Local caching architecture for RFID tag resolution

Abstract

Radio Frequency Identification (RFID) is set to become the standard in identification and tracking of items worldwide with projected numbers in tens of billions in a few years. With the reducing size and cost of RFID equipment, the use of RFIDs is being pushed well beyond just replacement of barcodes. Even though the deployment of RFIDs is growing at an exponential rate, the only global standard that currently exists for RFID networks is the EPCglobal Network proposed by EPCglobal, Inc. This thesis reviews the EPCglobal Network Architecture and provides a detailed explanation and critique of the proposed framework. Although this architecture has been designed as a very robust and scalable architecture, there are some limitations related to delay and reliability. The thesis proposes an enhancement to the EPCglobal Architecture to mitigate these issues. The new architecture suggests that RFID users maintain a local database that synchronizes with the Master database, so that under certain conditions, data can be retrieved locally instead of going over the network for identification of every single item. The thesis compares the two architectures using a cost function analysis. Also, as a means to showcase the advantages of the proposed enhancement as well as to enable future users and researchers to have access to a RFID test environment, a complete, scalable RFID simulator has been built from the ground up. The results show that the proposed solution demonstrates distinct performance improvements over the original architecture and also increases the reliability of the system.