WCCN Publishing, Inc.

The Case for a $15 Gen2 Tag

&

Non-EPC Compliant Object IDs

Synopsis

The academia vision of RFID-based Automatic ID Data Capture technology, encouraged by the Wal-Mart mandate and the EPCglobal Consortium, and complicated by the related intellectual property issues, have kept RFID deployment focused on Gen2-compliant UHF Passive Tags for essentially end-of-the-supply chain tracking applications.

The result is that all too many users and users-to-be simply assume that UHF passive is the tag type they should use, along with all that EPC represents. For many applications this is a correct assumption. But, for a goodly swath of applications, especially at other points in the supply chain, it is not.

The essential thesis of this discussion is that one should take a hard look at all tag type options, and the respective RF frequency alternatives. And all within the context of one’s overall automation systems application requirements – not just in treating data capture in isolation.

We more than suggest that there are instances when the $15 or $20 tag is far more functionality cost-effective than the proverbial, but elusive, 5¢ passive tag. And in a similar vein, for hard-core industrial object tracking, there are more appropriate alternatives to the EPC Object ID format.

In Search of the Ultimate Mobile Database

No need to convince Boeing of the benefits one can reap from the proper use of RFID. Boeing’s Ken Porad is fond of citing that circa 2004, the time required to build a plane once all the parts were assembled was 18 to 30 days. Then, using a mix of active, passive, and locator tags (all hardly standards compliant) reduced the time down to 72 hours. Ken doesn’t say it, but the resultant productivity gain falls beautifully and generously to Boeing’s bottom line. The average cost to Boeing for the active tags was $55 at that time.

Thus it was not surprising that, in 2006, Ken was willing to pay upwards of $15 per Gen2-compliant tag for use on the spare parts slated to be provisioned on Boeing’s new 787 Dreamliner plane. A startling premium over a tag that, at the time, was supposedly available for 8¢, and slated to go to 5¢, in the inlay form, and $4 in the tag form. But in the end, he settled on paying $5 per tag for a Gen2-compliant UHF passive tag with less functionality than he sought and for which he was willing to pay $15.

Boeing’s passive tag deployment in this case flies in the face of the vision of the Auto ID folks at MIT and other universities that adamantly contended that the pacing item for the widespread use of RFID was dependent on the availability of a passive tag inlay for 5¢ or less.

While no one can dispute that any technology has to be cost-effective before it can achieve widespread deployment, it must also provide meritorious functionality. Only in that context can one define what cost-effective means.

In the Boeing case, the functionality requirement was that the tag attached to each spare part on the airplane had to carry the life cycle history of the attached part. Details of the history vary depending upon the nature of the spare part. Details such as revision level, number of take-offs and landings experienced, ID and frequency of the maintenance person performing routine checks on the status of the parts, etc.

Obviously, this requirement mandated that the tag have sufficient data storage capacity. Boeing reckoned that to be 64K bytes of storage. To Boeing’s surprise, despite their willingness to accept a roughly six month lead time between placing a firm order and Boeing’s parts suppliers receiving the first batch of data storage-adequate tags, not a single major supplier could meet the requirement. And only one supplier, Intelleflex, could provide a Gen2 tag with a storage capacity of just 64K bits.

While Gen2-compliant, the Intelleflex tag was designed to be a battery-assisted UHF passive tag. The battery was primarily intended to support sensors attached to the Intelleflex tag. With the capability to store data about sensor activity a given, the Intelleflex tag had to be provisioned with data storage capacity. As a start-up, one could plausibly suspect that the anemic 64K bit capacity was a Model One version.

A Story of So Many Ignoring RFID’s Inherent & Market Required Attributes

Recognize what we have just related. After all the ballyhoo from the Auto ID visionaries’ and EPCglobal’s much heralded standards-making production, the commotion that Wal-Mart’s compliance requirements created, the imprudent spending of $280 Million by Symbol to gain a purported RFID capability that could not deliver, the massive amount of foreign funding Alien Technologies received, Intermec’s pursuit of royalties, the sums that Texas Instruments, Impinj, and the then Philips poured into producing UHF passive tags -- not one of these major players was able to provide a Gen2 tag that could function as substantially more than a provider of an (license plate) ID about the attached object (not much more than a bar code could provide).

The reason goes back to the original Auto ID vision. So intent was the focus on creating a 5¢ tag that, in the EPCglobal follow-through in drafting the Gen2 specification to this vision, critically attractive characteristics of RFID were excluded from the specification. Namely, the capability to maintain information on the tag about the life experiences of the object to which the tag is attached. Thus, the Gen2 specification just gave lip service to support for storage beyond that required for ID and ID related storage. While additional storage is optional in the Gen2 specification, there is no minimum, and no data storage security mechanism is spelled out, despite the draft writers acknowledging its necessity as the tag moves along the supply chain among different enterprises.

Stated another way, whether EPCglobal denies it or not, an implied premise behind Gen2 is that the tag’s usage, for all practical purposes, is as a license plate only tag.

Many more compromises were made in the name of arriving at a yet to be achieved 5¢ tag. For instance, not mandating support for a block read and write, and letting it be an optional feature, created an incredible burden on smart label printers when attempting to verify (at printer speed) something as simple as successful encoding of the license plate EPC ID on the smart label inlay. Another serious problem created in pursuit of a cheap tag was a fundamental lack of a sound RF approach for the UHF Band. This failure forced the EU Commission, of all entities, to step in and demand a fix. The result created a complicating level of post-installation support that tag suppliers were forced to accommodate – at no small cost to their customers. (Details of the RF deficiencies issues are best left for another discussion.)

Limiting, indirectly, the passive tag to an ID-only usage role has a commanding systems architectural following among certain network-centric persons. To these persons, there is a perceived elegance to having the details about a tagged object’s life resident somewhere on a network rather on the tag itself. And, it does satisfy an agenda of EPCglobal’s parent GS1. A component of GS1’s business strategy is to generate revenue by housing such data.

But this approach flies in the face of market requirements as dictated by enterprise users. They clearly have stated they want the data stored on the tag. Boeing again can make the point that a plane landing in an underdeveloped part of the world will not have network access, and thus requires that the critical data be tag-based. (To concerns about what would happen were the tag to become inoperative, the answer is to include a routine backup mechanism.)

Thus, the first point to be made within the context of this discussion is that the market requirement for tag usage is that the tag is expected to carry in its storage the life experience history of the attached object. The tag is expected to be the ultimate mobile database for transaction processing applications.

Thus, it follows that anyone who plans to deploy RFID passive-based systems is well advised to design and install systems that provide for tag data storage – even if the first deployment iteration is restricted to using the tag on a license plate basis only.

One Format to ID All Objects Ain’t Gonna Fly

A very related issue is the ID format to be employed. Here again, the Auto ID and EPCglobal visionaries got it wrong.

Their concept can be summarized in the cry to “Tag Everything”. As we approach the 22^nd Century that may prove to be the case, but we are a far cry from that potential eventuality. This early in the 21^st Century, the bar code is far from becoming extinct, and tagging everything makes no economic or functional sense. Nor does the idea that we have to be able always to physically identify each object separately from other identical objects.

It is overkill to be able to physically discriminate one bag of pretzels from a batch of identical bags of pretzels from a given supplier, or any other similar commodity. We may desire to recognize when the bag of pretzels reaches an expiration date. Or, we may wish to know the production lot or batch number in case of a recall. But the EPC ID format makes no direct accommodation for this.

When it comes to high priced objects, we appropriately may wish to be able to identify each individual object from a batch of like objects for asset control and fraudulent knock-off protection reasons. Or in certain cases, such as pharmaceuticals, the capability to individually identify like objects is justified for the same reasons, plus for tracking dispensing authorization.

Beyond the use of RFID for fast moving goods, the technology is perfectly suited for industrial goods. Here the capability to physically identify an individual object and keeping track of its life cycle experiences becomes important and a factor in cost management. Unlike the fast moving goods arena where the only change the object is likely to be subjected to is variation in packaging, an industrial object can be subjected to individual revisions or ECNs, to maintenance fixes, to usage and environmental violations, etc. -- all actions that should be recorded for service and other reasons. For such parts, it is impractical to assign an arbitrary EPC ID (as envisioned by the Auto ID folks) to such an object. The base ID required, beyond identifying the manufacturer, is to ID the object by its engineering part number. And then, to augment it with revision levels, etc. And in certain cases to change the part number entirely when the object is upgraded to operate like a new bill of material item – but with a reference to its birth ID maintained.

The point we are driving at here is that, increasingly, the use of Gen2-compliant tags does not automatically infer the use of the EPC ID format. Using the Boeing example, the entire aircraft industry has elected to empower its industry association to specify the ID format that will be used throughout the aircraft industry. The basic motivation is that its requirement is for a part number-based ID. The fact that this eliminates the need to pay GS1 for assigning an EPC formatted ID is a secondary, but not inconsiderable, reason.

Boeing’s motivation for supporting Gen2 tags was an attempt to un-complicate the systems requirements for its suppliers that have to support Gen2 for their commercial market products. But the dearth of memory incident caused the idea to be revisited.

As a consequence, recently Boeing teamed up with Fed Ex to experiment with the use of active tags on Fed Ex airplanes. Arguably, if Boeing felt a $15 passive tag was cost justified, then would the additional functionality of an active tag be cost justified at roughly $20 an active tag? The Fed Ex test will provide the answer. It’s not an easy answer, and it certainly will not be an unequivocal “No!”

Passive vs. Active Tags Trade-offs

Let’s drop the Boeing specificity example, and look at the passive-active tag trade-offs in generalities applicable to a wider range of users. In doing so, we must issue an important warning. The prices cited should not be viewed as absolute and cast in stone. They fairly represent the relative price differential at a point in time. They try to compensate for the vagaries of volume pricing and the dangerous practice of some suppliers to forward price. That is, to price based upon hoped-for cost savings and volume business they have yet to achieve. Again, the value of the prices cited is in establishing relative differentials between passive and active tags.

For illustrative purposes, we can view the cost of a UHF passive tag as between 8¢ and $4. The range differential represents whether the tag is delivered as an inlay or as a tag with the inlay housed (usually) in plastic. And the amount of memory provisioned will affect the pricing range.

The price for a battery-assisted Gen2-compliant tag ranges from $5 to $20; the difference depends on the amount of memory and the kinds of sensors employed. (The only noted supplier to date of Gen2-compliant battery-assisted tags is Intelleflex. The $5 tag can include a temperature sensor integrated on the tag chip. Since Intelleflex has not commented on pricing its tag with other than temperature sensors, the $20 figure is derived from estimations made by WCCN Publishing, Inc. A given is that only a temperature sensor can be integrated onto the tag chip. All others are tethered add-ons.)

Combinations of regular and battery-assisted UHF passive tags have an attractive cost of ownership advantage. They both can be supported by the same RFID LAN infrastructure.

The negative aspect of passive tags is that they do not have a radio. They can only “reflect” back an interrogation carrier signal received from a tag reader (after modulating the signal with the necessary response data). Thus, they cannot be self-actuating transmitters of information – even when an attached sensor may indicate a critical monitoring threshold has been reached. The fact can be stored, and transmitted only when the tag comes into range of a reader. However, for many applications, this limitation is acceptable, especially for fast moving goods.

Many in the media have harped on the range limitations of passive tag operation. In actuality, limited range is highly desirable in order to avoid the wrong tag reader accidentally engaging in communications with a tag, and thereby resulting in a wrong transactional recordation of what is happening to the tag’s attached object.

Passive Tag Bands

The reader of this discussion will note that, up to this point, the only passive tags referenced are those of tags that operate in the UHF Band and that are Gen2-compliant. While not intentional, this risks being misleading from two perspectives.

First, the UHF Band is not always the best band for passive tag operation. Especially for objects with sensitive radio wave absorbent characteristics, the HF Band is more appropriate, and for certain applications far more cost-effective. Additionally, there is a move by certain major systems suppliers to favor use of the 2.4 GHz Band for infrastructure reasons.

The point here is that while one is more likely to find the UHF Band the preferred band for one’s application, it should not be selected automatically – it should be a well-meditated selection. And there will be instances when one will have to build separate RFID LAN infrastructures capable of supporting mixed-band passive tags. A supermarket with an in-house pharmacy is a case in point.

The second possible misleading inference could result from referring to “Gen2” as a synonym for a standard. We did that for its easy recognition and convenience. However, any new RFID user of any passive or active tag solution is well advised to specify the appropriate ISO 18000 Standard when procuring tag and reader components. While ISO-18000-6c is a near-copy adoption of the Gen2 Standard, in time the likelihood of its divergence with Gen2 is very high. And it is the ISO Standard that one should hold one’s RFID suppliers accountable for.

To underscore the point, ISO refused to adopt Gen2 as the foundation for its 18000-6c Standard until EPCglobal agreed to incorporate a field for defining ID formats other than the EPC ID format. It was only when EPCglobal was convinced that ISO was serious about not putting an international standards approval on Gen2 that EPCglobal blinked on the issue. We expect that in time international demands will cause ISO to take the lead (via accredited standards making bodies), leaving EPCglobal with little option but to retro-upgrade its version of the standard.

The Independence a Radio Gives an Active Tag

The price of active tags, for illustrative purposes, can range from $15 to $100. The dramatically higher cost compared to passive tags is because the tag has a chip set that includes a bona fide radio incorporated in a tag housing, along with a battery. There is no deliverable inlay version of an active tag, as is the case with passive tags. The more rugged the housing in which an active chip set and battery are placed, the more memory it has; and the more expansive the sensors set with which it is provisioned, the higher the price.

The attractiveness of the active tag is that it can be self-actuating. It can originate a transmission at will if it has been so provisioned to do so. And because it is battery powered, its range can be far enough to be serviced by a far smaller population of readers than passive tags require.

Thus, the active tag is capable of issuing an alert when so directed by a sensor. Of course, it must be within range of a reader. It is also capable of informing its infrastructure system of its location – either by reader triangulation (known as RTLS active tags), by receiving location information from specialized transmitters known as Signposts or Exciters, or by its exhibited signal strength as is the case for 2.4 Band active tags that are able to communicate with the Wi-Fi Access Point infrastructure.

Active tags cannot piggyback on the passive RFID LAN infrastructure. The active tag system requires its own RFID LAN infrastructure. But a variety of active tags designed to operate in the 2.4 Band that are Wi-Fi-compliant are available, and thus can share the Wi-Fi infrastructure. A 2.4 Band active tag can be an attractive option when the composition of the tagged object can tolerate or accommodate the hostile RF characteristics of a tag operating so high up the spectrum. Especially when the transmission rate for tag-reader data exchange in the 2.4 Band is far, far higher than at lower spectrum frequencies.

An active tag band that increasingly will find deployment favoritism is the 433 MHz Band. The intellectual property for active tag use in this band belongs to Lockheed-Savi. Its 433 Band active tags are used by the US Department of Defense and NATO’s logistic applications. It is also used on ocean-going shipping containers. And the DOD has used its considerable influence to have suppliers other than Savi offer active tags in this band. The stated intent is to accelerate driving down the cost of the active tags in this band. (If the Boeing and Fed Ex trials prove successful, we speculate that Boeing will standardize on 433 Band active tags.)

In this context, it would not be surprising if the cost of 433 active tags drops sufficiently, then the DOD may feel justified in backing away from its practice of accepting deliveries from suppliers sporting UHF passive tags – requiring use of 433 active tags only. And that could have a ripple effect in a segment of the commercial marketplace.

Supporting this possible turn of events is the fact that when the active tags are used in applications where they can be reusable, the effective cost of the tag on a per usage basis (before the battery is depleted), can be quite competitive with the passive tag (not the inlay version). That is especially true if one adjusts the price to recognize the added functionality that an active tag can provide.

Final Driving Point

Despite all the hype, despite all the arm twisting, despite all the expense Wal-Mart put its suppliers and RFID equipment and software suppliers through, a recent study showed that the typical Wal-Mart buyer leaves the store having bought a lesser number of items than they came into the store to buy. They just couldn’t find the items despite all the courtesy the Wal-Mart salespersons attempt to extend to the shopper. The explanation of the problem involves describing a very complicated set of circumstances. But there is no denying that their IT people got wrapped up in the deployment of the (RFID) technology and ignored the underlying systems and organizational problems that made tracking merchandise on the selling floor and back room so imperfect. Those IT persons would have better served Wal-Mart had they put their energy into tracking and replenishment systems design rather than attempting to micro-manage the development of a passive tag standard based on a vision that has proved to be wanting.

Now that Wal-Mart has taken the ongoing deployment of RFID out of their hands, one should not be surprised to see some soul searching at Wal-Mart. Nor that active tags may find, however limited, a home in Bentonville – alongside the UHF passive tag. A passive tag that may ultimately have RF operating characteristics and storage capabilities never envisioned by the Auto ID folks, or specified by the EPC Consortium.

The final driving point is that how to use RFID has to be based on helping to solve an enterprise’s systems requirements in ways that achieve the next wave of productivity gains. When one comes at RFID deployment from this perspective, the results can be dramatically more rewarding for one’s company and one’s career.

About the Author

Tom Polizzi, the author-analyst of this discussion has been a fixture on the AIDC landscape for more than two decades. And before that as a marketer of data networks, and as an IT Director in both manufacturing and service companies. He is co-founder and executive analyst of WCCN Publishing, Inc. For sixteen years, until October 2006, the firm published The WCCN Letter, an 8-page monthly available by subscription only. The Letter was focused on enterprise automation with a wireless perspective, and carried no advertisements. WCCN has also published well-received Handbooks on wireless LANs and RFID in the Enterprise. Handbooks that treat data capture as but only a part of an enterprise’s overall automation system, and not as standalone technologies that one installs for compliance, or because someone produced a believable ROI for wireless LAN or RFID deployment. Currently, the firm is exclusively focused on customed consulting in the deployment of RFID.

Mr. Polizzi’s mantra is that it is insane to install RFID without first designing a well-thought out integrated RF tag and bar code LAN infrastructure. One that seeks to minimize multiple layers of muddleware software along with the necessity of housing such operationally complicated software between the tag reader/bar code scanner and the real time segmented bar code and RFID object tracking application systems. Systems that themselves also must be subjected to bona fide integration.

Fee-basis Follow-on Information

Interested readers can avail themselves of a Private Client Enterprise Automation Productivity & Futures Direction Report that WCCN Publishing provides its clients.

Within a prescribed scope, the report can be tailored to answer specific questions that the reader may have about RFID in general, and the report will offer specific suggestions on how to proceed relative to the reader’s issues and objectives. Included will be advisories for what to include in a RFP, and most important of all, strong suggestions about the systems architecture that should be employed. An architecture that includes effective methods for real-time detection and fixing of missed and wrongful recordation tag reads.

Anyone wishing more information about the report and how to avail oneself of the offer, please send the request to wccn@wccn.com.

Return to Home Page