Search

Exploratory factor analysis (EFA) is useful to identify the number and mapping of the hidden factors that could dominantly represent the features in the dataset. Principal component analysis (PCA) is the first step as part of the two-step procedure to conduct EFA, with the number of dominant principal components being the number of hidden factors and the entries for the features in the corresponding Eigenvectors serve as the initial values of the factor loadings. In this paper, we conduct EFA on a suite of 80 complex network datasets to identify the number and mapping of the hidden factors (expected to be less than four) that could dominantly represent the values incurred by the vertices with respect to the four major centrality metrics (degree: DEG, eigenvector: EVC, betweenness: BWC and closeness: CLC)., Comment: 8 pages, 4 figures

We propose a principal component analysis (PCA)-based approach to quantify (the node dissimilarity index, NDI) the extent of dissimilarity among nodes in a network with respect to values incurred for a suite of node-level metrics (like centrality metrics). We subject the dataset (n nodes and their values incurred for four commonly studied centrality metrics: degree, eigenvector, betweenness and closeness) to PCA and retain the m ( <= 4) principal components (with variance >= 1.0). We construct an n-node dissimilarity matrix whose entries are the absolute difference (if m = 1) or Euclidean distance (if M > 1) of the principal component coordinates of the corresponding nodes. We compute NDI (>= 1.0) to be the ratio of the principal Eigenvalue of the node dissimilarity matrix and average of entries in the node dissimilarity matrix. The larger the NDI, the greater the dissimilarity among the node-level metrics (centrality metrics) values considered for analysis., Comment: 7 pages, 4 figures

We perform factor analysis on the raw data of the four major neighborhood and shortest paths-based centrality metrics (Degree, Eigenvector, Betweeenness and Closeness) and propose a novel quantitative measure called the Core-Intermediate-Peripheral (CIP) Index to capture the extent with which a node could play the role of a core node (nodes at the center of a network with larger values for any centrality metric) vis-a-vis a peripheral node (nodes that exist at the periphery of a network with lower values for any centrality metric). We conduct factor analysis (varimax-based rotation of the Eigenvectors) on the transpose matrix of the raw centrality metrics dataset, with the node ids as features, under the hypothesis that there are two factors (core and peripheral) that drive the values incurred by the nodes with respect to the centrality metrics. We test our approach on a diverse suite of 12 complex real-world networks., Comment: 10 pages, 5 figures

We investigate the use of a recently proposed centrality tuple called the Neighborhood-based Bridge Node Centrality (NBNC) tuple to choose nodes for preferential vaccination so that such vaccinated nodes could provide herd immunity and reduce the spreading rate of infections in a complex real-world network. The NBNC tuple ranks nodes on the basis of the extent they play the role of bridge nodes in a network. A node is a bridge node, if when removed its neighbors are either disconnected or at least sparsely connected. We hypothesize that preferentially vaccinating such bridge nodes would block an infection to spread from a neighbor of the bridge node to an another neighbor that are otherwise not reachable to each other. We evaluate the effectiveness of using NBNC to reduce the spread of infections by conducting simulations of the spread of infections per the SIS (Susceptible-Infected-Susceptible) model on a collection of 10 complex real-world social networks. We observe the average fraction of infected nodes per round of the SIS simulations based on NBNC for preferential vaccination to be lower than that of the degree centrality-based preferential vaccination., Comment: 11 pages, 6 figures

Graph Isomorphism is one of the classical problems of graph theory for which no deterministic polynomial-time algorithm is currently known, but has been neither proven to be NP-complete. Several heuristic algorithms have been proposed to determine whether or not two graphs are isomorphic (i.e., structurally the same). In this research, we propose to use the sequence (either the non-decreasing or nonincreasing order) of eigenvector centrality (EVC) values of the vertices of two graphs as a precursor step to decide whether or not to further conduct tests for graph isomorphism. The eigenvector centrality of a vertex in a graph is a measure of the degree of the vertex as well as the degrees of its neighbors. We hypothesize that if the non-increasing (or non-decreasing) order of listings of the EVC values of the vertices of two test graphs are not the same, then the two graphs are not isomorphic. If two test graphs have an identical non-increasing order of the EVC sequence, then they are declared to be potentially isomorphic and confirmed through additional heuristics. We test our hypothesis on random graphs (generated according to the Erdos-Renyi model) and we observe the hypothesis to be indeed true: graph pairs that have the same sequence of non-increasing order of EVC values have been confirmed to be isomorphic using the well-known Nauty software., Comment: 9 pages, 4 figures; Proceedings of the Fourth International Conference on Advanced Information Technologies and Applications (ICAITA), pp. 1-9, Dubai, UAE, November 6-7, 2015

In this paper, we consider a random network such that there could be a link between any two nodes in the network with a certain probability (plink). Diffusion is the phenomenon of spreading information throughout the network, starting from one or more initial set of nodes (called the early adopters). Information spreads along the links with a certain probability (pdiff). Diffusion happens in rounds with the first round involving the early adopters. The nodes that receive the information for the first time are said to be covered and become candidates for diffusion in the subsequent round. Diffusion continues until all the nodes in the network have received the information (successful diffusion) or there are no more candidate nodes to spread the information but one or more nodes are yet to receive the information (diffusion failure). On the basis of exhaustive simulations conducted in this paper, we observe that for a given plink and pdiff values, the fraction of successful diffusion attempts does not appreciably change with increase in the number of early adopters; whereas, the average number of rounds per successful diffusion attempt decreases with increase in the number of early adopters. The invariant nature of the fraction of successful diffusion attempts with increase in the number of early adopters for a random network (for fixed plink and pdiff values) is an interesting and noteworthy observation (for further research) and it has not been hitherto reported in the literature., Comment: 7 papers; 5 figures

Our primary objective in this paper is to study the distribution of the maximal clique size of the vertices in complex networks. We define the maximal clique size for a vertex as the maximum size of the clique that the vertex is part of and such a clique need not be the maximum size clique for the entire network. We determine the maximal clique size of the vertices using a modified version of a branch-and-bound based exact algorithm that has been originally proposed to determine the maximum size clique for an entire network graph. We then run this algorithm on two categories of complex networks: One category of networks capture the evolution of small-world networks from regular network (according to the wellknown Watts-Strogatz model) and their subsequent evolution to random networks; we show that the distribution of the maximal clique size of the vertices follows a Poisson-style distribution at different stages of the evolution of the small-world network to a random network; on the other hand, the maximal clique size of the vertices is observed to be in-variant and to be very close to that of the maximum clique size for the entire network graph as the regular network is transformed to a small-world network. The second category of complex networks studied are real-world networks (ranging from random networks to scale-free networks) and we observe the maximal clique size of the vertices in five of the six real-world networks to follow a Poisson-style distribution. In addition to the above case studies, we also analyze the correlation between the maximal clique size and clustering coefficient as well as analyze the assortativity index of the vertices with respect to maximal clique size and node degree., Comment: 20 pages, 13 figures, 3 tables

In this paper, we show the evaluation of the spectral radius for node degree as the basis to analyze the variation in the node degrees during the evolution of scale-free networks and small-world networks. Spectral radius is the principal eigenvalue of the adjacency matrix of a network graph and spectral radius ratio for node degree is the ratio of the spectral radius and the average node degree. We observe a very high positive correlation between the spectral radius ratio for node degree and the coefficient of variation of node degree (ratio of the standard deviation of node degree and average node degree). We show how the spectral radius ratio for node degree can be used as the basis to tune the operating parameters of the evolution models for scale-free networks and small-world networks as well as evaluate the impact of the number of links added per node introduced during the evolution of a scale-free network and evaluate the impact of the probability of rewiring during the evolution of a small-world network from a regular network., Comment: 8 pages, 8 figures, Second International Conference on Computer Science and Information Technology, (COSIT-2015), Geneva, Switzerland, March 21-22, 2015

This paper presents the applications of Eigenvalues and Eigenvectors (as part of spectral decomposition) to analyze the bipartivity index of graphs as well as to predict the set of vertices that will constitute the two partitions of graphs that are truly bipartite and those that are close to being bipartite. Though the largest eigenvalue and the corresponding eigenvector (called the principal eigenvalue and principal eigenvector) are typically used in the spectral analysis of network graphs, we show that the smallest eigenvalue and the smallest eigenvector (called the bipartite eigenvalue and the bipartite eigenvector) could be used to predict the bipartite partitions of network graphs. For each of the predictions, we hypothesize an expected partition for the input graph and compare that with the predicted partitions. We also analyze the impact of the number of frustrated edges (edges connecting the vertices within a partition) and their location across the two partitions on the bipartivity index. We observe that for a given number of frustrated edges, if the frustrated edges are located in the larger of the two partitions of the bipartite graph (rather than the smaller of the two partitions or equally distributed across the two partitions), the bipartivity index is likely to be relatively larger., Comment: A paper with a similar idea has been published earlier. I was not aware of this before. Now that, I have come to know about this, I am withdrawing my paper from arXiv

We propose an Energy-aware Maximal Leaf Nodes Data Gathering (EMLN-DG) algorithm for periodic data collection and transmission in wireless sensor networks. For each round of data gathering, an EMLN-DG tree spanning the entire sensor network is formed based on the residual energy level available at the nodes and the number of uncovered neighbors of a node during tree formation. Only nodes that have a relatively larger number of neighbors as well as a higher energy level are included as intermediate nodes in the EMLN-DG tree. By maximizing the number of leaf nodes in a DG tree and considering the energy level available at the nodes while forming the tree, we reduce energy consumption per round as well as balance the energy level across all the nodes in the network. This contributes to a significantly larger network lifetime, measured as the number of rounds before the first node failure due to exhaustion of battery charge. Performance comparison studies with the well-known data gathering algorithms such as LEACH and PEGASIS illustrate that EMLN-DG can help to sustain the network for a significantly larger number of rounds and at the same time incur a lower, or if not comparable, energy loss, delay and energy loss*delay per round of data gathering., Comment: 14 pages, 5 figures, Journal of Theoretical and Applied Information Technology, Vol. 15, No. 1, May 2010

The high-level contribution of this paper is to establish benchmarks for the minimum hop count per source-receiver path and the minimum number of edges per tree for multicast routing in mobile ad hoc networks (MANETs) under different mobility models. In this pursuit, we explore the tradeoffs between these two routing strategies with respect to hop count, number of edges and lifetime per multicast tree with respect to the Random Waypoint, City Section and Manhattan mobility models. We employ the Breadth First Search algorithm and the Minimum Steiner Tree heuristic for determining a sequence of minimum hop and minimum edge trees respectively. While both the minimum hop and minimum edge trees exist for a relatively longer time under the Manhattan mobility model; the number of edges per tree and the hop count per source-receiver path are relatively low under the Random Waypoint model. For all the three mobility models, the minimum edge trees have a longer lifetime compared to the minimum hop trees and the difference in lifetime increases with increase in network density and/or the multicast group size. Multicast trees determined under the City Section model incur fewer edges and lower hop count compared to the Manhattan mobility model., Comment: 14 pages, 17 figures

With the phenomenal growth in the Internet, network security has become an integral part of computer and information security. In order to come up with measures that make networks more secure, it is important to learn about the vulnerabilities that could exist in a computer network and then have an understanding of the typical attacks that have been carried out in such networks. The first half of this paper will expose the readers to the classical network attacks that have exploited the typical vulnerabilities of computer networks in the past and solutions that have been adopted since then to prevent or reduce the chances of some of these attacks. The second half of the paper will expose the readers to the different network security controls including the network architecture, protocols, standards and software/ hardware tools that have been adopted in modern day computer networks., Comment: 21 pages, 19 figures

We propose the design and development of an energy-efficient Risk Notification Message Dissemination Protocol (RNMDP) for vehicular ad hoc networks (VANETs). RNMDP propagates Risk Notification Messages (RNMs) from their location of origin (called the Risk Zone) to vehicles approaching the Risk Zone. RNMDP assumes each node is aware of its current location in the network. The protocol works as follows: A RNM is broadcast in the neighborhood of the Risk Zone. A node receiving the RNM from another node waits for a Rebroadcast-Wait-Time before deciding to rebroadcast the message. The Rebroadcast-Wait-Time for a node is modeled based on the ratio of the distance between the node and the immediate sender of the RNM and the direction of movement of the node. Priority for rebroadcast is given for nodes farthest away from the sender and traveling towards the Risk Zone. Nodes that are traveling in lanes in direction away from the Risk Zone are also considered for rebroadcast, albeit with a larger Rebroadcast-Wait-Time. During the Rebroadcast-Wait-Time, if a node hears the same RNM again rebroadcast in the neighborhood, then the node stops from further broadcasting the message. If a node does not hear the RNM in its neighborhood during the RebroadcastWait-Time, the node broadcasts the message in its neighborhood. A RNM is considered to have been delivered to all the vehicles in the road, if the message reaches the Target Zone. The performance of RNMDP has been compared with that of the commonly used flooding strategy through extensive simulations conducted for highway networks with different number of lanes and lane density. Simulation results indicate that with a slightly larger delay (i.e., no more than 35% of the delay incurred for flooding), RNMDP can achieve the same message delivery ratio attained by flooding, but at a relatively much lower energy loss compared to flooding., Comment: 15 pages, 15 figures

We propose a two-layer mobile sink and static sensor network (MSSSN) architecture for large scale wireless sensor networks. The top layer is a mobile ad hoc network of resource-rich sink nodes while the bottom layer is a network of static resource-constrained sensor nodes. The MSSSN architecture can be implemented at a lower cost with the currently available IEEE 802.11 devices that only use a single halfduplex transceiver. Each sink node is assigned a particular region to monitor and collect data. A sink node moves to the vicinity of the sensor nodes (within a few hops) to collect data. The collected data is exchanged with peer mobile sinks. Thus, the MSSSN architecture provides scalability, extends sensor lifetime by letting them operate with limited transmission range and provides connectivity between isolated regions of sensor nodes. In order to provide fault tolerance, more than one mobile sink could be collecting data from a given region or a mobile sink could collect data from more than one region. In the later half of the paper, we discuss several open research issues that need to be addressed while implementing the MSSSN architecture., Comment: 11 pages, 1 figure

The high-level contribution of this paper is a simulation-based detailed performance comparison of three different classes of on-demand routing protocols for mobile ad hoc networks: stability-based routing, power-aware routing and load-balanced routing. We choose the Flow-Oriented Routing protocol (FORP), Min-Max Battery Cost Routing (MMBCR) and the traffic interference based Load Balancing Routing (LBR) protocol as representatives of the stability-based routing, poweraware routing and load-balancing routing protocols respectively. FORP incurs the least number of route transitions; while LBR incurs the smallest hop count and lowest end-to-end delay per data packet. Energy consumed per data packet is the least for LBR, closely followed by MMBCR. FORP incurs the maximum energy consumed per data packet, both in the absence and presence of power control. Nevertheless, in the presence of power control, the end-to-end delay per data packet and energy consumed per data packet incurred by FORP are significantly reduced compared to the scenario without power control. MMBCR is the most fair in terms of node usage and incurs the largest time for first node failure. FORP tends to repeatedly use nodes lying on the stable path and hence is the most unfair of the three routing protocols. FORP also incurs the smallest value for the time of first node failure., Comment: 8 pages, 12 figures. arXiv admin note: substantial text overlap with arXiv:1007.0409

A mobile sensor network is a wireless network of sensor nodes that move arbitrarily. In this paper, we explore the use of a maximum stability spanning tree-based data gathering (Max.Stability-DG) algorithm and a minimum-distance spanning tree-based data gathering (MST-DG) algorithm for mobile sensor networks. We analyze the impact of these two algorithms on the node failure times and the resulting coverage loss due to node failures. Both the Max.Stability-DG and MST-DG algorithms are based on a greedy strategy of determining a data gathering tree when one is needed and using that tree as long as it exists. The Max.Stability-DG algorithm assumes the availability of the complete knowledge of future topology changes and determines a data gathering tree whose corresponding spanning tree would exist for the longest time since the current time instant; whereas, the MST-DG algorithm determines a data gathering tree whose corresponding spanning tree is the minimum distance tree at the current time instant. We observe the Max.Stability-DG trees to incur a longer network lifetime (time of disconnection of the network of live sensor nodes due to node failures), a larger coverage loss time for a particular fraction of loss of coverage as well as a lower fraction of coverage loss at any time. The tradeoff is that the Max.Stability-DG trees incur a lower node lifetime (the time of first node failure) due to repeated use of a data gathering tree for a longer time., Comment: 16 pages, 11 figures

We propose a secure data aggregation (SDA) framework for mobile sensor networks whose topology changes dynamically with time. The SDA framework (designed to be resilient to both insider and outsider attacks) comprises of a pair-wise key establishment mechanism run along the edges of a data gathering tree and a distributed trust evaluation model that is tightly integrated with the data aggregation process itself. If an aggregator node already shares a secret key with its child node, the two nodes locally coordinate to refresh and establish a new pair-wise secret key; otherwise, the aggregator node requests the sink to send a seed-secret key message that is used as the basis to establish a new pair-wise secret key. The trust evaluation model uses the two-sided Grubbs test to identify outlier data in the periodic beacons collected from the child nodes (neighbor) nodes. Once the estimated trust score for a neighbor node falls below a threshold, the sensor node locally classifies its neighbor node as a Compromised or Faulty (CF) node, and discards the data or aggregated data received from the CF node. This way, the erroneous data generated by the CF nodes could be filtered at various levels of the data gathering tree and are prevented from reaching the root node (sink node). Finally, we assess the effectiveness of our trust evaluation model through a comprehensive simulation study., Comment: 20 pages

This paper presents the source code analysis of a file reader server socket program (connection-oriented sockets) developed in Java, to illustrate the identification, impact analysis and solutions to remove five important software security vulnerabilities, which if left unattended could severely impact the server running the software and also the network hosting the server. The five vulnerabilities we study in this paper are: (1) Resource Injection, (2) Path Manipulation, (3) System Information Leak, (4) Denial of Service and (5) Unreleased Resource vulnerabilities. We analyze the reason why each of these vulnerabilities occur in the file reader server socket program, discuss the impact of leaving them unattended in the program, and propose solutions to remove each of these vulnerabilities from the program. We also analyze any potential performance tradeoffs (such as increase in code size and loss of features) that could arise while incorporating the proposed solutions on the server program. The proposed solutions are very generic in nature, and can be suitably modified to correct any such vulnerabilities in software developed in any other programming language. We use the Fortify Source Code Analyzer to conduct the source code analysis of the file reader server program, implemented on a Windows XP virtual machine with the standard J2SE v.7 development kit., Comment: 16 pages, 16 figures

The interest in virtualization has been growing rapidly in the IT industry because of inherent benefits like better resource utilization and ease of system manageability. The experimentation and use of virtualization as well as the simultaneous deployment of virtual software are increasingly getting popular and in use by educational institutions for research and teaching. This paper stresses on the potential advantages associated with virtualization and the use of virtual machines for scenarios, which cannot be easily implemented and/or studied in a traditional academic network environment, but need to be explored and experimented by students to meet the raising needs and knowledge-base demanded by the IT industry. In this context, we discuss various aspects of virtualization - starting from the working principle of virtual machines, installation procedure for a virtual guest operating system on a physical host operating system, virtualization options and a performance study measuring the throughput obtained on a network of virtual machines and physical host machines. In addition, the paper extensively evaluates the use of virtual machines and virtual networks in an academic environment and also specifically discusses sample projects on network security, which may not be feasible enough to be conducted in a physical network of personal computers; but could be conducted only using virtual machines.

Communication protocols for mobile ad hoc networks (MANETs) follow either an Optimum Routing Approach (ORA) or the Least Overhead Routing Approach (LORA): With ORA, protocols tend to determine and use the optimal communication structure at every time instant; whereas with LORA, a protocol tends to use a chosen communication structure as long as it exists. In this paper, we study the impact of the ORA and LORA strategies on minimum hop routes and minimum connected dominating sets (MCDS) in MANETs. Our primary hypothesis is that the LORA strategy could yield routes with a larger time-averaged hop count and MCDS node size when compared to the minimum hop count of routes and the node size of the MCDS determined using the ORA strategy. Our secondary hypothesis is that the impact of ORA vs. LORA also depends on how long the communication structure is being used. Our hypotheses are evaluated using extensive simulations under diverse conditions of network density, node mobility and mobility models such as the Random Waypoint model, City Section model and the Manhattan model. In the case of minimum hop routes, which exist for relatively a much longer time compared to the MCDS, the hop count of routes maintained according to LORA, even though not dramatically high, is appreciably larger (6-12%) than those maintained according to ORA; on the other hand, the number of nodes constituting a MCDS maintained according to LORA is only at most 6% larger than the node size of a MCDS maintained under the ORA strategy., Comment: 17 pages, 19 figures, 6 tables

The objective of this research is to integrate an RFID (Radio Frequency Identification) reader into a Wireless Sensor Network (WSN) to authorize or keep track of people carrying RFID tags. The objective was accomplished by integrating hardware and software. The hardware consisted of two WSN nodes - the RFID node connected to one of the WSN nodes, and a computer connected to the other WSN node. For the RFID equipment, we used the SM130-EK kit, which included the RFID reader and the RFID tags; and for the WSN, we used the Synapse Network Evaluation kit, which included the two sensor nodes. The software consisted of a program module developed in Python to control the microprocessors of the nodes; and a database controlled by a simple program to manage the tag IDs of people wearing them. The WSN and RFID nodes were connected through I2C interfacing. Also, the work of sending commands to the RFID node, to make it read a tag and send it back to the computer, was accomplished by the Python code developed which also controls the data signals. At the computer, the received tag ID is evaluated with other existing tag IDs on the database, to check if that tag has authorization or not to be in the covered area. Our research has the potential of being adapted for use with secure real-time access control applications involving WSN and RFID technologies.

The high-level contribution of this paper is a simulation-based analysis to evaluate the tradeoffs between lifetime and hop count of link-disjoint, node-disjoint and zone-disjoint multi-path routes vis-\`a-vis single-path minimum hop routes for mobile ad hoc networks. The link-disjoint, node-disjoint and zone-disjoint algorithms proposed in this paper can be used to arrive at benchmarks for the time between successive multi-path route discoveries, the number of disjoint paths per multi-path set and the hop count per multi-path set. We assume a multi-path set exists as long as at least one path in the set exists. Simulation results indicate that the number of zone-disjoint paths per multi-path set can be at most 2, which is far lower than the number of node and link-disjoint paths available per multi-path set. Also, the time between zone-disjoint multi-path discoveries would be far lower than the time between node and link-disjoint multi-path route discoveries and can be at most 45% more than the time between single minimum-hop path route discoveries. However, there is no appreciable difference in the average hop counts per zone-disjoint, node-disjoint and link-disjoint multi-path sets and it can be only at most 15% more than the average minimum hop count determined using single-path routing. We also observe that even though the number of link-disjoint paths per multi-path set can be as large as 35-78% more than the number of node-disjoint paths per multi-path set, the time between two successive link-disjoint multi-path discoveries can be at most 15-25% more than the time between two successive node-disjoint multi-path discoveries, without any significant difference in the hop count per multi-path set., Comment: 17 pages

The high-level contribution of this paper is a simulation-based detailed performance comparison of three different classes of routing protocols for mobile ad hoc networks: stability-based routing, power-aware routing and load-balanced routing. We choose the Flow-Oriented Routing protocol (FORP), the traffic interference based Load Balancing Routing (LBR) protocol and Min-Max Battery Cost Routing (MMBCR) as representatives of the stability-based routing, load-balancing and power-aware routing protocols respectively. Among the three routing protocols, FORP incurs the least number of route transitions; while LBR incurs the smallest hop count and lowest end-to-end delay per data packet. Energy consumed per node is the least for MMBCR, closely followed by LBR. MMBCR is the most fair in terms of node usage and hence it incurs the largest time for first node failure. FORP tends to repeatedly use nodes lying on the stable path and hence is the most unfair of the three routing protocols and it incurs the smallest value for the time of first node failure. As we measure the failure times of up to the first five nodes in the network, we observe that LBR incurs the maximum improvement in the lifetime of the nodes and MMBCR incurs the least improvement beyond the time of first node failure., Comment: http://ijict.org/index.php/ijoat/article/view/performance-comparison-manet-routing-protocols