Simple hashcash implementation

barnesc at engr.orst.edu barnesc at engr.orst.edu
Tue Aug 10 03:46:21 EDT 2004


Hi!

Here's a simple hashcash implementation in Python.

28 lines of actual code.

Can be reduced to 17 lines for instructional purposes,
if you don't want clustering, and use xrange() instead
of irange().

I tried to make this as simple and elegant as possible.
Feel free to repost this code, if you can make the it
more simple.

I have made no attempt to optimize this code.

# ----------------------------------------------
# hashcash.py: Hashcash implementation
# ----------------------------------------------

"""
Hashcash is a "proof of work."

Example:

>>> import sha
>>> sha.new('denmark2890CF').hexdigest()
'000000cf89643370c24e413ec0886ab92bd7f6e8'

Here we have a 24-bit partial SHA collision against the zero string.

This proves to us that someone took the prefix 'denmark', and
tried about 2**24 different suffixes.  Thus we know that
someone has burnt around 2**24 CPU cycles on the prefix string
'denmark'.  Usually, 'denmark' will be a unique challenge string,
so old hashcash cannot be recycled.

For speed, this library takes the hash of the string 'denmark' before
doing the collision with the zero string.  Otherwise, it is identical
to the above example.

Library examples:

>>> import hashcash
>>> hashcash.make_token('Denmark', 22)
'59538D'
>>> hashcash.verify_token('Denmark', '59538D')
22
>>> t = hashcash.make_cluster('Denmark', 18)
>>> t
'BC48-D5A1-F8C2-27F0-9CC0-DD31-2F04-2052-835-FFF1-E319-0E91-A9D0-D359
-E996-70BA'
>>> hashcash.verify_cluster('Denmark', t)
18

Note that make_token() takes wildly varying amounts of CPU time.
The make_cluster() function concatenates 16 hashcash tokens to even
out the amount of CPU time spent.

"""

import sha, math, itertools

def trailing_zeros(n):
  """Number of trailing 0s in binary representation of integer n."""
  if n <= 0: return 0
  for i in itertools.count(0):
    if n & (1<<i): return i

def irange(n):
  """Implementation of xrange(n) that does not overflow."""
  i = 0
  while i < n:
    yield i; i += 1

def all_strings(charset='0123456789ABCDEF'):
  """Yields all strings in given character set, sorted by length."""
  m = len(charset)
  for n in itertools.count(0):
    for i in irange(m**n):
      yield ''.join([charset[(i//(m**j))%m] for j in xrange(n)])

def hash(s):
  """Hash function used by hashcash.  Returns an integer."""
  return int(sha.new(s).hexdigest(), 16)

def make_token(s, n, charset='0123456789ABCDEF'):
  """Makes hashcash token of value 'n' against basestring 's'."""
  h = sha.new(s).digest()
  for token in all_strings(charset):
    if trailing_zeros(hash(h + token)) >= n: return token

def verify_token(s, token):
  """Returns hashcash value of given token against basestring 's'."""
  return trailing_zeros(hash(sha.new(s).digest() + token))

def make_cluster(s, n, charset='0123456789ABCDEF'):
  """Makes hashcash cluster of value 'n' against basestring 's'."""
  return '-'.join([make_token(s+str(i),n-4,charset) for i in range(16)])

def verify_cluster(s, token):
  """Hashcash value of the given cluster against basestring 's'."""
  T = token.split('-')
  return min([verify_token(s+str(i), T[i]) for i in range(len(T))])+\
         int(math.log(len(T)) / math.log(2.0))


This document is in public domain (as are all of my past Usenet postings).

 - Connelly



More information about the Python-list mailing list