Source code adapted from dvf/blockchain and Falcon Blockchain. Primarily modified by adding two Trapdoor sampling algorithms, namely MALA for Lattice-Gaussian sampling (MALK) and HMC for lattice-Gaussian sampling (HMCK).
>>> python node.py
Note: Tested on Python 3.9.5. If you have poetry, you can use poetry install to install all necessary dependencies.
You will be prompted for a port number to listen for HTTP requests on, as well as whether to use FALCON (post-quantum signature scheme) or Ed25519 (using the PyNaCl implementation):
>>> Enter your port: [e.g. 5000]
>>> Use FALCON? (y/n): [e.g. y]
Note that if FALCON is chosen, there will be an additional prompt to check if you would like to utilise MCMC sampling with either the Independent Metropolis-Hastings-Klein algorithm (i) or Symmetric Metropolis-Klein algorithm (s), or to use the original FALCON signature scheme (o):
>>> Use independent MHK/symmetric MK/MALK/HMCK/original? (imhk/smk/malk/hmck/o) [e.g. malk]
The first time a transaction is carried out, there will be prompts for several parameters:
>>> Enter degree of n: [This is the ring degree of FALCON, e.g. 512]
>>> Retrieve polys from data/polys.txt? (y/n): ['n' if first time generating private key or if saved polys are in a different file location]
>>> Do you have saved polys? (y/n): ['n' to generate new polys for a new private key. Note: polys will be saved in data/polys.txt for future use]
>>> Enter file name: [Enter file name here if polys are saved in different file location]
The first time a transaction is carried out, there will be prompts for several parameters:
>>> Do you have a salt? (y/n): ['n' if first time generating salt for private key. Note: salt will be saved in /data/salt.txt for future use]
>>> Is it in /data/salt.txt? (y/n/raw) ['n' if salt saved in different file location; 'raw' to type in salt directly]
>>> Enter the name of the file: [Enter file name here if salt is saved in different file location]
>>> Enter your password: [Enter password here to generate unique private key with salt]
Send the following HTTP requests, e.g. using Postman, for the following interactions. Note: the request is preceded by the HTTP address, e.g. http://localhost:5000/mine:
- [GET]
/mine: Mine a new block using the POW scheme, adding all pending transactions to the block. - [GET]
/transactions/get: Get all pending transactions. - [POST]
/transactions/new: Make a new transaction. Required JSON fields:'recipients' <list>,'amounts' <list>. - [GET]
/transactions/verify: Verify a transaction. Required JSON fields:'id','output_index','block_index','amount','signature'. - [GET]
/chain/get: Get entire blockchain on node. - [GET]
/chain/valid: Check if blockchain is valid. - [GET]
/utxo/all: Get entire UTXO set. - [GET]
/utxo/user: Get UTXOs specific to a user, i.e., how many coins a user owns. Required JSON fields:'user'. - [GET]
/utxo/unmined: Get remaining unmined coins. - [POST]
/nodes/register: Register new list of nodes to current node. Required JSON field:'nodes' <list>. - [GET]
/nodes/resolve: Compare blockchain with other nodes to get longest chain (consensus scheme).
Utilising MCMC sampling, based on MALA and HMC in the trapdoor sampler of the post-quantum FALCON signature scheme. Current Python implementation (located in subfolder /falcon_mcmc) adapted from the original FALCON Python source code.
Main changes are made to SecretKey.sample_preimage() and SecretKey.sign() for the calculation of the mixing time and acceptance ratio of the MCMC sampling techniques.
Refer to FALCON README.md for detailed instructions. Summary of basic functions listed below:
>>> import falcon
>>> sk = falcon.SecretKey(128)
>>> pk = falcon.PublicKey(sk)
>>> sig = sk.sign(b"Hello", type_in='hmck', sigma_og=50, sigma_new=10,h=10, k_hmc=8,i_mix_sym=20)
>>> pk.verify(b"Hello", sig)
True
Additionally, note that to utilise MCMC sampling, several additional parameters would have to be passed to SecretKey.sign(). The input (and default) parameters are as follows:
>>> SecretKey.sign(message, type_in='', sigma_og=None, sigma_new=30, h=220, k_hmc=3, i_mix_sym=1000, overwrite=False, randombytes=urandom)
The additional parameters to take note of are:
type_in='': Whether to use IMHK ('imhk'), SMK ('smk'), MALK ('malk'), HMCK ('hmck') or no MCMC sampling, i.e., original FALCON, (default value of'').sigma_og=None: The original sigma to sample with, for IMHK and SMK. Recommended parameters are 65-75 for IMHK, and 60 for SMK (n = 512).sigma_new=30: The subsequent sigma to sample with for the SMK, MALK, HMCK, as part of the two-stage sampling process. Recommended parameter is the default value of 30 for SMK (n = 512) and 2-30 for MALK and HMCK (n = 128).h = 2300: The step size chosen for MALK and HMCK. Optimal values are different for each algorithm and deepends on other parameter values.k_hmc = 3: The number of steps that the Leapfrog integrator in the HMCK algorithm should perform.i_mix_sym=1000: Mixing time for SMK. Recommended parameter is the default value of 1000 (n = 512).
Therefore, to sign with original FALCON:
>>> sig = sk.sign(b"Hello")
To sign with IMHK:
>>> sig = sk.sign(b"Hello", type_in='imhk', sigma_og=75)
To sign with SMK:
>>> sig = sk.sign(b"Hello", type_in='smk', sigma_og=60, sigma_new=30, i_mix_sym=1000)
To sign with MALK:
>>> sig = sk.sign(b"Hello", type_in='malk', sigma_og=60, sigma_new=20, h = 2300, i_mix_sym=10)
To sign with MHCK:
>>> sig = sk.sign(b"Hello", type_in='hmck', sigma_og=50, sigma_new=20, h = 10, k_hmc =8, i_mix_sym=10)
Note: Tested on Python 3.9.5. If you have poetry, you can use poetry install to install all necessary dependencies.