Special Report: User backlash at SP100 meeting

April 1, 2007
CONTROL contributor Dick Caro, a member of the ISA SP100 committee devloping a set of wireless industrial networking standards, reports to us on SP100 activity. This is his latest progress report.
By Dick Caro, CMC Associates

The February ISA SP100 (Wireless Industrial Network Standards) meeting generated some strong end-user backlash—a sharp contrast to the previous meeting where cooperation seemed to be the mantra. The backlash reflects dissatisfaction with the process being used by SP100 as it moves toward the first draft of its standard.

The issue arose at a subcommittee meeting where Honeywell and the WNSIA group (Wireless Networks for Secure Industrial Applications) proposed adoption of a new physical layer using a narrow-band, frequency-hopping spread spectrum operating at 2.4 GHz as an alternative to the previously adopted physical layer based on the use of IEEE 802.15.4 chips using direct-sequence spread spectrum with channel-hopping.

The User Working Group was meeting separately when it received news that a vote on the alternative physical layer proposal was being taken. Most of the users left their meeting room to attend the physical layer meeting so that they could participate in the voting and express their opinions on the alternative proposal.

When the vote was taken in the PHY/MAC (physical and media access layer) task group, sufficient numbers of end users were present and voted against the motion to cause its narrow defeat. Subsequent questioning of those opposed to the proposition revealed the end users’ deep dissatisfaction with the process.

A few of the users who had heard the presentation of the alternative physical layer said their needs were not being heard. They had expected that an alternative physical layer based on one of the currently available physical layers operating at 900 MHz and using frequency-hopping spread spectrum would be among the first alternatives to be considered. This alternative would have met their requirements for an inexpensive, quickly available wireless network for data acquisition that could operate over distances typically found in refineries without a repeater.

After the meeting, the co-chair of the User Working Group resigned from the position, citing alleged harassment by one of the vendor companies, thus forcing SP100 management to consider changes in the process.

Many components make up the “wireless cloud” over process plants. (Courtesy,Honeywell Process Solutions)

What Actually Happened?
Development of a highly technical standard, such as SP100, requires the cooperation of the major industry suppliers and many other specialty firms to make complex, detailed and highly technical proposals. It also requires that end users make known their actual needs in terms familiar to them and communicate them clearly to the technical committees. In this case, the users did not circulate their draft of the User Requirements document, but held it as a working paper of the User Working Group. This document also was not widely circulated outside that working group to seek input from other users not able to attend committee meetings.

Until this hiccup, the development of the architecture for SP100 was proceeding on-plan. In previous meetings, the SP100.14 and SP100.11 subcommittees developing separate proposals for wireless networks for simple data acquisition and for control systems were combined because, except for the physical layers, the proposals were very similar. The combination was approved by SP100 so the committee could concentrate on a single standard with variation only in the physical and portions of the media access layers, SP100.11a. The “default” physical layer using IEEE 802.15.4 chips previously proposed by the Collaboration Group in its proposal for a data acquisition network had already been approved.

In this meeting, as planned, the WNSIA Group headed by Honeywell, was making its more detailed proposal based on the not-so-very-well-kept secret, second-generation wireless network code-named “Raptor.” As the meeting was going on, Honeywell was demonstrating Raptor to end users in another room. It also offered ts alternative physical layer to SP100 for use in control-oriented networks. WNSIA/Honeywell also offered its media access, network, transport and application layer proposals at the same meeting. The Collaboration Group also presented detailed proposals for the media access and network layers.

Development of a standard is a difficult process, mostly because the committee is made up of technically savvy contributors from many companies who are motivated to have their previous work written into the document. End users are involved to act as a sounding-board to make sure what is proposed meets their needs. Their presence is required in ISA standards committees to balance the suppliers. The most difficult obstacle to writing standards is that every one of the committee members has a full-time job that does not include the extensive time required to actually write the standard.

The alternative is to create a standard the way the IEC does. In that system, all proposed standards recognized by some recognized standards body can be accepted as “Publicly Available Specifications.” The result of this is a multi-part standard, such as IEC 61158 Fieldbus, which began with eight parts and soon will have more than 15, none of which interoperate.

The SP100 process, operating under ISA rules, seeks to develop a single standard designed to allow interoperability of wireless field devices for process control, a requirement users have strongly requested.

It would have been easy to quickly adopt currently available 900 MHz networks from Honeywell, Emerson, Accutech, Omnex and others, but that would have required either multiple non-interoperable protocols or creation of one network “winner” and three or more “losers.” This was never proposed to the SP100 committee for obvious reasons. Instead, the committee has proceeded down the path of developing a network that would offer several physical layers with common upper levels, allowing interoperability at the network/transport layer, but not necessarily at the field-device level.

This fine point was not sufficiently communicated to the end users on the committee. It seems that to understand this point required that explanations of the technology be written into the presentations and meeting minutes of committees other than that of the end users.

Management of SP100 has now taken the assignment of seeing that subcommittee decisions are translated into end-user language related to their written and implied requirements, and communicated to the User Working Group. The current User Requirements document must also be available to all SP100 subcommittees and be made available to a broader circulation of end users.

Meanwhile, Control magazine is also playing an important role in establishing user requirements by sponsoring an end-user survey to gather requirements not previously communicated to the SP100 committee. Take the survey here. All members of the SP100 committee have been given access to the uncompiled raw data from this survey.

Other Considerations
Standards committees work best when presented with a comprehensive solution in the domain of their work. Detailed invention by a volunteer group representing many different interests is a slow, laborious and often unproductive method for producing a standard. At the Phoenix meeting, Honeywell presented a comprehensive set of protocols for the physical, media access, network/transport and application layers, all taken from its Raptor system. Included in its presentation was the promise for detailed specifications for the physical, media access, network/transport and application layers.

The Honeywell demonstration was based on the use of a Bluetooth chip operating at 2.4 GHz and using frequency-hopping spread spectrum. During the presentation, Honeywell showed the results of its testing for this physical layer. Those test results are very encouraging in both the low rate of data loss and the achievable distances.

Raptor is also based on an architecture with field access points or routers each of which is connected to a wired or wireless intermediate level network using, for the present, a pre-release of IEEE 802.11s that is similar to Wi-Fi, but includes mesh networking. Since Raptor was only a concept demonstration, its details may change before the product is released.

The IEEE 802.15.4 standards committee recently approved two new standards: an alternate physical layer based on ultra-wide-band (UWB) technology, 802.15.4a, and a revised media access layer designed to work with alternate physical layers.

These two standards from another organization are of interest to SP100 now that it is considering use of an alternate physical layer. It should be noted that several members of SP100 have been deeply involved with the IEEE 802.15.4 committee work. ZigBee is based on this standard, and several suppliers that use ZigBee in their product lines are also represented on the SP100 technical committees.

The User Working Group has now published its requirements document for use by all committee members. When the Control magazine survey is completed, the document will again be updated to reflect new requirements. The User Working Group has also asked for additional end user members, especially from process industries other than petroleum refining, which is already well represented.

Where Are We Going, and When?
The schedule for the first release of SP100.11a is still scheduled for early in 2008. The committee has a long list of features that must be included in the standard, but has evaluated which features can be deferred to later releases, and which must be in the first release. One of the effects of delay is that some features tend to be removed from the first release list.

The committee is focused on retaining backwards compatibility with wired networks already installed in many process plants, and with wireless HART. The wired networks include HART, Foundation Fieldbus H1/HSE, Profibus-PA/DP and Modbus/RTU/TCP. Wireless HART coexistence is a great concern, since it is likely to use the same frequencies in the same area with a very similar, but not identical, protocol. Users have indicated a desire for SP100 to embrace wireless HART by allowing it to pass through the SP100 network in some way.

The WNSIA/Honeywell presentation for an application- layer protocol is highly independent of all lower layers and is based on use of EDDL, IEC standard 61804-3. This is now being balloted as ISA/ANSI 104. If this application layer is adopted, then SP100 will be highly compatible with HART, Foundation Fieldbus and Profibus, which also use EDDL. This means that SP100 field devices are expected to be “smart” to the same extent that HART, Foundation Fieldbus and Profibus-PA devices are smart.

Security Issues
Much of the technical work of SP100 has concentrated on its security aspects, since this is of great concern to the end users. Wired networks are not easily tapped unless they have a direct Internet connection, but wireless networks cannot be stopped at the plant fence. There is equal concern about unauthorized persons having access to data flowing on the network, as well as for disruption of network traffic by jamming methods. Many users have said that for these reasons, wireless will “never” be used for any critical services, such as safety or control.

Both aspects of security are being addressed within SP100. Part of the security is embedded in the complex protocols, such as frequency- or channel-hopping, or by use of either direct-sequence spread spectrum or orthogonal frequency division multiplexing, which spreads the network power over a broad spectrum, making random detection and broad spectrum jamming very difficult. Additionally, encryption will be used to make access to data a near-zero probability, even if the protocol is detected.

A Moving Target
Another problem in the development of SP100 is that low-power wireless networking is a moving target. Many companies are developing new chips and often new protocols in this area, and proposing them to IEEE 802 standards committees that are also developing new standards. Often some technology that initially appears to be unsuited to industrial automation use is modified in some way to become more appropriate. For example, IEEE 802.15.4—even with the ZigBee extensions—seemed inadequate for use as an industrial automation protocol. However, recent improvements in 802.15.4 now make it attractive for use with some protocol additions to the network/transport layers.

The recent approval of the use of ultra-wide-band (UWB) in the 802.15.4a standard may prove attractive in the future, once chips become available and it can be evaluated. Honeywell used Bluetooth chips for its Raptor demonstration, even though previous versions of Bluetooth seemed to be inadequate for industrial automation use. We expect an influx of new low-power chips from the development of WiMAX (IEEE 802.16e) for use in broadband (4G) wireless telephones in the next five years.

As the answer to the future development of new wireless technology, SP100 has adopted a philosophy that it must be agnostic to the actual radio used. All radios will need to operate with the same upper level of the media access control protocol adopted by SP100. This will allow smooth migration of SP100 devices to new technology as it becomes commercial; however, direct communication between field devices will not be possible unless they use the same radio technology. For this reason, all versions of SP100 will use wireless termination devices called access points or gateways. When multiple radio technologies are used in field devices, the access points/gateways will need multiple radios.

  About the Author

Dick Caro is principal of CMC Associates, is a recognized expert on fieldbus technology and a founding member and Chairman of the Fieldbus standards committees. You can reach him at [email protected].