If this is the first time you're hearing the word zkEVM, here's a quick refresher:
2/17
2/17
There are varying degrees to which a zkEVM can be compatible with Ethereum
Let's look at each of these:
3/17
Let's look at each of these:
3/17
Devs can basically copy-paste the existing code for their apps from the Ethereum mainchain to the zkEVM
Ethereum-equivalence also means that the zkEVM can draft on upgrades to Ethereum itself with minimal extra work needed
So what's the catch?
5/17
Ethereum-equivalence also means that the zkEVM can draft on upgrades to Ethereum itself with minimal extra work needed
So what's the catch?
5/17
Ethereum was not originally designed around ZK-friendliness
So the inefficiencies of Ethereum carry over to the zkEVMs too
Type-1 zkEVMs require a very high prover time, as no unique reworking is done to make proof generation faster
This makes them extremely costly!
6/17
So the inefficiencies of Ethereum carry over to the zkEVMs too
Type-1 zkEVMs require a very high prover time, as no unique reworking is done to make proof generation faster
This makes them extremely costly!
6/17
2. EVM Equivalent
Type-2 zkEVMs look exactly like Ethereum "from within", but they have some differences on the outside
This is done to facilitate development and speed up the proof generation
@0xPolygonHermez and @Scroll_ZKP are building Type-2 solutions
7/17
Type-2 zkEVMs look exactly like Ethereum "from within", but they have some differences on the outside
This is done to facilitate development and speed up the proof generation
@0xPolygonHermez and @Scroll_ZKP are building Type-2 solutions
7/17
Most applications that work on Ethereum would still work on Type-2 zkEVM rollups
And others would work with slight modifications to the code
So does Type-2 get us the best of both worlds?
Not exactly.
8/17
And others would work with slight modifications to the code
So does Type-2 get us the best of both worlds?
Not exactly.
8/17
These modifications significantly improve prover times, but they do not solve every problem
The slowness from the inefficiencies and ZK-unfriendliness inherent to the EVM still remains in Type-2
9/17
The slowness from the inefficiencies and ZK-unfriendliness inherent to the EVM still remains in Type-2
9/17
As is the case with Type-2, most Ethereum applications are compatible with Type-3 zkEVMs and others require slight modification
However, there will be some apps that would need to be rewritten entirely because they use features that the Type-3 zkEVM removes
11/17
However, there will be some apps that would need to be rewritten entirely because they use features that the Type-3 zkEVM removes
11/17
4. EVM compatible
Type-4 zkEVMs look & operate nothing like the EVM
Instead, they create a new Virtual Machine that makes development and proof generation extremely easy in the zk environment
@zksync and @StarkNetEco are building "EVM compatible" Type-4 solutions
12/17
Type-4 zkEVMs look & operate nothing like the EVM
Instead, they create a new Virtual Machine that makes development and proof generation extremely easy in the zk environment
@zksync and @StarkNetEco are building "EVM compatible" Type-4 solutions
12/17
A Type-4 system works by taking source code written in a high-level language like Solidity and compiling that to some language that is explicitly designed to be zkSNARK-friendly
By doing this, Type-4s drastically reduce costs and improve proof-generation time
13/17
By doing this, Type-4s drastically reduce costs and improve proof-generation time
13/17
Naturally, this comes at the cost of a few challenges:
- Most apps are not compatible and need to be rewritten entirely
- It may not be easy to carry over future Ethereum upgrades to Type 4 zkEVMs
14/17
- Most apps are not compatible and need to be rewritten entirely
- It may not be easy to carry over future Ethereum upgrades to Type 4 zkEVMs
14/17
As @VitalikButerin pointed out, no one type of zkEVMs is better or worse than the others
There is development happening in all 4 areas that will lead to more innovation in this space
15/17
There is development happening in all 4 areas that will lead to more innovation in this space
15/17
At any point, zkEVMs can decide to transition from one type to another by simply adding/removing certain features
In the future, Ethereum itself will undergo changes that will make it more zkSNARK friendly
16/17
In the future, Ethereum itself will undergo changes that will make it more zkSNARK friendly
16/17
Irrespective of which solution wins in the future, zkEVMs are a welcome development in solving the Ethereum scalability solution
I hope you learned something new from this thread. Help spread the word by retweeting the first tweet (linked below)
Follow me for more alfa!
17/17
I hope you learned something new from this thread. Help spread the word by retweeting the first tweet (linked below)
Follow me for more alfa!
17/17
Loading suggestions...