Passive Network Awareness for Adaptive Mobile Applications TR-UTEDGE-2006-002

In pervasive computing, adaptation is essential to applications. However, adaptation requires an up-to-the-minute understanding of the state of the environment, which often comes with a significant added cost in terms of computation and communication. In this paper, we explore possibilities for measuring the degree of mobility in dynamic ubiquitous networks. Existing measures of mobility degree are either global values (e.g., the average node speed) or local measures that require additional coordination (e.g., the relative speed of neighboring nodes). In contrast, we describe a completely passive approach to measuring the local mobility degree that uses knowledge about communication mechanisms to determine the network dynamics affecting a single node. This context information can subsequently be used by a node to adapt communication and application protocols to current conditions. In this paper, we define a passive metric that gives a node an approximate view of local network dynamics and compare its ability to capture dynamics with existing metrics.