Since its inception, the concept of
network coordinates has been proposed to solve a wide variety of
problems such as overlay optimization, network routing, network
localization, and network modeling. However, two practical problems
significantly limit the applications of network coordinates today.
First, how can network coordinates be stabilized without losing
accuracy so that they can be cached by applications? Second, how can
network coordinates be secured such that legitimate nodes'
coordinates are not impacted by misbehaving nodes? Although these
problems have been discussed extensively, solving them in
decentralized network coordinates systems remains an open
problem.
This research investigates new distributed algorithms
to solve the coordinates stability and security problems. For the
stability problem, we propose an error elimination model that can
achieve stability without hurting accuracy. A novel algorithm based
on this model is presented. For the security problem, we show that
recently proposed statistical detection mechanisms cannot achieve an
acceptable level of security against even simple attacks. We propose
to address the security problem in two parts. First, we show how the
computation of coordinates can be protected by a customized Byzantine
fault detection algorithm. Second, we adopt a triangle inequality
violation detection algorithm to protect delay measurements. These
algorithms can be integrated together to provide stable and secure
network coordinates.
Faculty: T. S. Eugene Ng,
Students: Guohui Wang
Stabilized Vivaldi algorithm:
you can find the code in /gnp/SVivaldiComp.cc under our configurable network coordinates system source code directory.
Video Demos: Coordinates movement traces for Vivaldi VS Stabilized Vivaldi systems.
On the Design Principles of Network Coordinates Systems; Guohui Wang; Master's Thesis, Rice University, April 2008
Distributed Algorithms for Stable and Secure Network Coordinates, Guohui Wang, T. S. Eugene Ng; Proceedings of ACM SIGCOMM/USENIX Internet Measurement Conference (IMC), Vouliagmeni, Greece, October 2008
This research is sponsored by the NSF under CAREER Award CNS-0448546.