- redesigned interfacce
- properly implemented client-side entropy sampling (crypto.getRandomValues)
- dapp functioning correctly
- updated smart contract code
- uploaded private net folder used for testing - extended from Vincent Chu's github project (https://github.com/vincentchu/eth-private-net)
- functionalities all added (and working properly) except for coding-theory-based verification
- for now, tests have been done for up to 6 parties, all with valid results (in various possible scenarios)
- added an implementation of SCRAPE as contract
- heavy computations are in the form of pure / view functions (which makes them effectively client-side, see "Solidity function types" for clarification)
- added a runtime transcript (log)
- added actual-elliptic-curve implementation of PVSS
- uses secp256k1 curve (to match major cryptocurrencies)
- added implementation for SCRAPE protocol
- added reconstructing party's proof in PVSS
- all implementations are for demonstration and do not involve creating dummy entities / simulating internet connections etc.
- added a simple implementation for Reed-Solomon code (for verification phase)
- mainly manipulates the matrices G and H, using SymPy library
- note that SymPy does not come with mpz support or Zp-typed elements, so it should only be used for small numbers
- computations now take place on an elliptic curve group with prime order
- new implementation class : dummyEC. It is used for giving small-sized demonstrations. Need to switch to an ECC library for practical use
- added some comments (function description)
- the protocol now adopts implementations of several algorithms in V.Shoup's book
to generate prime p with factorization of p-1, and find a generator
- PVSS works properly for seemingly arbitrary prime, but only some small pairs of n,t
- alternate usage: PVSS.py <p> <n> <t> <pre-secret>. If p is not prime, it will be replaced by a random p-bit prime.
- using very simple test for identifying generator. Only *usually* correct, but is nonetheless an improvement where there was previously no test
- using Generate-then-test method to find x(i) sets that evaluate all Lagrange terms into a desired pattern. If the generate-then-test yields a result, the PVSS will always work corectly. Otherwise, the search stops due to retries running out, and the protocol is considered terminated
- small-magnitude tests passed. PVSS working properly
- example usage: python PVSS.py <prime> <n> <t> <pre-secret>. Main flow:
- create PVSS object , given the prime
- call generateTestKeyPairs(n), which outputs the list of n (sk,pk) pairs
- call PVSSDistribute(n,t,secret), which outputs n shares of the secret as described by the DDH PVSS
- call decryptShare(share, secretKey) to obtain the decryption of secret share for one party
- call PVSSReconstruct(<list-of-t-parties>,<list-of-t-decrypted-shares>), which returns the reconstructed secret
The secret is written to the file "secret.txt"
- repository created
- initial code uploaded
- unfinished functions: decryptShare, PVSSReconstruct