[Gzz] None

[hemppah]

Tuomas, please ignore the previous message as I sent it too early (by accident).


Quoting Tuomas Lukka <address@hidden>:

> > Ok, then we should split this storage+p2p into two seperate
> presentations:
> > For vegai, Fenfire's storage and for me, p2p related things. Is this ok, or
> are
> > there other suggestions ?
>
> Not sure yet, before I see actual proposed abstracts for both.

Abstract proposal for p2p related stuff:

1. P2P and Fenfire
 -what is P2P  
 -P2P and Fenfire
   -why (location transparency, ad hoc, no central control etc.)
   -how (built on top of Storm, requires that Storm presentation is before this)
 -objectives
   -find data from global data repository --> blocks are building blocks of
    xu-document (enflilade)
   -xu-links, xu-transclusions
   -scalable (billions of users)
   -efficient
   -robust
 -explain pointers and block IDs (even if explained in Storm presentation)
   -block ID: globally unique identifier (SHA-1)
   -urn-5: consept to data mapping, several urn-5s can point to a single block,
   perhaps the newspaper example
 -urn-5 and block IDs: 
  -semantic-free 
  -location independent 
  -next generation Reference Resolution Services (DNS) ?
  -ref to recent publication by Shenker et al., IRIS-project
 -current p2p schemes
  -quick overview
  -loosely and tightly structured systems
  -currently, what is the best scheme for Fenfire (and why) 
 
2. Open problems
 -security issues
 -not only the problems of Fenfire

3. Future work
 -transclusions and xu-links from p2p
 -prototype

4. Published/sumbitted papers regarding this topic
 -HT'02 paper by Tuomas and Benja (published)
 -HT'03 paper by Tuomas, Benja, Toni and hemppah (submitted)
 -hemppah's MSc thesis


Again, comments ?


-Hermanni 

-------------------------------------------------
This mail sent through IMP: http://horde.org/imp/