[E-Lang] distributed resource discovery service

[zooko@zooko.com]

[Cc: e-lang, p2p-hackers, mojonation-devel]

 Someone wrote:
>
> There is a TR available at http://theory.lcs.mit.edu/~karger/chord.ps
> It is public; we hope soon to have a more official publication, but
> feel fee to tell other people about the tech report.  It is LCS TR
> 819.


This is very impressive.  I like it.  This is a nice implementation of the
"consistent hashing" approach[1], which I think should be added to Mojo
Nation's distributed meta tracking in the form of a
"ConsistentHashingHandicapper".  (Well, that will take care of the consistent
hashing part, but not of the "distributed" (i.e. transitive) consistent hashing
part.  Nonetheless, one step at a time is the way we get places...) 

FWIW, consistent hashing is I think the same thing as the "Mask Matching
Handicapper" that I proposed on p2p-hackers earlier[2].  It is the case for
distributed Meta-Tracking, as well as for block finding, that this
(*non*-transitive consistent hashing) can be cleanly plugged into Mojo Nation in
the form of a handicapper.

Oh!  Wait a second, we will have transitive as well in the form of "list meta
trackers".  Heh.  Coool.


Regards,

Zooko

[1] "Consistent Hashing and Random Trees: Distributed Caching Protocols for
    Relieving Hot Spots on the World Wide Web"; David Karger, Eric Lehman, Tom
    Leighton, Matthew Levine, Daniel Lewin, Rina Panigrahy; 1997

    http://citeseer.nj.nec.com/karger97consistent.html

[2] "proposal re: scalability of block publication and fetching"; Zooko

    http://zgp.org/pipermail/p2p-hackers/2001-February/000019.html