Mobility support and bandwidth aggregation in heterogeneous wireless networks

• Main Author: Zhong, Feng
• Other Authors: Lee Bu Sung, Francis
• Format: Theses and Dissertations
• Language: English
• Published: 2012
• Subjects: DRNTU::Engineering::Computer science and engineering::Computer systems organization::Computer-communication networks
• Online Access: https://hdl.handle.net/10356/48907
• Institution: Nanyang Technological University

The future wireless communication system is envisioned to be an IP-based, ubiquitous system with heterogeneous wireless networks integrated, where multiple wireless technologies converge to a common infrastructure. The unified system can bring users many benefits such as the expansion of coverage and the aggregation of multiple-path bandwidth. However, before enjoying these benefits, a lot of issues need to be solved. This thesis focuses on proposing solutions for two issues: mobility management and bandwidth aggregation. According to the service ranges, mobility management can be classified into two categories: intra-domain mobility and inter-domain mobility. Intra-domain mobility handles user's reachability within a single domain while inter-domain mobility handles outside domain. Efficient Framework for Local Mobility Management (EFLoM) is a inter-domain mobility which employs a IP swapping technology to deliver traffic. Specifically, EFLoM employs three network entities: Mobile Node (MN), Local Anchor Router (LAR) and Wireless Access Gateway (WAG). MN is the mobile device held by users. LAR is the gateway and it holds the location information of MN. WAG is an access router of the network in the domain which regulates the intra-domain connection. When a data packet arrives, LAR updates the destination IP header with MN's current location and sends to MN. On reception, MN restores the destination address and pass on to applications. Because IP swapping can remove the tunnel header, EFLoM outperform other mobility protocols in terms of route hops, protocol signaling cost, handover delay and traffic overhead. Except EFLoM, this thesis proposes a inter-domain mobility framework named Traffic Distributor Inter-Domain Mobility (TDIDM). Through placing a new network entity named Traffic Distributor (TD) in the aggregate network, TDIDM interconnects neighboring local domains. When MN moves out of one domain into another, TD is notified. TD then tunnels the traffic to the new domain. After the traffic arrives at the new domain, the local mobility protocol (e.g., EFLoM) takes over the responsibility to deliver the traffic. Moreover, TDIDM develops an algorithm to select a TD set. The TD set covers MN's traffic in all paths while reducing the redirection time. Experiments have been conducted to compare it with Neumann's proposal which is another proposal to handle inter-domain issues. Results show that our TDIDM is a feasible alternative for inter-domain handover, and TDIDM outperforms Neumann's proposal in terms of binding cache entry number, transmission delay and handover delay. The third part of the thesis is a multipath scheduling algorithm, named Adaptive Load Balancing Algorithm (ALBAM). Targeting at solving adverse impacts on TCP flows, ALBAM employs four techniques. Firstly, ALBAM takes advantage of the bursty nature of TCP flows and performs scheduling at the flowlet granularity where the packet interval is large enough to compensate for the different path delays. Secondly, ALBAM develops a packet number estimation algorithm (PNEA) to predict the buffer usage in each path. With PNEA, ALBAM can prevent buffer overflow and schedule the TCP flow to a less congested path before it suffers packet loss. ALBAM can greatly reduce if not prevent buffer overflow and schedule the TCP flow to a less congested path before it suffers packet loss. Thirdly, ALBAM employs virtual flowlet division (VFD) to divide the flow into several virtual flowlets. The VFD enables ALBAM to improve the scheduling performance when the flowlets are missing. Lastly, to obtain more accurate path delay measurement, ALBAM makes use of the Access Points (AP) to provide statistical delay measurements in each path. Simulation results show that ALBAM can yield a more accurate delay reading. Through these techniques, ALBAM can provide better performance for TCP connections than its counterparts, namely, Flowlet Aware Routing Engine (FLARE) and Opportunistic Multipath Scheduling (OMS).