[mnet-devel] FEC, blocksizes, chunking

[Hauke Johannknecht]

On Wed, 16 Apr 2003, Zooko wrote:

just for the record ...

> > Any reason not to just use blockid[byte-range slice]? Is there any
> > reason for a blocklet to be an independently addressable blob of data?

> Note that we aren't implementing any sub-block-transfer
> thingies at this point.

all my talk about this has been purely hypothetical, based on exp
with other p2p systems, most notably watching the mldonkey people
try to get a reliable resume for blocks (chunks) ...

> I suspect that Tschechow specified blocklets in order to get cryptographic
> authentication on each transferred unit.

yes and no.
i think we need authentication for the smallest transferred unit.
at the moment, these are called blocks, and are authenticated by
their sha1 identifier.

if we need to get sub-block transfers at some point, basic auth
methods could be

a) download keyed by blocklets and a id-list
b) download keyed by block-offset and a hash-tree

i have mostly talked about (a) because it would have the
advantage of using known-to-work technology we allready have
(blockstores) instead of implementing yet another partial
download mechanism with its own resume-, auth- and
tempfile-management just for the blockfetching stage.

but, these are mostly implementational details, and we will cross
that bridge when we come to it. ;)

Pythonizing,
  Hauke

-- 
Hauke Johannknecht        Berlin / Germany        HJ422-RIPE
Use PGP ! -> lynx -dump http://www.ash.de/ash.asc | pgp -kaf



-------------------------------------------------------
This sf.net email is sponsored by:ThinkGeek
Welcome to geek heaven.
http://thinkgeek.com/sf
_______________________________________________
mnet-devel mailing list
mnet-devel at lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/mnet-devel