Fanglei Sun

Design of SNACK mechanism for wireless TCP with new snoop

TCP is the most widely adopted transport layer communication protocol. In heterogeneous wired/wireless networks, however, the high packet loss rate over wireless links can trigger unnecessary execution of TCP congestion control algorithms, resulting in performance degradation. TCP performs poorly on wireless links with bursty losses, when it is forced to rely on limited information available from batched acknowledgements, (i.e., multiple packets are acknowledged with one acknowledgment packet). In this paper, a selective negative acknowledgement (SNACK) mechanism is designed to overcome the limitation of batched acknowledgments. A new link layer retransmission protocol, called, SNACK-NS (new snoop), is proposed. Through the detection and retransmission functions that are provided by the two protocol components of SNACK-NS, namely, SNACK-snoop and SNACK-TCP, the transmission performance of TCP over wireless network is greatly enhanced in both fixed host (FH) to mobile host (MH) and MH to FH transmissions.

Modified bipartite matching for multiobjective optimization: Application to antenna assignments in MIMO systems

Based on the Hungarian algorithm, the Kuhn-Munkres algorithm can provide the maximum weight bipartite matching for assignment problems. However, it can only solve the single objective optimization problem. In this paper, we formulate the multi-objective optimization (MO) problem for bipartite matching, and propose a modified bipartite matching (MBM) algorithm to approach the Pareto set with a low computational complexity and to dynamically select proper solutions with given constraints among the reduced matching set. In addition, our MBM algorithm is extended to the case of asymmetric bipartite graphs. Finally, we illustrate the application of MBM to antenna assignments in wireless multiple-input multiple-output (MIMO) systems for both symmetric and asymmetric scenarios, where we consider the multi-objective optimization problem with the maximization of the system capacity, total traffic priority, and long-term fairness among all mobile users. The simulation results show that MBM can effectively reduce the matching set and dynamically provide the optimized performance with different quality of service (QoS) requirements.

Multiobjective optimized bipartite matching for resource allocation

The Hungarian algorithm can provide the maximum weighted bipartite matching for assignment problems. However it can only solve the single objective weight optimization problem. In this paper, a modified bipartite matching (MBM) algorithm is proposed to solve the weighted bipartite matching problem with multiobjective optimization. In addition, our MBM algorithm is applicable to asymmetric bipartite graph, which is common in resource allocation problems. We illustrate the application of MBM to antenna assignments in wireless multiple-input multiple-output (MIMO) systems for both symmetric and asymmetric scenarios. The simulation results show that MBM enjoys low computational complexity and maximizes the system capacity, while keeping the fairness among mobile users.

Modified bipartite matching for multiobjective optimization

In graph theory, the Hungarian algorithm can provide the maximum weighted bipartite matching for assignment problems. In this paper, a modified bipartite matching (MBM) algorithm is proposed for multiobjective optimization. This algorithm can be widely used to solve the weighted bipartite matching problem with multiobjective optimization. We illustrate the application of MBM to antenna assignments in wireless MIMO system. The simulation results show that MBM enjoys low computational complexity and maximizes the system capacity, while keeping the fairness among mobile users.

An Adaptive TCP Algorithm to Support Handoffs in Heterogeneous Overlay Networks

In this paper, an adaptive TCP algorithm is proposed to support handoffs in heterogeneous overlay networks. The algorithm is implemented as a collection of units. The main body, namely the TCP implementation unit, considers the optimization of TCP parameters over different wireless networks. This unit is decomposed into three modules: transmission control, congestion control, and error detection. Based on this module decomposition, several TCP versions are re-constructed, and the different combinations of their congestion control and error detection modules are studied. Aided by other units in our adaptive TCP algorithm, such as the realtime report and throughput evaluation units, when a mobile user experiences acute changes in access parameters due to handoffs, the adaptive TCP algorithm will be triggered to reconfigure TCP parameters or TCP module combinations in the TCP implementation unit, thus optimizing the TCP performance over heterogeneous wireless networks

A New Cross-Layer Designed Multipolling Mac Protocol Over WLANs

This paper develops a multipolling MAC protocol which exploits cross-layer information to support delay-sensitive multimedia services over WLANs. A user detection module and a multi-rate adaptation module are proposed in the physical layer to assist in link differentiation. With these two modules, our new multi-polling MAC protocol, named PALD-MPMP not only reduces the polling overhead, but also provides an effective polling scheduler by allocating transmission priorities to users according to their delay requirements or levels and channel gains. Simulation results show that our proposed protocol outperforms the standard point coordination function (PCF) for delay-sensitive services. All the performance improvements are mainly contributed by the awareness of cross-layer channel state information, and the consequent multi-rate adaptation schemes.

Clustered-loss retransmission protocol over wireless TCP

Transmission control protocol (TCP) performs well in traditional wired networks where the packet loss rate is low. However, in heterogeneous wired/wireless networks, the high packet loss rate over wireless links may result in excessive invocation of the congestion control algorithm, thus deteriorating the performance of TCP. In this paper, a novel localized link layer retransmission protocol, called clustered-loss retransmission protocol (CLRP), is proposed. CLRP consists of three protocol components, namely, TCP-FH deployed on a fixed host, TCP-MH deployed on a mobile host and CLRP-BS deployed on a base station. CLRP can provide not only explicit distinction between congestion and packet corruption losses, and effective multiple wireless loss information for retransmissions, but also better retransmission control for wireless losses. Thus it is well suited to wireless networks, in which packet loss and bursty packet corruption is a serious problem. Moreover, CLRP does not require any modifications to TCP deployed on fixed hosts