EPCglobal and Auto-ID Center Building Blocks
May 24, 2004
The MIT Auto-ID Centers, now part of EPCglobal, developed four key elements to help realise The Internet of Things in its ultimate form of uniquely identifying thousands of trillions of items in an open system modelled on barcode numbering. These are:
1. The Electronic Product Code (ePC), which is intended to uniquely identify physical objects.
2. The Object Naming Service (ONS), which acts as the 'glue', linking the ePC with its associated data file.
3. The Physical Mark-up Language (PML), which is intended to be the standard in which networked information about physical objects is written.
4. Filters, which is distributed application software.
We now look at these systems, and their options, in more detail.
The Electronic Product Code (ePC)
The Auto-ID Center at MIT describes its system as "A numbering scheme that can provide unique identification for physical objects, assemblies and systems. Information is not stored directly within the code - rather, the code serves as a reference for networked (or Internet-based) information, in other words, the code is an 'address' - it tells the computer where it should go to find information on the Internet.
The ePC requires relatively few parameters to determine the design:
Number of bits - i.e. How much information is needed to provide a unique identity in every single product manufactured, sold and consumed in the global supply chain?
Bit "partitions" - i.e. What is the best way to organise - or "break up" - the numbers/figures so that we achieve as many unique combinations as possible, while also expediting Internet searches?
Consider this an exercise in determining the best "search hierarchy" - like a postal address - which goes from country to city, to zip code, to street, to house and individual. As the detail or level of the hierarchy increase, the speed and accuracy of the search will likewise increase, but the possible combinations of unique numbers will decrease."
To be precise, MIT Auto-ID Center call for 96 bits in the smart label for The Internet of Things and, in some presentations, far more. The data are to be employed in four zones mimicking barcode numbering systems but running to far more numbers because barcodes rarely identify anything uniquely and the Auto- ID Centers headquartered at MIT intend to facilitate tagging of even greater numbers of items than barcodes today.
Object Naming Service
The MIT Object Naming Service (ONS) 'tells computer systems where to locate information on the Internet about any object that carries an ePC (electronic Product Code).
ONS was developed at the Massachusetts Institute of Technology by Dr David Brock, Professor Sanjay Sarma and Joseph Foley. ONS is similar to - and (in part) based on - the Internet?s existing DNS (Domain Name System), which allows Internet routing computers to identify where the pages associated with a particular Web site are stored.
The DNS is used every time a Web site is accessed.
The ONS will be used every time information is needed about a physical object. It is likely that the ONS will be many times larger than today's DNS.
Although conceptually simple, designing ONS was a challenge. The system must be capable of quickly locating data for every single one of the trillions of objects that could potentially carry an ePC code in the future. The ONS must serve as a lightning-fast post office that, on a daily basis, receives and delivers millions (if not billions) of letters.'
Product Mark-up language of MIT
Product mark-up language, or PML, is billed to be a standard "language" for describing physical objects. It will be based on the extensible Mark-up Language (XML).
Today, HTML (Hyper Text Mark-up Language) is the common language on which most websites are based, allowing individuals to surf the Internet from their desktops regardless of the type of computer or operating system used. Where HTML tells a computer how information should be displayed (e.g. what colour and size it should be) - XML goes a step further, telling the computer what kind of information it is viewing (e.g. an address or a telephone number). The PML will go even further, building in layers of increasingly specific data in order to describe physical objects, their configuration and state. In the end, PML:
- Should translate or contain static data such as dosage, shipping, expiration, advertising and recycling information.
- Should provide instructions for machines that ?process?or alter a product, such as: microwaves, laundry appliances, machine tools and industrial equipment.
- May need to communicate dynamic data: information that changes as a product ages or as it is consumed, such as: volume, temperature, moisture and pressure.
- May need to include software, or programs, which describe how an object behaves, for instance: a PML file may contain the program which describes how fast the tyres on your car will wear before they need to be replaced, or how fast an object may burn in case of a fire.
Filters (formerly called Savant)
In a world where every object has an RFID tag, interrogators will report a continual stream of ePC's codes. Managing and moving data is a difficult problem and one that must be overcome if a global RFID network is to be of value. The Auto-ID Center has made software technology called Filters, formerly Savant, to act as the nervous system.
The Filter system is different from most enterprise software. It is not one overarching application. Instead, it uses a distributed architecture and is organised in a hierarchy that manages the data. There will be Filter software running in stores, distribution centers, regional offices, factories, perhaps even on trucks cargo planes, according to the Auto-ID Centers. Filters at each level will gather, store and process information and interact with other Filters. For instance, a system at a store might inform a distribution center that more product is needed. The Filter system at the distribution center will inform the store Filter that a shipment was despatched on time. Some of the tasks the Filters will handle are given below:
- Data smoothing - The Filter software at the edge of the network - the interrogators - will smooth data. Not every tag is read every time. Sometimes a tag is read incorrectly. By using algorithms the system is able to correct these errors.
- Reader co-ordination - If the signals from two readers overlap, they may read the same tag producing duplicate ePCs. One of the Filter System's tasks is to analyse reads and delete duplicate codes.
- Data forwarding - At each level, the Filter has to decide what information must be forwarded up and down the chain. For instance, Filter Software in a cold storage facility might forward only changes in the temperature of stored items.
- Data storage - Existing databases cannot handle the data of an Internet of Things fast enough, so another job of the Filter system is to maintain a time in-memory event database. In essence, the system takes the ePC data that is generated in real-time and stores it intelligently, so that other enterprise applications have access to information, but databases are not overloaded.
- Task management - All Filters, regardless of their level in the hierarchy, feature a Time Management System (TMS), which enables them to perform management and data monitoring using customisable tasks, for example, Filter Software running in a store might be programmed to alert the stockroom manager when product on the shelves is below a certain level.