Posted by Lori Ayre on May 29, 2010

I've been focused recently on developments in UHF technology and how it might help libraries improve on the RFID products we currently use (which are  based on HF tags).  My interest, of course, is having something that fits our library applications best and if UHF is a better fit, I'd rather know now than later after we have an even greater investment in the HF technology.

So, several of us have been discussing this very thing and I feel we finally got a pretty definitive response from Alan Butters who has been one of the people talking quite a bit about UHF  and it's promise (he certainly peaked my interest).  The response, co-authored with Paul Chartier (they are co-authors of ISO 28560-2) is provided here (with their permission):
As co-Project Editors of ISO 28560-2, we thought it useful to provide some input to the list.

Let's start with the installed base of RFID systems.  At a conference a few months ago, one of the major vendors of chips (they produce HF and UHF) estimated that there were 2500 libraries that had implemented RFID systems using high frequency technology.  The count for UHF was less than 1% of this number.  Our guess is that the 2500 might be a slightly low estimate.

When the work started on ISO 28560, we had to consider the installed base and the infrastructure and associated investment that had already been made.  Simply disenfranchising all the libraries that had adopted national models or proprietary solutions would have been unthinkable – the aim was to increase interoperability between existing systems.

As those on this list probably know. ISO 28560 is a three-part standard as follows:

* Part 1 defines the data element – almost certainly more than necessary for a single library, but trying to cover all aspects of circulation, collection management and the acquisition of library material.

* Part 2 defines a flexible encoding system based on ISO/IEC 15962, which is increasingly being used in a number of applications and is independent of frequency and air interface protocol.

* Part 3 is a fixed structure data model, effectively standardising the Danish national model.

These three standards together will allow libraries to move away from proprietary data models or systems providing an increased degree of interoperability. Furthermore, it may be possible to intermix old and new tag formats in the same library system.

Both Part 2 and Part 3 specify the use of ISO/IEC 18000-3 Mode 1 tags (also known as ISO/IEC 15693 when used as a smart card).  This is a mature technology that is tried and tested, but is probably no longer the best RFID technology (other than its extensive installed base) with a wide choice of vendors for chips, tags and reading devices. However, it does still have some strong features that are not necessarily supported by other tags and one particular feature is the capability of selective memory locking.

Now, let's consider the position of UHF which has failed to make extensive market penetration in any library sector as was expected by some at the time that we started work on ISO 28560. UHF technology could not have been used universally a few years ago, because it was only recently approved by radio regulators in different regions of the world. There are still significant differences in performance capabilities based on regional and national radio regulations.

Although there are other UHF technologies, I am assuming that the discussions are about ISO/IEC 18000-6C tags (also known as EPCglobal Class 1 Gen 2). If we consider some of the features not mentioned in Lori's e-mail, we might identify some of the challenges that libraries will face in selecting this technology:

* Some tags only support a 96-bit identifier with no additional user memory, so there is no step forward from this tag being an electronic bar code.

* Even those tags that have additional memory currently do not support selective locking.

* Not all the tags support a unique serial number of the chip itself.

There is a gradual move away from some of these constraints, but libraries and specialist vendors need to be aware that these constraints will be around in some products for years to come.  So selecting the tag without an understanding of its capabilities could cause problems for libraries.

Because Paul is heavily involved in ISO standardisation of the technology, the committee developing ISO 28560 has cognisance of the fact that there is the possibility of a new 18000-3 Mode 3 high frequency tag which has some of the features of the 18000-6C tag.  This has just been approved for publication and the library community needs to wait for products and a proper evaluation. So, even if we acknowledge that HF might not be the ideal platform for library RFID systems, UHF might not turn out to be the ideal replacement anyway – there may be other options to consider.

Now to some comments based on the heading of the CityU report.

* Performance will be better for UHF based on the fact that this is a new technology.  But 18000-3 Mode 3 will have bit transfer rates across the air interface that are up to four times (even eight times, if an optional feature is supported by the tag) than that of the present 18000-3 Mode 1 tags.

* An extensive read range is not always a plus feature.  When UHF was introduced in the logistics field, cross talking between interrogators at different dock doors in warehouses caused significant operational problems until fixed.  The same also applied for airline baggage handling, where the read range enabled tags on other luggage on other conveyors to be read.  Certainly, in different parts of the world, there are privacy concerns about the read range where the RFID tag might be used for tracking an individual.  These concerns are such that legislation has already been passed – expect more in Europe.

The discussion on security gates is interesting, because the more extensive read range of UHF may have a positive impact, but in reality the extended read range may create a significant no-go zone for library users which could impact heavily on space utilisation. It is also necessary to consider UHF in the context of the three security systems discussed in ISO 28560, which are:

* The main solution is to toggle between two different values of the tag’s AFI. This is possible with an 18000-3 Mode 1 tag while still permitting selective locking of the primary item identifier – a situation desired by many libraries. This strategy is not possible with an 18000-6 Type C tag, nor an 18000-3 Mode 3 tag.  This is because the AFI is an integral part of what is known as Memory Bank 01 and this has to be locked or unlocked in its entirety.

* Some of the vendors have developed proprietary EAS solutions for the 18000-3 Mode 1 tag.  This second solution is not available in 18000-6C and it is not clear yet what will happen with 18000-3 Mode 3 tags.

* The final security measure is to use a unique chip ID, which is an essential component of an 18000-3 Mode 1 tag that is at the core of the "hand-shaking" process with the communications protocol.  This unique chip ID is optional in 18000-6C and 3 Mode 3, and is not used for communication purposes.

It was interesting to read that "CityU plans to adopt the data model of the US (ISO 15962) with slight modifications but is likely to follow ISO 28560, once adopted, which supports backwards compatibility and flexibility."  As co-Project Editor for ISO 28560-2 and Project Editor for ISO/IEC 15962, Paul is happy to assert that this solution is NOT A STANDARD, it is just another proprietary solution that will create problems of interoperability.

So, let's take stock of what might happen in the future with ISO 28560, while we are still awaiting the final Yes/No ballot for the three standards. These standards will be implemented to different degrees by libraries with existing RFID systems, and certainly by libraries yet to embark on RFID. When the standards were being developed, the experts on the ISO committee were obviously aware of both 18000-6C (a reality) and the impending 18000-3 Mode 3 standards. One likely outcome is for the next generation of library RFID standards to be based on 18000-3 Mode 3 tags, which will require either a new part to ISO 28560 or enhancements to existing parts. The tags will be different but, because of the significant installed base, there might be a degree of interoperability in readers that can be achieved with an engineering upgrade.

Because of the laws of physics, adopting UHF will certainly require different equipment, but it might still be possible for a future part of ISO 28560 to support UHF technology. However, before this can be done, the library community needs to be clear not just on aspects of the technology, but also on potential business and competitive issues that might exist with UHF technology. To support a business case, not only does the solution need to be based on suitable technology but the supplier community must build systems that take advantage of the strengths of the technology platform in such a way as to provide a compelling case for its purchase. Particularly is this so in the context of a market heavily dominated by HF solutions which themselves continue to evolve.

As soon as libraries embark on adopting standards, the standards makers have to take into account infrastructure issues and the installed base. A balance needs to be struck between the benefits of interoperability with existing technology versus switching to new technology. This is why most railway track gauges around the world have not changed for a hundred years, because the cost of infrastructure changing is too high compared to the benefits. The GS1 bar code (previously known as EAN and UPC) that we see on most retail products including books, is based on a technology developed in 1973.  37 years later, the same technology is being used even though there are significantly "better" bar code symbologies available.  So while GS1 has added a number of new bar code symbologies, the old symbol might still make its 50th birthday!

Also, even if one is of the view that UHF is the “right” technology for libraries, the corollary that HF is the “wrong” technology does not necessarily follow. This is an important point. To miss this point is to suggest that project success and ROI in libraries cannot be achieved with HF platforms. There are simply too many successful library RFID projects around the world for this to be the case. This is not to say that library RFID systems couldn’t be better – of course they could and my own investigations suggest that implementing a UHF based platform could contribute some useful improvements. But are all library RFID systems based on HF really so bad that we should just push back and refuse to embrace the technology until something new comes along? Libraries will make their own call but I think not.

So, whatever we do in the future with the new generation HF or the existing UHF technology, recognition of the millions of Dollars (or Euros) of investment already made by libraries in RFID technology must be made. The door continues to stay open, but the work of the ISO committee must proceed with full acknowledgement of the extensive investments made by libraries over the past 12 or so years.

Kind regards,

Alan Butters and Paul Chartier
My take-away is this:  don't be hesitant about moving to RFID because of potential developments with UHF.  Yes, the technology has some benefits over HF (that it didn't have even a few years ago) but it still has some drawbacks that are potentially bigger drawbacks than the devil we already know (HF tags).

But.  Let's keep demanding something better.  There is very definitely room for improvement.  I hope Alan and Paul are right that we will continue to see more improvements  in existing HF-based systems without sacrificing interoperability while maintaining some measure of backwards compatibility.  From their lips to....