The variation of bandwidth in wireless networks imposes significant challenges to the operation of congestion control protocols, especially, those relying on estimations of link bandwidth. For example, TCP CUBIC probes the end-to-end available link bandwidth while XCP and VCP require an explicit knowledge of the available link bandwidth at intermediate nodes. Thus, these protocols are subject to oscillatory behavior and serious performance deterioration in wireless networks without properly compensating against the fluctuations of the bandwidth. In this paper, we propose a bandwidth estimation scheme utilizing Extended Kalman Filtering (EKF) to which we refer as EBE. Rather than directly measuring bandwidth in real-time, EBE monitors either per flow states at a sender or persistent queue sizes to predict the available bandwidth. It is utilized to efficiently cope with bandwidth variations thereby stabilizing and improving the performance of congestion control protocols in wireless networks. We implement EBE in NS2 and integrate it with XCP, VCP, TCP CUBIC, and few other TCP variants. Through extensive simulation studies, we demonstrate significant performance improvements of these protocols in wireless networks as the result of using EBE.

In this paper, we present UAV-aided Cross-Layer Routing Protocol (UCLR) that aims at improving the routing performance of a ground MANET network with aid from an Unmanned Aerial Vehicle (UAV). The UAV is added to a connected backbone formed by a collection of designated nodes in order to combat link failures and wireless link effects detected at PHY/MAC layers before the routing table adapts to the changes. In the context of the UCLR protocol, we introduce a UAV-aided cross-layer routing scheme, an associated cross-layer routing metric, and a UAV load-balancing algorithm. We implement UCLR using Linux Quagga routing suite along with OSPF MANET Designated Routing (MDR) and demonstrate its performance improvements compared to the original MDR through emulation studies.

Router Assisted congestion control Protocols (RAPs) appear to be the most efficient solutions to the TCP performance degradation issue in high Bandwidth Delay Product (BDP) networks. Global deployment of RAPs such as XCP, VCP, and MPCP however, has been challenging due to their need for router support. In this paper, we propose RAPPEP a framework for deploying RAPs on potential congestion zones such as satellite links that are locally utilizing the architecture of TCP Performance Enhancement Proxy (PEP). Such a marriage allows for an immediate deployment of RAPs without the need for global router support, while still being able to take advantage of sophisticated RAPs. Beyond the deployed congestion zone, RAPPEP is completely transparent to the rest of the network including end nodes and other routers. Adapting from two implementations of RAPs and an implementation of TCP PEP (PEPSal), we implement and integrate RAPPEP in the Linux kernel and demonstrate its performance improvement compared to PEPSal through emulation studies.

Aside from the introduction of many new schemes, the use of TCP-based AQM schemes and in specific RED is anticipated to continue in foreseeable future as the de-facto standard of network congestion control. Therefore, conducting extra research work aiming at improving the performance of RED is still a topic of high interest. In this paper, we present an analytical study aiming at the fine tuning of the RED parameters. Utilizing a statistical analysis approach, we formulate an optimization problem aimed at addressing the loss and delay tradeoff of the RED queuing discipline. We provide a two-phase iterative solution to the problem in order to identify the settings of the RED parameters. We discuss the convergence characteristics of our solution and investigate its low complexity characteristics. Through extensive NS2 experiments, we illustrate the advantages of our proposed optimization approach by comparing its results to those of adaptive RED as well as standard RED with recommended parameter settings.

Wireless link effects impose unique challenges to routing protocols in Mobile Ad Hoc Networks (MANETs). In this paper, we propose Hybrid Cross-Layer Routing (HCLR) protocol designed and implemented based on the premise of being able to leverage a pair of proactive and reactive routing schemes. While the former is used to avoid long routing discovery latencies, the latter is used to compensate against wireless link effects thereby allowing for making more intelligent routing decisions. Compared to standalone proactive or reactive routing protocols, HCLR provides a two-tier improvement logic by performing local on-demand search in an n-hop neighborhood utilizing a 2-tuple cross-layer routing metric. We implement HCLR in Linux Quagga suite and demonstrate its performance improvements compared to OSPF MANET Designated Routing (MDR) through emulation studies.

The performance of prevalent wireless MAC protocols is a function of network contention level and the capabilities of the underlying network nodes. While contention-based MAC protocols such as CSMA experience significant performance degradation under high contention levels, slot-based MAC protocols such as TDMA perform in the opposite way. In this paper, we propose a hybrid MAC protocol to which we refer as Load-Adaptive MAC (LA-MAC) protocol for MANETs formed by a collection of Multiple-Input Multiple-Output (MIMO) equipped nodes. By adaptively switching its mode of operation between CSMA and TDMA, LA-MAC achieves a high channel utilization and a reasonable delay profile under both high and low contention levels identified by the measured collision rate of the MAC frames. As a cross-layer protocol, LA-MAC relies on the information received from the physical layer in order to distinguish between corrupted frames due to actual collisions as opposed to those corrupted by wireless link effects such as fading. We describe the design of LA-MAC protocol and report on its implementation in a MANET testbed formed by a collection of MIMO Universal Software Radio Peripheral (USRP) Software Defined Radio (SDR) nodes. Through analytical and experimental studies, we compare the performance of LA-MAC with that of CSMA and TDMA under different traffic conditions. The results show the improvements achieved by LA-MAC in comparison with other alternatives.

Many of the recently proposed high performance congestion control protocols rely on an estimation of the available link bandwidths. Specially, XCP and VCP need an explicit knowledge of the link bandwidth. However, wireless networks are characterized by bandwidth variations due to the openness of air links. Experiments have clearly shown that operating over variable bandwidth wireless links can lead to a significant performance degradation of congestion control protocols. Such degradation is typically measured in terms of oscillatory behavior. In this paper, we propose the use of an Extended Kalman Filter (EKF) to filter out the impact of bandwidth variations from the operation of wireless congestion control protocols. Our EKF-based Bandwidth Estimation (EBE) scheme can predict link capacity by monitoring the persistent queue size of a wireless link thereby eliminating the need for direct measuring of the real-time bandwidth. We implement EBE in NS-2 and integrate it with XCP and VCP protocols. Through extensive simulation studies, we demonstrate significant performance improvements of both protocols as the result of using EBE.

Recently, Variable-structure Congestion-control Protocol (VCP) has emerged as a great practical alternative of deployment for congestion control by requiring the use of only two ECN bits in the IP header. However, VCP suffers from a relatively low speed of convergence and exhibits biased fairness in moderate bandwidth high delay networks due to utilizing an insufficient amount of congestion feedback. Our previous work Double-Packet Congestion-control Protocol (DPCP) addressed this problem by increasing the amount of the feedback distributed over two ECN bits in the IP header of a pair of packets. However, DPCP faces deployment obstacles in encrypted wireless networks due to the fact that it relies on partial information in the TCP header and the TCP header information is lost when crossing encryption boundaries. Furthermore, wireless networks are characterized by both error- and congestion-caused loss. Our previous work has revealed that the efficiency of DPCP, and for that matter any congestion control protocol, over wireless networks may be reduced as the result of not being able to differentiate between two types of loss. In this paper, we propose an alternative congestion control protocol to which we refer as Distributed Congestion-control Protocol for Encrypted Wireless (DCP-EW) networks. DCP-EW is capable of efficiently operating in encrypted wireless networks while preserving all of the benefits of DPCP for wired networks. It does so by passively utilizing the IP Identification field of a packet header instead of the TCP header in conjunction with a heuristic algorithm to differentiate between different sources of loss. We implement DCP-EW in NS-2 and the Linux Kernel. We demonstrate the performance improvements of DCP-EW compared to DPCP and VCP through simulation and experimental studies.

Every congestion-control protocol operating in wireless networks is potentially faced with two major challenges of performance degradation. These sources are: 1) the coupling of fairness and efficiency; and 2) not properly differentiating between congestion-caused loss associated with network buffering and error-caused loss associated with fading effects. In this paper, we provide a Variable-structure Congestion-control Protocol (VCP)-based cross-layer framework of congestion control that can address both challenges noted. As a part of our framework, we introduce a loss differentiation heuristic algorithm that can be used with a variety of congestion-control protocols. Then, using analysis, simulation, implementation, and emulation, we profile the performance of a number of congestion-control alternatives in wireless networks. We describe the first implementation of VCP as a collection of loadable kernel modules along with fine-tuned implementations of XCP and TCP/AQM + ECN in Linux. We utilize NS2 as our simulation tool and a wired Linux testbed emulating wireless link effects as our experimental tool. We implement a finite-state Markov chain in both NS2 and our testbed in order to model error-caused loss over wireless links. We further use link-layer forward error correction (FEC) codes on a per-packet basis to compensate against such loss. Our profiling results demonstrate that VCP equipped with our loss differentiation heuristic and link-layer FEC represents a well-performing yet practical alternative of wireless congestion control. We also identify some of the shortcomings of VCP, including its oscillatory behavior in the presence of link estimation errors and poor fairness characteristic in multibottleneck networks.

The performance of prevalent MAC protocols in MANETs relies on the level of contention in networks. While contention-based MAC protocols such as CSMA suffer from inefficiency under high contention, slot-based MAC protocols such as TDMA perform in the opposite way. In this paper, we propose a hybrid protocol to which we refer as Load-Adaptive MAC (LA-MAC) protocol for MANETs. By adaptively switching its running mode between CSMA and TDMA, LA-MAC achieves high channel utilization under both high and low contention. We report our implementation of LA-MAC on a MANET testbed formed by a collection of Multiple-Input Multiple-Output (MIMO) Universal Software Radio Peripheral (USRP) software defined radio nodes. We program the PHY layer of USRP nodes using GNU Radio and integrate LA-MAC with the PHY layer implementation of USRP. Through experimental studies, we demonstrate the performance improvements of LA-MAC relative to CSMA and TDMA.

Following the design philosophy of XCP, VCP is a router-assisted congestion protocol that intends to balance the efficiency and the fairness control in high Bandwidth-Delay Product networks. While both VCP and XCP achieve comparable performance, VCP represents a more practical alternative of deployment as it only requires the use of two ECN bits in the IP header. However, the use of two ECN bits only allows for establishing three levels of congestion notification signaling. Our previous work reveals that VCP suffers from relatively low speed of convergence and exhibits a biased fairness behavior in moderate bandwidth high delay networks due to utilizing an insufficient amount of congestion feedback. In this paper, we propose a distributed ECN-based congestion control protocol to which we refer as Double-Packet Congestion Control Protocol (DPCP). DPCP is capable of relaying a more precise congestion feedback compared to earlier proposed Variable-structure Congestion-control Protocol (VCP) yet preserving the utilization of the two ECN bits. By distributing (extracting) congestion related information into (from) a series of packets, DPCP is able to circumvent the limitations of VCP related to the use of three congestion levels encoded into two ECN bits. We implement DPCP in Linux and demonstrate its performance improvements compared to VCP through experimental studies.

We describe a cross-layer study that attempts at profiling the performance of end-to-end congestion control protocols in high BDP wireless networks. Packets transmitted over such networks are assumed to face temporally correlated random bit errors caused by fading and/or blocking. We utilize a finite-state Markov chain to model bit error characteristics and apply per packet link layer FEC codes in order to compensate for such errors. Utilizing our model along with loss expressions, we profile the end-to-end performance of XCP, VCP, and TCP/AQM+ECN congestion control protocols. Based on our profiling results, we observe that (1) the performance of any congestion control protocol highly degrades without properly protecting its data and signaling information against random bit errors, (2) utilizing multiple antenna nodes improves the transient characteristics of the congestion control protocol in terms of utilization and packet error rates, and (3) XCP and VCP represent the best alternatives of congestion control for high bandwidth moderate delay and moderate bandwidth large delay environments, respectively.