Thread by @orbisproject

#Cardano requires scalability to achieve mass adoption

But scalability has to be achieved without compromising decentralization & security

Rollups are one of the best solutions we have now to achieve this goal

But which one?

Let's compare, "optimistic" and "zk-rollups"🧵👇

So what are rollups?

They are L2 solutions

•That bundle hundreds of transactions into a single transaction

•Rollup transactions get executed outside of L1

•But the transaction data gets posted to L1

•Which allows rollups to be secured by native L1 security

Rollups are general-purpose

Meaning

•You can run a copy of L1 (eg: #Cardano) inside a rollup

•Allowing existing dApps & native assets to migrate to rollups

•With almost no need to write any new code

This is how they work in general:

1. Users ''Bridge'' their funds from L1 to the rollup (L2) via smart contracts

2. Users can use these funds to transact or interact with the dApps inside the rollup ''cheaper and faster''

3. Rollups bundle, or “roll-up” the transaction data

4. This data is handled by network participants,

Known as 'sequencers' or 'validators'

5. Who then submits batches of highly compressed transaction data back to the main chain

Those batches contain the minimum information needed to verify whether the transactions are ''valid''

Fundamentally,

There are two approaches to rollups

Two approaches with very different ''security models''👇

•zk-rollups &

•optimistic rollups

Let's first understand the concept of "optimistic rollups"

Optimistic rollups assume all transactions to be valid by default

Then it rolls up transactions into batches

Which gets submitted to L1 in a single transaction

The transaction can be challenged if fraud is suspected

In the event of a ''challenge''

The optimistic rollup executes a ''fraud-proof''

meaning,

if someone can publish a proof to chain, proving that the batch was computed incorrectly

The contract verifies the proof and reverts that batch of transactions and all batches after it

''Fraud Proof'' comes with a tradeoff

You cannot bridge back the funds/assets on the rollup to L1 immediately

The withdrawal of funds needs to be delayed even ''7 days''

To give time for someone to publish a ''fraud-proof''

And cancel the withdrawal if it is fraudulent

This delay destroys the ''Composability'' between the dApps in L1 & L2

What does it mean in the context of DeFi?

Composability is the ability of dApps and tokens to be seamlessly interoperable with one another Like the Lego,

Where every piece can connect to every other piece

One of the main problems with TradFi is the lack of composability

Its almost always the case

That you have to wait for a couple of days to transfer the funds from one bank to another

Composability is one of the features that make DeFi unique

If you take away composability from the DeFi

You are eventually going to have dApps on L1 and different L2s

Which are siloed into different worlds,

Disconnected with a delay in moving funds between each other

Which affects the capital efficiency of dApps in a negative way

So how are "zk-rollups" different?

Instead of just bundling & compressing the transactions

zk-rollups allow L1 to settle the transactions off-chain

And the results are published on the L1

Using a clever piece of cryptography known as zero-knowledge proofs (ZKP)

"zk-rollups" bundle transactions & executes it off-chain

It then generates a zero-knowledge proof (ZKP)

For the hundreds of transactions, it has settled off-chain

And then this proof gets submitted on-chain/L1 for validation

So what's the fundamental difference between the two?

''Optimistic'' Rollups assume all transactions to be valid by default

And assume that everyone acts in good faith

Whereas "zk-rollups" ensures it is actually the case

This allows for immediate bridging back of funds to L1

The delay in transferring funds from "zk-rollups" to L1 depends on only one factor

The frequency of proof generation on L2 & it's submission on L1

Orbis aims to submit proofs in every block on #Cardano

This way dApps on Orbis would be perfectly composable with #Cardano

One of the prevalent criticism of "zk-rollups" while comparing it to "optimistic rollups"

Is about the computationally intensive validity proof generation of "zk-rollups"

Which may make it expensive to use the "zk-rollups"

Especially for dApps with little on-chain activity

ZKPs, especially the field of zk-SNARKs

Is an area of research, which is advancing at a very fast rate

The R&D is advancing so fast,

That we are getting more secure, faster & cheaper ways of developing ZKPs

Eventually, making it possible for every dApps to use "zk-rollups"

Many of the problems that development of "zk-rollup" on #Ethereum

Results from the notion of the "global state" of EVM

Dealing with the "global state" on an L2 is a nontrivial issue

This makes the development of "zk-rollups" on #Ethereum a challenge

But #Cardano is built differently

It profits from the local state of EUTxO ledger model

This allows for a "local state" instead of a "global state" of "Ethereum"

Which makes #Cardano a perfect L1

For the implementation of an L2 scaling solution like "zk-rollups"

TL;DR -

• There are two approaches to rollups based on ''security models''- "optimistic-" & "zk-rollups"

• "zk-rollups" is an L2 solution, with clear advantages over "optimistic rollups

•When it comes to "composability" of dApps

•And suits the EUTxO model of #Cardano

If you are looking to learn more about zero-knowledge proofs and "zk-rollups",

Here’s a similar thread on "zk-rollups" that you might find useful:

https://twitter.com/orbisproject/status/1537072470246703105?s=20&t=Z-FZEcmPwf8Ew_olAlNDEg