Skip to content

Latest commit

 

History

History
54 lines (37 loc) · 2.45 KB

limitations.md

File metadata and controls

54 lines (37 loc) · 2.45 KB

Limitations

There are currently a number of limitations to the code that users should be aware of:

Security

The code is still undergoing a security audit and therefore security should not be assumed.

ZK-SNARKs

The code makes use of ZK-SNARKS for transaction hiding, this approach is maximally efficient but carries the normal requirement for trusted setup and the knowledge of exponent assumption. We hope to address the former of these soon.

Hash and Key lengths

Most hashes are sha256 based but truncated to 216 bits for convenience in the generation of proofs. Keys are also 216 bit. Whilst we do not believe this has a significant impact on security in itself, we mention it so that you can make up your own mind.

Storage of secrets

Secrets are stored in the clear in your local Mongodb instance. Therefore this database should be appropriately protected.

Persistence of Whisper messages and Events

We have not implemented persistence of Whisper messages (beyond the time to live) and therefore if a user is not logged in they may miss a transaction. If this is a problem, we recommend using a mailserver to offload whisper messages as described in the Whisper documentation. Likewise we do not currently search for old blockchain events.

Use of ZoKrates Docker image

We use a ZoKrates Docker image for proof generation. This is stored separately from the main ZoKrates github so that changes to ZokRates do not break this code. However this means that future improvements and bugfixes that the ZoKrates team implement will not automatically carry through into this code.

No batching or aggregation of proofs

We currently treat each ZKP transaction as a separate event. In the future we expect to aggregate proofs to improve efficiency of verification.

Analysis of transactions

Zero Knowledge transactions are exactly that, an observer learns exactly nothing about the transaction. However, this does not mean that analysis of the blockchain will in all circumstances tell one nothing. For example if I mint 123456789 coins and send them directly to Bob, who then converts them back into conventional ERC-20, it indicates with high probability that I just transacted with Bob because of the disappearance and reappearance of the same unusual value. The situation is worse with non-fungible tokens. Therefore transactions should be made with care, making use of transactions that change fungible token values and not burning non-fungible tokens commitments.