SNSF ISJRP: Trust-based search in P2P Networks - ETH - Chair of Systems Design

Topology Adaptation for Trust-based search in Peer-to-Peer Networks

This project is related to our research line: Design and analysis of socio-technical systems.

Duration: 7 months (October 2011 - May 2012)

Funding source: Indo Swiss Joint Research Programme

Research partners: Chair of Systems Design of ETH Zurich (host group), Professor N. Ganguly, IIT Kharagpur, India.

Peer-to-Peer (P2P) networks have enabled easy content sharing over the web through the users personal computers and hence have become hugely popular. However, side by side, content download from unauthentic sources has led to a surge in fake content and malware that are disseminated through these files. This motivates the need to design sophisticated search mechanisms in P2P networks that will not only guarantee low search latency and low bandwidth consumption but also ensure trustworthy sources, i.e. that are known to provide authentic files.

One way of designing such technique is to suitably organise the topology of the overlay network which interconnects peers. In this line, we aim at developing a suitable topology adaptation mechanism in P2P networks so that peers providing authentic content of a particular category such as music, movies, ebooks etc. can be grouped together to form a community, i.e. highly trusted peers in a community will be densely connected. This implies that the search queries can be disseminated quickly to these highly trusted peers.

Although this broad picture looks simple, to achieve these objectives, we need to address several issues. P2P networks are subject to heavy churn with peers continuously leaving and joining the systems, thus making the overlay unstable. Content is constantly updated by the users, making the discovery of content itself a challenging problem. However, the most important component needed to ensure the success of such algorithms is the choice of suitable trust ranking metrics that a node can use to rate the trust of its neighbours.

At the Chair of Systems Design, which is the host group of this project, we have been actively engaged in developing distributed trust management mechanisms for trust-based networks. In particular, we proposed an elegant trust ranking mechanism for web-based recommender systems. The principal task of this project is to customize the trust metric in such a way that it can be used in dynamic P2P networks. The research group at IIT, Kharagpur is actively involved in P2P research that involves enhancing the performance of P2P systems in terms of search latency, network coverage and traffic reduction. The results of this research have been published in top-tier conferences and journals.

Selected Publications

A tunable mechanism for identifying trusted nodes in large scale distributed networks				[2012]
Chandra, Joydeep; Scholtes, Ingo; Ganguly, Niloy; Schweitzer, Frank
Proceedings of 11th IEEE International Conference on Trust, Security and Privacy in Computing and Communications (TrustCom 2012)				more» «less

DOI:

10.1109/TrustCom.2012.63

Abstract

In this paper, we propose a simple randomized protocol for identifying trusted nodes based on personalized trust in large scale distributed networks. The problem of identifying trusted nodes, based on personalized trust, in a large network setting stems from the huge computation and message overhead involved in exhaustively calculating and propagating the trust estimates by the remote nodes. However, in any practical scenario, nodes generally communicate with a small subset of nodes and thus exhaustively estimating the trust of all the nodes can lead to huge resource consumption. In contrast, our mechanism can be tuned to locate a desired subset of trusted nodes, based on the allowable overhead, with respect to a particular user. The mechanism is based on a simple exchange of random walk messages and nodes counting the number of times they are being hit by random walkers of nodes in their neighborhood. Simulation results to analyze the effectiveness of the algorithm show that using the proposed algorithm, nodes identify the top trusted nodes in the network with a very high probability by exploring only around 45% of the total nodes, and in turn generates nearly 90% less overhead as compared to an exhaustive trust estimation mechanism, named TrustWebRank. Finally, we provide a measure of the global trustworthiness of a node; simulation results indicate that the measures generated using our mechanism differ by only around 0.6% as compared to TrustWebRank.

arxiv

|link

|bibtex

|view

|download

@inproceedings{Scholtes2012,
    author = "Chandra, Joydeep and Scholtes, Ingo and Ganguly, Niloy and Schweitzer, Frank",
    publisher = "IEEE Computer Society Washington, DC, USA",
    doi = "10.1109/TrustCom.2012.63",
    title = "A tunable mechanism for identifying trusted nodes in large scale distributed networks",
    url = "http://dl.acm.org/citation.cfm?id=2360332",
    abstract = "In this paper, we propose a simple randomized
protocol for identifying trusted nodes based on personalized
trust in large scale distributed networks. The problem of
identifying trusted nodes, based on personalized trust, in a
large network setting stems from the huge computation and
message overhead involved in exhaustively calculating and
propagating the trust estimates by the remote nodes. However,
in any practical scenario, nodes generally communicate with
a small subset of nodes and thus exhaustively estimating the
trust of all the nodes can lead to huge resource consumption.
In contrast, our mechanism can be tuned to locate a desired
subset of trusted nodes, based on the allowable overhead, with
respect to a particular user. The mechanism is based on a
simple exchange of random walk messages and nodes counting
the number of times they are being hit by random walkers
of nodes in their neighborhood. Simulation results to analyze
the effectiveness of the algorithm show that using the proposed
algorithm, nodes identify the top trusted nodes in the network
with a very high probability by exploring only around 45\% of
the total nodes, and in turn generates nearly 90\% less overhead
as compared to an exhaustive trust estimation mechanism,
named TrustWebRank. Finally, we provide a measure of the
global trustworthiness of a node; simulation results indicate
that the measures generated using our mechanism differ by
only around 0.6\% as compared to TrustWebRank.",
    mendeleytags = "SG-Publication",
    year = "2012",
    keywords = "SG-Publication",
    booktitle = "Proceedings of 11th IEEE International Conference on Trust, Security and Privacy in Computing and Communications (TrustCom 2012)",
    arxivid = "1204.5314"
}