© 2008 OPNET Technologies, Inc.
1. Cognitive Radio Networks: MAC Layer Sensing Issues
With the J.Mitola's PhD Dissertation (2000) entitled as "Cognitive Radio: An Integrated Agent Architecture for Software Defined Radio", the idea of intelligent, reconfigurable and adaptive radios has become an attractive alternative for wireless communications. In the widest sense, cognitive radio (CR) communications is a new paradigm which proposes the use of the spectrum band dynamically depending on the requirements/needs of the application and the users, and can make some autonomous learning and adaptation activities to have an efficient communication. In a more realistic and simplistic way, CR facilitates the use of spectrum more dynamically and so in a more efficient way.
The research challenges are numerous, though not very different from the ad-hoc or sensor communications. The first issue is the determination of frequency opportunities in which cognitive users (also called secondary users) can transmit as long as the primary license holders appear. This problem is called Primary user detection. Spectrum Sensing is applied to determine the existence of a primary signal in the band of interest. Spectrum sensing is based on the foundations of statistical signal analysis and detection/estimation theory. Hence, it is mostly considered as a physical layer issue. However, it is not independent of other protocol layers. Simply, sensing is done in time slots specifically allocated to it. Therefore, it can also be considered as a MAC layer issue. To put in another way, allocation of specific time slots for spectrum sensing requires planning in the MAC layer. Basic considerations are channel sensing order, sensing slot allocation planning. These problems are generally called MAC layer sensing problem.
In this research, MAC layer sensing planning considering the channel characteristics (more specifically: PU traffic characteristics) and cognitive radio limitations (i.e. single antenna, energy limitations) is tthe main focus with the aim of designing cognitive network friendly MAC protocols. Briefly, effects of predicting the secondary traffic characteristics in channel sensing and swiching are examined in our work.
2. Reliable Data Delivery in Wireless Sensor Networks
Congestion is a major source of data loss in wireless sensor networks which adversely affects the reliable data delivery and consequently the success of the sensor network application. In event triggered applications, regarding the dense deployment of sensors, many sensors detect the same event and create burst of reports destined to the sink which creates transient local congestions in the network. Moreover, shortest distance based classical single path routing schemes increase the severity of the congestion by concentrating the traffic on some preferred sensors, resulting with an unbalanced traffic distribution in the network. We worked on two cross-layer geographic forwarding schemes which aim to increase the reliable data delivery in wireless sensor networks by avoiding congestion using distributed and dynamic load balancing approaches. The main characteristic of these schemes is the utilization of the buffer occupancy levels of the neighbor sensors in the decision of data delivery. Both schemes utilize a modified version of SMAC where the packet structure as well as the operation of SMAC is modified for the accurate monitoring of the buffer occupancy conditions of the neighbors. The first scheme try to provide local load balancing where a sensor dynamically determines the next hope among the alternative neighbors providing positive advancement towards the sink by considering the balance of their buffer occupancy levels at the time of delivery. The second scheme combines local and direction-based load balancing approaches to provide more reliable and faster data delivery by benefiting from the advantages of the both approaches. The performance of the forwarding schemes are compared using simulation in OPNET with two geographic routing schemes where one applies no load balancing and the other applies direction-based load balancing. The results show that both schemes provide more reliable data delivery as compared to other schemes, whereas the second scheme is more reliable and faster as compared to the first scheme by benefiting from both local and spatial load balancing approaches. In addition, the second scheme provides more energy efficient video delivery in terms of energy expenditure per successfully delivered data unit to the sink as compared to the first scheme and the other two schemes.
3. QoS-aware MAC Protocol for Wireless Multimedia Sensor Networks
Wireless Multimedia Sensor Networks (WMSNs) have entered the class of Wireless Sensor Networks (WSNs) to meet the multimedia requirements of new emerging applications, such as surveillance and telepresence. Combining the traditional scalar sensors used in WSN nodes with tiny cameras and microphones, WMSNs commonly carry heterogeneous traffic with different quality of service (QoS) requirements. In order to deliver heterogeneous traffic with different requirements in highly resource constrained sensor networks, QoS provisioning and service differentiation become unavoidable. In our work, we discuss the QoS provisioning in sensor networks and evaluate the efficiency of existing QoS-aware MAC protocols. As a result of this evaluation, we design and implement a QoS-aware MAC protocol for WMSNs, Diff-MAC, which integrates different methods to meet the requirements of QoS provisioning to deliver heterogeneous traffic and provides a fair all-in-one QoS-aware MAC protocol. Diff-MAC aims to increase the utilization of the channel with effective service differentiation mechanisms while providing fair and fast delivery of the data. Performance evaluation results of Diff-MAC, obtained through extensive OPNET simulations, show significant improvements, in terms of latency, data delivery and energy efficiency, compared to two other existing protocols. Implementation of Diff-MAC on Imote2 platform also reveals that the protocol with moderate complexity can be easily implemented on the resource constrained motes.
4. Lifetime Optimization for Heterogeneous Networks Using Concentric Rings
Lifetime constraint is the limiting feature for Wireless Sensor Networks. Since the data dissemination in sensor networks is towards the sink, the data communication is realized by the relay of the packets of the intermediate nodes in wireless sensor networks. As a result the nodes with high packet relay burden are the ones to go offline first, which are the nodes close to the sink. Moreover; terminations of these nodes means also that, some nodes depend on them to carry their traffic get disconnected, which defines the end of the useful lifetime of the network.
In order to alleviate the non-homogeneous energy drainage pattern, in this thesis study we propose to use different energy capacity sensor nodes and deploy them in a concentric ring shaped areas. The nodes with high battery capacity are located near the sink and those nodes with the weakest capacity are located in the outmost ring. The overall performance of the network will be then maximized and nodes arrive to end of their lifetime at about the same time.
For verifying the obtained results, simulations were conducted on OPNET Modeler. The proposed network were modeled and simulated using the wireless module of OPNET. Different scenarios were devised and the lifetime of the network were compared.
5. Application Specific Lifetime Measurement for Wireless Sensor Networks
Network Lifetime is the most important performance metric in evaluating the Wireless Sensor Network designs. Existing trend is to measure the lifetime with simplistic metrics, such as the time till the first node death. However, almost every aspect of the WSN design should be considered in an application dependent context and lifetime measurement is no exception. In this work, we propose a general framework to enable the network lifetime measurement be carried out in an application dependent manner. We show, through OPNET simulations, that metrics that do not reflect the requirements of the application lead to incorrect lifetime results and that with our proposed framework lifetime can be measured more realistically.
6. Quality of Service in Wireless Metropolitan Area Networks
As an wireless alternative to xDSL and Cable modem systems, the importance of WMAN systems has increased significantly in the recent years. The most important standard in WMAN systems is IEEE 802.16 or WiMAX. IEEE 802.16 is expected to provide detailed QoS support to support different application types. The basis operating mode of WiMAX, the point-to-multipoint (PMP) mode has a detailed QoS mechanism. On the other hand, the second operating mode, the mesh mode, only provides a simple QoS management mechanism. We developed a Mesh mode model of IEEE 802.16 using the centralized scheduling method using OPNET. With this model we are developing more detailed QoS mechanisms for the Mesh mode of IEEE 802.16 and provide performance evaluations for these mechanisms. So far, we have developed a method for incorporating the QoS mechanism of the PMP mode of IEEE 802.16 to the centralized scheduling method in Mesh mode. Our current task is to extend this method to the distributed scheduling method in the Mesh mode.
7. QoS with OFDM for B3G Wireless Systems
OFDM is chosen as a comprehensively acknowledged kernel technique for implementation of physical layer of B3G systems due to its capability of high data rate transmission in multi path environment, high spectrum efficiency, and simple implementation. Utilization of orthogonal frequency division multiplexing technology for B3G may avoid intercarier interference, and increase system resistance to channel impairments thus provides better and adaptive QoS support. The main objective of the proposed work is to come up with a model for B3G wireless that utilizes OFDM and provides adequate and adaptive QoS support for different services with different needs. To model the architecture we propose, we have chosen OPNET modeler. We will use MPLS capabilities of OPNET at core network model, and develop our own physical and mac layers for air interface extending existing wireless models. When the model is fully implemented over OPNET, the strengths and weaknesses of the model will be examined with simulation and analysis tools of OPNET.
8. Packet Traffic Modeling for Intrusion Detecting WSN
Since sensor nodes have limited battery capacities, research on wireless sensor networks (WSN) focus on energy-efficient algorithms. However, potential application areas of WSN show very different characteristics that obstructs development of common algorithms that fits to all potential application areas. For instance, military applications may need very fast responding WSN protocols, whereas precision agriculture protocols may trade-off these delay sensitivity with energy conservation. Likewise, because of the different application characteristics, although a proposed communication protocol performs in a very energy-efficient way when used for an application, it may performs poorly in another one.
One reason for these differences is the data traffic generated by the WSN application. When communication protocols are developed without considering properties of the data traffic generated by the application, they may behave inefficiently. For instance, in the WSN literature, the performance evaluations of the proposed protocols are generally done with a constant bit rate (CBR) data traffic. Moreover, certain types of variable bit rate (VBR) data traffic is also used such as Poisson. However, event-driven applications such as target detection and tracking generate bursty traffic which cannot be modeled as either CBR or Poisson. Therefore, one needs to define a realistic packet traffic model for event-driven applications to model these applications successfully.
In this work, we investigate the packet traffic model generated by intrusion detecting WSN systems since intrusion detecting WSNs are widely used but are lack of a realistic packet traffic model. We hope that this work will be a starting point for packet traffic models of other event-driven WSN applications, also. Moreover, we plan to show how a realistic model makes a difference by comparing the performance results of a WSN MAC protocol for three different packet traffic models using simulations.
9. Handover Management Algorithms in LEO Satellite Communication Networks
One of the major problems with LEO satellites is their higher speed relative to the terrestrial mobile terminals, which move at lower speeds but at more random directions. This high mobility characteristic of LEO satellites causes mobile users to hand over between footprints of adjacent satellites very frequently. Therefore, handover management in LEO satellite networks becomes a very challenging task for supporting global mobile communication. The thesis discusses the basic handover management schemes used in LEO satellite networks and points out their drawbacks. By taking into consideration the inconvenience and incompleteness of the current proposals in literature, this work proposes a new mobility management scheme for LEO satellite communication networks.
10. VoIP Performance over Satellite Networks with On-Board Processing Capabilities
In this study, an adaptive routing policy utilizing the real-time network information of a two-layered satellite network is introduced. In a satellite network, depending on the requirements and properties of services provided, various kinds of satellites from different orbits can be employed. Geostationary Earth Orbit (GEO) systems are not suitable for Voice over Internet Protocol (VoIP) applications due to long end-to-end delay values about 250-270 ms. Non-Geostationary Earth Orbit (NGEO) systems consisting of Low Earth Orbit (LEO) and Medium Earth Orbit (MEO) satellites can satisfy the performance requirements of VoIP applications. Moreover, a two-layered system of LEOs and MEOs can outperform single plane satellite networks. However, due to the dynamic topology of these networks and nonuniform traffic distribution over the Earth, terrestrial packet based routing algorithms cannot perform well. The proposed routing scheme dubbed as Adaptive Routing Protocol for Quality of Service" (ARPQ) prevents the congestion on some bottleneck links by distributing the traffic over the entire network. Furthermore, link capacities can be efficiently used. Additionally, delay and jitter sensitive voice traffic is processed in a prioritized way to prevent long queueing delays.
By a set of simulations, we showed that proposed mechanism performs better than nonadaptive routing mechanisms and therefore can enable VoIP applications over satellite networks.We used OPNET Modeler 10.5 A to model our scenarios. OPNET Wireless Module provides extensive simulation and modeling tools. Using the Transceiver Pipeline, link properties (e.g. delay, bit error rate) can be controlled. Link properties are modeled using different settings at each pipeline stage. Random error and burst error models are simulated. Various performance parameter statistics are collected and analyzed.
11. Service Differentiation for QoS-aware Multicast Routing in Mobile Ad Hoc Networks
Recent developments in wireless mobile telecommunications and portable computing devices, as well as the increasing importance of multimedia applications on the Internet have stimulated a conceptual shift in users' expectations towards group-oriented, mobile multimedia communication. For this new generation of networks, mobility, quality of service (QoS) support, and multicast routing are essential components. To this end, the ad hoc QoS multicast (AQM) routing protocol has been developed recently (see below). AQM covers the admission control, resource reservation and routing aspects of the ad hoc QoS multicast problem. At this point, it is necessary to combine AQM with priority assignment and fair scheduling mechanisms, such that the final design also covers the service differentiation aspect of QoS.
The purpose of this work is to design a service differentiation component for AQM. The protocol in its final form will differentiate between QoS and best-effort traffic coexisting in the network, and apply intelligent queuing and scheduling strategies to meet the QoS requirements of both real-time multimedia as well as bursty data flows. Performance metrics such as throughput, delay, packet delivery rate, overhead and multicast efficiency will be measured. It is strongly recommended by IETF MANET that these performance metrics should be considered in the experiments during the evaluation of the proposed schemes. Simulations will be based on realistic network dynamics and mobility patterns. The simulations will be conducted on the OPNET Modeler, which is a verified and widely accepted network simulator with many ad hoc networking capabilities, in order to evaluate the performance of the final QoS multicast routing protocol.
12. Quality-of-Service-Aware Multicast Routing for Multimedia Applications in Mobile Ad Hoc Networks
The conceptual shift in the expectations of wireless users from voice towards multimedia, from availability towards acceptable quality, and from stand-alone towards group-oriented computing has a significant impact on today's networks in terms of the need for mobility, quality of service (QoS) and multicast routing. Ad hoc networks, being independent of any fixed infrastructure, can provide mobile users with these features. However, it is imperative for ad hoc networks to combine QoS and multicast routing strategies in order to utilize the wireless medium efficiently.
This work defines the ad hoc QoS multicast routing (AQM) protocol, which achieves multicast efficiency by tracking the availability of resources for each node within its neighbourhood. Computation of free bandwidth is based on reservations made for ongoing sessions and the requirement reported by the neighbours. Current QoS status is announced at session initiation and updated periodically to the extent of QoS provision. Nodes are prevented from applying for membership if there is no QoS path for the session. When nodes wish to join a session with certain service requirements, a three-phase process consisting of request, reply and reserve steps ensures that the QoS information is updated and used to select the most appropriate routes. The allowed maximum hop count of the session is taken into account in order to satisfy the delay requirements of the multimedia applications. To cope with the continuous nature of streaming multimedia, AQM nodes check the availability of bandwidth within their neighbourhood not only for themselves but within a virtual tunnel of nodes. Objection queries are issued prior to reservation to avoid excessive resource usage due to allocations made by nodes which cannot detect each other directly.
New performance metrics are introduced to evaluate the efficiency of AQM regarding the satisfaction level of individual members as well as the success rate of sessions. The simulations are based on realistic membership dynamics and mobility patterns. They are conducted using the OPNET Modeler, which is a verified and widely accepted network simulator with many ad hoc networking capabilities. Simulation results show that, by applying novel QoS management techniques, AQM significantly improves multicast efficiency for members as well as sessions.
13. Quantifying Energy Saving using Multi-Hop Links in Wireless Sensor Networks
E. Ilker Oyman
Wireless sensor networks consist of small sensor elements which are distributed randomly over the area under investigation. Sensor nodes are equipped with very limited computational resources and battery power. Therefore, the underlying network architecture must provide with power aware techniques as a primary design constraint.
In this work, we present the use of multi-hop communication links. The gain in total energy is calculated using analytical techniques. We show that shorter links must be employed whenever possible, in order to prolong the network lifetime. The model is built on the OPNET Modeler. The sensor network consists of several sensor nodes and one destination node. Each sensor node is deciding on the output power level of its transmitter circuitry and therefore is adjusting its transmission range and hence its power consumption. According to scenarios, the amount of energy saving is calculated, and the expected network lifetime is compared among the scenarios.
14. Large Scale Multiple Sink Wireless Sensor Network Design
E. Ilker Oyman
Wireless sensor networks consist of small sensor elements which are distributed randomly over the area under investigation. After deployment of sensor nodes to the environment, the network is organized so that the data can be propagated to the control center.
Each sensor node senses the environment and extracts the necessary data. Thereafter, this data is sent to the next sensor node which is responsible to forward the data packet towards the control center. In large scale networks, the network should be divided into smaller subnetworks to achieve scalability. At the center of these subnetworks, some special nodes should be used as concentrators. The sensors redirect their data packets to these concentrators. These nodes are mainly collecting the data from the sensors, and forward them to the control center. They are powerful gateway elements which can communicate with sensors, with each other, and also with the global network.
This work focuses on finding locations of concentrators on large scale static wireless sensor networks. The objective is to minimize the number of concentrator nodes as their cost is much higher than sensor nodes. On the other hand, they must be sufficient to maximize total network life optimizing power usage of the sensor nodes. We have implemented a self-organizing map approach to this well known concentrator location problem using the OPNET Modeler. The results achieved by our network simulations are proven to be of high quality.
15. Routing In A LEO Network Using Min Flow Max Residual Path In A Routing Set
Low Earth Orbit (LEO) satellite networks will be widely used within the next few years. There are many satellite projects and theoretical studies being worked on. Many of them deal with power consumption, orbit selection, inclination angle, etc. However, the routing/switching of these communication paths over the satellite network is a very important issue which will result in higher utilization and less congestion on the satellite networks.
In this work we designed a LEO network routing protocol that will find a more reasonable path for a packet on the network using Min Flow Max Residual (MFMR) path in a boundary called Routing Set (RS). The network is modeled as a finite state automaton. The satellite network has a finite number of states that are countable and nodes are assumed to be stable during a state. The routing information is relative to the satellite position.
The proposed network and protocol are modeled using the OPNET Modeler. The nodes are capable to discover their surrounding nodes and their links. They are aware of the network architecture and the flows on each link. The routing of the packets or the path selection for a circuit is done by choosing the min flows and then max residual capacity on the candidate paths. As a result of the project, max flows on the links and the packet delays are measured and reported for different scenarios.
16. Fuzzy Logic Based Congestion Control for Wireless Multimedia Sensor Networks
Congestion is one of the important issues that should be handled while transporting video data over Wireless Multimedia Sensor Networks (WMSNs). It is occurred when the data traffic exceeds the available capacity of network. In WMSNs congestion causes buffer drop, packet loss and delay increase. Since it directly degrades the QoS, it should be mitigated. In wired networks congestion control is usually done with end to end manner, and congestion decision is done with respect to the packet drops and increased delays. However in WMSNs end to end approach is not suitable. Because of the multi hop topology of sensor networks, packet drops and increased delay are not an only indicator for congestion.
In this research, a fuzzy logic based congestion estimation model is proposed. In WMSNs there are several factors which can cause congestion. It is difficult to say that the network is congested or not. As a result fuzzy logic is a good solution for estimating and mitigating congestion. In fuzzy logic based systems there should be inputs in order to construct fuzzy inference system. These inputs are very critical since they directly effect the congestion estimation result. Fuzzy inputs may be provided by different layers of the sensor node’s communication protocol. Gathering values from different layers requires a cross layer solution. The following values are the most common inputs for the fuzzy logic congestion estimation models;
For data loss recovery, transport protocols generally use ACK/NACK based models. Also they use forward error correction (FEC) to recover corrupted packets. However in our case, neither ACK/NACK based model nor FEC method is applicable. In VWSNs, where nodes are equipped with cameras and can transmit video data, 100% reliability is not necessary. Recovering lost packets with ACK/NACK messages increases the newer packets delay, an d video frames may received by the sink unacceptably late. Also sensor nodes have limited battery, so protocols should be implemented as simple as possible. FEC or similar error correction mechanism brings an extra overhead to the sensor nodes. As a result in this work such data loss recovery mechanisms will not be implemented.
1. CmpE 523 - Performance Evaluation of Computer Networks
Introduction for computer networks performance evaluation. Modeling of traffic flows. Delay and loss models for computer networks. Networks of queues. Performance evaluation of multiple access methods and local area networks. Measurement and simulation of computer networks.
Instructor: Cem Ersoy, firstname.lastname@example.org, http://www.cmpe.boun.edu.tr/~ersoy
2. CmpE 524 - Computer Network Design
Principles of computer networks design, network design algorithms, centralized network design. Application of minimum spanning tree and shortest path algorithms to problems in network design. Static and dynamic routing algorithms. Network reliability analysis. Linear and integer programming techniques. Heuristics, evolutionary algorithms. Distributed network design. Case studies: Multimedia network design, multihop lightwave network design, cellular network design, satellite network design, ad hoc network design problems.
Instructor: Cem Ersoy, email@example.com, http://www.cmpe.boun.edu.tr/~ersoy
3. CmpE 470 - Computer Performance Evaluation
Systematic approach to computer performance evaluation. Measurement techniques and tools. Analytical techniques. Queueing models. Simulation. Experimental design and analysis. Case studies.
Instructor: Cem Ersoy, firstname.lastname@example.org, http://www.cmpe.boun.edu.tr/~ersoy
4. CmpE 591 - Broadband Wireless Networks
This course aims introducing concepts and research topics in emerging wireless broadband networks. We will cover wireless networks ranging from WLANs to WiMAX, Cognitive Radio and molecular communication networks with emphasis on mobility, quality of service (QoS), and seamless operation. Most of the course will focus on Cognitive Radio Networks. We will also learn how much networks can be analytically modelled
Instructor: Tuna Tugcu, email@example.com, http://www.cmpe.boun.edu.tr/~tugcu
5. CmpE 58C - Wireless Sensor Networks
This course aims introducing concepts and research topics in wireless sensor networks. We will cover topics ranging from MAC, routing, and transport layer protocols to topology, sensor/actuator networks, and new research topics in the field.
Instructor: Tuna Tugcu, firstname.lastname@example.org, http://www.cmpe.boun.edu.tr/~tugcu
6. CmpE 581 - Communications Engineering for Mobile/Wireless Networks
Special topics selected to reflect recent technologies and trends in communication networks as dictated by faculty member expertise and students' interests. All seven-layer of computer communication networks will be explored for the MS and PhD students with no background in communications & networks. Course objectives and related program educational objectives are: 1. To be familiar with the wireless/mobile market and the future needs and challenges, to apply mathematical models of radio wave propagation, to strengthen students' knowledge in mobile/wireless communication systems, to design mobile/wireless communication systems, to develop students' writing and research skills.
7. CmpE 582 - Satellite Networks
Special topics selected to reflect recent technologies and trends in Satellite Networks as dictated by faculty member expertise and students' interests. All seven-layer of Satellite Communication Networks will be explored for the MS and PhD students. Course objectives and related program educational objectives are: To be familiar with the satellite networks market and the future needs and challenges, to apply mathematical models of satellite networks, to strengthen students' knowledge in satellite communication systems, to design satellite communication systems, to develop students' paper writing and research skills.
NETLAB, August 3, 2011
OPNET Technologies, Inc. is a leading provider of solutions for managing networks and applications. OPNET's best-in-class solutions address application troubleshooting, application monitoring, network monitoring, network configuration management, network planning , and network simulation. OPNET’s solutions have been operationally proven in thousands of customer environments worldwide, including corporate and government enterprises, government and defense agencies, network service providers, and network equipment manufacturers. For more information about OPNET and its products, visit www.opnet.com.