Maznah Kamat

Towards next-generation routing protocols for pocket switched networks

A pocket switched network (PSN) is dynamically formed by people who carry portable handheld devices. Interest in PSNs is driven by the increasing number of handheld devices, the several wireless interfaces they possess, as well as their ability to store, carry and forward data. The lack of fixed network topology distinguishes PSNs from traditional networks, and unlike other types of mobile networks, nodes in PSNs closely follow human movement patterns. As a result, PSNs are faced with new challenges especially in the aspect of routing. Although various routing protocols have been proposed, most of them focus on optimizing the performance of networking primitives for traditional networks such as unicast, broadcast and multicast. However, these primitives themselves appear to be insufficient due to new application opportunities presented by PSNs. This paper adopts a user scenario based approach to determine the current state of PSN routing protocols. Specifically, four modes of data transfer are established from six generalized PSN user scenarios. Due to the wide range of existing routing proposals, a new taxonomy is proposed to facilitate analysis of their compatibility with the established modes of data transfer. The analysis provides new insights into application based routing approaches for realizing next-generation PSN routing protocols. A detailed characterization of PSN user scenarios and respective data transfer modes.An exhaustive analysis of existing PSN routing solutions.A novel classification for PSN routing complementing gaps of existing taxonomies.A scenario-driven set of design guidelines for next-generation PSN routing solutions.An instructive list of future directions towards real-world PSN implementations.

Quantifying the severity of blackhole attack in wireless mobile Adhoc networks

Blackhole attack is one of the severe attacks in MANET routing protocols. Generation of this attack is simple, which does not require specific tools or sophisticated attack techniques. However, this attack seriously corrupts the routing tables of nodes in the network. Even worse, the attack could increase the chances of losing confidential data and deny network services. Many researchers have proposed a variety of solutions to prevent this Conventional Blackhole Attack (CBA). Unfortunately, none of them has measured the severity of this attack. Filling gaps in measuring the severity of this attack, this research proposes a new security metrics, namely Corruption Routing Table (CRT), Compromising Relay Node (CRN) and Compromising Originator Node (CON). In addition, this research also introduces a new blackhole attack, namely Hybrid Black Hole Attack (HBHA) with two variants — Independent and Cooperative HBHAs. The proposed metrics proved effective in measuring the severity of both CBA and HBHAs. Simulation using Java in Time Simulator/Scalable Wireless Adhoc Network Simulator (JiST/SWANS) showed that Independent HBHA is the most severe attack compared to Cooperative HBHA and CBA. In addition, Cooperative HBHA is the most efficient attack than Independent HBHA and CBA.

Modified Hilbert Space-Filling Curve for Ellipsoidal Coverage in Wireless Ad Hoc Sensor Networks

Sensor nodes are deployed to sense an environment for observing the physical world. The sensing field (an area of interest) can be represented as an ellipse. This paper proposes an algorithm for covering an elliptical shaped area of interest within wireless sensor network fields by modifying the Hilbert-space filling curve. Conventionally, an area is partitioned into 2n times 2n square grid (Hilbert cells). Hilbert space-filling curve is modified to map only specific area of interest. To cover the entire area of interest, the algorithm requires a mobility pattern for a sink node in order to collect data for a specific mission. This is done by traversing the area from an entry point and finishing at an exit point from the square grids. Simulation results show that modified Hilbert space-filling curve achieves its best Hilbert-covering for an ellipse that is almost circular as opposed to a narrow ellipse.

Optimized-Hilbert for Mobility in Wireless Sensor Networks

In wireless sensor networks (WSNs), mobilizing sink node for data collection can minimize communication and maximize network lifetime. Several mobility algorithms have been proposed to control sink nodes movement. However, they do not consider mobility for coverage area problem. Hilbert space-filling curves technique is modified to navigate mobile sink in traversing specfiic area of the sensor field. The algorithm is called optimized-Hilbert, which provides mobility pattern of a sink node in covering an ellipsoidal monitoring area during data collection for a specific mission by traversing the entire area from an entry point and finish at an exit point of square grids. The optimized-Hilbert requires less number of steps as compared to conventional Hilbert. Additionally, it is also compared with circular mobility, and result indicates the efficiency of optimized Hilbert.

Location Service Management Protocol for Vehicular Ad Hoc Network Urban Environment

Location based service is used in vehicular ad hoc networks (VANET) to locate a node’s position before the start of any communica- tion. The existing location services proposed for Mobile Ad hoc Networks (MANET) suffer from low scalability in VANET environments. Protocols proposed for VANET do not consider load balance on location servers, and do not consider realistic information for selecting location servers which affect the protocols efficiency. This paper proposes a Quorum- Based Location Service Management Protocol (QLSP) which is designed for urban area topology utilizing specific node information such as the distance to intersection centre point, and speed in selecting main loca- tion server. Formation of quorum location servers is achieved by the main location server through the nomination of other nodes located at the in- tersection based on their direction of movement. QLSP shows excellent performance in reducing overhead of control packets, provides a high de- livery ratio of packets to destination, and reduces the end-to-end delay of routing packets. The performance of the protocol is then compared to other existing location service protocols.

Spatial locality in pocket switched networks

The current Internet design falls short in scenarios characterized by intermittent connectivity and absence of supporting infrastructure. Pocket switched networking allows communication between portable handheld user devices in such environments through the store-carry-forward paradigm. This involves storing data in device memory, exploiting user movement to physically transport the data, and forwarding through available short-range wireless interfaces. In this regard, user devices can be leveraged as data mules for point-to-point communication in two main scenarios: (i) conveying data between major infrastructures in urban areas (e.g., transporting sensor data to gateways in smart cities); and (ii) extending wireless coverage to provide connectivity in developing regions (e.g., bridging user devices and remote Internet access points). User movement in such scenarios exhibit spatial locality, which is the preference to a geographical location than others. This feature influences the data transfer opportunities that exist between different sets of users. This paper investigates the impact of locality on routing in pocket switched networks by: (i) evaluating the performance of existing routing strategies for point-to-point communication over different degrees of locality; and (ii) providing insights towards enhancing routing performance for realizing the aforementioned application scenarios through locality awareness.

Collecting Sensed Data with Opportunistic Networks: The Case of Contact Information Overhead

The rising human population in urban environments drives the mission towards smart cities, which envisions a wide deployment of sensors in order to improve the quality of living. In this regard, opportunistic networks (OppNets) present an economical means of collecting delay tolerant data from sensors to their respective gateways for providing various Smart City services. Due to the distributed nature of the network, encounter-based routing protocols achieve acceptable throughput by requiring nodes to exchange and update contact information on an encounter basis. Unfortunately, sufficient insight into the associated overhead is lacking in the literature. Hence, we contribute by modelling contact information overhead and investigating its impact on OppNet routing, particularly in terms of data exchange success and energy consumption on portable handheld devices. Our findings reveal that the expected contact information overhead in Smart City scenarios significantly reduces data exchange success and increases energy consumption on portable handheld devices, thereby threatening the feasibility of the technology. We address this issue by proposing an algorithm that can be incorporated into encounter-based routing protocols to reduce contact information overhead without compromising throughput. Simulation results show that our proposed algorithm reduces the average contact information overhead, increases throughput and reduces average energy consumption.

Fine-granularity clustering in wireless sensor networks

The capability of sensor nodes to collect and aggregate ambient data enables a mobile user to receive assistance from the network in charting a safe path through a potentially dangerous area. However, the information would be meaningless without localization. Applying an existing coarse-grain location-aware cluster (virtual infrastructure i.e. VI) for object detection provides only approximate location of dangerous objects as the objects are further away from the center. This distance-dependent location-based cluster creates various sizes of localized clusters, which cause inconsistency in size and thus may lead to inefficient path planning. The main contribution of this paper is to propose a fine-granularity clustering technique, that we call eVI that enhances the performance of previously-proposed VI by dividing clusters into smaller sub-clusters. Utilizing the eVI for autonomous navigation enables mobile sink to detect location of dangerous object effectively and thus result shorter safe travelled distance compared to other sector-based clusters as demonstrated in the simulation result.

Sink node mobility for ellipsoidal area coverage for efficient data collection in Wireless Sensor Networks

In this article, we explore the usage of Hilbert space filling curves in solving area covering for a target field, and also ensuring accurate gathering of periodic data collection. In covering the area, the mobility pattern that is based on modified Hilbert space filling curves suited best for ellipsoidal bounded area of interest. Comparative study demonstrates those circular and conventional Hilbert traversals are rejected due to non-adaptiveness to monitoring area shapes.

Granular location-based cluster in wireless sensor networks

The capability of sensor nodes to collect ambient data makes environmental monitoring as one of the prime applications of wireless sensor networks. Mobile sink traverses an area of interest in collecting data. With the help of virtual infrastructure (VI), an area is divided into cluster. However, the nature of the structure causes various sizes of localized cluster; coarser as a cluster is farther from the center (distance dependent). This causes inconsistency and approximate in cluster-based localization. Distance independent clustering provides consistent granularity which creates almost equal size of clusters called enhanced VI (eVI) is proposed. Applications that can benefit from this granularity include accuracy in aggregation function, and shorter safe path charting. Our previous work intelligently handled out-of-bound sensor nodes that reside outside an area of interest. However, the excluded region varies due to the structure. Distance independent clustering can enhance the exclusion process further for efficient data collection of the queried area through dividing the cluster into smaller sub-cluster appropriately. Additionally, utilizing the eVI for autonomous navigation enables mobile sink to detect more precise location of dangerous object and thus result shorter safe travelled distance. Simulation result demonstrates that the granular clusters perform better than coarse-grained clusters.