mTor: A multipath Tor routing beyond bandwidth throttling

One of the main obstacles that impede further expansion of Tor, the most popular anonymous communication system, is its large performance variance. The problem becomes worse when bandwidth-intensive applications, such as video streaming, contend with latency-sensitive applications, such as web browsing, for the scarce resources. Most of the existing solutions involve circuit-scheduling techniques to prioritize interactive traffic over bulk traffic or to completely throttle traffic of bandwidth-intensive applications. However, these approaches not only rely on accurate detection of traffic types but also adopt detection strategies that are easy to game. In this paper, we propose a different approach by exploring new capabilities of Tor to support bulk data transfers without degrading the performance of interactive traffic. Based on our observations that a large portion of low-bandwidth relays are under-utilized, we develop a multi-path Tor (mTor) routing algorithm to cater to bandwidth-intensive applications by constructing multiple circuits across low-bandwidth Tor relays. We present a self-adaptive “pulling” scheduling technique to dynamically allocate cells across multiple circuits, and an active congestion detection scheme to prevent slow circuits from becoming a bottleneck of the entire tunnel. Based on the results from experiments on the live Tor network and simulations over the Shadow simulator [1], we conclude that mTor not only achieves a desirable performance for bandwidth-intensive applications by utilizing multiple low-bandwidth relays, but also benefits latency-sensitive applications by reducing the load on high-bandwidth relays.

[1]  Olivier Bonaventure,et al.  MultiPath TCP: From Theory to Practice , 2011, Networking.

[2]  Ian Goldberg,et al.  Enhancing Tor's performance using real-time traffic classification , 2012, CCS.

[3]  Ian Goldberg,et al.  An improved algorithm for tor circuit scheduling , 2010, CCS '10.

[4]  Micah Sherr,et al.  Exploring the potential benefits of expanded rate limiting in Tor: slow and steady wins the race with Tortoise , 2011, ACSAC '11.

[5]  Nicholas Hopper,et al.  Recruiting new tor relays with BRAIDS , 2010, CCS '10.

[6]  Paul F. Syverson,et al.  LIRA: Lightweight Incentivized Routing for Anonymity , 2013, NDSS.

[7]  Nikita Borisov,et al.  A Tune-up for Tor: Improving Security and Performance in the Tor Network , 2008, NDSS.

[8]  Paul F. Syverson,et al.  Locating hidden servers , 2006, 2006 IEEE Symposium on Security and Privacy (S&P'06).

[9]  Tao Wang,et al.  Congestion-Aware Path Selection for Tor , 2012, Financial Cryptography.

[10]  Ieee Staff,et al.  2013 IEEE Conference on Communications and Network Security (CNS) , 2013 .

[11]  Nicholas Hopper,et al.  Shadow: Running Tor in a Box for Accurate and Efficient Experimentation , 2011, NDSS.

[12]  Ian Goldberg,et al.  The Path Less Travelled: Overcoming Tor's Bottlenecks with Traffic Splitting , 2013, Privacy Enhancing Technologies.

[13]  Nicholas Hopper,et al.  Throttling Tor Bandwidth Parasites , 2012, NDSS.

[14]  Ian Goldberg,et al.  DefenestraTor: Throwing Out Windows in Tor , 2011, PETS.

[15]  Nick Mathewson,et al.  Tor: The Second-Generation Onion Router , 2004, USENIX Security Symposium.

[16]  Roger Dingledine,et al.  Building Incentives into Tor , 2010, Financial Cryptography.

[17]  Harsha V. Madhyastha,et al.  LASTor: A Low-Latency AS-Aware Tor Client , 2012, IEEE/ACM Transactions on Networking.

[18]  George Danezis,et al.  Low-cost traffic analysis of Tor , 2005, 2005 IEEE Symposium on Security and Privacy (S&P'05).

[19]  Micah Sherr,et al.  Users get routed: traffic correlation on tor by realistic adversaries , 2013, CCS.

[20]  Dirk Grunwald,et al.  Shining Light in Dark Places: Understanding the Tor Network , 2008, Privacy Enhancing Technologies.

[21]  Micah Adler,et al.  An Analysis of the Degradation of Anonymous Protocols , 2002, NDSS.