-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1

Ian Clarke wrote:
> But then the claims discuss using this technique to retrieve things over
> a network.  Now, one might argue that simply applying a common computer
> science technique to a distributed situation is not novel (I don't
> believe you can get a valid patent simply by combining two other things
> you didn't invent), but it would be really useful to find some robust
> examples of requesting files by their hashes over a network that
> pre-date October 1997.

The PRR paper that's the basis of Pastry and Tapestry might cover this.

C. Plaxton, R. Rajaram, and A. Richa. Accessing nearby copies of
replicated objects in a distributed environment. In Proceedings of the
9th Annual ACM Symposium on Parallel Algorithms and Architectures, pages
311-320, June 1997.

http://cs.brown.edu/courses/cs296-2/papers/plaxton.ps

"Accessing shared objects. Consider a set _A_ of m objects being shared
by a network G, where several copies of each object may exist." (page 1)

"Each object A has a unique (log m)-bit identification." (page 3)

"Most of the earlier work in this area, e.g., emulations of PRAM on
completely-connected distributed-memory machines (e.g., [9, 16])
or bounded-degree networks (e.g., [14]), and algorithms for providing
concurrent access to a set of shared objects [12], assume that each of
the nodes of the network has knowledge of a hash function that indicates
the location of any copy of any object." (page 3)

The last quote seems to point to even older prior art, but I haven't
followed up the references.

Cheers,
Michael
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.4.5 (GNU/Linux)

iD8DBQFE8YRQyua14OQlJ3sRArxvAKDFii6PTuHZ49lnVTwP5j6xSjT5cQCeN0WI
8/1JldDNNp+yU5a47Jpjs3A=
=E5jK
-----END PGP SIGNATURE-----

Reply via email to