[freehaven-cvs] revision of slides based on talking to nick

[arma@seul.org (Roger Dingledine)]

Update of /home/freehaven/cvsroot/doc/econp2p03
In directory moria.mit.edu:/home/arma/work/freehaven/doc/econp2p03

Modified Files:
	slides-econp2p03.mgp 
Log Message:
revision of slides based on talking to nick


Index: slides-econp2p03.mgp
===================================================================
RCS file: /home/freehaven/cvsroot/doc/econp2p03/slides-econp2p03.mgp,v
retrieving revision 1.1
retrieving revision 1.2
diff -u -d -r1.1 -r1.2
--- slides-econp2p03.mgp	3 Jun 2003 07:52:37 -0000	1.1
+++ slides-econp2p03.mgp	3 Jun 2003 21:44:22 -0000	1.2
@@ -68,6 +68,19 @@
 %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
 %page
 
+What do I mean by anonymity?
+
+We want to prevent an adversary from linking sender to receiver
+
+Even an adversary who can watch a lot of the network, etc
+
+Generally done by using multiple nodes for a transaction
+
+Not steganography, not just leaving your name off
+
+%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
+%page
+
 Many people need anonymity
 
 %leftfill
@@ -127,7 +140,7 @@
 %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
 %page
 
-But trust bottlenecks can break everything
+Beware of trust bottlenecks
 
 Nodes with more traffic must be more trusted
 
@@ -160,8 +173,8 @@
 An Economics of Anonymity
 
 	Systems need cover traffic (many low-sensitivity users) to attract the high-sensitivity users
-	Most users do not want (know they want) anonymity
-	Weak security (small mix batch, no-delay proxy) can mean more users
+	Most users do not value anonymity much
+	Weak security (fast system) can mean more users
 		which can mean 
 %cont, font "italic"
 stronger 
@@ -175,48 +188,48 @@
 %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
 %page
 
-Facets of Reputation
+Reliability is critical for anonymity systems
 
-Reputation as a signal:
-Tool to predict the future based on past behavior
+	Traditional accountability (eg contract) doesn't work
 
-Reputation as a sanction:
-Tool to change the future by giving people incentive to behave well
+	Since we don't know full network state, transactions tend to be unreliable
 
-Tool for risk management
+	With many nodes, each node won't interact with everybody often
+      Free riding, abuse, anonymity attacks
+
+	Maybe reputation can help...
 
 %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
 %page
 
-Reliability is critical for anonymity systems
-
-	Traditional accountability (eg contract) doesn't work
+Facets of Reputation
 
-	Since we don't know full network state, transactions tend to be unreliable
+Reputation as a signal:
+Tool to predict the future based on past behavior
 
-	With many nodes, each node won't interact with everybody often
-      Free riding, abuse, anonymity attacks
+Reputation as a sanction:
+Tool to change the future by giving people incentive to behave well
 
-	Shared reputation can maybe help reliability?
+Tool for risk management
 
 %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
 %page
 
-Reputation is problematic for anonymity systems
+Reputation is problematic in anonymity systems
 
-Attacker or freeloader can cheaply throw away bad-reputation nyms.
+Attacker or freeloader can cheaply throw away bad-reputation nodes.
 
 Hard to detect/verify a node's behavior while maintaining anonymity.
 We had to redesign p2p systems to support this!
 
-Reputation information can be exploited to subvert anonymity.
+Reputation information can be exploited to subvert anonymity...
 
 %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
 %page
 
 Reputation can be exploited
 
-	A node can't measure all nodes. If he measures only some, he gives away who he might use.
+	A node can't measure all nodes. If he measures only some, he gives away which ones he might use.
 
 	A central reputation server can give different info to different people. Must replicate and coordinate?
 
@@ -276,7 +289,7 @@
 More open topics
 
 	Are dynamic p2p systems that need reliability and don't allow verifying claims doomed? Altruism, other factors?
-	Can we model reputation as currency? May allow easier decentralization.
+	Can we implement reputation with currency? May allow easier decentralization.
 	Incentives: If anonymity for all requires each user doing similar things, how do we deal with users who don't want as much anonymity?
-	Do we have to abandon statistical rigor in the face of this uncertainty?
+	Do we have to abandon statistical rigor in the face of dynamic systems and adversaries?
 

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