Thi Mai Trang Nguyen

Evaluation of throughput optimization and load sharing of multipath TCP in heterogeneous networks

Todays laptops and mobile terminals are equipped with several network interfaces belonging to different access technologies. Users can benefit from multi-homing to access network services anywhere, at any time and from any network. Multipath TCP (MPTCP) is a modified version of Transmission Control Protocol (TCP) which allows a TCP connection to simultaneously send data over multiple paths. One of the objectives of MPTCP is to improve the throughput of TCP connection. A MPTCP connection over multiple paths should achieve a no worse throughput than a single TCP connection over the best constituent path. In this paper, we evaluate the performance of MPTCP in terms of throughput optimization and load sharing. The measurements show that the context of heterogeneous networks (Ethernet, Wifi and 3G) has a great impact on the MPTCP performances and reveal the need of an intelligent algorithm for interface selection in MPTCP.

Evaluation of multipath TCP load sharing with coupled congestion control option in heterogeneous networks

The evolution of technologies, particularly the emergence of multi-interface terminals (3G, WLAN, WMAN, etc), totally changes the vision of services offered to users in the future. Multi-homing and multipath transmission support becomes more and more indispensable to satisfy user demands. Multipath TCP (MPTCP) is a modified version of Transmission Control Protocol (TCP) which allows using simultaneously multiple paths for data transmission. It aims to improve the throughput of TCP connection in a manner that a MPTCP connection over multiple paths should achieve at least a no worse throughput than a single TCP connection over the best constituent path. In a previous work, we have evaluated the load sharing of MPTCP without the coupled congestion control option. In this paper, the couple congestion control of MPTCP is enabled. The measurement results show that MPTCP with coupled congestion control provides a better performance. However, the heterogeneous networks environment (Ethernet, Wifi and 3G) still has a great impact on MPTCP throughput and reveals the need of an intelligent algorithm for interface selection in MPTCP.

Towards a New Architectural Framework – The Nth Stratum Concept

Current architectures and solutions are about to reach the limits of sustainable developments. Over the years, many new requirements have emerged, and there are observations pointing to an ever-increasing diversity in applications, services, devices, types of networks at the edge and the access. Meanwhile, the infrastructures for internetworking, connectivity, and also management remain fairly the same. A new paradigm is needed that can support a continuous high pace of innovations in all the different parts and aspects of a communication system, while at the same time keeping costs of deployment and maintenance down. This new paradigm has to embrace current trends towards increased heterogeneity, but on the other hands provide support for co-existence and interoperability between alternative and various solutions all residing within a global communication system. This paper presents a new architectural framework called the Nth Stratum concept, and which takes a holistic approach to tackle these new needs and requirements on a future communication system.

Dynamic cluster-based over-demand prediction in bike sharing systems

Bike sharing is booming globally as a green transportation mode, but the occurrence of over-demand stations that have no bikes or docks available greatly affects user experiences. Directly predicting individual over-demand stations to carry out preventive measures is difficult, since the bike usage pattern of a station is highly dynamic and context dependent. In addition, the fact that bike usage pattern is affected not only by common contextual factors (e.g., time and weather) but also by opportunistic contextual factors (e.g., social and traffic events) poses a great challenge. To address these issues, we propose a dynamic cluster-based framework for over-demand prediction. Depending on the context, we construct a weighted correlation network to model the relationship among bike stations, and dynamically group neighboring stations with similar bike usage patterns into clusters. We then adopt Monte Carlo simulation to predict the over-demand probability of each cluster. Evaluation results using real-world data from New York City and Washington, D.C. show that our framework accurately predicts over-demand clusters and outperforms the baseline methods significantly.

Performance–Cost Trade-Off Strategic Evaluation of Multipath TCP Communications

Todays mobile terminals have several access network interfaces. New protocols have been proposed during the last few years to enable the concurrent use of multiple access paths for data transmission. In practice, the use of different access technologies is subject to different interconnection costs, and mobile users have preferences on interfaces jointly depending on performance and cost factors. There is therefore an interest in defining “light” multipath communication policies that are less expensive than greedy unconstrained ones such as with basic multipath TCP (MP-TCP) and that are strategically acceptable assuming a selfish endpoint behavior. With this goal, we analyze the performance-cost trade-off of multi-homed end-to-end communications from a strategic standpoint. We model the communication between multi-homed terminals as a specific non-cooperative game to achieve performance-cost decision frontiers. The resulting potential game always allows selecting multiple equilibria, leading to a strategic load-balancing distribution over the available interfaces, possibly constraining their use with respect to basic MP-TCP. By simulation of a realistic three-interface scenario, we show how the achievable performance is bound by the interconnection cost; we show that we can halve the interconnection cost with respect to basic (greedy) MP-TCP while offering double throughputs with respect to single-path TCP. Moreover, we evaluate the compromise between keeping or relaxing strategic constraints in a coordinated MP-TCP context.

Strategic evaluation of performance-cost trade-offs in a multipath TCP multihoming context

Todays mobile terminals have several access network interfaces. In practice, the use of different access technologies is subject to different interconnection costs, and mobile users have preferences on interfaces jointly depending on performance and cost factors. There is therefore an interest in defining “light” yet rational multipath communication policies less expensive than greedy ones such as with basic Multipath TCP (MP-TCP). We analyze the performance-cost trade-off of multi-homed end-to-end communications from a strategic standpoint. We model the communication between multi-homed terminals as a multi-criteria non-cooperative game so as to achieve performance-cost decision frontiers. The resulting potential game always allows to select multiple equilibria, which correspond to a strategic load-balancing distribution over the available interfaces, possibly constraining their use with respect to basic MP-TCP. We specify how the resulting model may be in practice implemented by users willing to jointly control the interconnection cost and the performance, based on user Quality of Experience (QoE) assessments. By simulation of a realistic 3-interface scenario, we show how the achievable performance is bound by the interconnection cost; we show that we can halve the interconnection cost with respect to basic (greedy) MP-TCP under a reasonable trade-off, while offering double throughputs with respect to single-path TCP.

Distributed Opportunistic and Diffused Coding in multi-hop wireless networks

In wireless mesh networks, network coding can be adapted and deployed for performance enhancement. Some early results such as COPE, BEND and DCAR show that the use of network coding can outperform the original 802.11 networks. However, these network coding architectures have two limitations: (i) the coding chances inevitably depend on the established route, (ii) traffic is strictly coded based on the two-hop coding pattern. In this paper, a novel and enhanced network coding architecture, Distributed Opportunistic Diffused Encoding (DODE), is proposed. DODE combines (i) the diffused gain from BEND and (ii) the generalized coding condition from DCAR to completely solve the limitations present in the three previous propositions. In addition, we propose a routing metric called Shortest Path with Enriched Neighborhood routing Metric (SPENM) which allows DODE to follow the most enriched coding chance path and also to enjoy the diffused gain. We implement DODE system in NS-2. The simulation results show that DODE outperforms 802.11, COPE, BEND and DCAR.

Kalman filter based bandwidth estimation and predictive flow distribution for concurrent multipath transfer in wireless networks

More and more terminals come equipped with multiple network interfaces, usually connected to different wireless networks. Connected multiple wireless networks have varying link characteristics. Because of unreliable wireless links, we need an efficient bandwidth estimation to describe link characteristics. The Kalman filter is an efficient recursive method, which not only estimates and corrects the current system states but also predicts even the future states based on the latest state. In this paper, we use the Kalman filter to estimate available bandwidth in wireless networks with different loss rates. The predictive arrival time is calculated for each packet before it is transferred. Based on this, we proposed a predictive flow distribution algorithm for concurrent multipath transfer in wireless networks. The simulation results show that our predictive flow distribution algorithm improves congestion window growth pattern by reducing the out-of-order packets. As a result, the total throughput increases under different random wireless link loss conditions.

Concurrent multipath transfer performance optimization using Kalman filter based predictive delay estimation in wireless networks

Todays researchers are witnessing the evolution of multi-interface terminals. The simultaneous utilization of multiple interfaces to improve network performances is facing a great challenge. The multiple interfaces are usually heterogeneous and wireless with different link characteristics. Because of unreliable wireless links, we need a good description of link characteristics. Kalman filter is an efficient recursive method which not only estimates and corrects the current system states but also predicts even the future states based on the latest state. In this paper, we use Kalman filter to predict packet transmission delay and use this estimation result for the flow distribution algorithm of concurrent multipath transfer. The simulation results show that our proposal with Kalman filter reduces the reordering degree and improves congestion window growth pattern. As a result, the total goodput increases.

Adaptive redundancy control with network coding in multi-hop wireless networks

Recent research shows that random linear network coding is adapted into current network systems to provide transmission reliability. In lossy environments, like wireless networks, a redundancy control is also needed to mitigate losses. Some implementations suggest adding a fixed amount of redundant packets per batch. This solution poses the risk of wasting bandwidth by over-redundancy or unsuccessful decoding due to packet insufficiency, dramatically degrading the performance. In this paper, we present the Adaptive Redundancy Control for Network Coding (ARC) that adds redundant packets dynamically to mitigate losses. First, ARC is based on the use of MAC acknowledgements as a feedback on the state of link quality. Second, instead of determining the correct number of redundant packets, we decide to find the suitable time to send the redundant packets to cover losses. We design the multi-batch coding scheme that continuously streams new packets while interpreting the received MAC acknowledgements as the feedback on the state of link quality. The state of link quality will be rechecked before every transmission to deduct if it is appropriate for pumping the redundant packets or transmitting the new data. We implemented and compared ARC with 802.11 and FRC - Fixed Redundancy Control in NS-2. Results show that ARC outperforms 802.11 and FRC.