[Author Prev][Author Next][Thread Prev][Thread Next][Author Index][Thread Index]
[freehaven-cvs] crunch it back down to 20 pages
Update of /home/freehaven/cvsroot/doc/sync-batching
In directory moria.mit.edu:/home2/arma/work/freehaven/doc/sync-batching
Modified Files:
badnodes.eps badnodes.pdf model.tex sync-batching.pdf
sync-batching.tex
Log Message:
crunch it back down to 20 pages
also commit an updated badnodes.eps
Index: badnodes.eps
===================================================================
RCS file: /home/freehaven/cvsroot/doc/sync-batching/badnodes.eps,v
retrieving revision 1.4
retrieving revision 1.5
diff -u -d -r1.4 -r1.5
--- badnodes.eps 26 Jan 2004 03:55:52 -0000 1.4
+++ badnodes.eps 25 Jun 2004 00:33:08 -0000 1.5
@@ -1,651 +1,648 @@
-%!PS-Adobe-3.0
-%%Title: entropy.xls
-%%Creator: PScript5.dll Version 5.2
-%%CreationDate: 1/24/2004 1:14:17
-%%For: shmat
-%%BoundingBox: (atend)
-%%Pages: (atend)
-%%Orientation: Landscape
-%%PageOrder: Special
-%%DocumentNeededResources: (atend)
-%%DocumentSuppliedResources: (atend)
[...1266 lines suppressed...]
+%%Trailer
+%%BoundingBox: 12 12 599 779
+%%DocumentNeededResources:
+%%+ font Helvetica-Bold
+%%+ font Helvetica
+%%DocumentSuppliedResources:
+%%+ procset Pscript_WinNT_ErrorHandler 5.0 0
+%%+ procset Pscript_FatalError 5.0 0
+%%+ procset Pscript_Win_Basic 5.0 0
+%%+ procset Pscript_Win_Utils_L2 5.0 0
+%%+ procset Pscript_Win_GdiObject 5.0 0
+%%+ procset Pscript_Win_GdiObject_L2 5.0 0
+%%+ procset Pscript_Text 5.0 0
+%%+ procset Pscript_Encoding256 5.0 0
+%%+ procset Pscript_Win_Euro_L2 5.0 0
+Pscript_WinNT_Incr dup /terminate get exec
+ehsave restore
+%%Pages: 1
+(%%[LastPage]%%) =
+%%EOF
Index: badnodes.pdf
===================================================================
RCS file: /home/freehaven/cvsroot/doc/sync-batching/badnodes.pdf,v
retrieving revision 1.4
retrieving revision 1.5
diff -u -d -r1.4 -r1.5
Binary files /tmp/cvs9nDfSY and /tmp/cvsyfcVIN differ
Index: model.tex
===================================================================
RCS file: /home/freehaven/cvsroot/doc/sync-batching/model.tex,v
retrieving revision 1.11
retrieving revision 1.12
diff -u -d -r1.11 -r1.12
--- model.tex 3 May 2004 05:42:33 -0000 1.11
+++ model.tex 25 Jun 2004 00:33:08 -0000 1.12
@@ -139,10 +139,9 @@
is how the targeted message moves through the network, resulting in
different mixing distributions $\PP_1\ldots\PP_N$.
-\subsubsection{Assumptions.}
-
-We assume that the adversary observes the edge of the network and thus
-knows the first mix chosen by the targeted message. This means that the
+%\subsubsection{Assumptions.}
+We assume the adversary observes the edge of the network and thus
+knows the first mix chosen by the targeted message---so the
randomness of mix selection is ignored for the first hop. Formally,
we make probability $\PP_i$ conditional on selection of a particular
first mix. Instead of computing
Index: sync-batching.pdf
===================================================================
RCS file: /home/freehaven/cvsroot/doc/sync-batching/sync-batching.pdf,v
retrieving revision 1.21
retrieving revision 1.22
diff -u -d -r1.21 -r1.22
Binary files /tmp/cvsXoEy20 and /tmp/cvsqTBvYU differ
Index: sync-batching.tex
===================================================================
RCS file: /home/freehaven/cvsroot/doc/sync-batching/sync-batching.tex,v
retrieving revision 1.57
retrieving revision 1.58
diff -u -d -r1.57 -r1.58
--- sync-batching.tex 25 Jun 2004 00:06:31 -0000 1.57
+++ sync-batching.tex 25 Jun 2004 00:33:08 -0000 1.58
@@ -198,7 +198,7 @@
\section{Related work}
\label{sec:related}
-\subsection{Sync-batching timing model and protocol}
+\subsection{Synchronous batching timing model and protocol}
Dingledine et al.~present in~\cite{mix-acc} a mix network that uses
synchronous batching. We refer to that paper for a detailed discussion
@@ -363,7 +363,7 @@
its input-output relationships, while the mix network as a whole still
protects linkability with high probability.
-Clearly much work has been done to address blending attacks; each
+Clearly much work has been done to address blending attacks. Each
topology seems to have some plausible partial solutions.
%solutions apply to synchronous free-route networks
%as well as cascade networks.
@@ -378,7 +378,7 @@
Section~\ref{subsec:random-adversary} for more discussion on this point).
Along with being able to control some of the nodes, our adversary can
observe messages from senders to the mixnet and from the mixnet to
-receivers, but for our initial analysis we assume he cannot observe the
+receivers, but our initial analysis assumes he cannot observe the
links between honest nodes in
the mixnet (in Section~\ref{subsec:average-actual} we argue that
with high probability, observing these links will not yield much
@@ -407,21 +407,20 @@
%path free route.)
%We use the same number for all mixnets to
%have a reasonably fair comparison of available resources.
-Besides being a tractable size for analysis, 16 nodes is also a fairly
-good approximation of actually deployed mixnets. (Mixminion currently
+Besides being a tractable size for analysis, 16 nodes also
+approximates deployed mixnets. (Mixminion currently
has between 20 and 30 active nodes.) One might argue that a 4 node mixnet
-provides better security, because all the messages are mixed together
+gives better security, because all messages are mixed together
in any topology. We assume a larger network is needed because 1) the
bandwidth of a single node may be insufficient to handle all the traffic;
2) a single path may not include as many choices for jurisdiction as
some users want; and 3) a single path is not very robust, either to
network attacks or to nature.
-Messages proceed through the network in \emph{layers}; all of the
+Messages proceed through the network in \emph{layers}; all the
nodes in a layer process messages of one mixnet batch at the same time.
In general we describe networks as $w$x$\ell$, where $w$ is the number
of nodes at each layer and $\ell$ is the number of nodes in a path.
-
We consider three basic topologies: a 4x4 cascade mixnet in which all
messages pass through four cascades of length four; a 4x4
\emph{stratified} mixnet, in which all messages pass through four
@@ -429,10 +428,10 @@
one layer to any node at the next layer; and a 16x4 free-route mixnet,
in which all nodes may receive messages at all layers. Note that
because free-route nodes are reused, `16x4' does not mean 64 nodes.
-
-Examples of a 2x2 cascade mixnet, a 2x2 stratified mixnet, and a 4x2
-free-route mixnet are illustrated in Figures~\ref{fig:casc-2x2},
-\ref{fig:sa-2x2}, and \ref{fig:free-4x2}, respectively.
+Examples of the three topologies are illustrated below.
+%Examples of a 2x2 cascade mixnet, a 2x2 stratified mixnet, and a 4x2
+%free-route mixnet are illustrated in Figures~\ref{fig:casc-2x2},
+%\ref{fig:sa-2x2}, and \ref{fig:free-4x2}, respectively.
\begin{figure}
\begin{minipage}[h]{3.5cm}
@@ -466,8 +465,8 @@
\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}}
+\caption{Entropy vs probability of compromise for each node (16 nodes)}
\label{fig:badnodes}
\end{figure}
@@ -706,8 +705,9 @@
time and is active at each layer. The cascade and stratified networks
require a larger capacity from each node: they must handle 32
messages at once (128/$w$), but they are idle for all but one hop in the
-batch. One could imagine a \emph{systolic network} where $t_\mathrm{batch}
-= t_\mathrm{hop}$ and 32 messages are let in every 45 minutes---in this
+batch. One could imagine a \emph{systolic} or \emph{pipelined} network
+where $t_\mathrm{batch} = t_\mathrm{hop}$ and 32 messages are let in
+every 45 minutes. In this
case the capacity of nodes in cascade and stratified networks would also
be 8, and indeed the latency could be cut to between 3 hours and 3 hours
45 minutes---but the expected entropy would be cut by a factor of $\ell$.
***********************************************************************
To unsubscribe, send an e-mail to majordomo@seul.org with
unsubscribe freehaven-cvs in the body. http://freehaven.net/