Silicon Valley Musings

As RFID tags are introduced into the supply chain, I feel it will become necessary to have RFID event agents at the edge processing the fine-grain raw RFID data into meaningful events.

Lets look at the example of a shipment of toys from Hong Kong as it makes its way to a Palo Alto based toy store. The Hong Kong based manufacture packages dozens of items per case, and then puts thousands of cases into a shipping container for export. At every checkpoint or supply chain transition, the shipment would generate tens if not hundreds of thousands of RFID events.

One of the things the use of RFID tags it supposed to provide is the reduction of merchandise lost due to theft or errors in loading/routing. This can only be detected by the absence of one or several RFID events. One solution for this would be a central server that receives all the fine grain RFID events and matches them with logical groupings of product RFID tags.

Another approach would be to build software/hardware agents that are associated with each logical product grouping: the cases, the shipping container, the ship, the store shelf etc. These agents would be aware it product?s RFID tags. It could produce consolidated or ?roll-up? events like: my items is completely accounted for, the following items are missing: x, y, ?, item x is moving to the next transaction state (e.g. the case is being unpacked, or a toy is being taken off the store shelf).

These agents would produce asynchronous real-time events as well as report consolidated events on demand. RFID readers could be programmed to receive either the raw RFID tag data or the RFID event agent events.

TIBCO’s Hawk application monitor provided a very similar service for application-level events in the financial networks that were designed by Reuters and its clients. Hawk agents would consolidate or roll-up application events so that that a single IT administrator could monitor an entire global trading infrastructure for a single desktop workstation.