Radio Frequency Identification (RFID), specifically the Electronic Product Code (EPC, the next generation of barcode), have improved the way the world works and lives. Basing EPC as an identification, GS1 developed standard framework for common visibility of supply chain data called EPCglobal. EPC information system(EPCIS) is one of the standards which encompasses interface for sharing and capturing data triggered by different events. Because the amount of data captured and stored in each event type is huge and the query interfaces are general, it is difficult to get traceability data easily in reasonable time. Even for some traceability queries like finding products at a location in certain time are too complex to fulfill with the current EPCIS standard implementation. In this paper, by extending EPCIS query interface we have proposed an efficient, well-organized, and scalable traceability architecture. In order to effectively collect and process events in real time using our stream processing, we first proposed an efficient pub/sub based EPCIS. Then, to retrieve information effectively we proposed an adequate column family based data modeling. Finally, we have implemented our system and developed a RESTfull service interface to query track and trace information. Experimental result shows that our traceability architecture with the column based data modeling process traceability-queries more efficiently and scales gracefully.