WLL link layer protocol for QoS support of multi-service internet

Abstract

Wireless Local Loop (WLL) such as LMDS typically has time-varying and high Bit Error Rate (BER) channel characteristics [1-3]. To provide Internet service over a such link, error protection mechanisms at link layer are necessary [4-6]. However, since different services provided by higher layers require different link layer behavior, a multi-service link layer architecture is needed, as has been proposed in [7]. The later assigns transmission bandwidth in proportion to network layer allocation independent of arbitrary overhead rates. As a different approach, we propose a WLL link layer architecture and protocol that assign transmission bandwidths according to services' priorities and their actual bandwidth needs (taking in to account network layer allocation as well as any overhead occurred). This new model keeps track of the current status of each link service, such as link service overhead rates, and informs the network layer. We argue that our link layer architecture and protocol more effectively support QoS services implemented at higher layers, especially those that require hard QoS guarantees (such as guaranteed service in IntServ model [9] and Expedited Forwarding (EF) service in DiffServ [10]).