Low-resource routing attacks against tor

Abstract

Tor has become one of the most popular overlay networks for anonymizing TCP traffic. Its popularity is due in part to its perceived strong anonymity properties and its relatively low latency service. Low latency is achieved through Tor's ability to balance the traffic load by optimizing Tor router selection to probabilistically favor routers with highbandwidth capabilities.We investigate how Tor's routing optimizations impact its ability to provide strong anonymity. Through experiments conducted on PlanetLab, we show the extent to which routing performance optimizations have left the system vulnerable to end-to-end traffic analysis attacks from non-global adversaries with minimal resources. Further, we demonstrate that entry guards, added to mitigate path disruption attacks, are themselves vulnerable to attack. Finally, we explore solutions to improve Tor's current routing algorithms and propose alternative routing strategies that prevent some of the routing attacks used in our experiments.

Metrics for Security and Performance in Low-Latency Anonymity Systems

“…For this reason, it was assumed and subsequently demonstrated [3,4] that by observing both ends of a connection, timing patterns are enough to confirm a suspected link between sender and receiver. This attack has also been shown to work even if only a small proportion (e.g.…”

“…This was initially adequate, but as heterogeneity of node bandwidth capabilities increased, path selection was changed to take into account bandwidth capacity of nodes. The basic algorithm is described by Bauer et al [3], and our updated description is in Section 4.…”

“…These include guard nodes [4] and bandwidth/uptime caps [3]. One further notable paper is by Snader and Borisov [7], in which two proposals are made.…”

“…Bauer et al [3] demonstrated that by artificially inflating bandwidth claims, an attacker could compromise 46% of connections while controlling only 6 out of the 66 nodes on a private Tor network. In this paper we will expand on the results of Bauer et al, simulating this attack while varying path selection algorithm.…”

“…The former reflects a world in which many sites contribute malicious nodes with bandwidth, such as a botnet 2 , and the latter in which a single site is able to use diverse network addresses over a single link of fixed bandwidth, such as in a co-location centre 3 . These represent extremes in strategy, but allow us to evaluate the effectiveness of different path selection algorithms inside, and outside, of their assumptions.…”