π Execution Flow with FHE Module
To better understand the workflow of the FHE rollup, we highlight the integration of FHE functionalities in the rollup created by Cyferio SDK.
Cyferio SDK Execution Flow and DA overview
1. Transaction Posting
- The sequencer posts a data blob to the Data Availability (DA) layer.
- The transaction is sent to the full node, allowing expensive computations (e.g., FHE operations) to begin early, reducing overall execution time.
2. Data Finalization on DA Layer
- Once the data blob is finalized on the DA layer, it becomes logically final.
- Full nodes wait for the finalization to trigger the next state root update.
3. State Root Update
- After the DA finalizes the block, full nodes update the new rollup state root.
- The block is considered subjectively final once execution is complete.
4. Proof Generation
- Prover nodes in the zkVM generate cryptographic proofs after computation.
- These proofs are posted on-chain to confirm the validity of the executed rollup batch.
5. FHE Computation & Decryption
- FHE computations are performed during this phase.
- If decryption is required, the Threshold Service Network is invoked, which executes the Secure Multi-party Computation (MPC) protocol to decrypt data securely.
6. Finality of Proofs
- Once a proof for a batch is posted on-chain, that batch becomes final for all nodes, including light clients.