Related work:

• https://www.notion.so/Tavloid-towards-Simple-Verifiable-Spreadsheets-and-Databases-8696bd318d424325a76123277eaf0397
• https://hackmd.io/@liamzebedee/BydanTDSi

Goliath seeks to create the first decentralized database cryptonetwork.

The core innovation is using recursive ZK proofs to achieve horizontal scalability.

Currently the main problem is choosing an efficient representation for the database, which can be proven recursively.

There are three problems really:

1. Implementing a basic database index structure.
   1. Process database writes, updating a set of indexes.
   2. Ideas: Vector commitments scheme.
      1. Kate polynomial commitments [1] [2].
2. Proving this database efficiently on-chain, so we can build an economic model.
   1. Nodes prove database work, in exchange they receive fees.
   2. Ideas: StarkNet contract which verifies proofs
3. Scaling the communication and computational costs of the database network.
   1. Recursive proofing is O(log N) communication (aggregating proofs) and O(log N) computational efficiency.
   2. Proving database work is ??
   3. Ideas: recursive proofing system

Idea:

• We could implement the system in Cairo, which is the most feature complete environment.
• We process maybe one tx per proof.
• And then we send the proofs around the network, verifying using Giza.
• This is still O(N^2) cost.
• There’s another party which aggregates proofs.
  • We could proof recursively using this.