Thread by @Alex_Ausmus1

Let's talk about @dYdX and their decision to leverage the Cosmos SDK. This thread is a summary of a convo I had with a friend outside of crypto who was curious...

2/ What's dYdX?

An orderbook-based DEX built on StarkWare providing features attractive to institutional traders. They are heavyweights in the Etherum DEX niche and the team is composed of OGs in the broader crypto space

3/ So why did they choose StarkWare?

Frankly, it provided the best environment for dYdX to build their vision at the time. StarkEx brings STARK proofs to scale on-chain, pairing perfectly with exchanges such as Immutable X & dYdX.

4/ What changed?

Nothing really, the dYdX team has constantly been iterating to provide the best product they can. StarkWare has been a great platform for them but they want to grow further and to do so they need more control over their product.

5/ So Cosmos?

Yep, the Cosmos stack offers tooling to build sovereign, secure, performant decentralized networks. dYdX finds community involvement and ability to tweak network parameters important in this next stage, which Cosmos allows out of the box

6/ They can also tap into some other goodies that the Cosmos ecosystem provides, such as IBC, Interchain Accounts, and maybe even Layered Security from other Cosmos chains down the road.

7/ Also, the Cosmos community is a community of builders helping builders. Many core contributors to Cosmos projects help throughout the ecosystem creating a tight-knit community - even for those not developing. It seems that everyone wants to help everyone in the Cosmos

8/ Alright, so how can they build something more decentralized on Cosmos today?

Well, it’s important to know a bit more about dYdX’s architecture first. Currently, there are decentralized and less decentralized parts to it.

9/ The orderbook and matching engine are running on centralized servers allowing for the throughput needed to handle the high amount of order additions/cancellations per second. Then there are the smart contracts on StarkWare and Ethereum.

10/ The contracts on Ethereum are as decentralized as Ethereum is, but those on StarkWare are not.

11/ All transactions that occur on StarkWare (or any rollup for that matter) must pass through a centralized sequencer which may be subject to censorship.

12/ Well, why not wait until a performant and secure decentralized sequencer is built?

Let me answer with a question. Why make the evolution of your product rely on an external team?

13/ They can build a similarly decentralized and secure product while improving other aspects via an appchain rather than waiting for StarkWares decentralized sequencer. It's not just about decentralizing the sequencer, there are other things dYdX wants to achieve with this move

14/ Okay, I get that but isn’t a validation set of 50-100 pretty damn centralized?

If you’ve spent time in the Cosmos ecosystem you have heard this before. It’s not about the number of validators, it's about the decentralization of stake and the value of that stake.

15/ Sure, it would be ideal to have more validators, but it’s all about tradeoffs. Finding the balance between sufficiently decentralized stake and network and application performance is what’s important here.

16/ To be clear, the move to a Cosmos-based chain is not solely about decentralization. It's about pushing boundaries and making logical tradeoffs for a better product.

17/ Okay, what about security? 50-100 validators doesn’t sound secure.

You’re right, a fresh blockchain with few validators even with well distributed stake won’t be extremely secure, so they encourage security with other means as well

18/ Enter Game Theory - one interesting benefit of an appchain is that all the validators are generally incentive aligned for the application’s success. So long as it is expensive to act maliciously, a small validator set will be fine for now.

19/ But let’s not ignore the fact that ideally, the validator set grows larger as the technology improves. Again, it’s all about tradeoffs.

20/ So what are they building?

A decentralized off-chain orderbook network paired with a Cosmos-SDK based blockchain for trade execution. Validators will store their own orderbook, the blockchain doesn't need to come to consensus on the orderbook.

21/ Are there risks?

Totally! This is a large technical endeavor and it will not be easy, but this is the team to do it.

22/ Another is external dependencies. To reach their goal there needs to be improvements to many other pieces in the tech stack that they will not be working on. One in particular is bridging technology for a better UX.

23/ And finally, the issue of a network halt due to 1/3 of network stake not signing blocks. The implications of this sort issue are very real. Excited to see how the dYdX chain approaches the inherent pain points in PoS systems.

24/ What’s next?

Well, contribute. Join their discord, ask questions, create content, offer help, what ever you can do. And at least just watch for new announcements.
discord.com/invite/dydx

25/ It will be really interesting to see dYdX evolve into its v4 form. The push back on it is also awesome! It sparks great discussions and will only help dYdX and the entire crypto ecosystem grow together