Re: [Sks-devel] gossip specification.

[Yaron Minsky]

There's no real specification outside of the code, I'm afraid.  I'd be
happy to help out in getting the required information spec'd out, but
it's worth noting that there are some other difficulties standing in
front of an implementation.  Here are the basic algorithms that you
need to implement to do the efficient set-comparison algorithm that
SKS depends on:

- You need integer-arithmetic-mod-p for p around 2**128.
- You need to build a polynomial arithmetic on top of that Z_p package
- You need to implement polynomial-interpolation and polynomial
factoring algorithms

You also need to build a not-terribly-complicated tree-based data
structure for storing the reconciliation data.  The one complication
is that it's fairly big (around 90M), and so you'd prefer to store it
mostly on disk. The algorithm (without an explicit discussion of the
disk-storage issues) is described in the "Practical Set
Reconciliation" paper.

I'd be happy to help you figure out what would be required for
interoperability, but I want to make sure you're up for doing the rest
of what is necessary.  it's not a small project.

y


On Tue, 28 Sep 2004 22:54:20 +0100, Jonathan McDowell <address@hidden> wrote:
> Is the SKS gossip implementation documented anywhere? I've checked CVS
> and the Wiki, but all I can find are the abstract papers rather than any
> details about how things work for PGP. I'd like to implement gossip in
> onak (another OpenPGP keyserver), but would rather not delve into the
> SKS source to learn how to do so.
> 
> J.
> 
> --
> ] http://www.earth.li/~noodles/ []  "I will not carve gods." - Bart's  [
> ]  PGP/GPG Key @ the.earth.li   []             Blackboard              [
> ] via keyserver, web or email.  []                                     [
> ] RSA: 4DC4E7FD / DSA: 5B430367 []                                     [
> 
> _______________________________________________
> Sks-devel mailing list
> address@hidden
> http://lists.nongnu.org/mailman/listinfo/sks-devel
>