Medium access control protocols for multi-hop wireless ad hoc networks.

摘要

Wireless networking has become an increasingly active research area over the past decade. Recent advances in silicon manufacturing technology have enabled a steady increase in wireless communication capabilities yet still reducing form-factor. For instance, highly integrated, small foot-print, single-chip CMOS radios that can support very high data-rates (up to 480Mbps) are now commercially available [1] and the industry is moving towards multi-gigabit data rates using advanced coding/modulation techniques and new technologies [56,65].;Such technological advances will enable new classes of applications for MANETs that require high application-level throughput and quality-of-service (QoS). Some examples include multimedia streaming, large content transfer, wireless storage area networks, etc. Consequently, the current trend is that new physical layer (PHY) technologies have been moving the fundamental limits challenging development and deployment of high data-rate, QoS-sensitive applications from the PHY to the medium access control (MAC) layer.;On the other hand, these advanced PHY techniques often result in higher energy requirements. For instance, the use of multiple RF chains and advanced decoding techniques to improve PHY performance increase power consumption of the receive operation significantly. To support better energy management, recent radios provide finer-grained power modes that selectively turns on portions of the radio transceiver. Radio power mode control at the MAC layer is thus critical in improving the battery life of the node. The state-of-the-art in medium access for MANETs is far from addressing the requirements imposed by upcoming applications and PHY technological advances.;Our research focus is to bridge this gap and develop a medium access scheduling framework that: (1) is application aware by adapting to changing traffic patterns and satisfying QoS requirements, (2) is self-organizing by adapting to current node state, and connectivity, (3) improves channel utilization and spatial re-use by multi-channel usage and collision-avoidance, and (4) is energy efficient.;The initial motivation for our research on MAC protocols is derived from our work on reliable multicast transport in MANETs. In this work, we introduce the Reliable Adaptive Congestion-controlled Transport protocol, or ReACT, that demonstrates the importance of optimizations at the MAC layer to improve the reliability at the transport layer. ReACT, combines source-based congestion- and error control with receiver-initiated localized recovery. While the latter attempts to recover localized losses (e.g., caused by transmission errors), the former is invoked only for losses and congestion that could not be recovered locally (e.g., caused by global congestion). Loss differentiation is an important component of ReACT and uses medium access control (MAC) layer information to distinguish between different types of losses. Through simulations, we evaluated ReACT's performance and compared it with RALM, a strictly source-based protocol. As our simulation results indicate, significant improvement in throughput and reliability could be achieved by using feedback from the MAC layer to differentiate congestion losses from other losses.;The Traffic-Adaptive Medium Access (TRAMA) protocol [42,43] was the first proposal to implement energy-aware schedule-based medium access. TRAMA reduces energy consumption by ensuring that unicast and broadcast transmissions incur no collisions, and by allowing nodes to assume a low-power, idle state whenever they are not transmitting or receiving. TRAMA assumes that time is slotted and uses a distributed election scheme based on information about traffic at each node to determine which node can transmit at a particular time slot. Using traffic information, TRAMA avoids assigning time slots to nodes with no traffic to send, and also allows nodes to determine when they can switch off to idle mode and not listen to the channel. TRAMA is shown to be fair and correct, in that no idle node is an intended receiver and no receiver suffers collisions. An analytical model [43] to quantify the performance of TRAMA is presented and the results are verified by simulation. The performance of TRAMA is evaluated through extensive simulations using both synthetic- as well as sensor-network scenarios. The results indicate that TRAMA outperforms contention-based protocols (CSMA, 802.11 and S-MAC) and also static scheduled-access protocols (NAMA) with significant energy savings.;FLAMA [41] avoids explicit traffic information exchange and employs a much simpler election algorithm than TRAMA. FLAMA does not require explicit schedule announcements during scheduled access periods. Alternatively, application-specific traffic information is exchanged among nodes during random access to reflect the driving application's specific traffic patterns, or flows. This allows FLAMA to still adapt to changes in traffic behavior and topology (e.g., node failure). FLAMA uses flow information to establish transmission schedules for each node. Additionally, FLAMA achieves traffic adaptiveness by assigning slots to a node depending on the amount of traffic generated by that node. This is accomplished by assigning node weights based on the incoming and outgoing flows. Nodes with more outgoing flows are given higher weights (i.e., more slots); the net effect is that nodes that produce/forward more traffic are assigned more slots.;FLAMA is simple enough so that it can be run by nodes with limited processing, memory, communication, and power capabilities. We evaluate the performance of FLAMA through simulations and test-bed experimentation. Simulation results indicate that, in terms of reliability, queuing delay and energy savings, FLAMA outperforms TRAMA, the first traffic-adaptive, schedule-based MAC proposed for sensor networks, and S-MAC, a contention-based energy-efficient MAC. FLAMA achieves significantly smaller delays (up to 75 times) when compared to TRAMA with significant improvement in energy savings and reliability, demonstrating the importance of application-awareness in medium access scheduling. Our simulation and test-bed results show that FLAMA achieves better end-to-end reliability with significant energy savings compared to S-MAC.;All previously mentioned protocols are designed to work with a single channel. Given that most commercially available radios to-date provide multiple orthogonal channels, protocols should make use of this feature to schedule parallel transmissions within a two-hop neighborhood, thus improving channel utilization. We introduce the Multi-Channel FLAMA (or mFLAMA), that extends the scheduling algorithm of FLAMA to support multiple channels. We compare the performance of mFLAMA with that of FLAMA by simulations, to illustrate the benefit in channel utilization when multiple channels are used for communication.;Finally, we present a new framework for energy-efficient channel access. One of the main features of the proposed framework, or DYNAMMA for DYNAmic Multi-channel Medium Access, is its ability to accommodate and adapt to different application traffic pat terns in an efficient fashion, i.e., minimizing protocol overhead and delivery delay. This is an important contribution as it addresses a drawback inherent to scheduled-access MAC protocols. In DYNAMMA's current implementation, traffic adaptation is done by explicit traffic announcements (in a considerably more efficient way, than existing scheduled-access protocols such as TRAMA [42]). Besides "explicit" adaptation, the flexibility provided by DYNAMMA's framework allows it to accommodate "implicit" traffic adaptation strategies, for example, using learning algorithms, which will further reduce protocol overhead.;We evaluate the performance of DYNAMMA by extensive simulations for different application scenarios. The results from our simulation study shows that DYNAMMA achieves significantly lesser queueing delay than TRAMA and provides high channel utilization and energy savings when compared to TRAMA and 802.11. We also present a generic MAC development test-bed for evaluating scheduled-access MAC protocols using UWB physical layer and we evaluate the performance of DYNAMMA using our FPGA-based test-bed.;In future work, DYNAMMA framework can be used for incorporating traffic prediction to establish the flow information. This can potentially reduce the queueing delay introduced due the scheduling. Another direction of work is to improve the transmission scheduling algorithm presented in the DYNAMMA framework to provide guarantied delivery delays.