Advanced Resource Allocation and MAC Protocols

From Access Wiki

Contents 1 Overview 2 Radio Resource Management for WLAN 3 Mesh Networks 4 Ad hoc and Cooperative Networks 5 Sensor Networks 6 MAC Simulator for Wireless Networks "MACSWIN" 7 PHY-MAC Resource Allocation in Wireless Systems

[edit] Overview

As wireless telecommunication systems evolve, heterogeneous applications are becoming ubiquitous, and QoS is becoming ever more important to those applications. Because they consume so many resources (e.g. bandwidth), such applications need resource management to ensure quality and to regulate resource contention for equitable resource sharing. An RRM entity is the responsible for the correct use of the air interface resources in order to guarantee the quality to the offered services.

For the next generation systems, new and specific radio resource management algorithms are required to guarantee four main goals: maximize spectral efficiency, prevent network congestion, satisfy required QoS, and achieve low signaling. Resource management is a broad-scope concept and includes critical issues and challenges such as power control, handover, call admission control and scheduling.

[edit] Radio Resource Management for WLAN

As for centralized WLAN systems, Radio Resource Management hand-off procedures based on Cross-Layer issues will be studied. Previous results indicate that when adaptive-rates scheduling algorithms are applied in a distributed queues MAC protocol (DQCA), its efficiency boosts in comparison to that without Cross-Layer. The efficiency of the aforementioned protocol has been studied in indoor environments for different types of traffic and QoS requirements. We will focus our research on cellular WLAN systems and we will address the AP association problem. Association with the AP that has the highest RSSI (Received Signal Strength Indicator) is not the appropriate decision factor and cannot always indicate high communication rates in the network due to the overload of the AP while other APs remain unused due to the slightly longer distance for the majority of the nodes. Therefore, new association algorithms should be designed and evaluated. Association decisions will be based on resource availability in each AP, on the resources that have already been assigned and on resources required in order fulfill QoS parameters such as delay and throughput. To be more precise, the RSSI, the position of the node in the data queue of the MAC protocol, the average time in order to get access to the medium and the average transmission time during the last transmitted packets will be combined in cost functions to get appropriate association decisions. Moreover, an analysis of the trade-off between overhead and enhanced performance in the aforementioned association approaches will be done.

It has been proved that throughput limits exist in 802.11-based systems due to MAC and PHY overheads. One of the enhancements in the recently approved 802.11n draft standard is the support of data aggregation techniques. Therefore, frame aggregation permits the encapsulation of more than one MAC frame using a single header. Consequently, the number of overhead packets and interframe spaces are reduced significantly. Aggregated frames can be addressed in single or multi- destination, while single or multi- rates per aggregate frame can be used. Therefore data ACK can be sent immediately after each transmission or in a single delayed group frame. It is evident that new MAC strategies and RRM schemes are needed in 802.11n systems. Although longer frame may achieve higher system’s throughput the maximum length of the frame should be limited. The aim of the aggregation is not to build huge frames but reasonable sized ones that offer relevant fairness to the system. In any case, the maximum frame length should be limited by the ARQ window. On the other hand, the minimum length should be also longer than a threshold. The consideration of QoS measures such as throughput, delay, data rate and effective bandwidth increases the necessity for the design of efficient access strategies. The thought of finite or dynamic data queue length poses additional features in the problem. The combination of aggregation mechanisms with beamforming, antennas selection schemes and advanced cross-layer PHY-MAC techniques can further increase the efficiency of the systems. On the other hand, there is a trade-off between the efficiency of the aggregation mechanisms and the robustness and complexity of the system that should be also taken into account in the design of the algorithms.

[edit] Mesh Networks

Despite recent advances in wireless mesh networking, many research challenges remain in all protocol layers. Our intention is the development of advanced Data Link Control protocols and resource management algorithms in order to guarantee the QoS of wireless mesh networks. We will address both single channel and multi-channel scenarios. The research work will start considering single channel protocols and we will address two approaches. Firstly existing MAC protocols based on the 802.11s will be further evaluated. Secondly, innovative self-organized MAC protocols based on distributed queues will be proposed. As far as multi-channel mesh networks are concerned our research will be focused on multi-channel multi-transceiver and on multi-radio networks. In multi-channel multi-transceiver a MAC protocol coordinates several simultaneous channels supported by baseband modules. The main topic that we will address includes negotiation and selection of the available channels based on certain criteria such as maximum throughput and fairness among the users. Furthermore, attention will be given to the exposed nodes problem produced in this type of networks. In multi-radio networks a MAC protocol in order to manage the communication among the different radio technologies is required. Critical issues such as neighbor discovery, switching channels, and hidden nodes will be considered.

[edit] Ad hoc and Cooperative Networks

In ad-hoc networks, work will be based on an already defined architecture (DQMAN) that contains a MAC protocol with distributed queues and a clustering algorithm. Nodes are divided into either disjoint or overlapping clusters in order to operate in a synchronized manner in a hierarchical architecture. Hence, master-slave architecture is defined whereby nodes share the responsibility of pretending as access points. Preliminary results obtained by computer simulations are very promising in a relevant range of scenarios, so it can be considered as a good starting point for research. Moreover, it has arisen that some aspects related to the clustering mechanism can be improved in order to increase the overall performance in terms of throughput and delay.

We have remarked that clustering overlapping may cause difficulties in the spatial channel reuse and nodes out of the coverage of a cluster head cannot communicate with a slave node even it lies in its transmission range. Our research will try to solve these problems by proposing suitable solutions for each one of the previous problems such as a three-hop master slot mechanism, an active listening scheme and a master cooperation request scheme. To improve the overall performance, re-clustering can be periodically triggered in order to minimize the presence of blocked nodes and nodes out of coverage. The re-clustering algorithm should ensure fairness when sharing the Master's responsibilities among all nodes in the system. Furthermore, analysis of the trade-off between overhead and enhanced performance will be done. Our ongoing work addresses specific multihop scenarios such as tandem and lattice, while our future work will include multihop scenarios. Both simulations and analytic results will be presented.

Node cooperation in wireless networks is mainly motivated by the broadcast nature of the wireless medium. Due to the fact that every network node within the coverage area of a particular source can listen to its transmission, those nodes could help to improve the system’s efficiency. Significant effort and huge amount of innovative results for cooperation in the physical layer have been presented in the literature but few MAC protocols have been developed to exploit the advantages of cooperation. In that sense, the Multiple Relay Access Control (MRAC) protocol, called Relay Carrier Sensing Multiple Access (RCSMA), for ARQ cooperative scenarios that are formed by a single transmitter and a group of faraway located potential receivers has been proposed and initial results have been obtained. It is worth mentioning that a wide range of applications in growing interest are characterized by low SNR values between transmitter and receiver, such as sensor networks, heterogeneous 4G networks, energy-constrained ad hoc networks etc.

Our future work deals with the MAC optimization, in terms of maximizing throughput and minimizing the delay, as well as the promotion of efficient power consumption mechanisms for the proposed cooperative scenario. The RCSMA MRAC protocol will be extended in order to cover various schemes of cooperative relaying and various options for relay nodes (decode, amplify and forward, modify the information…). The previous scenarios will be studied both with fixed and mobile relays. Furthermore, schemes that encourage mobile users to share their resources in order to cooperate will be proposed and evaluated.

Security is an important enabling technology aspect for cooperative networks. It is usually treated as a separate issue from cooperative communications. However, security can be used as one way to provide incentives and guarantees for cooperative systems. In the recent years, much of the research has been focused on networking aspects of cooperative communication systems. On the other hand, significant works have been presented in the literature in MAC misbehavior for 802.11-based systems. However, it is not adequate for cooperative MAC protocols due to the protocols´ operation. In such a kind of systems malicious or misbehavior users can take profit from the protocols’ operation to monopolize the channel and to waste the battery of the other users. In our work, we will classify security threats for cooperative MAC protocols and subsequent security requirements. Next we will enhance the previous works in order to improve the security in some of the proposed in the literature MAC protocols for IEEE 802.11-based cooperative systems.

[edit] Sensor Networks

Because of the distinguishing characteristics of Wireless Sensor Networks, the required algorithmic and technological solutions differ significantly with the respect to ad hoc networking. New algorithms and protocols should be designed and studied. Since currently the IEEE 802.15.4 framework appears to be the most suitable protocol for wireless sensor networks, we will further analyze this standard. The main areas of research will be the design of a MAC protocol that will contain power management techniques, cross-layer definitions and scheduling mechanisms. Quality of service will also be considered in order to enable a far more efficient performance of applications using wireless sensor networks. Further, some means to allow sensors to transmit to multiple sensors at the same time (i.e. multicasting) at different rates and to increase their data throughput without putting their battery life at risk will be analyzed. Moreover, self-configuration and interference mitigation challenging topics within wireless sensor networks will be studied and adapted to the proposed MAC protocol model. Our proposed solutions will be optimized to cover specific issues of target scenarios and the relevant applications. We will focus our work on Body Sensor Networks and we will analyze the interaction with Ambient sensing networks. For that purpose, the optimization of the presented MAC protocol enhancements will include an especially designed MAC functional definition for the efficient management of channel resources and the overall minimization of energy consumption in order to prolong their battery lives. Moreover, for such a kind of applications, secure key management schemes are going to be introduced to guarantee data integrity and authentication among sensors for a real efficient use of the multiple receiver solutions.

[edit] MAC Simulator for Wireless Networks "MACSWIN"

The MAC Simulator for Wireless Networks (MACSWIN) is an object-oriented (OO) platform programmed in Visual C++ .NET. Any MAC protocol can be easily implemented by modifying a very specific class of the whole architecture. The simulation core has been integrated with a graphic user interface (GUI) with a twofold objective:

• First, all simulation parameters required to define the network or scenario to be simulated can be introduced by the user in a user-friendly interface.

• Secondly, The simulation can be visualized by using a user-defined time scale. Continuous observation of a simulated network can be helpful in understanding the run-time operation of the system and in designing mechanisms to improve the network performance. To do so, a full time control is a must. In MACSWIN simulation time can be stopped and resumed as needed in order to keep track of a protocol operation’s details.

The simulator is expected to integrate all the advanced algorithms that will be studied in the framework of the Resource Management and MAC Protocols research line:

• Advanced clustering schemes for the DQMAN.
• Additional cooperative scenarios with MRAC.
• The cooperative version of DQMAN.
• The DQCA MAC protocol for infrastructure-based WLAN.
• All the Radio Resource Management algorithms based on Cross-Layer schemes that we studied during the last two years.
• WLAN cellular (multi-AP) scenarios for hybrid networks deploying both DQCA and 802.11 MAC protocols.
• As a reference scenario, the MAC in the 802.11 standard will be integrated, including both the 802.11e and the recently approved draft of the 802.11n PHY descriptions.

In addition, the incorporation of specific application scenarios such as military communications, vehicular communications, sensor networks, search and rescue operations, office environments is seen as an essential tool for network designers and operators.

[edit] PHY-MAC Resource Allocation in Wireless Systems

The problem of resource allocation in multi-user MIMO wireless networks has been a wide topic of research in the recent years. For instance, it is well known that the Successive Interference Cancellation (SIC) coding/decoding techniques are optimal and maximize both, uplink and downlink multi-user MIMO channel capacities. Unfortunately, it is in general assumed that their implementation in practical systems is rather difficult because of their high complexity and therefore, sub-optimal less complex solutions have been also investigated. As an example, the waterfilling solution is known to be the optimal power allocation in the MIMO broadcast channel with a Zero Forcing beamformer. In this case, the optimal solution is given by performing waterfilling power allocation over all the possible sets of users and choosing those users that present the maximum Shannon's information-theory sum rate.

Clearly, a scheduling problem comes into play because an optimal user set selection is to be performed. Typically such user set selection is carried out through an exhaustive search because of the non-degenerative properties of the MIMO channel, opening the door to investigate suboptimal user selection (or scheduling) algorithms. Furthermore, since scheduling implies that not all users are served simultaneously some users could be unfairly served and thus, requirements in terms of fairness and QoS arise.

In consequence, a completely information-theoretical (PHY layer) resource allocation problem has been inherently converted into a more intricate cross-layer resource allocation problem where resource allocation implies PHY layer issues such as power control and multi-antenna processing and also MAC layer issues such as scheduling and resource access control.

The three main objectives for this research line are:

• To develop a general framework for PHY-MAC resource allocation in MIMO wireless networks.
• To investigate optimal PHY-MAC space-time scheduling strategies
• To develop a real-time resource allocation testing platform focused on WLAN and WIMAX technologies.