sks-devel
[Top][All Lists]
Advanced

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: [Sks-devel] Optimum number of peers


From: Ari Trachtenberg
Subject: Re: [Sks-devel] Optimum number of peers
Date: Thu, 21 Apr 2011 09:26:26 -0400

Gossip protocols exhibit a thresholding phenomenon.  If everyone talks to 
greater than a certain
fraction of their peers, data will propagate to everyone in the network.  If 
everyone talks to less than
this fraction, then very few network members will get all the data.  
Unfortunately, this fraction depends
on many parameters of the network ...we are working on some research that may 
give some more
concrete answers ... but it will take a bit.

best,
        -Ari

On Apr 21, 2011, at 9:20 AM, Robert J. Hansen wrote:

>> In theory an optimal value could be computed but there are
>> too many parameters (e.g. network delay and bandwidth...
> 
> Additionally, these parameters are in effect constants.  The SKS algorithm 
> will have roughly logarithmic speed in propagating keys through the network: 
> everything else (including how many peers you're connected with) affects it 
> only by a constant factor.
> 
> O(lg N) speed is *fast*.  Like, really, really fast.  So my suggestion: make 
> sure you have at least two peers and you should be golden.  Trying to make it 
> even faster is kind of like trying to put racing stripes and aftermarket body 
> kits on a Saturn V rocket: you can do it if you really want to, but there's 
> not much point.  :)
> 
> 
> _______________________________________________
> Sks-devel mailing list
> address@hidden
> https://lists.nongnu.org/mailman/listinfo/sks-devel

--- 
Ari Trachtenberg, Boston University 
http://people.bu.edu/trachten mailto:address@hidden 
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ 
Random quote of the day: 
"Only a nation that can defend its freedom truly deserves it." - Ehud Olmert. 
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ 




reply via email to

[Prev in Thread] Current Thread [Next in Thread]