## The state of stateless
---
![](https://storage.googleapis.com/ethereum-hackmd/upload_dbd1fa52deb9f970928191b7a7619dac.png)
---
#### What verkle does better than binary with hashes ?
* reduce state/slows down state growth
* only store commitments
* Binary tree: 15 x as many internal nodes, 4x less siblings in proofs
* compatibility with SNARKS
* State expiry
---
#### What verkle does better than hashing? (2)
* smaller proofs
* binary: ~500kb fixed + witness
* verkle:
![](https://storage.googleapis.com/ethereum-hackmd/upload_8a4717bcff7aa49c63b1b907b476eba2.png)
---
#### What binary does better than verkle?
* hashing speed (zk & IRL)
* quantum resistance
* compatibility with STARKs
---
#### What would change with binary tries?
* proof
* crypto
* commitment scheme: polynomial -> hash
* key computation: pedersen hash -> some type of hash
---
#### What won't change?
* gas costs
* single-tree structure
* conversion
* including preimage distribution
* sync (?)
---
#### Binary extension-and-suffix tree
```graphviz
digraph D {
node [label="H(1 || stem || suffix_tree)"]
root -> extension
root -> suffixes
extension [label="stem"]
suffixes [label="suffix_tree"]
suffix0127 [label=""]
suffix128255 [label="..."]
suffix064 [label=""]
suffix64127 [label="..."]
suffix128192 [label="..."]
suffix192255 [label="..."]
suffix032 [label="..."]
suffix3264 [label="..."]
suffix02 [label=""]
suffix24 [label="..."]
val0 [label="value_0"]
val1 [label="value_1"]
suffixes -> suffix0127
suffixes -> suffix128255
suffix0127 -> suffix064
suffix0127 -> suffix64127
suffix128255 -> suffix128192
suffix128255 -> suffix192255
suffix064 -> suffix032
suffix064 -> suffix3264
suffix032 -> suffix02
suffix032 -> suffix24
suffix02 -> val0
suffix02 -> val1
}
```
---
#### Future-proofiness
* STARKS have shown great hashing performance
* security is conjectured
* still not there, performance wise
* SNARKS might catch up, it's unpredictable
* Quantum resistance: "a chance in the 2030s"
---
![](https://storage.googleapis.com/ethereum-hackmd/upload_35b637f187ea1f67a18392f77e08c0aa.png)
---
#### Ajtai's hash
* Very similar to verkle, but quantum resistant
* Homomorphism
* Faster computations
* Con: large commitments (Kb)
---
### TL;DR
* Verkle has properties desirable _now_
* Done > Perfect
* The annoying parts remain no matter what
* More improvements unfolding
* "last-minute swap" fairly easy to do
* Quantum still realistically far off

{}