Dr. Klaus Kursawe, Vega Protocol BFT and blockchain researcher, led discussions during San Francisco Blockchain Week on how many decentralized trading protocols are sidestepping or paving over four major technical issues that will preclude mainstream adoption of DeFi products.
Dr. Klaus Kursawe, Vega Protocol BFT and blockchain researcher, led discussions during San Francisco Blockchain Week on how many decentralized trading protocols are sidestepping or paving over four major technical issues that will preclude mainstream adoption of DeFi products.
As outlined by Vega’s research, these systemic issues present both old and previously unknown impossibility results that can disrupt current decentralized trading applications built on both Proof-of-Work (PoW) and Proof-of-Stake (PoS) protocols:
Fairness: While many DeFI products do not care too much about fairness – in Bitcoin, for example, it is part of the system that one can pay the miners for preferred processing – trading applications need to take more care about this. The desired property would be, for example, that if all honest validators in a trading platform see a trade from Alice happen before a trade from Bob, then Alice should be processed before Bob. Unfortunately, this is impossible to guarantee even if a single validator is corrupted. While work is still ongoing, PoS networks for decentralized trading have the potential to bring us closer to these ideal results.
Beyond the ⅓ threshold: In an asynchronous system, agreement is only possible if less than one-third of all participants are corrupt. This problem does account for coordinated malicious failures, as well as independent or arbitrary ones. While PoW systems usually assume a 51% honest majority, the actual corruption resilience in an asynchronous network is not that simple. At present, PoW protocols get around this issue (as well as the following ones) by forgoing finality, and by introducing some implicit timing assumptions. Vega’s approach is to use a scheme that extends fixed thresholds with properties, e.g., taking a party’s geography into account, and can be modified without a hard fork in response to threats to the integrity of the network.
Scalability: Scaling the number of validators in a PoS protocol requires a (worst case) quadratic increase in communications, which caps the number of validators at 100-1000 nodes. There are several approaches around this which still need to stand the test of time. Vega’s proposed architecture resolves the issue by being flexible enough to adapt to different solutions in vivo while using financial incentives to steer the number of validators to a manageable size, which can be defined by an on-chain vote.
Agreement is impossible, anyhow: It has already been shown in 1985 that in an asynchronous system, there is always an action the adversary can perform to prevent termination with an agreement. PoW is a cheat to consensus — it never terminates. PoS generally cheats this by either introducing some timing assumptions or by using a randomized protocol that terminates with probability one. While all solutions have drawbacks, the latter is more suitable for a financial markets, which depend on low latency and transaction finality; nevertheless, it is still vital to understand where the compromise was taken and to assure it does not affect higher level protocols.
While both PoW and PoS protocols are affected by these issues, both use fundamentally different approaches to navigate their results. Vega’s strategy is to use a PoS solution, which is better suited to address the unique needs of financial markets. Vega’s solution is a PoS-based protocol, designed for financial markets. In addition, their protocol’s proposed architecture provides the flexibility needed to adapt to future changes, including portability to other existing platforms or the ability to replace the consensus protocol on the fly through an on-chain vote.
“PoW protocols have had a great impact on Byzantine fault-tolerant applications, and have enabled a large number of exciting applications. However, for the case of trading, we believe that the latency, finality and fairness goals strongly prefer a voting based (i.e., PoS) set of protocols, even if that poses new challenges in keeping the network open and permissionless,” said Dr. Klaus Kursawe.
Vega’s decentralized trading protocol resolves the issues above for posterity, presenting a better or alternative protocol for the development of, and porting over of existing DeFi products and exchanges. As it stands, DeFi applications being built on smart contract platforms, depending on their use case, could potentially encounter obstacles as these impossibility results arise.
Decrypt reported, “there’s now more Ethereum (ETH) tied up in DeFi apps than ever before.” The legacy blockchain today hosts an all-time high of 2.36 million ETH in lockup for DeFi apps valued at more than $420 million. That number is likely to increase as more and more blockchain projects latch onto the latest “killer use case” for cryptocurrency.
These numbers become more alarming if we consider the existing limitations of decentralized exchanges. The need for liquid trading, and an eventual stable medium of exchange, are major components for the adoption DeFi DApps. Proper infrastructure is required to architect a decentralized financial ecosystem for the trading of both crypto and non-crypto products.
Vega Protocol plans to dedicate significant research and development resources to tackle the obstacles in successfully architecting the decentralized finance (DeFi) world. Through collaborative efforts, the Vega team plans to craft viable solutions for the mainstream adoption of decentralized technologies.
