On Wed, Oct 10, 2001 at 11:07:51AM +0200, Oskar Sandberg wrote:
> On Wed, Oct 10, 2001 at 12:19:46AM -0500, thelema wrote:
> > On Wed, 10 Oct 2001, Oskar Sandberg wrote:
> < > 
> > > It hardly matters whether retrieval success is 90% or 99%, if you are
> > > trying retrieve 100 parts without redundancy you are still fucked (2e-5%
> > > or 36% success). In a system where retrieval cannot be guaranteed,
> > > redundancy is necessary when splitting so as to offset the normal
> > > exponential decay when a file is split into many parts. It is ridiculous
> > > that we should sacrifice aspects of the network to achieve 1 in 1000
> > > failure rather than something acceptable like 1 in 50 or 100, so that
> > > files split into 100 parts "only" should fail 1 time in 10. 
> > > 
> > I'm just suprised that everyone now seems to agree that we need
> > redundancy when before everyone seemed to be saying "hell no, keep that
> > redundancy away", and I had to compromise with a system that allowed
> > both redundant and non-redundant usage.  What about the argument that
> > goes sorta like "when you insert more data into freenet that'll cause
> > the pieces of other files to fall out faster"?  Or the arguments about
> > piece popularity being not as high when there's redundant pieces, and
> > that causing pieces to fall out faster.
> 
> I don't believe there has been any significant questioning of redundant
> splitting since more than a year ago. The hit against network
> performance because there is more data cannot compare with the
> exponential numbers if you split without redundancy - and if Freenet is
> bordering on holding more data than it can carry it is in trouble
> anyways. The second is true, which is why we need clients to behave well
> and request different patterns randomly.
But what clients will do, most likely, is request all pieces and use whatever
comes first, improving latency.
> 
> > I'd like to clarify my position, which probably seems to be against
> > redundancy if you just read the above, but I really think that we should
> > try non-redundant splitfiles and if they really can't be requested
> > successfully, we should try adding a *little* bit of redundancy and see
> > how that improves things, and not to go overboard with the redundancy.
> 
> I promise that you success rate is not even 90% ATM. Do the math
> yourself.
> 
> > 
> > Thelema
> > -- 
> > E-mail: thelema314 at bigfoot.com        If you love something, set it free.
> > GPG 1536g/B9C5D1F7 fpr:075A A3F7 F70B 1397 345D  A67E 70AA 820B A806 F95D
> 
> 
> 
> -- 
> Though here at journey's end I lie 
>   In darkness buried deep,          above all shadows rides the Sun
> beyond all towers strong and high,    and the Stars forever dwell:
>   beyond all mountains steep,       I will not say the Day is done,
>                                       nor bid the Stars farewell.
> (JRRT)
> 
> Oskar Sandberg
> oskar at freenetproject.org
> 
> _______________________________________________
> Devl mailing list
> Devl at freenetproject.org
> http://lists.freenetproject.org/mailman/listinfo/devl

-- 
The road to Tycho is paved with good intentions

_______________________________________________
Devl mailing list
Devl at freenetproject.org
http://lists.freenetproject.org/mailman/listinfo/devl

Reply via email to