Wireless local area networks (WLANs) designed as wireless ATM systems to extend the services of fixed ATM networks to mobile users appear best suited to provide a guaranteed quality of service (QoS) for wireless IP networks. HiperLAN/2 is an ETSI/BRAN standard providing convergence layers for both IP and ATM classes of service. Besides a description of HiperLAN/2 and its Home Environment Extension, the performance for IP traffic flows is presented from analysis and from simulating a prototype implementation. Coexistence with the IEEE 802.11a WLAN is discussed and the ability of HiperLAN/2 to guarantee QoS even when coexisting is analyzed.Ad hoc networking of HiperLAN/2 is analyzed and two possible extensions of the system are introduced and their performance evaluated, namely, adaptive antennas and wireless base stations.

No abstract

No abstract

In general, the users of wireless ATM terminals request the same functionality and Quality of Service as users of wired terminals. These user requirements can be transfomed into the demand for building an ATM multiplexer around the air interface which is characterized by a radio channel inside the ATM layer. The main difference between this logical ATM multiplexer around the air interface and a normal ATM multiplexer is the distribution of the multiplexing function between wireless terminals and the base station. For the uplink this requires a frequent notification of the ATM cell scheduler in the base station about the status of the incoming queues inside the wireless terminals. This paper focuses on different transmission methods for transmitting capacity request messages (also designated as reservation request) over the uplink.

One of the major problems in multihop ad hoc ATM networks is the establishment and maintenance of virtual connections. Therefore, PNNI, IP and ZRP have been investigated as possible routing protocol candidates. All three have been modelled and analysed in terms of signalling overhead. The analytical model showed that ZRP is the most promising candidate as it is able to be adapted to any kind of network topology. As ZRP has been developed for IP it has been changed for ATM routing. Computer simulations show that the analytical model is accurate and the delays due to topology changes are very limited.