[freehaven-cvs] finally solve the badnodes scaling issue (it is now ...

[arma@seul.org (Roger Dingledine)]

Update of /home/freehaven/cvsroot/doc/sync-batching
In directory moria.mit.edu:/home2/arma/work/freehaven/doc/sync-batching

Modified Files:
	sync-batching.tex 
Log Message:
finally solve the badnodes scaling issue (it is now bigger and takes
up less space, and we're at 20 pages!)
start cleaning 7.2


Index: sync-batching.tex
===================================================================
RCS file: /home/freehaven/cvsroot/doc/sync-batching/sync-batching.tex,v
retrieving revision 1.44
retrieving revision 1.45
diff -u -d -r1.44 -r1.45
--- sync-batching.tex	26 Jan 2004 15:11:23 -0000	1.44
+++ sync-batching.tex	26 Jan 2004 15:39:40 -0000	1.45
@@ -447,9 +447,13 @@
 \label{sec:graphs}
 
 \begin{figure}[t]
+\unitlength=1in
 \centering
+\begin{picture}(5.0,2.0)
+\put(2.5,.75){\makebox(0,0)[c]{\epsfig{angle=270,figure=badnodes,width=5in}}}
+\end{picture}
 \caption{Entropy vs probability of compromise for each node (16 nodes)}
-\mbox{\epsfig{angle=270,figure=badnodes,width=4in}}
+%\mbox{\epsfig{angle=270,figure=badnodes,width=4in}}
 \label{fig:badnodes}
 \end{figure}
 
@@ -735,22 +739,24 @@
 
 \subsection{Robustness of Message Delivery}
 
-% somewhere, cite the 1990 paper about cascades that can skip nodes
-
-Better anonymity is achieved by longer routes. For example, in a 1x16
-cascade or a 16x16 free-route, the entropy is virtually identical.
-There is almost no falloff in entropy until each node has a ninety
+Better anonymity is achieved by longer routes. For example, if we
+formed our 16 nodes into a 1x16
+cascade or a 16x16 free-route, there is almost no falloff in entropy
+until each node has a ninety
 percent chance of being compromised. But this ignores robustness of
-message delivery.  (Robustness of anonymity is discussed below.) For
-the free-route mixnet with only a single node failure (and for
+message delivery. % (We discuss robustness of anonymity to active attacks
+%below.)
+For the free-route 16x16 mixnet with only a single node failure (and for
 randomly chosen routes through this mixnet) nearly two thirds of
 messages will be undelivered (because they will need to pass through
-it at some point). Such a network is very brittle. For the cascade, it
+it at some point). Such a network is very brittle. For the 1x16 cascade, it
 is even worse: a single node crash blocks all message delivery.
-Parallel cascades can be added to the network, however unlike the
-free-route, this will a priori reduce the entropy of an input message
-for the same size mixnet batch.  This illustrates that robustness can
-be an important question.
+(We might take advantage of schemes to bypass a single failed node
+\cite{pfitzmann85}, but this does not address the real problems.)
+Parallel cascades can be added to the network, but unlike the
+free-route, this will \emph{a priori} reduce the entropy of an input
+message for the same size mixnet batch. Thus robustness can be an
+important question.
 
 
 \begin{table} \caption{Percent of messages delivered}

***********************************************************************
To unsubscribe, send an e-mail to majordomo@seul.org with
unsubscribe freehaven-cvs       in the body. http://freehaven.net/