Time synchronization in ad hoc networks
Proceedings of the 2nd ACM international symposium on Mobile ad hoc networking & computing - MobiHoc '01
Ubiquitous computing environments are typically based upon ad hoc networks of mobile computing devices. These devices may be equipped with sensor hardware to sense the physical environment and may be attached to real world artifacts to form so-called smart things. The data sensed by various smart things can then be combined to derive knowledge about the environment, which in turn enables the smart things to "react" intelligently to their environment. For this so-called sensor fusion, temporal
... lationships (X happened before Y) and real-time issues (X and Y happened within a certain time interval) play an important role. Thus physical time and clock synchronization are crucial in such environments. However, due to the characteristics of sparse ad hoc networks, classical clock synchronization algorithms are not applicable in this setting. We present a time synchronization scheme that is appropriate for sparse ad hoc networks. These examples indicate that temporal ordering and other realtime 1 issues play an important role in such environments. As we will see later, neither logical time [12, 14] nor classical physical clock synchronization algorithms [3, 13, 16, 17] can be used to solve this problem in general. We will suggest an algorithm that solves the temporal ordering problem and other real-time issues in environments sketched above. AD HOC NETWORKS Ad hoc networks  are networks of mobile wireless computing devices. Due to the limited communication range of wireless technology (about 10 meters for Bluetooth ), nodes of the network form spontaneous connections when they are brought within the communication range of each other, providing typically a symmetrical communication link where message exchange is possible in both directions. The limited communication range and the mobility of the nodes lead to frequent reconfiguration of the network topology. The left hand side of figure 1 shows the configurations (topologies) of an ad hoc network consisting of three nodes at two points in time ¢ ¡ ¤ £ ¥ § ¦ . At ¡ nodes 1 and 2 are able to communicate with ¡ Throughout the paper the term real-time refers to UTC. ¦ nodes 2 and 3 are able to communicate with each other. This may result from the following physical setting: nodes 1 and 3 are out of the communication range of each other. At ¡ node 2 is within the communication range of node 1, then node 2 is moved out of the communication range of node 1 into the communication range of node 3. Assuming this setting, there is no point in time ¢ ¡ ¡ ¡ ¦ where communication between node 1 and 3 (either directly or indirectly via node 2) is possible. This example shows an important property of ad hoc networks: the frequent temporary existence of network partitions, especially in sparse ad hoc networks with only a few nodes distributed over a large area (relative to the communication range) in contrast to dense ad hoc networks.