Comparison of the second Ethereum drug zkSync and Starkware from 6 points of view

@Dareal_sisyphus recently published a comparative article on the two-tier Ethereum solutions zkSync and Starkware, examining it from 6 perspectives: team, technology, data on hand, finances and sponsors, current products and information, and personal feelings. Here is the compilation and finalization of everything.

To say that the ZK accumulation will be too large would be an understatement.

Two beasts have been working on this device in the background for many years. The time is coming when they will shine. So what are the opportunities in this area? What is the difference between these two positions? When do they give out coins?

Here is a comparison of zksync and StarkWare.

I will say before that this is a big point, and I should know it to the best of my ability, and there may be some short or small mistakes in the description. I accept any correction or clarification in the comments.

There are several programs available that explain what ZK stacking is and how it differs from optimism.

Here we are only focusing on the basics after zk rollup and the differences between the two main players: zk games-zksync and StarkWare.

Basics:Ethereum rn is what we call a "microcontroller". Indeed, approvals, achievements and data are available everywhere on the ETH blockchain. This is unstable and therefore cannot be measured. This is because it means that each of them has to do these 3 tasks. Eth must be "modular" to measure. This means that the blockchain does what it does best: use it only for approval and the off-chain execution and data acquisition (DA) load. But how do you move the kill and DA in a way that secures the whole of L1?

zk rov:With zk rollup, thousands of tx are bundled into a single zk proof, and only that proof is stored on the chain.

Compare zksync and @StarkWare below. We will try to control it clearly and follow these benchmarks.

start with the team

Technology

have information

Finances and Sponsors

Current products and plans

personal impressions and opportunities

start with the team

Starkware: By May 2018. The team is made up of world-class cryptographers and scientists. The key staff member is the former Director of Research at Zcash who has been the new leader in the zk field for many years. They use Starknet to turn educational theories into real products. They are arguably the most famous team in the game.

Siegsink: Released in December 2019. In fact, there is a "second best" in the zk arena, and it was the Matter Labs team that developed ZkSync. This team was created by @ gluk64. I can't find detailed information on the genius behind this bunch, but the new technology brought by ZkSync 2.0 is good enough to explain the problem. They have the breath of an outsider and go fast.

Technology

The two projects have similarities, with the smart contract connected to the Eth blockchain keeping the zk proof of the L2 state exchange. There are also two types of off-chain actors who have a strong network.

Proof:A small number of nodes play an important role. They count all tx and end with zk certificates. It works on special materials and can be turned into a black box. What we do know is that we cannot tamper with false ZK evidence.

Validators:More responsible for the prevention of censorship. They validate the certificate delivered by the authenticator. No special hardware required. Anyone can run.

In addition, two projects are expected to compete to overcome the major challenges of developing the ZK Universal Certification System.

This is pure grail, because every ZK line so far needs a special application. This means that there is only one ZK rollup for each application and no EVM compatibility.

zksync and StarkWare were both successful, but the technology was different.

Starkware: The encryption technology they use is based on STARK. They are the pioneers of this technology and have two advantages over SNARK (used by zkSync).

"T" stands for "transparent" which means that the system should not be dependent on offices. The production speed is 10 times faster than SNARK. But the problem with STARK is that the tech release is not like SNARK, and if you allow Turing to be complete, it will be difficult to fit it into EVM.

Starkware also developed Cairo, a specialized programming language that can run programs illegally by STARK. However, not many people want to learn new programming languages ​​to create smart contracts, so we are currently working with @nethermindeth to develop a Warp Interpreter that can replace these smart contracts for Cairo to make them compatible with @StarkWare EVM.

zksync: The Zk system used is based on SNARK (especially PLONK). Therefore, the security of the entire system will depend on the reliability which was launched in the fall of 2019. This includes many well-known cryptocurrency players, including Vitalik Buterin. If at least one member is honest the system can turn out to be safe so I like it very much and I don't think the word "trust" can be used for zksync.

Although the zkSync team has improved, SNARK is still slower than STARK. But their system allows better EVM integration than STARK. In fact, smart contracts can be (almost) interpreted one by one by the zkSync compiler, resulting in a solid “first public” version of zkSync without the need for middleware or modifications.

Industrial data (DA)

The cumulative Zk unloads the calculated txs in L1 and can measure ETH for 2-3K tps. This is good, but not enough. With the "rollup only" method, the DA load is always on L1 and tx data is written to L1 over the call data.

DA is very important. This is the key that allows you to see what happened to the Etherscan tx. Without it, the execution of tx will become a black box. This is possible if you want a cheaper tx instead of tracking directly from L1. Thus, the cumulation zk must be attributed to the user of his choice. Your tx file is written to L1, but key tx with a slightly higher or lower value and trust L2 to make sure the tx file is present. To do this, L1 can be released from the load of DA cumulative, and ZK can be extended to 20-30k TPS.

Starkware: Solve the DA challenge with the Volition system. Volition allows end users to choose between tx roll-up (on-chain DA) session and tx-phase (off-chain DA) access. For efficient payment, the off-chain DA is accredited by the Data Availability Committee (DAC), which is made up of reputable cryptocurrency entities.

It sounds like centralization, but it impacts ultra-low cost TX that end users can choose from.

zksync :The DA solution is a zk cumulation of zkSync and zkPorter, an incompatible functional connection. Consider that the zkSync rollup is shard0 which guarantees the DA of the chain. Different partitions can then select different DA rules and store their data off the chain.

The main difference between StarkWare's Volition and zksync zkPorter is that for Volition the payment option is tx-based (user selects DA for each tx) and for zkPorter it is account-based (zkPorter account is tx as DA). chain).

In addition, for zkPorter, the off-chain DA system is more decentralized. Indeed, the DA is recognized by the "guardian" network generated by the native zkSync token rather than by the root "DAC".

Finances and Sponsors

Starkware :

A funding round of $ 6 million was completed in May 2018 (Pantera / Naval / Vitalik).

In October 2018, he completed a $ 30 million Series A (Paradigm / Sequoia / Cb Ventures) financing.

Completed $ 75 million in Series B funding in March of last year (Paradigm / 3AC / Alameda)

Last November, it completed a $ 50 million Series C funding (Paradigm / 3AC / Alameda).

Currently, StarkWare is worth around $ 2 billion. It is a global financial system with well-known investors. The conglomerates and the Ethereum Foundation are one of them. Vitalik himself has commented on most of the products developed by Starkware. Shu and Sam are also on the same boat. How to achieve this goal?

zksync :

In March of last year, she finalized a Series A (Binance / Cb Ventures / AAVE / Balancer / Curve) funding worth $ 6 million.

Last November, she completed a $ 50 million Series B funding (Horowitz / Placeholder / Crypto.com, etc.).

Fewer board members and it looks like the financial giant of the Defi / CEX cryptocurrency family. It doesn't look good, but the success of the ZK rollup is very important, as it largely depends on the direct integration of the Defi protocol with CEX. So I'm very excited about the rapid integration of the zksync ecosystem.

Current products and plans

StarkwareThey made a lot of sense because they turned the best crypto data into the world of innovation. Their plans are listed below.

They were the first to develop StarkEx for the "planetary" phase and were able to develop special licensed zk rollups powered by Cairo and STARK. If you are not sure you can think of @dydxprotocol, @Immutable or @deversifi. It works on StarkEx products because there are 3 main applications.

To date, StarkEx has processed over 5 million transactions and over $ 250 billion in applications. StarkEx's performance has been recognized and StarkWare will soon enter the “constellation” phase of its process. On November 29, we released the alpha version of StarkNet to the Mainnet. StarkNet is not licensed, many zk rollup apps and goals have arrived.

Initially, StarkNet was used by the Base Validator, and applications were distributed regularly through a free list, such as Optimism. Therefore, at the “center” level of the road, we plan to develop the ecosystem and gradually decentralize StarkNet.

zkSyncThe roadmap is written in four stages.

The first phase corresponds to zkSync 1.0 released in June 2020, which is estimated to correspond to the zk rollup, but without a smart contract, allows you to send and receive tokens. Although no configuration has been done, contractual commitments have been submitted and all metrics show interest in zkSync 1.0.

The second phase starts with the mainnet deployment of zkSync 2.0, which is all we need, including full EVM compatibility of the Zk rollup and smart contract composability.

zkSync 2.0 was originally slated for release on the mainnet in August, but has been delayed due to performance issues. These issues are now being solved in experimental networks, and in October of last year zksync announced that the technology process was almost complete and that it was using a network like AMM to prove EVM quality compatibility.

Matter Labs' delay in ensuring LLVM / Integration can be stressful at first, but it helps zksync free up more time using all tools and Rely can still be integrated with zkSync 2.0.

personal impressions and opportunities

StarkwareIt is very impressive. This is the best construction project I have seen in a long time. The All-Star Group, the most innovative and profitable group in the world. The surprising thing is that they don't just create a ZK stack. They think of everything. We are talking about Cairo and Wharf. But they also work hard to bring true randomness to Eth, from incident to L1 to L2 communication using Veedo, long-term batch flash loans, and other ingenuity. The use of "distributed AMM" to resolve AMM fragmentation of L1 and L2 is a real encouragement. They see the world.

One of the technical issues I'm seeing with StarkWare right now is the integration of the Warp Adapter. Starknet was first optimized for Cairo and a Solidity transponder was added to it. I'm not a very smart person, but I'm concerned that “collaborative interpretation” is not fair in all cases and may lead to relationship issues for some smart contracts.

The other question I see is: where are the public opportunities? StarkWare is already worth around $ 2 billion, but currently has no plans to develop a fair token.

On the other hand I also likeSiegsink“Emphasize the community approach and 'safety as a public priority'. They all strive to build on the latest EVM experience without commercialization, which can be beneficial as an adoption. zksync is also tightly integrated with the current Defi / CEX environment, making it beneficial for future access and adoption.

Currently, zkSync has announced that it has a local token and will release token token and drop capabilities at some point in the future. It's good for the community, but I think the Zk rollup root token will be outdated as everyone and their team are waiting for it.

We also want StarkWare to place local tokens in the “settings” level. Because one day it will be necessary to decentralize the network and encourage provers.

Overall performance is better in terms of StarkWare zk rollup, but I think the community / open source vibe makes zksync more appealing.

This is my last me.

In terms of opportunities, we need to shift our thinking from tokens at the bottom of stacks to new ones that will see the user base for stacks. ETH is a flexible exchange, and the task of picking up chain orders was quickly replaced by the AMM pattern that fits this new space. Likewise, the ZK cumulation is a change for Eth. For example (probably) copying the AMM in a race is a mental concept.

These are all temporary examples, but our guess is that you might be in a project that uses these rollups to do something impossible in L1 rather than a project that publishes existing ones.

If you've come far, celebrate.