The emerging trend towards seamless monitoring of all business processes via comprehensive sensor networks - in particular RFID readers - creates new data management challenges. In addition to handling the huge volume of data generated by these sensor networks the information systems must support the efficient querying and analysis of both recent data and historic sensor readings. In this paper, we devise and evaluate a data staging architecture that consists of a distributed caching layer and a data warehouse. The caches maintain the most recent RFID events (i.e., sensor readings) to facilitate very efficient OLTP processing. Aged RFID events are removed from the caches and propagated into the data warehouse where related events are aggregated to reduce storage space consumption. The data warehouse can be utilized for business intelligence applications that, e.g., analyze the supply chain quality.