Agenda:
- BESU follow up
- Fabric Private Chaincode – Bruno, Michael and Marcus
Attendees:
- Mic Bowman
- Hart Montgomery
- James
- Bruno Vavala
- Michael Steiner
- Prakash Moorthy
- Vipin Bharathan
- Binh Nguyen
- Dawid Tomczyk
Fabric Private Chaincode
Extends privacy/confidentiality in Fabric, moves peer host out of the TCB so that confidentiality can be preserved even from the peer hosts.
Fabric integration challenged by the signing processes in the peers... change set is not designed to be attested by unique participants (in this case the enclaves).
Hyperledger Labs project... Intel & IBM primary contributors but open to other contributors. Prototype functional but several things still missing from the implementation.
One example.. Fabric 2.0 will allow for multiple implementations (in different languages even- so definitely not able to produce the same hash) of the same chaincode so long as they generate the same results (change set). FPC will still require that all chaincode be consistent and it will enforce this through the MAC+attestation. Really just needs to extend the Fabric metadata to make it work.
Positioning HL Avalon... clearly the projects are related. Avalon is more about "off chain" (that is, computations managed by Avalon do not currently change the state of data stored on the ledger) while FPC focuses on "on chain" in that the transactions update ledger state.
Q: what about GDPR? A: FPC might provide some tools that would be helpful for addressing the requirements for GDPR (though complete solutions will not be purely technical)- maybe only if they miimic the PCC implementation of segregating private data (right now it seems to send encrypted data to all validators)
Q: what is the performance impact? A: not known, but target is "a few percent"
most of the speedups in fabric have been through sharding or scaling of specific components (see fastfabric,, accelerator, etc.) Integration of FPC into these solutions may be acceptable as performance impact will be on a "speeded up" fabric...