| Providing instant communications to mobile users is becoming one of the key driving forces behind mobile computing. This dissertation mainly strives to answer the questions why, how, and for what Mobile IP, the dominant network-layer mobility support mechanism, should be used.; First, we put the network-layer mobility support mechanism into perspective amid controversies surrounding Mobile IP, by analyzing the essential components of mobility support schemes and comparing Mobile IP with other higher-layer mobility support mechanisms. We establish the need for network-layer mobility support and the importance of having a unique identifier at the routing layer for supporting mobility.; Second, we describe two new flow-oriented mechanisms, in conjunction with the use of Mobile IP, to ensure a mobile host's robust and efficient communication with other hosts in a mobile environment. One mechanism supports the use of multiple packet delivery methods simultaneously by a mobile host, allowing it to select adaptively the most appropriate method to use (such as regular IP or Mobile IP) according to the characteristics of each traffic flow. The other mechanism enables a mobile host to make use of multiple network interfaces simultaneously and to control the selection of the most desirable network interfaces for both outgoing and incoming packets for different traffic flows.; Third, we demonstrate that it is not always a disadvantage for traffic to travel via the home network in Mobile IP. By making use of this extra level of indirection, we can maintain better control over traffic to a mobile host to protect the potentially weak communication link to the mobile host. We introduce a framework that incorporates Mobile IP as a basic building block to support instant communication to a mobile host, a promising application for Mobile IP, and illustrate that this scheme can balance controllability and the capability of being directly reachable in peer-to-peer mobile instant communication. |
