Abstract
Accessing wireless services and application on the move has become a norm among casual or business users these days. Due to societal needs, technological innovation, and networks operators’ business strategies, there has been a rapid proliferation of many different wireless technologies. In many parts of the world, we are witnessing a wireless ecosystem consisting of wide-area, low-to-medium-bandwidth network based on access technologies such as GSM, GPRS, and WCDMA, overlaid by faster local area networks such as IEEE 802.11-based Wireless LANs and Bluetooth pico-networks. One notable advantage of wide-area networks such as GPRS and 3G networks is their ability to provide access in a larger service area. However, a wide-area network has limited bandwidth and higher latency. 3G systems promise a speed of up to 2Mbps per cell for a non-roaming user. On the other hand, alternative wireless technologies like WLAN 802.11and Personal area network (PAN) using Bluetooth technology have limited range but can provide much higher bandwidth. Thus, technologies like WWAN and WLAN provide complementary features with respect to operating range and available bandwidth. Consequently, the natural trend will be toward utilizing high bandwidth data networks such as WLAN, whenever they are available, and to switch to an overlay service such as GPRS or 3G networks with low bandwidth, when coverage of WLAN is not available. Adding to the existing public networks, some private institutions (i.e., universities) have joined the fray to adopt wireless infrastructure to support mobility within their premises, thus adding to the plethora of wireless networks. With such pervasiveness, solutions are required to guarantee end-user terminal mobility and maintain always-on session connections to the Internet. To achieve this objective, an end device with several radio interfaces and intelligent software that would enable the automatic selection of networks and resources is necessary (Einsiedler, 2001; Moby Dick, 2003).