How does the “open standards” approach create this problem?
The premise behind the International Standards Organisation (ISO) is to create so-called “open standards.” Generic solutions. An ISO standard is designed to be a sort of “cook book recipe” that any manufacturer may follow to create a product that complies with the standard. ISO does not police compliance. Compliance is entirely voluntary. This approach to standardization works extremely well for modem protocols and paper sizes, and it is also suitable for identification of dairy cows in closed loop systems and the like, but it is flawed if the premise is to generate unique ID numbers for use in open loop systems.

ISO 11784/85 "STANDARD" WITH BLEMISH
A discussion of the ISO standad for RFID: its provenance, feasibility and limitations

The ISO transponder-based standard was originally intended for agricultural use to identify agricultural equipment and livestock. During the evolution of the standard, it was expanded to include companion and exotic animals as well as endangered species. The standard has a number of serious flaws which make it unsuitable for its intended applications.

Background
In the barcode market, the product being supplied is a commodity: ink on paper, with the associated production means (printers) and readers. The readers and printers of different manufacturers may provide different features. However, they are by and large interchangeable in that they can all produce/read the desired barcode. This type of interchangeability has not been forthcoming in RFID, due to the fact that the technologies involved are significantly more complex and highly proprietary. There are currently a number of systems on the market, which are based on a range of mutually incompatible technological approaches. No parallel can be drawn between barcode and RFID: RFID does not lend itself to standardisation in the conventional sense.

Approaches to Standardisation
A number of standardisation efforts have been initiated for animal applications by and on behalf of various user groups. Two distinct approaches have been taken by all of these past and current standardisation efforts. In a nut-shell these are:

• Transponder-based standard: where all manufacturers choosing to compete in a particular market manufacture transponders conforming to one set of specifications, utilising technology and protocols in the public domain. Transponder-based standards are based on OPEN STANDARDS.
  
• Reader-based standard: where all manufacturers competing in a particular market cross-license their reader interfaces, so that each manufacturer can provide readers that read the other manufacturers' transponders.

Companion animal user organizations, AIM USA and other user groups are supporting the reader-based standard to ensure that the existing installed base of microchipped animals is protected.

The essential requirement for RFID to do its job for companion animal identification is that positive identification of each animal must be assured. On a global basis this can only be achieved by unique, unduplicated ID codes. The uniqueness of the code numbers must be assured. In both read-only and read-write systems, the code number must provide unique and positive identification of animals in the manner of a license plate number. The presence of duplicate code numbers compromises the integrity of the identification system. RFID with duplicate numbers provides very little in the way of improvement over existing identification methods for animals (such as tattooing). The existence of duplicate code numbers opens the way to rampant fraud and record-keeping problems.

Transponder-based standard
OPEN STANDARD
Methodology: one company's system is selected (FECAVA's approach for the European companion animals), or an entirely new system is specified (ISO's approach for livestock and, now, companion animals). All relevant technology, including the interface and protocols, is placed in the public domain. Interested companies can then manufacture transponders complying to the new open standard, without having to pay royalties or only minimal royalties, to the patent owner.

Intended benefit: an unlimited number of manufacturers can produce transponders independently, without having to pay any royalties or only minimal royalties to patent holders. Permits multiple-sourcing of transponders and readers.

Implementation: An OPEN STANDARD relies on an honour system, the implication being, that somehow all manufacturers will agree on who manufactures which numbers, to prevent duplicates.

Caveat: Without legal barriers to the uncontrolled use of the technology (in the form of patents and trademarks) there is no way to enforce compliance by various manufacturers.

A case study: the FECAVA "open standard"
FECAVA proposed the Destron/Hughes system as the basis of its "open standard." Destron has placed its interface and protocols in the public domain, and any company can manufacture any transponder ID number using the Destron protocol quite legitimately. The transponders can all be read by Destron readers and look exactly like Destron transponders. They are indistinguishable clones. Currently there are several companies manufacturing product or preparing to manufacture transponders complying with this OPEN STANDARD. Destron is programming from one end of the numbers spectrum and AVID, another manufacturer subscribing to the standard, is programming from the other end. It will be a while, but at some point the programmed numbers will begin to overlap and duplicates will begin to appear. A third company has declared that it will sell Destron-clone transponders. (1) It has announced that, for a price, it will manufacture any code number the user desires. They call this service the Designer Chip. There will undoubtedly be further entrants if the market proves lucrative. Clearly, no "orderly" approach can be maintained under these circumstances. It is likely that duplicate numbers have already occured.

The reason people want to use RFID in their pets is to make sure they can be identified for life. Therefore, the ID chip of a champion cocker spaniel should not be duplicated in a mongrel dog or in a goldfish. This unique RFID number is associated to a single animal's record in the breed registry and/or a national recovery database. The uniqueness of the ID numbers is essential for any national database concept to work. The presence of a multitude of manufacturers, all building transponders under a single protocol, opens the door to duplication of numbers, whether on purpose or by accident.

The ISO "open standard"
The technology was developed specifically for livestock identification, based on a proposal to WG3 of the ISO by the Group of Four (AEG, Datamars, Nedap and Trovan) and Texas Instruments. Subsequently expanded to companion animals, the ISO "open standard" has been five years in the making, and is now in legal limbo. Two companies have begun asserting patents in the market in a manner that is non-conformant to the ISO patent policy. It is doubtful whether, under these circumstances, the ISO "open standard" is a viable approach to standardisation.

This problem, while well-known to the ISO officials, was never disclosed by them to the voting members prior to solicitation of their yes vote for the standard.

There are a number of other problems that would adversely impact implementation of the standard in its present form.

The legal problems
ISO/DIS 11785 is impacted by a Destron patent and an AVID patent, which effectively violate the ISO patent policy which states:

Clause A.2 (b) "if the right holder does not provide a [Letter of Intent to comply with ISO Patent Policy], the technical committee or sub-committee concerned shall not proceed with inclusion of an item covered by a patent right in the international Standard without authorization from ISO Council or IEC Council."

Destron, the holder of U.S. Patent Number 5,211,129, pertaining to glass encapsulated implantable transponders which are the subject of the ISO/DIS 11785 standard, has in the past made demands from their competitors amounting to $3 million up front license fees and 7% royalties (for a single regional license). Their recent demand, calculated on wholesale prices, was 22.4% royalties. These demands should be of the utmost concern to user groups, because they would eliminate many potential vendors from the market at the discretion of the patent holders.

There are two more patents issued to AVID under U.S. Patent Number 5,235,326 and U.S. Patent Number 5,499,017. In this case, AVID has clearly stated in their letters to the ISO that they will not conform to the ISO's Patent Policy. The patent conflicts affecting ISO/DIS 11785 have not been resolved.

The problem of duplicate numbers
Like the FECAVA "open standard," the ISO standard is a transponder-based "open standard." The ISO "open standard," like FECAVA, is based on technology that is currently in the public domain. Because there are no legal "teeth" (in the form of patented technology), there is no means to interdict the production of unsanctioned transponders or to prevent their being imported into individual European countries. The ISO "open standard" by its nature depends upon an honour code. It is susceptible to compromise by manufacturers, whose cooperation cannot be enforced.

There are three ways that the uniqueness of ID codes can be undermined under any "open standard":

• Pre-programmed/Factory-programmed ICs.
  Chips can be ordered ex works, factory-programmed, with the desired ID number. There have already been companies offering "made-to-measure," factory programmed ICs. (see ID Boutique's "designer chip") for a number of years.

More recently, several manufacturers are selling OTP or re-programmable transponders that conform with ISO 11784/85. These are ISO-compliant transponders which are field programmable by the user. Physically, they are indistinguishable from the factory-programmed IC.

• OTP/WORM ICs
  OTP = One-time-programmable. WORM = Write-once-read-many.
  These can be programmed to the desired code number in the field, by the end user. This is the configuration which as a matter of fact was discussed in the ISO SC19, WG3 meetings as the preferred configuration, as it allows manufacturers to minimise the number of transponders held in inventory.
  
• WMRM ICs
  WMRM = Write-many-read-many.
  These can be reprogrammed as many times as desired. The code could potentially even be changed once it is implanted in the animal. Several such products are already on the market today or on the verge of being introduced.
  
  Some WMRM type transponders may mimic pre-programmed transponders, because they can be temporarily "locked," and accessed for reprogramming through the use of a password code known only to privileged users.
  
  One possible scenario would be that animals from an embargoed country (due to a highly infectious disease like BSE) could have their code numbers changed en route to their destination. The point of origin of such animals could be completely obscured.

Due to the fact that the ISO standard is an open standard, even obtaining transponders from selected manufacturers and filtering all these ISO-conforming transponders through a single hub would not prevent ISO-conforming transponders with duplicate ID numbers from entering the market.

The ISO "open standard" by its nature depends upon an honour code because the ISO organisation does not enforce compliance with its standards. As a matter of fact, it appears that three companies have already declared that they will make codes to order when the ISO standard is implemented - without going through the Brussels bureaucracy to have numbers assigned.

In the ISO system, as currently specified, corruption of the ID numbering system is virtually built-in. While duplicate ID numbers are of no consequence in the small, closed-loop dairy herd and feed lot applications the ISO standard originally targeted, they would wreak havoc in national database programs. As of this writing, several manufacturers have entered the market with custom-programmed, OTP and read/write transponders. ICs of all three types are now readily available on the market. The problem of duplicate numbers is inherent to transponder-based "open standards," including the ISO standard.

Furthermore, the ISO 11784 stipulates that, even for manufacturers who adhere strictly to the ISO honour code, identification numbers can be "recycled" every 33 years. The recycling of numbers poses no problem in the original, intended target market, as cattle and pigs lives ordinarily do not exceed five to seven years. (Agricultural use for livestock and equipment was the original, intended target application of the ISO "open standard.") Some companies have lobbied heavily to have the I.S.O. standard expanded to small animals, in spite of its obvious lack of suitability. In long-lived animals, such as exotics, endangered species in controlled breeding projects, the 33 year time frame may pose some problems.

The problem of manufacturers' accountability
A pervasive though little-discussed problem in all "open standards" is the accountability of individual manufacturers. Seen from the outside, an implantable, glass-encapsulated transponder is virtually indistinguishable from other makes of the same size. The problem becomes even more acute once the transponder is implanted. All "open standard" transponders can be read by a common reader protocol. The idea of requiring each participating manufacturer to program an assigned manufacturer's code into his transponders was put forward, in order to give the user a means of distinguishing among different transponder makes, in the event quality control problems should arise. In this way, so the thinking went, it would be possible for the user to identify inferior product, or manufacturers who didn't strictly control their numbering schemes, and avoid the offending manufacturer's product in the future.

The problem with RFID, however, is that a defective transponder is a non-functional transponder. Consequently, it would not be possible to read the manufacturer's code in those instances where it would be most urgent to do so. The owner of a prize Korat cat, or even of a beloved household pet, would hardly suffer his pet to be subjected to surgery to remove a transponder, so that said transponder could be submitted for a lengthy and costly laboratory analysis to ascertain the manufacturer's code.

Even if the manufacturer's code of a defective (failed) transponder could be ascertained, which is highly doubtful, there is no guarantee that it is, in fact, the code of said manufacturer. Some manufacturers, in their desire to preempt FECAVA's situation, have requested a Rome-based institute (ICAR) to take upon itself the job of allocating code numbers to interested manufacturers. It must be noted, however, that ISO 11784 and ISO 11785 do not assign this role to ICAR or to any other party. Neither ICAR nor ISO can enforce their respective recommendations. In an environment where there is uncontrolled manufacture of transponders with ID codes made to order, manufacturer's codes are no guarantee for product origin, quality and manufacturer's accountability.

In the world of the ISO open standard, there is no guarantee that a manufacturer's three-digit ID number ensures that that this manufacturer has in fact made the transponder in question. Some of the available custom-programmed, OTP and read/write transponders already duplicate ID numbers provided by Destron, Datamars and Texas Instruments.

The problem of transponder performance
Neither ISO 11784 nor ISO/DIS 11785 stipulates any minimum transponder performance requirements for microtransponders (the size of transponder that is suitable for use in companion animals, exotics etc.) Therefore, a transponder reading at "touch" reading distance (a matter of 1 cm or less) would be fully ISO-compliant. Companion animal veterinaries around the world have repeatedly expressed their strong reservations about systems with short reading distances (2). Users in the livestock business require even longer read ranges in order for RFID to be effective for them. ISO-compliance is therefore no guarantee to the veterinary of suitability of a given RFID product for use in animal applications.

The problem of feasibility: can the product be manufactured cost-effectively, in large quantities?
It should be understood by the consumer that the ISO process is a committee-based process. Compromise is the key tool to arriving at any solution. Like all committee-based approaches involving manufacturers with conflicting interests in a high-stakes market, this is an extremely lengthy and political process. The hybrid solutions that result from such committee-based engineering efforts are based on political compromises and not on performance considerations, cost control or technical feasibility.

• Two standards in one: a political compromise
  Currently there are two fundamentally different design principles being championed in the passive, low-frequency RFID market: the full-duplex approach and the half-duplex approach.
  
  By way of a quick back-ground explanation, the transponder has no energy source of its own but receives its power from the interrogation signal that is transmitted by the transmitter/receiver antenna. The resonance circuit of the transponder oscillates at the same frequency as that of the interrogation signal and charges the power capacitor.
  
  At a certain voltage level on the power capacitor, the transponder oscillates at the same frequency as that of the interrogation signal and charges the power capacitor. At a certain voltage level on the power capacitor, the transponder electronics are ready to start operating and transmit the transponder code (return signal) via the coil of the resonant circuit.
  
  With a so-called full-duplex approach (FDX) the return signal initiates as soon as the beginning of the interrogation signal is received and the smoothing capacitor has been charged. The return signal is received repetitively and without interruption for as long as a continuous interrogation signal is maintained. An FDX-transponder therefore does not have to store energy to be able to return its entire code.
  
  With a half-duplex approach (HDX) the return signal starts only after the end of the interrogation signal has been received and only after the storage capacitor has been fully charged. The return signal is then only sent once, since the transponder has emptied its storage capacitor after it has sent its code.
  
  The full-duplex and half-duplex approaches are fundamentally incompatible.
  
  Currently only one company is marketing a half-duplex system. Every other company in the business today is marketing a full-duplex system of one type or another. In spite of the fact that a preponderance of the manufacturers has proceeded with the full-duplex approach, a political compromise has been reached in the ISO WG3 which stipulates that the new ISO standard has incorporated both half-duplex and full-duplex technology.
  
  As conceptualised in the ISO 11784 and 11785 documents, ISO-compliant transponders can be either full-duplex transponders or half-duplex transponders, and the readers will have to read both types of transponders. In essence, this will require two readers in one box.
  
  A reader capable of detecting both HDX and FDX transponders would involve the disadvantages inherent in the HDX-system degrading the performance of FDX. This degradation, unfortunately, cannot be avoided.
  
  Of course it is technically possible to design and manufacture a reader that can read two systems i.e. HDX and FDX. In the case of a reader compatible with HDX, an FDX transponder can easily be read during the time that the power capacitor of an HDX transponder is being charged, in order for the IC to give its code at a later moment in time.
  
  The drawback of such a reader is that its timing is made dependent on the slower HDX-system. Thus, the read speed performance of FDX, will be reduced to that of HDX. Repeated reliable readings with moving transponders or readers can thus not be performed.
  
  Needless to say, the resulting readers will be less efficient and more costly than readers designed to read only a single type of technology. In the envisioned core application area, livestock, and now companion animal identification, cost of RFID has already proven to be a major sticking point in acceptance of the technology. It remains to be seen whether the users will find this political compromise, resulting in higher costs to them, palatable.
  
  Strictly speaking, this "compromise" has resulted in a standard that is no standard, since it embraces the co-existence of two mutually incompatible systems.
  
• The issue of transponder design: a "high-Q" product
  Both the FDX and HDX components of the ISO "open standard" are conceptualised as so-called "high-Q" transponders. Reading distances for high-Q systems are accomplished primarily by transponder design and require careful tuning of both the transponder and the reader. The result is an expensive and cumbersome manufacturing process with comparatively lower yields, resulting in a more-costly transponder.

The conceptual standard proposed to ISO WG3 by the Group of Four and T.I. was intended for use in livestock and agricultural equipment. The code structure of the standard reflects this pedigree.

Premises Identification = Encumbrance