The Key Technology Research Of Rfid Middleware In Data Storage Management

With the scale of RFID application expanding constantly, the amount of RFID data generated by the sensors are explosively growing, how to effectively managing these huge data becomes a difficult problem in RFID technology. At present, there are mainly two different RFID event processing models, one is data-centered model, another is event-centered model. Data-centered model stores RFID data in the database firstly, and then supports event processing based on the database. Event-centered model directly compare RFID event stream with predefined rules, if it matchs, reports RFID event to the application.After analysis the characteristics of RFID data, we proposed the RFID data multi-level storage model. We designed and implemented RFID data cache based on T linked tree, we designed and implemented RFID data warehouse based on BerkeleyDB database. We stored the current RFID information in RFID data cache, to improve the efficiency in RFID event processing. We stored the history RFID information in RFID data warehouse. Aiming to the batch moving characteristics of RFID label, we designed a bulk movement approach for RFID data warehouse. This model stores the RFID path information separately, multiple tags can share the same path information. This model not only reduces the storage space, but also improves the query efficiency.In order to handle the RFID event effectively, this paper designed and implemented the updating algorithm for RFID data cache, also designed and implemented two different data transmission algorithm for RFID data warehouse. First is single tuple-based data transmission algorithm based on single RFID event, the second is multiple tuple-based data transmission algorithm based on multiple RFID events, and we proved the second data transmission algorithm more efficient than the first data transmission algorithm through the experiment.