Bankless | Why are dapps leaving Ethereum? What trade-offs has dYdX made?

Last week, dYdX, a perpetual derivative exchange built on Ethereum's L2, announced that it will migrate its protocol to the Cosmos ecosystem in the upcoming v4 version and develop an application chain based on Cosmos SDK. This news has sparked discussions in the crypto community, as it may be the first time a prominent native DeFi protocol on Ethereum has chosen to leave, raising questions about the slow development progress of L2. Crypto research firm Bankless analyst Donovan Choy also shared his views on this matter. This translation and summary are provided for reference. For any uncertainties, please refer to the original article.

Reasons for dYdX Leaving Ethereum

According to the official article, dYdX's main motivation for moving to Cosmos is to expand transaction speed while maintaining decentralization:

"The fundamental problem with every L1 and L2 network on which we could build is that they don’t have the throughput to handle a first-class order book or matching engine. For reference, the existing dYdX products process around 10 trades per second and 1,000 orders or cancels per second. We are targeting several orders of magnitude improvement."

Due to the current slow speed of dYdX and the lack of suitable alternative networks, along with the setup of StarkEx L2 forcing them to deploy a centralized order book system rather than achieving decentralization through a network of validation nodes, dYdX made the decision to leave.

However, why Cosmos?

Donovan stated that the SDK provided by Cosmos allows developers to build their own blockchain with full control over their validation nodes, greatly enhancing scalability. According to the official article, the network will match orders in real-time and submit generated transactions to the chain in each block. This will enable dYdX V4 to provide high throughput for the order book while maintaining decentralization.

Additionally, dYdX's current setup on StarkEx relies on a centralized sequencer, Sequencer, which, although efficient, leads to excessive centralization as only one participant can batch submit blocks to the Ethereum mainnet. In theory, having control over their own validation nodes in the Cosmos environment would make the protocol more decentralized.

Trade-offs by dYdX

The major advantage of deploying on L2 Rollups like StarkEx is the improved scalability compared to the Ethereum mainnet, while still relying on Ethereum's robust security, avoiding the high costs of enhancing protocol security.

However, since dYdX chose to leave, it will inevitably lose these advantages and face other challenges.

Donovan stated that as dYdX's plan is to build an independent L1 blockchain, this means it will not rely on Cosmos Hub's validation nodes but rather on its own centralized design, even if they intend to decentralize the validation nodes. Therefore, after migration, users not only lose the security guarantees of Ethereum but also have to trust a new set of validation nodes on the dYdX chain.

How many validation nodes will there be? What will be the majority stake parameters? Currently, these are unknown.

Furthermore, on the future dYdX chain, users will pay transaction fees with DYDX tokens instead of ETH. Validators on dYdX will also be paid in DYDX, fundamentally changing the nature of the token and incurring security costs as some form of token needs to be sent to validators as a cost for the protocol.

Additionally, with at least 50% of DYDX allocated to internal personnel, there is an issue of token concentration.

However, the most complex issue has yet to arise. By setting up a new L1 network chain with its own validation nodes, dYdX will need to use cross-chain bridges for users to migrate to the dYdX chain. Over the past year, cross-chain bridges have caused significant fund losses, which will bring more risks to dYdX.

"In conclusion, dYdX's security will depend less on whether users trust the Ethereum chain and more on whether users trust the operators of the new dYdX chain," said Donovan.

The familiar blockchain impossible trilemma reappears, with dYdX being decentralized and secure on StarkEx but lacking scalability. The new direction for dYdX is sacrificing security for scalability.

dYdX Regulatory Issues

While scalability seems to be the main reason for dYdX's move away from Ethereum, Donovan believes regulatory issues are also a significant factor.

Donovan stated that dYdX is particularly cautious about regulatory issues in the United States and does not offer its protocol to U.S. residents. Additionally, in an interview with Forbes, founder Juliano publicly stated that the protocol has been in contact with the Commodity Futures Trading Commission and other government regulatory agencies since early 2018 and has submitted multiple comment letters to regulatory agencies.

"Having a set of decentralized validation nodes on a new blockchain will prevent regulatory agencies from classifying the platform as a centralized exchange," said Donovan.

In summary, dYdX's choice may stem from scalability and regulatory issues, and its departure has sparked community reflection on the current state of L2. However, whether this is bullish or bearish for the future of dYdX, Starknet, and Cosmos remains to be seen. How will it affect dYdX's growth? Will users be willing to migrate to the new chain? Will their competitors like FutureSwap or Perpetuals Protocol benefit from this? Further validation is needed as time progresses.