Files in this item

FilesDescriptionFormat

application/pdf

application/pdfDARIR-THESIS-2019.pdf (2MB)
(no description provided)PDF

Description

Title:Privacy-preserving network congestion control
Author(s):Darir, Hussein
Advisor(s):Dullerud, Geir
Department / Program:Mechanical Sci & Engineering
Discipline:Mechanical Engineering
Degree Granting Institution:University of Illinois at Urbana-Champaign
Degree:M.S.
Genre:Thesis
Subject(s):Cyber-physical
Distributed cyber-physical networks
networking
communication
Internet
Privacy
Anonymity
Tor network
Congestion
load-balancing
Path selection
differential-privacy
Abstract:Cyber-physical technology is applied in various domains and connect computation with physical processes. Distributed cyber-physical networks have had a major impact in the area of networking and communications, more specifically on the Internet. The Internet nowadays is an important tool of communication, however one of the main challenges facing users on the Internet is maintaining privacy, especially when facing widespread surveillance. Anonymity networks have emerged as a solution to this problem by allowing users to conceal their identities online. The most successful anonymous communications network to date is currently the Tor network. It is operated by volunteers around the world and has many users worldwide. However, the trade off between performance and anonymity has always been a major problem for this type of networks. Users' traffic in Tor is routed across a series of servers; each user's path going through the network transits three of them. This process of path selection creates a load balancing problem that could lead to network congestion. Congestion may deteriorate the network performance and service quality resulting in queuing delay, data packet loss and the blocking of new connections. In this work, we study the problem of load-balancing in path selection in anonymous networks such as Tor. We first find that the current Tor path selection strategy can create significant imbalances. We then develop a (locally) optimal algorithm for selecting paths and show, using flow-level simulation, that it results in much better balancing of load across the network. Our initial algorithm uses the complete state of the network, which is impractical in a distributed setting and can compromise users' privacy. We therefore develop a revised algorithm that relies on a periodic and differentially private summary of the network state to approximate the optimal assignment. Our simulations show that the revised algorithm significantly outperforms the current strategy while maintaining provable privacy guarantees.
Issue Date:2019-07-03
Type:Text
URI:http://hdl.handle.net/2142/105636
Rights Information:Copyright 2019 Hussein Darir
Date Available in IDEALS:2019-11-26
Date Deposited:2019-08


This item appears in the following Collection(s)

Item Statistics