In the last year, Radio Frequency Identification (RFID) caught the imagination as the umbrella beneath which a set of next generation technologies is being developed that should significantly improve supply chain operations. But it is only recently that RFID architectures have emerged so that the technology can be introduced into the applications landscape.

Fortunately, understanding the basics of an RFID architecture isn’t arduous and there are emerging standards aplenty to help you.

Unfortunately, this is a rapidly expanding area of interest and so what you see today may not be quite how an RFID architecture looks like in the future. What’s most frustrating is getting hard facts about how companies are faring in their implementations.

You can take one of two views: RFID is over-blown and over-hyped, or it is so successful as a cost saver that companies won’t talk about it because of the competitive advantage they’ve gained. Nevertheless, enough is known for us to lay out the bones of an RFID architecture.

Using a model outlined by EPCglobal, a joint collaboration among academics and about 100 supply chain and RFID suppliers, the architecture consists of:

•Passive RFID tags, which transmit Electronic Product Codes (EPC) when exposed to an RFID reader.

•RFID readers which activate the tag and reads its response

•An Object Naming Service (ONS) database, which acts like a reverse telephone directory - it receives a number and produces a network address.

•A Product Markup Language (PML) server which stores comprehensive data about manufacturers' products. It recognises the incoming EPCs as belonging to specific items. Since it knows the location of the reader which sent the query the system now also knows which plant produced the item.

•Integrations to existing supply chain, warehousing, conveyor and ERP applications. This can be an applications server that uses SOAP across HTTPS but might equally be an enterprise EAI layer

The Electronic Product Codes element is where the guts of the information lies but of itself, it tells you little more about a package, or pallet, than the number plate tells you about your car. In an industrial setting, systems need to access and distribute a lot more data about items for that information to be useful. In response, EPCglobal is developing a series of emerging standards along with a raft of technologies designed specifically for RFID. At present, the favoured EPC is 96-bit, held on a 128-bit chip, but that could easily expand in the future.

A core component is the Savant server. Savant software uses a distributed architecture and is organised in a hierarchy that manages the flow of data. Savants will run in stores, distribution centres, regional offices and factories. Savants at each level will gather, store and act on information and interact with other Savants. Apart from resolving tag-specific issues, like read errors, Savants perform two critical tasks.

First, they maintain a real-time, in-memory event database (RIED) to overcome the problems associated with very high levels of data processing associated with large numbers of RFID tags. The system will take EPC data that is generated in real time and store it intelligently, so that other enterprise applications have access to the information but without overloading enterprise databases.

Second, all Savants, regardless of their level in the hierarchy, feature a Task Management System (TMS), which enables them to perform data management and data monitoring using customizable tasks. For example, a Savant running in a store might be programmed to alert the stockroom manager when product on the shelves drops below a certain level.

This is all very well but in order to be truly useful, this technology bundle needs to be tied into applications. Despite the dearth of hard facts, it is recognized that RFID holds the potential for exposing complexity. This is because CPG and retail organisations – the current main markets for this technology - have typically created bespoke application architectures that drive their businesses. In response, vendors like Manhattan Associates and Provia are attempting to shield companies inside the supply chain from complexity by offering ‘RFID in a box'.

Accenture is working on a similar idea and SAP recently announced a specific solution for its applications. But these solutions are only any good if they work easily with existing supply chain execution (SCE) systems, or where SCE is being implemented on the back of RFID. Those projects are likely to be in the minority but at least the solutions on offer will allow companies to get pilots off the ground.