Catálogo de publicaciones - libros

In WDM mesh networks, low-speed data streams from individual users are combined, using the techniques of traffic grooming, for efficient utilization of the high bandwidth of a lightpath. The objective of traffic grooming is to minimize the cost of the network and/or to maximize the network throughput. Proposed solutions for this optimization problem are computationally intractable, even for networks of moderate size. In this paper, we have presented efficient Integer Linear Program (ILP) formulations for traffic grooming on mesh WDM networks, one for minimizing the congestion and the other to maximize the throughput of the network, with an assumption that the logical topology is specified. Unlike previous formulations, our formulations can be used for practical sized networks. We have simulated our formulations with networks up to 30 nodes, and with hundreds and even thousands of low-speed data streams and have shown that the formulations are able to generate optimal solutions within a reasonable amount of time.

IPv4 architecture is well entrenched with Network Address Translation (NAT) boxes, which cause well-known problems for Peer-to-Peer (P2P) applications. IPv6 would enable end-to-end connectivity when deployed, but the industry has been slow in transitioning to IPv6. IPv4+4 has been suggested as an alternative NAT-extended addressing architecture, where the idea is to assign 64-bit end-to-end globally unique addresses for nodes on private address realms by concatenating the 32-bit globally routable IPv4 address of the realm (border) gateway with the 32-bit private IPv4 addresses of the nodes. While IPv4+4 addressing proposal is neat, existing IPv4+4 implementations require changes to all border gateways and end-hosts, which hinders its deployment. In this paper we show how the IPv4+4 addressing architecture can be implemented by using a modified version of the standard IPv4 Loose Source Record Route (LSRR) option. Our proposal requires no changes to existing IPv4 infrastructure (assuming all IPv4-compliant nodes implement LSRR as required by RFC 791), thus enabling seamless end-to-end communication for P2P applications. We demonstrate packet forwarding with the 64-bit IPv4+4 addresses, and illustrate how the widely-used P2P voice over IP protocol, the Session Initiation Protocol, can make use of our proposal for seamless end-to-end communication.

Mobile SCTP (mSCTP) has been proposed to support transport layer mobility management. In this paper, we proposed a dynamic handover mechanism named Contention-based mSCTP (C-mSCTP) in the contention based wireless network. The handover strategy in C-mSCTP takes wireless bandwidth and the contention probability into account to set the primary path before transmitting data. The simulation results show that the C-mSCTP achieves much better performance in terms of transmission delay and throughput than mSCTP does.

Multiple non-overlapped channels are available in IEEE 802.11 but are rarely used today in wireless multi-hop networks. Wireless mesh network is a special type of multi-hop ad hoc network and is envisioned to provide high capacity and large coverage. In this paper, we propose a 2-hop clustering based multi-interface, multi-channel network architecture and design a novel channel assignment algorithm and routing metric. Channel assignment is composed of Inter-cluster Static Assignment and Intra-cluster Dynamic Assignment. Since traditional routing metrics, such as hop-count, may not perform well in multi-channel wireless networks, we propose the CDM routing metric, which combines hop-count, channel diversity and channel switching capability together. Simulation results show that our algorithms achieve up to 3.3 times higher end-to-end throughput.

In this paper, we apply the hierarchical concept to the Care-of Prefix (CoP) scheme as HCoP and enhance HCoP with a novel Binding Update Tree (BUT) structure as HCoP-B for NEtwork MObility (NEMO) management of the nested mobile network. As compared to schemes such as Reverse Routing Header (RRH), Route Optimization using Tree Information Option (ROTIO) and HCoP with numerical performance evaluations, HCoP-B achieves the shortest handoff latency and significantly reduces the consumed network bandwidth of global binding update messages for route optimizations (RO) of all correspondent nodes (CN) after the nested mobile network hands over to a new AR. Consequently, HCoP-B resolves the RO storm problem.

The surface area which is defined as the number of vertices at a given distance from a base vertex of a graph is considered to be as one of the most useful yet abstract combinatorial properties of a graph. The applicability of surface area spans many problem spaces such as those in parallel and distributed computing. These problems normally involve combinatorial analysis of underlying graph structures (e.g., spanning tree construction, minimum broadcast algorithms, efficient VLSI layout, performance modeling). In this paper, we focus on the problem of finding the surface area of a class of popular graphs, namely the family of WK-recursive and swapped networks. These are attractive networks due to their useful recursive structures.

This paper addresses the issue of multicast communication in scalable interconnection networks, using path-based scheme. Most existing multicast algorithms either assume a fixed network size, low dimensional networks or only consider the latency at the network level. As a consequence, most of these algorithms implement multicast in a sequential manner and can not scale well with the network dimensions or the number of nodes involved. Furthermore, most of these algorithms handle multicast communication with low throughput. In this paper, we propose a multicast algorithm for multidimensional interconnection networks, which is built upon our Qualified Groups QG multicast scheme for ensuring efficient communication irrespective of the network sizes/dimensions or the number of the destination nodes. Unlike the existing works, this study considers the scalability and latency at both the network and node levels so as to achieve a high degree of parallelism. Our results show that the proposed algorithm considerably improves the multicast message delivery ratio, throughput and scalability.

A wireless sensor network consists of many micro-sensor nodes distributed throughout an area of interest. Each node has a limited energy supply and generates information that needs to be communicated to a sink node. The basic operation in such a network is the systematic gathering and transmission of sensed data to a base station for further processing. During data gathering, sensors have the ability to perform in-network aggregation (fusion) of data packet routes to the base station. The lifetime of such a sensor system can be defined as the time during which the sensor information is gathered from all of the sensors and combined at the base station. Given the location of the sensors, the base station and the available energy at each sensor, the main interest is to find an efficient manner in which data can be collected from the sensors and transmitted to the base station, so as to maximize the system lifetime. We address the zone based data aggregation scheduling scheme for maximizing network lifetime. The experimental results demonstrate that the proposed protocol significantly outperforms other data aggregation protocols in terms of the energy saving and system lifetime.

Wireless sensor networks are widely deployed for a wide range of applications for gathering information about the environment, monitoring huge building etc. However, the limited energy of the sensor nodes requires efficient gathering of information so that the network lifetime is increased. In literature it is proved that this efficiency can be achieved by clustering the sensor nodes in the network. Previously we proposed a single hop genetic based clustering protocol (GCA) for sensor networks. However, multi-hop routing techniques are known to be a practical approach to solving the problem of routing in sensor networks. In this paper, we present a new robust clustering based multi-hop routing protocol for wireless sensor networks. The proposed genetic clustering algorithm (M-GCA) employs evolutionary techniques to form an efficient virtual backbone. This backbone is used to support routing messages to the base station. Simulation results show that the proposed multi-hop routing protocol outperforms GCA, a single-hop clustering protocol in several scenarios.

This paper proposes a novel resource management framework that integrates Call Admission Control (CAC) and aggregate reservation of bandwidth for mobile networks in a scalable fashion. Our proposal avoids per-user reservation signaling overhead and takes into account the expected bandwidth to be used by calls handed off from neighboring cells within a prediction interval through the Trigg and Leach Method (an adaptive exponential smoothing technique). Our scheme is compared through simulations with the ACR (Adaptive Channel Reservation) scheme, a dynamic reservation-based proposal that uses GPS systems to extrapolate users’ movement and to trigger reservations in the next predicted cell. The simulation results show that our proposal provides the best performance in terms of handoff dropping probability and can achieve similar levels of call blocking probability as compared to ACR. In addition, our proposal can grant an upper bound on handoff dropping probability even under very high loads.