In recent years topics related to
mobile wireless networks have evolved into a major research interest fueled by
the latest advances in radio technologies, and the special appeal of wireless
devices to the end users. Quality of service research in mobile ad-hoc networks
is a specifically difficult topic due to the network dynamics, variations in
radio link quality, limited capabilities of mobile nodes, and lack of central
authority.
This research proposes the PYLON-Lite QoS model that deals with the cross-domain QoS connectivity issues. A PYLON-Lite gateway operates between the ad-hoc network, with its unique characteristics on one side, and the fixed topology access network on the other side. The fundamental differences between these two networks represent the heterogeneous environment of the PYLON-Lite model. The PYLON-Lite QoS model provides a, seemingly, homogeneous cross-domain QoS solution. PYLON-Lite has a supple model design that facilitates a lightweight implementation to benefit from the already existing QoS models.
The development of PYLON-Lite is motivated by the need of mobile nodes in the ad-hoc network to access the Internet. PYLON-Lite relies on the QoS models implemented on each side of the gateway to provide detailed services in the relevant network, while it remains focused on the QoS concatenation issues. The PYLON-Lite gateway design follows the principles of cascaded network services. No comparable models are found in the literature to cover the defined problem to this date; therefore, PYLON-Lite is unique.
This research identifies the challenges in designing the cross-domain QoS model. PYLON-Lite presents specific mechanisms to deal with those challenges while maintaining the lightweight approach. It defines methods to interact with other QoS models on both sides of the gateway, methods to police traffic, and conditions to guarantee model scalability. This research provides intensive analysis and evaluation of the model performance and behavioral characteristics. PYLON-Lite is shown to consistently improve the QoS provided for real-time traffic with limited impact on best-effort traffic.