Thread by @CryptoExplain23

1/ Here we go... ZK World pt. 3 all about @zksync 

If you need a preface please check out pt. 1 on L2 scaling and pt. 2 on zero-knowledge proofs

It’s TIME! 🧵👇

cryptoexplainere60.substack.com/p/zk-world-pt-…

2/ @the_matter_labs: founded in late 2018 by @gluk64 & Alexandr Vlasov to focus on Eth L2 scaling; recieved grant from @ethereum foundation in March 2019

ML is "engineering team passionate about liberty, blockchain, and math"

Creator of @zksync

matter-labs.io

3/ Investors: ML has raised two major rounds

Series A: $6m raised in March 2021 led by @usv; inc. @dragonfly_cap & @placeholdervc. Strategic partners: @AaveAave & @1inch, etc

Series B: $50m raised in Nov. 2021 led by @cdixon/@a16z. Strategic partners: @blockchain & @OKEx, etc

4/ Products: in 2019 ML launched prototype zk-rollup (zkr), followed by zkSync 1.0 in July 2020

zkSync 1.0 is app specific zkr L2 w/ focus on payments. It provides:

- low txn fees
- instant txn confirmation
- txn finality within minutes
- withdrawals to L1 after ~15 mins

5/ zkSync 2.0 is EVM compatible zkr - "zkEVM" which was originally supposed to go live on mainnet in August 2021 but delayed due to technical difficulties

Currently there is zkSync 2.0 @Uniswap fork live on testnet

blog.matter-labs.io/unisync-a-port…

6/ Environment: despite not yet having zkEVM live on mainnet, zkSync has experienced parabolic user growth over the past couple months

zkSync now has over 2x as many users as Optimism and is closing in on Arbitrum 

Now let's get into the weeds 👇

7/ EVM compatibility: zkSync 1.0 was based on Groth16 SNARK proof system which limited it to app specific use cases bc trusted setup would need to be repeated for each new app (and app update)

In PLONK proof system trusted setup is done once = universal & updateable SNARK

8/ zkSync 2.0 & EVM: zkEVM aims to run 99% of contracts from Solidity & maintain behavior of contracts (not tech. 1:1)

LLVM compiler converts code to SNARK friendly code

this is done for Eth (solidity, vyper, yul) & zkSync native Zinc

result = EVM compatible & composable

9/ How zkr work: rollups move execution off-chain while maintaining per txn data on-chain; result is scale, EVM compatibility, and L1 security

Merkle tree cryptography is used to maintain and update rollup

10/ In Merkle tree each txn in block is hashed, then each pair of txn's is concatenated and hashed together until there is one hash for block; the Merkle root

11/ in rollup, smart contract on L1 maintains the Merkle root for the state of the rollup (state root) and is updated each new rollup batch

There are two ways to verify this process - fraud proof (optimistic rollup) & validity proof (zk-rollup)

12/ zkr grant ability to prove honest computation w/o providing inputs to that computation using zero-knowledge proofs (zkp) - SNARKs (zkSync) or STARKs (StarkWare)

zkp for verification is based on prover/verifier relationship discussed in part two

cryptoexplainere60.substack.com/p/zk-world-pt-…

13/ the prover does complex computational work that results in a cryptographic proof so that the verifier can easily verify the integrity of the work using only the proof itself

14/ zkr ex: user deposits funds L1 --> L2 (zkSync) then does txn on @ZigZagExchange

cumulative txn's from users are grouped into batch by prover

to update rollup state prover submits to L1:

- root hash of new state for rollup
- zkp (SNARK)
- delta in rollup state a calldata

15/ huh?

- root hash reps. hashed batch of txn's; this combined w/ Merkle root in L1 smart contract to update rollup state 

- zkp is cryptographic verification the root hash being submitted is accurate

- change in rollup state -> L1 so any user can prove ownership to L1 on own

16/ zkr efficiencies:

- publish hashed txn's in form of state root instead of all txn data

- to update rollup state root prover does computation off-chain then L1 can efficiently verify the proof itself (instead of all the computation)

17/ efficiencies cont…

- calldata is cheapest way to store data on L1 (and could become cheaper w/ EIP-4488) = data availability (DA) on-chain

net result = 100x cheaper than Eth L1 txn's

18/ zkSync 2.0: zkr are incredible but scalability is linear due to DA being on-chain (calldata sent to L1)

to exponentially scale and offer even cheaper fees (tradeoff lower security) DA can be conducted off-chain

zkr w/ off-chain DA = validium

19/ Validium in zkSync 2.0 = zkPorter

zkPorter maintains DA off-chain using a PoS system similar to many L1s

- Validators batch txn's & compute zkp then submit

- Guardians stake zkSync tokens and protect against censorship in off-chain DA

20/ Volition: zkSync 2.0 uses zkr & validium (zkPorter) to provide users choice

zkr = higher fees (but still 100x lower than L1) but more secure (L1 secure)

validium = lower fees (multiples lower than zkr) but less secure than zkr (PoS security not L1 security)

USER HAS CHOICE

21/ Progressive decentralization: currently ML does prover work but once zkSync 2.0 launches and zkSync token goes live there will be progressive decentralization until the ML team is no longer relied upon (self-sufficient system)

22/ zkSync v StarkNet: philosophically Alex G. has been focused on digital freedom and decentralization while stating he sees StarkWare as building proprietary scaling solutions 

reddit.com/r/ethereum/com…

23/ technical perspective differences:

- SNARK v STARK

- Validium: zkSync uses PoS (decentralized); StarkNet uses DAC (centralized)

- EVM: zkSync focused on EVM compatability then native language (Zinc); StarkNet more focused on native language (Cairo) then EVM compatibility

24/ Ecosystem & Roadmap: in prep for zkSync 2.0 launch very much in infrastructure build-out = working with CEXes, L2 direct fiat on-ramps, & infrastructure integration

apps: very limited; mainly zigzag and zkNFT

roadmap: zkSync 1.0 -> zkSync 2.0 -> privacy -> decentralization

25/ Conclusions:

- @zksync has strong strategic partners w/ dApps & CEXes; fiat -> L2 and CEX -> L2 on-ramps are big deal. combined w/ dApp partnerships & infra. integrations prime zkSync 2.0 to hit the ground running

26/ conclusions cont…

- zkSync's focus on EVM compatability in ST and native Zinc in LT seems like winning strategy as it is paramount to build user base (strong user base leads to more developer engagement leading to more user engagement = flywheel effect)

27/ conclusions cont…

- Promise of decentralized future is driving user growth. Users want to be early to protocol and have opportunity to own (airdrop). Hype around zkr and promise of zkSync token creates excitement

28/ conclusions cont…

- Will zkSync's ST advantages of growing user base and ease of EVM compatibility hold up in LT compared to what many consider superior STARK technology implemented by StarkWare?