RFID Applications and Network Standards and ProtocolsPage 1

RFID System Protocols and Standards Overview

By Joshua Rhodes

For CPET 384 – Wide Area Networks

RFID Overview

Much has been written about radio frequency identification (RFID), from how it is currently used to what it could be used for, but the facts are that RFID has been slow to take off. The problem has been in the lack of accepted standards being in place for easy implementation of RFID applications. Cost has been another issue impacting the liftoff of RFID applications. Because of the lack of accepted standards and protocols in place, costs for RFID networks, components and expertise have stayed high.

Another issue regarding RFID applications is the privacy issues. Are consumers ready for RFID? Privacy concerns have caused some companies to remain spectators in the RFID space. The reason is because the proposed use of RFID’s on clothing, food, etc. can be a sticky issue with consumers, as it could become very easy then for governments to, in turn, begin tracking people. However, vendors and governments claim that the technology is far away from being ready for that type of usage. (11)

RFID is not really a new technology-it has been around since the 1970’s, with the fundamental idea starting during World War II, in the 1940’s. The Japanese, Americans and British were all using radar during the war as a warning of approaching planes. However, they couldn’t tell if the planes approaching were friendly or not. The Germans discovered that if they rolled their planes when approaching that this altered the radio signal reflected back and thus a way for the base to be alerted to approaching plane. This would have been the first passive reader system.

The British then countered with their own identification system called identify friend or foe (IFF). A transmitter was placed on each British plane, so that when the plane received the signal it could respond and thus base would know that this plane was friendly. This would have been the first type of active system. A crude system, but obviously useful in this context and it created the beginnings of the RFID networking we know today. (7)

How Does RFID Work?

RFID uses electronic tags placed on items to track either their locations or item descriptions. RFID tags are really nothing more than tiny microchips that can, in some cases, be smaller than a pinhead or grain of sand. The chip is attached to a tiny antenna which allows it to communicate and transmit information. Below (Figure 1) is a blown up view of a simple RFID tag.

Figure 1. RFID Tag Example (6)

These tags come in three different types – passive, semi-passive, and active. Passive tags are sleeping until they are “awakened” by a RFID readers’ radio wave. These passive tags do not have a power source and so need the radio wave energy (electromagnetic radiation) from the reader in order to be operable. These tags are particularly appealing because they’re able to draw this power wirelessly, and therefore can be very small and very cheap. Also, these tags can be placed on almost anything, since constant power is not needed in order to keep them operable. Semi-passive tags have a power source but do not actively communicate until requested to, and when requested it uses the radio wave power to transmit the information up to the reader. (14) Active tags have a power source and actively communicate with readers without the need of radio wave power to communicate.

Most RFID tags used in the market are of the passive kind and rely on the reader for its energy. RFID readers usually have a RF (Radio Frequency) module that allows them to transmit and receive messages. They also are usually manufactured with additional interfaces (e.g. RS 232, RS485) to allow for connection to PC’s, etc. Below (Figure 2) is a simple diagram showing the communication between the RFID reader and the tag (or transponder). The ‘application’ shown in the diagram can, and most often times does, include an enterprise network infrastructure. (6)

Figure 2. Simple RFID Diagram. (6)

Tags can usually hold a unique identification code, or number, of 8 bytes in length, along with other small pieces of information. Active and passive tags can differ in regards to the types of information they store. Passive tags usually store only object identification information, while active tags can store object descriptions and transportation history in addition to the identification information. (2)

To avoid interference with other devices the power radiated by RFID readers cannot be too strong, therefore readers often have to be very close to the tagged item in order to read them. The distance varies based on surroundings, available technology, and local standards and laws, but a common distance is usually within a meter or two. (11) This does impose some limitations on the use of RFID, as it may not be physically possible to have readers within distance to constantly keep up to date on tagged items, say for an inventory count. Obviously, there is still some manual intervention that may be needed in some cases. This also drives up cost since, if desired, more readers might be required and distributed liberally.

Initially implemented to track entire pallets or cases of items, RFID is slowly making its way onto individual items. Tagging individual items will offer many benefits for retailers looking to monitor consumer spending trends and habits. Along with the benefits, though, comes an increase in network requirements.

Most RFID tags are static and usually cannot be altered after manufacture or configuration. However, writable tags are available. Three different procedures are available to write to RFID tags. EEPROMs (electrically erasable programmable read-only memory) are the most common. EEPROM memory capacity usually ranges from 16 bytes to 8 kilobytes. However, it has disadvantages in that it has high power consumption during the writing process. FRAMs (ferromagnetic random access memory), read power consumption is lower than in EEPROM, but has experienced manufacturing problems in the past – which has had an impact on its market acceptance. Based on available information, it appears to have similar limits in memory capacity. SRAMs (static random access memory) are used especially in microwave systems and have very high write cycles. In order to retain the data, the power supply needs to be uninterrupted so an auxiliary battery or some other power source would be required for the tag, which obviously would limit its usefulness. SRAMs memory capacity ranges from 256 bytes to 64 kilobytes. According to some experts RFID tags can be read and written to, up to 10 billion times before performance dips, so the investment appears to be a solid one.

RFID Network and Connectivity Requirements

The biggest fear in the growth of RFID technology is the exponential need for RFID readers because of their limitations in read distances. This in turn is feared to put quite a strain on enterprise networks as they are stretched to support these RFID applications. Further, today’s machine-to-machine communications alter the networking landscape as we see the majority of data flow happening at the edge of the network instead of in the central. Obviously, this can have negative impacts on a network.

Another issue is just how to implement reader systems in order to provide the power requirements to the readers. A couple of options were provided in [11], one was wired Ethernet and the other wireless LANs. Each, of course, has their individual advantages and disadvantages. The advantage with wired Ethernet is that it is a straightforward and proven tool. Also, Ethernet could provide readers with the needed power using standard 802.3af power. This option could save companies money, as readers will often have to be placed in areas where electrical outlets are not normally found (e.g. ceilings). This allows companies to avoid having to pay electricians to add electrical outlets throughout their buildings. Also available are wireless LANs, which offer much flexibility, allowing the readers to be placed wherever coverage is required. However, these would still require AC power and thus possibly more AC outlets to support the amount of readers necessary. With wireless LANs being used the chances for interference at 2.4 or 5 GHz, however, are minimal. (11) Soon almost all readers will be outfitted with TCP/IP stack, as well as Ethernet, power-over-Ethernet, and 802.11 wireless LAN connections.

In addition to the above, soon most reader devices will connect to the enterprise network using TCP or UDP. While today’s devices typically, as mentioned earlier, support RS232 or RS488 serial communication protocols, future releases will more than likely support multiple LAN and WAN technologies. This would include Ethernet (10Base-T), Wi-Fi (802.11x), GPRS and EGDE (cellular), Bluetooth and infrared among others, using IPV6 addressing. (9)

Much has been written about the perceived strain that RFID will place on enterprise networks. There are many different views about what is the best way to structure a network in order to handle RFID. One suggestion is to implement a layered approach using services like DHCP and multiple databases, while trying to utilize the existing enterprise security. Also, the thought in using a layered approach is that it will allow new RFID applications to be written and implemented by the company, without having to worry about “reinventing the wheel”, in regards to the network infrastructure. (12)Further, filtering and data aggregation can go a long way in making the move into RFID easier for companies with solid network infrastructures already in place. (15)

The layered approach seems to be well-prepared for large-scale RFID applications. Many vendors, including Cisco, IBM, and upstart Reva Systems, are jumping into this market with the tools ready to provide companies the keys to a successful RFID implementation. Therefore, network planning is critical to a successful implementation of RFID. This would include eliminating single points of failure by creating redundancy in the network. Another thing to keep in mind would be to give the RFID application proper security and priority in the network. RFID applications could carry time sensitive data and need top priority in order to be processed quickly and correctly. This could be implemented using network QoS (Quality of Service) software to ensure network priority.

Another issue inherent in all network systems is how to manage network collisions. While many of the protocols mentioned above have their own ways of managing collision, there might be ways to improve. Collisions can cause the delivery of meaningless messages to the reader and therefore can wreak havoc on time-sensitive RFID applications. In (1), the writers here were trying to prove that a Tree Slotted Aloha protocol is superior in its handling of collisions in this type of network structure (i.e. RFID system). The claim is that the Tree Slotted Aloha protocol provides better performance as it consistently outperforms other protocols in avoiding tag collisions. Outside of Tree Slotted Aloha the basic principles in communication between reader and tags are half duplex procedure (HDX), full duplex procedure (FDX) and sequential systems (SEQ).

In HDX, the data transfer from the tag to the reader alternates with data transfer in the opposite direction, from the reader to the tag. This is most often used with the load modulation procedure or modulated reflected cross-section procedure. These procedures all influence the magnetic or electromagnetic field generated by the reader, and are therefore known as harmonic procedures.

In FDX, the data transfer from the tag to the reader takes place at the same time as the data transfer from the reader to the tag. The data is transmitted from the tag at a fraction of the frequency of the reader (i.e. subharmonic), or at a completely independent(i.e. anharmonic), frequency. Both FDX and HDX procedures require the transfer of energy from the reader to the tag to be continuous.

In SEQ, the transfer of energy from the tag to the reader takes place for a limited period of time only (i.e. pulsed system). Data transfer from the tag to the reader occurs in the pauses between the power supply transfers to the tag. Figure 3 shows a diagram of how full duplex, half duplex and sequential systems transmit power and information between RFID tag and reader.

Figure 3. Diagram showing how Full Duplex, Half Duplex and Sequential Systems work. (6)

RFID Protocols and Standards

RFID readers use tag-reading algorithms that allow for hundreds of tags to be read every second. Readers usually fall within two different categories in regards to the frequency used, High Frequency (HF) and Ultra High Frequency (UHF) (ISO 18000-3 and 18000-6). The following table (Figure 4) shows some guidelines currently in place in regards to these two categories. Low frequency is also sometimes used for RFID; much depends on the needs of the applications.

Figure 4. RFID HF and UHF system comparison (14)

Not shown on this table is the breakdown of microwave systems for using RFID. Microwave systems have much higher read ranges, reaching highs of 15 meters. However, the good comes with some bad as microwave systems often need auxiliary power backups, in the forms of batteries or some other power; as the reader is usually not able to provide the tag with enough power to complete the task.The interference resistance in microwave systems is much better, so the microwave systems are better in assembly production lines as lower frequency systems are more susceptible to interference from heat and electromagnetism, among other impacts.

RFID ranges vary based on different factors:

  • Positional accuracy of tags
  • Distance between several tags that are within same general area
  • Speed of the tag

Of course, the application’s need will drive what kind of frequency standards will be used, and in turn drive the decisions in setup of the RFID system. For example, in a system used for public transport ticketing a range of 5 to 10 centimeters is perfect. If the range is any larger, the reader may have a problem differentiating between several tags (i.e. train passengers). So the RFID system should be designed at the maximum required range. Also, the tags in the field should be setup within the range so that only one tag is read at a time, avoiding collision. The RFID handbook highlighted the following standards of interrogation zones commonly used in RFID systems. See Figure 5.

Figure 5. Interrogation Zones (6)

There are also procedures regarding how tags send back information to the reader. Directly from (6) are those procedures:

  • Use reflection or backscatter (the frequency of the reflected wave corresponds with the transmission frequency of the reader. Frequency ratio 1:1)
  • Load modulation (the reader’s field is influenced by the transponder. Frequency ration 1:1)
  • Subharmonics (1/n fold) and the generation of harmonic waves (n-fold) in the tag.

Another interesting note about RFID systems is that they are classified as radio systems, because they generate and radiate electromagnetic waves. Therefore, they fall into a category that requires that the systems do not disrupt or interfere with other radio services. Some of these radio services are for entertainment (e.g. radio, television, mobile radio), but others are for public safety (police, security, emergency, etc.). Because of this, the available frequency ranges for RFID systems are restricted.Thus, RFID systems have to use frequency ranges that have been reserved specifically for industrial, scientific or medical applications (ISM Frequency).The diagram below (Figure 6) shows the available ranges for RFID systems.

Figure 6. Available FrequencyRanges for RFID Systems

Much has been done over the last two years to propose standards in RFID applications and implementations. Since modern RFID applications started over 6 years ago, most implementations have been proprietary in regards to network and protocol. Standardization of theseprotocols in this arena have started but never seemed to finish. Some that have been used are Airespace’s Lightweight Wireless Access Point Protocol (LWAPP) or the Control and Provisioning for Lightweight Access Points (CAPWAP) protocols for WLANs; or a simple lightweight RFID reader protocol (SLRRP). (11)

A few years ago the Auto-IDCenter proposed standardized protocols for use in RFID systems. The plan was to implement a series of classes for the tags. Each class would be more sophisticated then the next. Eventually what was implemented was Class 0 and Class 1 tags. The problem with this implementation was that the classes were not compatible (used different protocols) and therefore multiple readers would be needed to read the tags. In addition to this these tags were incompatible with ISO standards. This overhead watered down the boom that this technology was supposed to have in the market. (7)

However, that should change; in 2004 EPC Global began work on a second generation protocol, called EPCGlobal Gen 2. The goal was to create a single, global standard for RFID communication that would more closely align with ISO standards. Gen 2 was approved in December 2004. (7) Some of the requirements for Gen 2, as listed on their home page, are as followed:

  • Convergence to One Global, Interoperable Standard
  • Increased Speed and Ease of Global Adoption – adaptive techniques implemented for readers; compatibility with regional RFID regulations.
  • Increased Functionality and Performance – better read/write rate performance.
  • Increased Production and Competition – increased inventory of tags available and reduced costs for tags and readers.

By implementing these standards, companies can have all the confidence that their partners around the world can easily read these tags, giving them real value.