be expensive with mining and exchange fees. Worse, it would take time to do the transfer, potentially costing traders significant amounts of money as the market moves. Day trading depends on speed, and waiting for the blockchain and exchange is simply not an option for a lot of them.
And so we have these central points of failure. We have our Mt. Goxes and and the dozen or so other exchanges that have failed over the years. But we still have them because a large portion of the bitcoin ecosystem needs them and needs to use them in a way that puts coins at risk.
That was underlined earlier this week when a fairly obscure South Korean exchange Coinrail was allegedly hacked for around $30M USD of different altcoins.
There have been attempts to fix this problem. There are decentralized exchanges out right now. But they aren’t popular for a host of reasons. They are slow. They have issues with pairing orders. It just isn’t easy or quick to trade without a trusted third-party.
Decred has proposed a new decentralized exchange, one that will solve many of the problems associated with current attempts to decentralize how we trade decentralized currencies. Importantly: It goes far beyond the Atomic-Swap system utilized by today’s decentralized exchanges
There are several features to the proposed system from the Decred post:
It facilitates exchange between only cryptocurrencies, not fiat currencies.
It is architected as a simple client and server, without a corresponding token or a blockchain.
Server operators never take custody of client funds.
It uses on-chain transactions for order fulfillment and rule enforcement.
Server operators collect no fee for matching orders.
Adding support for coins is a straightforward matter of adding the corresponding atomic swap support.
Orders placed on the exchange can be internally regulated via rules enforced by the clients and the server.
Malicious clients are managed using a reputation system based on Politeia.
There is an upfront fee to create a client account on a server, to discourage malicious behavior.
Order matching occurs pseudorandomly within epochs.
Order sizes on both the buy and sell side of a trading pair have standardized lot sizes.
Limit orders and cancels are broadcast by clients via the server, but market orders are routed from client-to-client.
Near-instant exchange for smaller orders can be achieved through a related off-chain LN-based network which uses atomic swaps.
Servers can connect via a mesh network to allow cross-server order matching.
External services, e.g. wallets, can access a simple client API on the server that provides a data feed, ability to place orders, and other services.
Rather than use an Ethereum Token or a new blockchain to facilitate Atomic-Swap trades between users, the proposed system would use a client-server system to match orders.
Essentially, users would decide to set up their own mini-exchange on the Politeia registry. Buyer and sellers would use that exchange to enter a smart contract that will protect the funds from all participants, and then transfer the funds when the contract’s conditions are met.
Any malicious behavior would be recorded on their corresponding blockchains and presumably the Politeia registry. That, along with a small sign up fee, should make any attempts at malicious behavior unprofitable. The fee could also be used to offset the expenses of running and maintaining a DEX server.
In principle, it is a bit like my favorite 1990s file-sharing software, Hotline Connect. That service enabled users to set up their own file repositories that had their own rules governing who could access what. Servers were found in centralized lists, but each server was an entity in of itself and if one server went down, it had little to no effect on the system as a whole.
This proposed system seems to work similarly, except that the different servers can communicate and match orders.
It seems to be a novel solution that goes beyond the “blockchain for every problem” solution that has permeated the space. It eliminates the need for a Trusted Third Party by decentralizing that trust to the crowd and ensures speed by not weighing it down with an unnecessary blockchain.
There are sure to be unforeseen problems with this proposed system. Which is what makes open-source software so great. Those problems will have solutions eventually, even if that solution is something different than what Decred has proposed. But the centralized exchange problem has not been solved with slow decentralized exchanges, so seeing a proposal for a faster decentralized exchange is very exciting news for traders and potentially bad news for regulators.