es.ants.felixgm.trmsim_wsn.trm.eigentrust
Class EigenTrust

java.lang.Object
  es.ants.felixgm.trmsim_wsn.trm.TRModel_WSN
      es.ants.felixgm.trmsim_wsn.trm.eigentrust.EigenTrust

public class EigenTrust
extends TRModel_WSN

This class models EigenTrust algorithm used by a client in a P2P, Ad-hoc or Wireless Sensor Network, in order to find the most trustworthy server offering a desired service.

It needs some parameters to be passed as a EigenTrust_Parameters object. To do this, a file can be given following the next structure:

    ####################################
    # EigenTrust parameters file
    ####################################
    windowSize=5
    epsilon=0.1
    preTrustedPeersPercentage=0.3
    preTrustedPeersWeight=0.25
    zeroTrustNodeSelectionProbability=0.2
 

This file can be downloaded here. But if any of the parameters can not be successfully extracted from the file, they are set to a default value.

For more information regarding EigenTrust algorithm, please check the following reference:

• Kamvar, S., Schlosser, M. and Garcia-Molina, H., "The EigenTrust Algorithm for Reputation Management in P2P Networks", WWW03: Proceedings of the 12th international conference on World Wide Web, pp. 640-651, 2003

Since:

0.2

Version:

0.4

Author:

Félix Gómez Mármol, Gregorio Martínez Pérez and Antonio Bernárdez

Field Summary

protected double	MAX_SATISFACTION Maximum satisfaction value: 1.0
protected double	MIN_SATISFACTION Minimum satisfaction value: -1.0

Fields inherited from class es.ants.felixgm.trmsim_wsn.trm.TRModel_WSN

trmParameters

Constructor Summary

EigenTrust(EigenTrust_Parameters eigenTrust_parameters)
Class EigenTrust constructor

Method Summary

GatheredInformation	gatherInformation(Sensor client, Service service) This method collects or gathers the required information from the network needed to evaluate each sensor offering a given service in order to determine whether to have a transaction with it or not
Network	generateRandomNetwork(int numSensors, double probClients, double rangeFactor, java.util.Collection<java.lang.Double> probServices, java.util.Collection<java.lang.Double> probGoodness, java.util.Collection<Service> services) This method generates a new random network specific for this trust and reputation model
static java.lang.String	get_name() Returns the name of this model, i.e., "EigenTrust"
Network	loadCurrentNetwork(java.lang.String fileName) This method loads a network from a XML file and creates the specific network corresponding to this trust and reputation model
Outcome	performTransaction(java.util.Vector<Sensor> path, Service service) This method actually requests a desired service to a specified server and evaluates the satisfaction of the client with the actually received service
Outcome	punish(java.util.Vector<Sensor> path, Outcome outcome) EigenTrust does not implement this method, since it does not apply any specific punish step
Outcome	reward(java.util.Vector<Sensor> path, Outcome outcome) EigenTrust does not implement this method, since it does not apply any specific reward step
java.util.Vector<Sensor>	scoreAndRanking(Sensor client, GatheredInformation gi) This method computes a trust and/or reputation value for each reachable server from the specified client and returns either a sorted list of all the reachable servers or a path leading directly to the most trustworthy and/or reputable one

Methods inherited from class es.ants.felixgm.trmsim_wsn.trm.TRModel_WSN

get_TRMParameters, set_TRMParameters

Methods inherited from class java.lang.Object

clone, equals, finalize, getClass, hashCode, notify, notifyAll, toString, wait, wait, wait

Field Detail

MIN_SATISFACTION

protected final double MIN_SATISFACTION

Minimum satisfaction value: -1.0

See Also:

Constant Field Values

MAX_SATISFACTION

protected final double MAX_SATISFACTION

Maximum satisfaction value: 1.0

See Also:

Constant Field Values

Constructor Detail

EigenTrust

public EigenTrust(EigenTrust_Parameters eigenTrust_parameters)

Class EigenTrust constructor

Parameters:

eigenTrust_parameters - Parameters needed for the algorithm, as described before

Method Detail

get_name

public static java.lang.String get_name()

Returns the name of this model, i.e., "EigenTrust"

Returns:

The name of this model, i.e., "EigenTrust"

gatherInformation

public GatheredInformation gatherInformation(Sensor client,
                                             Service service)

Description copied from class: TRModel_WSN

This method collects or gathers the required information from the network needed to evaluate each sensor offering a given service in order to determine whether to have a transaction with it or not

Specified by:

gatherInformation in class TRModel_WSN

Parameters:

client - The client applying this trust and reputation model in order to find the most trustworthy and reputable server offering a given service

service - The service requested by the given client

Returns:

The required information from the network needed to evaluate each sensor offering a given service in order to determine whether to have a transaction with it or not

scoreAndRanking

public java.util.Vector<Sensor> scoreAndRanking(Sensor client,
                                                GatheredInformation gi)

Description copied from class: TRModel_WSN

This method computes a trust and/or reputation value for each reachable server from the specified client and returns either a sorted list of all the reachable servers or a path leading directly to the most trustworthy and/or reputable one

Specified by:

scoreAndRanking in class TRModel_WSN

Parameters:

client - The client applying this trust and reputation model in order to find the most trustworthy and reputable server offering a given service

gi - The gathered information fro the network needed in order to compute the trust and/or reputation value for each reachable server

Returns:

Either a path leading to the most trustworthy server found, or a sorted list of all the reachable servers

performTransaction

public Outcome performTransaction(java.util.Vector<Sensor> path,
                                  Service service)

Description copied from class: TRModel_WSN

This method actually requests a desired service to a specified server and evaluates the satisfaction of the client with the actually received service

Specified by:

performTransaction in class TRModel_WSN

Parameters:

path - Path of sensors leading to the most trustworthy and reputable server or sorted list of all the reachable servers

service - Service requested by the client running this trust and reputation model

Returns:

Outcome with the satisfaction of the client with the received service

reward

public Outcome reward(java.util.Vector<Sensor> path,
                      Outcome outcome)

EigenTrust does not implement this method, since it does not apply any specific reward step

Specified by:

reward in class TRModel_WSN

Parameters:

path -

outcome -

Returns:

Outcome with the updated transmitted distance of the messages sent

punish

public Outcome punish(java.util.Vector<Sensor> path,
                      Outcome outcome)

EigenTrust does not implement this method, since it does not apply any specific punish step

Specified by:

punish in class TRModel_WSN

Parameters:

path -

outcome -

Returns:

Outcome with the updated transmitted distance of the messages sent

generateRandomNetwork

public Network generateRandomNetwork(int numSensors,
                                     double probClients,
                                     double rangeFactor,
                                     java.util.Collection<java.lang.Double> probServices,
                                     java.util.Collection<java.lang.Double> probGoodness,
                                     java.util.Collection<Service> services)

Description copied from class: TRModel_WSN

This method generates a new random network specific for this trust and reputation model

Specified by:

generateRandomNetwork in class TRModel_WSN

Parameters:

numSensors - Number of sensors composing the network

probClients - Probability of a sensor to act as a client requesting services

rangeFactor - Maximum wireless range of every sensor. It determines the neighborhood of every sensor

probServices - A collection of probabilities of offering a certain service, one per service

probGoodness - A collection of goodnesses about offering a certain service, one per service

services - All the services offered by the generated Network

Returns:

A new random network specific for this trust and reputation model

loadCurrentNetwork

public Network loadCurrentNetwork(java.lang.String fileName)
                           throws java.lang.Exception

Description copied from class: TRModel_WSN

This method loads a network from a XML file and creates the specific network corresponding to this trust and reputation model

Specified by:

loadCurrentNetwork in class TRModel_WSN

Parameters:

fileName - Path of the XML to load the network from

Returns:

The loaded and generated specific network for this trust and reputation model

Throws:

java.lang.Exception - If the XML file given does not have the appropriate structure, or if a sensor links to an undefined sensor, or if a sensor links to itself