HiFi: Network-centric query processing in the physical world

Advances in wireless sensors, RFID technology, and mobile devices have enabled the development of information systems that monitor and react to events in the real world. When deployed on a large (e.g., national) scale, these systems assume a high fan-in architecture, in which vast numbers of events measured at the edges of the network are continually refined, summarized, augmented, and aggregated as they flow towards the interior. High fan-in systems present a wealth of new research problems reflecting the different concerns and priorities at each level of the system as well as the interactions among the levels. The solutions will require insights from recent efforts in data stream processing, sensor databases, event systems, data warehousing, and spatio-temporal data management. In this talk I will identify the key characteristics and challenges presented by high fan-in systems, and argue for a uniform, querybased approach towards addressing them. I will then present our initial thoughts on the design of HiFi, the system we are building to embody these ideas, and describe an initial proof-of-concept prototype that is capable of combining data from RFID readers and clusters of sensor motes.