Skip to content

RFID and Auto-ID

Abstract

A radio-frequency identification tag is a small circuit that answers when a reader calls. The passive kind has no battery: it runs on the energy of the reader’s own signal and replies by reflecting part of it. The idea is older than the transistor’s commercial use, described by Harry Stockman in 1948 and patented in workable form by Mario Cardullo in 1973. For fifty years it was a niche technology for toll roads and cattle. Then, in 1999, a group at MIT proposed putting a five-cent tag with a unique number on every manufactured object, and coined the phrase “Internet of Things” for the result. The article follows the technology from wartime radar to the transit card in a wallet, including the card whose secret cipher was read off the chip with a microscope.

Reflected Power

The principle came from radar. In the Second World War both sides fitted aircraft with transponders that answered a radar pulse with a coded reply, so that operators could tell their own planes from the enemy’s (identification friend or foe). A transponder is an active radio. The passive variant has a stranger ancestor: in 1945 the Soviet inventor Leon Theremin built a listening device, later called “the Thing”, that had no power source at all. It worked only while it was illuminated by a radio beam from outside, and it modulated the reflection with the sound in the room.

In October 1948 Harry Stockman published “Communication by Means of Reflected Power” in the Proceedings of the IRE. A target that varies how much of an incoming radio wave it reflects is transmitting information without a transmitter. Stockman said that considerable development work remained before the idea became useful, and it took about twenty-five years.

The First Tags

Mario Cardullo wrote a business plan for a passive transponder with a memory in 1969 and demonstrated a prototype with 16 bits of memory to the New York Port Authority in 1971, as a device for collecting tolls. His patent was granted on 23 January 1973. In the same year Steven Depp, Alfred Koelle, and Robert Freyman at Los Alamos demonstrated tags that worked by modulated backscatter at 915 MHz and carried 12 bits. Charles Walton, a California inventor, received in 1983 the first patent to use the abbreviation RFID.

The early uses were those where a tag costing several dollars was no obstacle: railway wagons, livestock, access badges, cars passing a toll gantry. A printed barcode cost nothing and identified a product type. A tag cost money and could identify a single object, without a line of sight, and many at once.

The Auto-ID Center

In 1999 Kevin Ashton, a brand manager at Procter & Gamble whose team’s Oil of Olay lipstick shade kept disappearing from store shelves, founded the Auto-ID Center at MIT with the researchers David Brock, Daniel Engels, Sanjay Sarma, and Sunny Siu. Procter & Gamble and Gillette were among the funders. The plan inverted the engineering of the tag. Earlier designs stored data on the chip. The Center’s tag would hold only a number, the Electronic Product Code, and everything known about the object would be kept in a database reachable over the Internet. A chip that stores only a number can be very small and very cheap. Ashton’s title for a 1999 presentation on the idea, “Internet of Things”, outlived the project (see The IoT Revolution).

The Center was dissolved in October 2003. Its standards work went to a new body, EPCglobal, and its research to the Auto-ID Labs, a network of seven universities. EPCglobal approved the second-generation air interface for UHF tags, known as Gen2, in December 2004, and the International Organization for Standardization adopted it as ISO 18000-6C in 2006. That is the tag now sewn into clothing labels and stuck on pallets.

What happened when retailers tried to impose the tag on their suppliers, Wal-Mart’s mandate of 2003 and the Metro Group’s Future Store in Germany, is told in the Dead End section of Barcodes and the UPC. The short version is that by 2011 a passive tag still cost about nine cents, and a barcode still cost nothing.

Cards and Passports

The other branch of the technology works at a few centimeters and carries a processor or at least some memory. In 1994 the Austrian company Mikron, in Gratkorn near Graz, introduced a contactless card chip for ticketing and named it MIFARE, for “Mikron fare collection”. Philips bought Mikron in 1995. The card became the basis of public transport ticketing in many cities, among them the Oyster card in London and the OV-chipkaart in the Netherlands, and of door-access systems everywhere (see The Smart Card). According to the manufacturer, since 2006 called NXP, ten billion of its contactless chips have been sold.

Passports followed. Malaysia issued the first passport with an embedded chip in 1998, and the United States began issuing them in the mid-2000s.

⚠️ Dead End: Crypto-1

MIFARE Classic protected its contents with a cipher called Crypto-1, which its maker kept secret. The key was 48 bits long. In December 2007 Karsten Nohl and Henryk PlΓΆtz told the Chaos Communication Congress in Berlin that they had recovered part of the algorithm from the chip itself (see The History of Hacking). In March 2008 the Digital Security group at Radboud University Nijmegen announced that it had reverse-engineered the cipher completely and could clone an OV-chipkaart and alter its contents.

NXP went to court to stop the Nijmegen group from publishing. The court in Arnhem refused the injunction on 18 July 2008, and “Dismantling MIFARE Classic” was presented at the ESORICS conference on 4 October 2008. Transport operators had to replace a card system that had just been rolled out; London moved the Oyster card to a newer chip with a published cipher in 2011.

Cryptographers had argued since the nineteenth century that a cipher should stay secure when everything about it but the key is known. Crypto-1 was a test of the opposite policy at a scale of billions of cards, and the secret lasted thirteen years.

The Chip Under the Skin

In October 2004 the US Food and Drug Administration approved the first RFID chip for implantation in humans, a glass capsule from VeriChip intended to link a patient to a medical record. Two years later the journalist Annalee Newitz and the engineer Jonathan Westhues demonstrated that the signal of an implanted chip could be read and cloned by someone standing nearby. The same objection had been raised in Germany against tags in shop goods and loyalty cards: a tag answers any reader, and its bearer does not notice. Gen2 tags therefore have a command that disables them for good, and passport chips answer only after the reader has proved that it has seen the printed page.

πŸ“š Sources