Data-Collection Capacity of IEEE 802.11-like Sensor Networks

Data-collection is an important application of sensor networks. This paper considers the scenario in which data generated from all sensors are to be forwarded to a single "data center" for processing. Although there have been many studies on this many-to-one communication scenario, it has generally been assumed that the data-collection capacity is upper-bounded by the link capacity L. We show that when the IEEE 802.11 protocol is used, the data-collection capacity has a tighter analytical upper-bound of 3L/4, and simulated throughput of 0.601L. In deriving our results, we introduce the notion of "canonical networks", which is a class of regularly-structured networks whose capacities can be analyzed more easily than unstructured networks. We argue that the capacities of canonical networks serve as a good benchmark for general networks in that the maximum possible capacity of any given network is unlikely to exceed the upper-bound capacity established for canonical networks.