Every time I start reading a table comparison mapping out tradeoffs in the L2 design space when I'm in the metro station, I snap out of in a completely foreign part of the town. This took me to "Bredäng". https://t.co/aOwnDJYbGj
I'd be interested to know what the throughput would be if the new StarkEx stark was used within a ZKrollup rather than within Validium. Anyone have any insight? That way would have a ZKrollup with no trusted setup
This is absolutely possible, although STARKs currently have relatively large proof sizes—verification of a single proof takes something \~5m gas, if I'm not mistaken. So the rollup capacity of StarkRollup would be lower as compared to a zkRollup on PLONK (\~500k gas for verification).
I was wondering if the new Starkware stark (Ethstark I think it might be called?) managed to reduce this proofsize. I think they are giving an overview of it later this month so hopefully should find out a bit more then
Many new L2 solutions are coming up to scale ETH. Thus ETH will definitely be a powerful ecosystem capable enough to host other ecosystems with more oriented utilities. Speaking of which VID.camera succeeds in the creation of Value Income opportunities in collaboration with businesses. We need more such projects in the ethereum ecosystem.
The PLONK setup is universal and updateable, which basically means that it's very easy to have a setup that has hundreds of participants; the AZTEC one that Matter Labs is using had something like 180+ (including myself)
For zkSync, Matter Labs is reusing the Ignition ceremony from last fall with 202 participants, many of whom are public figures including Vitalik Buterin, Aztec protocol, Matter Labs, and others: [https://www.aztecprotocol.com/ignition](https://www.aztecprotocol.com/ignition)
What is the chance that everybody colluded and not a single one was honest? Up to you to decide.
Many of what we call trustless systems have much stronger security assumptions. Many have weaker ones. It's a spectrum, not a binary option.
202 participants doesn't really mean anything. 201 of them just ran a docker image somebody pushed to github, and Vitalik wrote his own python script (which I assume pulled in 500 megs of unaudited package dependencies) and ran it on his (highly targeted by intelligence agencies) laptop.
It's "probably" fine, but we don't really know and we have no way of verifying. Hence the *trusted*.
This is a valid concern and we hope that the next ceremony (PLONK for BLS12-381 elliptic curve) will attract even more community interest and more independent implementations.
Most people don't verify the code of their wallets or full nodes either, and certainly don't verify the hardware circuits of their machines. There is always \*some\* degree of trust involved in any cryptographic system.
So what I got from this is Zkrollups using plonk is by far the best L2 solution imo, from a security and UX perspective. It remains decentralised and doesn't require you to deposit ETH into a seperate L2 compatible wallet before using it. You can use it directly from your current wallet.
I've been on the fence about L2 solutions for years now because lightning network completely ruined the idea for me. I hated the complexity and felt that even with years of improvement the fact you have to load a LN wallet with funds before being able to use the protocol wasnt great.
Zkrollups are genuinely amazing and I now understand why vitalik is so enthusiastic about them.