Does RFID demand new middleware? The architect members of the Auto ID Center, now EPC Global have always contended that purpose-built middleware was required to meet the new demands of RFID-enabled supply chains. Their Savant architecture has various supporters lead by OAT Systems and Sun Microsystems.
As far as I can tell, there are three main drivers for building specialized RFID middleware:
1. Very high transaction volumes
2. Unprecedented information sharing between trading partners in a supply chain
3. The competitive need to displace existing middleware
In theory, the broad adoption of RFID for tracking products will create vast distributed networks with very high transaction volumes. In practice, the transaction volumes are not high enough to exceed the capacity offered by existing technology. For example, Wal-Mart receives about 2 billion boxes of merchandise per year. Even if each box is tracked 10 separate times, that only adds up to an average transaction rate of about 600 transactions per second. Today’s state of the art transaction processing systems can easily support 10 to 30 times that volume.
The unprecedented sharing of information is just that – unprecedented. Once again let’s look at Wal-Mart – who has emerged as a leader in RFID adoption. Wal-Mart is requiring its top 100 suppliers to provide cartons and palates of their merchandise with RFID tags attached. Beyond that, there’s no requirement for the suppliers to actually track their merchandise, let alone share tracking information across the supply chain.
Ironically, the third driver in the list – the competitive need to displace existing middleware – might create the strongest market force of all. Various investors seem to be eager to invest in RFID middleware right now because it appears to be the next big thing. Unfortunately, this competitive driver does not itself create a need. Existing middleware companies appear to have adequate technology – for now. And the new business information sharing promised by RFID visionaries does not appear to be imminent – for now.
7:48:48 PM 
|
