The industrial ethernet protocol wars fieldbus revisited

��

��

��

��

�� !�� "��#��$� %

��#��&�'!�$��'��(�� )

'��*'��"�!��"��+�� ,�+*+�� -�.�� /

'01��#!'01�+�� 23

1�� 1�� +��(-�� (��4��5�� '�� 2�

��#�-��$��6�� 2%

'�7��1�� 28

�� !��"��

$5&�6WUDWHJLHV��2FWREHU��

��$5&ZHE�FRP��&RS\ULJKW��$5&�$GYLVRU\�*URXS�

��

��

� ��

��

��

��

'3

3$

��

��

� �!

�

�"��#��

##

# 0HPRUDQGXP�RI�8QGHUVWDQGLQJ

�

��

� � � �$��%��&��'��#��'��()

1HWZRUN��7UDQVSRUW

��

�

/LQN��3K\VLFDO

8VHU�/D\HU

$SSOLFDWLRQ

��* +�� ,�� -��

��.��.�.�

!�. $��%��&��,��'�� +*��'�

��.)

��%��

(WKHUQHW

7&3�,3��8'3

+773 )73 6073�HWF�

'HYLFH�3URILOHV��2EMHFW�

0RGHO�RU�/LEUDU\

&RQILJXUDWLRQ&RQILJXUDWLRQ��

0DQDJHPHQW�6RIWZDUH

0HVVDJLQJ

��

� $5&�6WUDWHJLHV��2FWREHU��

&RS\ULJKW��$5&�$GYLVRU\�*URXS��$5&ZHE�FRP��/�

�� !!��"��#�!��$!%�&$�� !!��'� ��'�(�� '��

��)$��*��$�� &�'��(+

�"��$��%�

Industrial Ethernet continues its march to lower levels of the plant hierarchy as both standards organizations and automation suppliers address criticisms of its suitability for plant floor applications. While Ethernet is an interna-

tional standard, IEEE 802.3 specifies only the physical and data link layers of the 7-layer network stack. TCP/IP adds the further but minimal guarantee that two devices can exchange data with one another. In the industrial

automation space, however, the lack of standard application and user layers translates to ongoing headaches concerning which devices may connect to the network, how devices interoperate on the network, and what level of

functionality is supported.

Ethernet’s rise in prominence comes after a protracted, decade-plus long bat-tle to standardize the process Fieldbus certification, and now many fear that

industrial Ethernet protocol standardization will result in “Son of Fieldbus.” Judging by today’s market profile, this fear of Fieldbus revisited is well founded. The only solace lies in

availability of common physical and network/transport layers and the need to reconcile just three or four upper-level Ethernet protocols compared to the ten or more Fieldbus protocols.

While the Fieldbus wars were fought all the way down to the level of com-peting network media or physical layers, availability of a common Ethernet TCP/IP stack has largely moved the protocol wars to the higher-level appli-

cation or “user” layers. This upper-layer functionality represents the new battleground where the strategies of competing industrial Ethernet protocols such as EtherNet/IP, PROFInet, IDA, and Foundation Fieldbus diverge.

Both IAONA and OPC are stepping in as third parties to try and mitigate the potential for still another fieldbus war. After its initial founding as still an-other industrial Ethernet specifying body, IAONA has positioned itself as a

neutral umbrella organization for the disparate industrial Ethernet factions. The OPC foundation, on the other hand, has announced its intention to ex-tend the existing OPC DA (Data Access) specification to allow run-time

interoperability across systems based on disparate industrial Ethernet net-works.


0��$5&ZHE�FRP��&RS\ULJKW��$5&�$GYLVRU\�*URXS�

�&�'��(��!�,��!�� "��*$� ��'��(�*��!�$��'�(��

�*��$��'�� #��*��'��$��'

��'$��+

&��'��(��) !��

User layer functionality such as common device profiles and their associated object models is supposed to ensure device interoperability and interchange-ability. Common EtherNet/IP device profiles, for example, are designed to

ensure that a replacement device is configured to produce or consume the same basic set of I/O data and exhibits the same network behavior as the original. This upper-layer functionality represents the new battleground

where the strategies of competing industrial Ethernet protocols such as EtherNet/IP, PROFInet, IDA, and Foundation Fieldbus diverge.

The competitive landscape parts even further at the network configuration

level. While each protocol specifies some configuration parameters, vendors augment this functionality with their own proprietary network configuration tools that often add incremental functionality to their own devices or systems

versus those of competitors. As more and more of the industrial network is standardized in first the industry standard Ethernet TCP/IP protocol, then the higher-level industrial Ethernet protocol specifications, these vendor-

specific network configuration tools and associated software assume even greater importance in control system selection.

When considering industrial Ethernet for control applications, varying inter-

pretations of often-broad protocol specifications are leading some end users to implement homogenous vendor environments even when a common higher-level protocol is employed. This strategy stems from the bottom-line

need for correct interpretation of control messages among plant floor devices and concern that devices from different vendors, even those supporting the same protocol, may not interoperate.

In order for true interoperability to occur, all suppliers must im-plement a common set of communication services in the same manner.

Similar to the findings in the device network arena, the combination of these factors will result in most machines or control systems being implemented with a single Ethernet protocol, e.g., EtherNet/IP or PROFInet, rather than a

mix and match of products supporting differing protocols. Consequently the value proposition for industrial Ethernet will initially stem not from an abil-ity to mix and match devices from different suppliers, but rather the

commonality and incremental functionality detailed in ARC’s recent report on the Industrial Ethernet Value Proposition. Even at the supposedly com-


&RS\ULJKW��$5&�$GYLVRU\�*URXS��$5&ZHE�FRP��1�

�� ,��,��* +&�)�

��#��&�' "��#��9��#��

1�� 1�� +��(-� 1�� 1�� +��(2

�"� #��:�"�;

'01�� '01�+��"'9'01�+��'�9�-*�

�� !��

mon physical layer a standard industrial Ethernet connector may prove elu-sive, although the efforts of standardization bodies such as EIA/TIA

(Electronics Industry Association/Telecommunications Industry Association) may result in a standard physical layer connector suitable for industrial ap-plications.

&�� *��+�%��,��

In spite of efforts designed to push Ethernet to the lowest levels of the auto-mation hierarchy, the reality is that most manufacturers will end up with a

continued cascade of automation networks from the supervisory through control and ultimately device levels. Traditional device networks such as PROFIbus DP and DeviceNet have established a firm foothold in the mar-

ketplace and in the near term many manufacturers will continue to specify these networks for their low cost, real-time, multi-drop capabilities. This is particularly true as long as the cost of an Ethernet interface, plus power to

the device if necessary, exceeds the cost of a similar configuration using a standard device network. The star topology employed with Ethernet’s hub-and switch configuration also sends

some customers to device networks for their multi-drop capability and resulting lower wiring cost. With the notable

exception of the emerging IDA interface, most of the competing industrial Ethernet protocols have complementary

device networks as part of their overall architecture.

Interestingly enough and as a reversal of the prior Fieldbus landscape, the

process industries will likely end up with a single Ethernet standard – Foun-dation Fieldbus HSE and its associated H1 non-Ethernet device level network. The discrete or machine control side of the automation business, on

the other hand, has several competing protocols. Global PLC market leader Siemens is promoting the PROFInet/PROFIbus combination while Rockwell Automation and Omron, among others, are moving forward with

EtherNet/IP and its sister CIP-based DeviceNet and ControlNet interfaces. Long associated with the Modbus TCP technology that is widely adopted in



third party Ethernet products but does not incorporate a true application layer, Schneider Electric is now positioning itself for the future by allying

itself with the IDA organization. GE Fanuc, on the other hand, has assumed the same passive “pull” stance evidenced in the device network realm where they waited to see what the market and their projects demanded.

Many of the industry leaders have pledged support for third party efforts designed to foster commonality across the competing protocols. EtherNet/IP and IDA have signed a Memorandum of Understanding to pursue common-

ality through the umbrella IAONA organization, for example, and the sponsoring organizations for EtherNet/IP, Foundation Fieldbus HSE, and PROFInet have all lined up behind the OPC DX specification. In the near

term, however, particularly as the fruits of these efforts take shape, single vendor industrial Ethernet environments will remain the norm.

��+�-�� ,��&�� .��)��

Official EtherNet/IP products first hit the market in the year 2000, but in re-ality Rockwell had already been shipping like products under the ControlNet

over Ethernet moniker for some time. As this last label implies, EtherNet/IP represents migration of the upper-level CIP or Control & Information Proto-col common to the ControlNet and DeviceNet networks onto the Ethernet

physical media. As a media-independent protocol, CIP is capable of migrat-ing to further media such as FireWire or wireless networks.

While EtherNet/IP is positioned as the

information-oriented supervisory inter-face designed to serve plant floor information to higher-level systems, the

protocol is designed to accommodate real-time transfer of control messages as well as non time-critical data transfers.

This capability is evident in Rockwell’s introduction of not only control-level EtherNet/IP products, such as its new

ControlLogix line of processors, but also I/O level products such as Flex I/O and MicroLogix controllers. The high-"�� ##�� $�%�

�� %��

,(((��

,3

8'3 7&3

(QFDSVXODWLRQ

��

:LUHOHVV��)LUHZLUH��

HWF�

#��

!�. $��%�� %�-�� !�. $��%�� %�-��

��

&RPPRQ�0HVVDJLQJ��3URGXFHU�&RQVXPHU

&RPPRQ�'HYLFH 3URILOHV �$SSOLFDWLRQ 2EMHFWV

&RQWURO1HW7UDQVSRUW��

'DWD�/LQN�/D\HU

'HYLFH1HW7UDQVSRUW��

'DWD�/LQN�/D\HU

'HYLFH1HW3K\VLFDO�/D\HU

&RQWURO1HW3K\VLFDO�/D\HU



bandwidth producer/consumer network also lends itself to incorporation of sensors with high data content. Cognex, for example, recently announced a

family of EtherNet/IP-compatible networkable vision sensors.

EtherNet/IP is able to accommodate both real-time and non time-critical messages through its encapsulation of both the UDP and TCP protocols.

Critics contend that this approach creates significant protocol overhead, but it does allow control messages and data transfers to be handled through the same network interface. Implicit messaging is used for real-time control and

interlocking via UDP, while non time-critical uploads and downloads use explicit messaging over TCP. All set-up functions and incidental messaging also employ the TCP protocol. While EtherNet/IP can accommodate real-

time control, applications requiring strict determinism and user-defined scheduling are steered toward ControlNet.

*�� $�/��)�� !�� (��$*0��1�The EtherNet/IP SIG (Special Interest Group) within ODVA, or the Open DeviceNet Vendor Association, is the keeper of EtherNet/IP-specific objects and device profiles that go beyond the basic CIP specification. CIP device

profiles contain the object model for the device type in question, along with object configuration data and the public inter-faces to that data. Sample control object types

include Digital Input Point, Digital Output Point, etc., which are similar to the objects specified in other industrial network protocols. Electronic

Data Sheets (EDS) are used to provide the infor-mation necessary to access and alter the configurable parameters of a device. At this

point the specification does not include XML-based EDSs, which would be a logical next step, but common Internet protocols such as HTTP,

FTP, and SNMP are supported in the existing specification.

Along with ODVA, sister group ControlNet International (CI) is a primary

promoter of EtherNet/IP, while the Synergetic-sponsored Industrial Ethernet Association is also on board. Rockwell Automation, strategic partner Om-ron, and Cutler-Hammer are the leading players in the DeviceNet camp,

although numerous other suppliers also belong to ODVA and/or CI and of-fer compatible products. ODVA also signed a memorandum of

��&��'(��

(QWHUSULVH

6XSHUYLVRU\

&RQWURO

,�2

'HYLFH

*XDUDQWHHG�'HWHUPLQLVP��,QWULQVLF�6DIHW\

,QIRUPDWLRQ�2ULHQWHG��6XSHUYLVRU\�,QWHUIDFH

:LUH�5HSODFHPHQW��'LDJQRVWLFV

&RQWURO ,QIRUPDWLRQ

1(7:25.�/(9(/

1(7:25.�)81&7,21



��-��!� �� '.��%�$��.�'�(��$��'��

�� '��!��/��&$��%"�&.&�� $��"��

��0��'��*��(�*��$��+

understanding with IDA and the IAONA group in late 2000 concerning in-dustrial Ethernet. This activity will not impact the existing EtherNet/IP

specification but instead is designed to address areas of potential compatibil-ity that are not currently addressed in the competing protocols.

��+�-�� &��$ ��$+��-��EtherNet/IP’s physical layer specification directly states that the use of COTS Ethernet components may result in unsatisfactory performance in industrial control applications. A bayonet-style IP 65/67 industrially rated connector,

really an encapsulated RJ 45, is specified instead. The group has offered this connector to IAONA and EIA/TIA for consideration as a possible common industrial Ethernet connector.

��2�� !��*�� *��2� ��%�� !�

IDA, or the Interface for Distributed Automation, is a pure-play industrial Ethernet specification that promises to marry a real-time, distributed, web-

based automation environment with an integrated safety architecture. As a pure-play protocol, the IDA industrial Ethernet vision encompasses all levels of the automation hierarchy, including the device level. This strategy differs

from other competing industrial Ethernet protocols that have an accompany-ing, non-Ethernet-based, device network component.

The IDA group was formed in the year 2000, soon after forma-

tion of IAONA. In August of that year the two groups announced an intent to merge which was subsequently called off. While IAONA opted to pursue evaluation and rationaliza-

tion of existing industrial Ethernet protocols, the original founders of the IDA group chose to actually develop an indus-

trial Ethernet specification. Like ODVA and its EtherNet/IP SIG, IDA has

since signed a Memorandum of Understanding with IAONA whereby the groups will work towards defining areas of commonality among the compet-ing industrial Ethernet protocols.

�*��'�� &�34456��7��,�� The IDA protocol is based on the architectural elements included in the IEC 61499 Part 1: Architecture draft function block standard, but replaces por-


&RS\ULJKW��$5&�$GYLVRU\�*URXS��$5&ZHE�FRP��

tions of the IEC model with IDA architecture. Along with support of the full suite of Ethernet TCP, UDP, and IP-related web services, the IDA protocol

specification to date incorporates RTPS (real-time publish/subscribe) mid-dleware based on RTI’s NDDS (Network Data Delivery Service), IDA communication objects, and real-time and safety APIs. Real time applica-

tions are executed on UDP using the RTPS middleware.

�� IDA’s original founders include European suppliers Jetter, Kuka, Phoenix

Contact and Sick. Turck, AG-E, Lenze, Innotec, middleware technology pro-vider RTI and Schneider Electric have since joined IDA, with late arrival Schneider as the “big fish” that every group of this nature needs as a mini-

mum requirement for success. For Schneider, IDA offers an easy migration path to a distributed web-based architecture on the Ethernet network, a ca-pability their popular Modbus TCP protocol does not provide due to its lack

of true user layer functionality.

Representatives from these companies staff the five working groups cur-rently formed within the IDA organization: architecture, web, real-time

communication, safety, and marketing. Activities of the architecture commit-tee are currently focused on infrastructure development in the following areas: Application Model, Engineering Model, Presentation Model, Process

Model, and HMI model. The web group is defining XML-based style sheets and data structures for web-based monitor-ing, visualization (HMI), diagnosis, control, and device

management, as well as input and output. The real-time group is charged with development of the object model and device conformance requirements, while the safety group is

working on the definition of protocols and architectural ele-ments that support architectural safety features. Inclusion of a safety API in the IDA specification is designed to preclude

the need for a separate machine safety architecture.

IDA issued its first white paper/specification earlier this year detailing the architecture definition to date and its intended direction. Proponents of the

group point to the architecture’s emphasis on distributed, web-based capa-bilities and leveraging of the technology already available in today’s marketplace as the rationale behind development of still another industrial

Ethernet protocol. This is in contrast to non pure-play competing protocols that continue to employ a hierarchical structure and rely on non-Ethernet

��



protocols for execution of real-time applications. It does, however, result in development of still another object library since the group has developed

their own object model and accompanying object library.

The protocol and its supporters lean heavily toward motion control and as-sociated applications in areas such as robotics, packaging, and machine

control in general, so much of the object library and its associated functional-ity are expected to likewise emphasize real-time motion applications. The group states that their goal is not to define device profiles but instead their

associated containers and XML-based interfaces.

�*��$�� 1�� ' ��7��The most likely scenario where manufacturing customers will encounter the

IDA protocol is when buying machinery from European, particularly Ger-man, OEM machine builders who are not using PROFInet and its associated networks. When complete, the IDA architecture will allow leading-edge ma-

chine builders to implement distributed, web-based machine control architectures with integral safety features. Inklings of this were already evi-dent at this year’s Hannover trade fair where IDA co-founder Kuka

demonstrated IDA-based material handling robots and “Control Web” soft-ware for cell control applications that utilized the IDA architecture. Phoenix Contact (Ethernet hubs and switches), Lenze (intelligent standalone drives),

and Bosch (integrated welding stations) all demonstrated supporting prod-ucts.

��

PROFInet is the industrial Ethernet communications profile from Profibus International and joins PROFIbus DP, PA, PROFIdrive, PROFIsafe, and other

application and communications profiles in that stable of automation net-works. Unlike these other profiles, however, the PROFInet specification did not emerge as the expected PROFIbus protocol ported onto the Ethernet me-

dium. Instead, PROFInet is a non real-time control-level architecture that bears no resemblance to existing PROFIbus networks. The new protocol lev-erages commercial Ethernet, Microsoft component technology, and other

commercial elements such as XML. While this marked departure from the legacy PROFIbus architecture may come as a surprise for some, in reality it



dovetails nicely with the Component-based Automation strategy currently pur-

sued by primary Profibus supporter Siemens.

PROFInet supporters contend that they

are not part of the industrial Ethernet “fieldbus wars” because PROFInet is a high-level interface designed for peer-to-

peer or controller-to-HMI connectivity and is not targeting the device or I/O level. PROFInet is however the Ethernet-

based portion of an overall automation architecture that invokes the lower-level PROFIbus networks for real-time activity and is therefore included in this discussion of industrial Ethernet networks.

��$��PROFInet is designed to provide an environment that allows machine build-ers to develop an object-based distributed automation system using standard

PROFInet engineering tools and vendor-specific programming and configu-ration software. The Ethernet portion of the PROFInet architecture employs TCP and UDP along with IP, plus the DCOM wire protocol for communica-

tion between automation objects created by the various vendors.

The system developer or machine builder will use the PROFInet engineer-

ing model to create COM-based automation objects out of devices devel-oped in controller programming

software. The tool will also be used to configure PROFInet-based automation systems, including those containing

automation objects developed by others. Device descriptions developed using vendor-specific development tools are

given COM interfaces and exported via XML for use by other developers in their own systems. PROFInet supporters

claim that use of their vendor-independent object and connection editor can reduce engineering time up to 15 percent relative to traditional methods.

3K\VLFDO

��

'DWD�/LQN

��


$SSOLFDWLRQ

�8VHU�/D\HU�

56��,(&��)LEHU

3URILEXV 'DWD�/LQN

3URILEXV 3URILOHV��'3��3$��)06��HWF�

(WKHUQHW

(WKHUQHW

7&3�,3��8'3

'&20��53&

&RPSRQHQW�(QJLQHHULQJ�0RGHO

��)��*�� #(��)(��

:LUH�&RQFHQWUDWRU

��5HDO�WLPH�&RQWURO

1RQ�7LPH�&ULWLFDO

��

��

� ��

��

��

� �� % ,��

��

��

� �!

�

��#8��

��)�� '(��+� �� +� ��'��#��)(��'��



Within the system itself, OLE and Active X are employed for object handling in the engineering and HMI environments.

Machine builders or other system developers will use the programming or configuration software that comes with their controllers to create the actual automation objects and their functionality. For example, in the Siemens envi-

ronment a machine builder would use Step 7 software to create object functionality specific to their machine and then the PROFInet engineering tool to turn it into a COM-based object for use in the PROFInet environment.

As this example illustrates, the PROFInet concept hinges on implementation of the Microsoft-defined object interfaces rather than definition of the actual library of automation objects.

PROFInet extends its component-based ap-proach to the device or I/O level by incorporating devices on non-Ethernet

automation networks, including real-time device networks, via proxy. With this con-cept, each I/O-level device is made into a

COM-based automation object that can be recognized and manipulated in the PROFI-net environment. Ability to incorporate

device networks by proxy is said to extend to both PROFIbus and non-PROFIbus device

networks alike but to date this capability has not been demonstrated.

��1�� !� ��$��With the development work on Stage 1 of the PROFInet protocol just com-pleted, the market will likely have to wait until the end of this year to see the

first compatible products actually shipping. One of the earliest Siemens products announced by Siemens is their iMAP engineering software. iMAP marries PROFInet and Siemens’ Component-based Automation strategy and

is a key part of the company’s emphasis on reducing engineering costs for builders of complex machines and others with similar engineering require-ments.

Component-based Automation is a subset of the Totally Integrated Automa-tion strategy announced by Siemens several years ago. Component-based automation targets system modularization through the use of Microsoft-

based component technology. Execution of this strategy in the machine con-trol segment is evident as the company goes to market with the PROFInet

��)�� +� ��)(��(�� &�%�" ��"��#��,��(-��

��

/LQNLQJ�3UR[\�'HYLFH

RIORIO�5HDO�WLPH�

�1RQ�WLPH�FULWLFDO�


&RS\ULJKW��$5&�$GYLVRU\�*URXS��$5&ZHE�FRP��6/�

architecture, accompanying SIMATIC and other controllers, and now iMAP engineering tools. Targeted at distributed machine control architectures,

Siemens is positioning iMAP and PROFInet as the vendor-independent envi-ronment that joins distributed applications into an integrated solution and enables Component-based Automation.

�� .��"� ��.��8��7�!��&��

While Foundation Fieldbus H1 was designed to connect field-level process instrumentation in harsh or intrinsically safe environments, Foundation

Fieldbus HSE (High Speed Ethernet) was introduced to the market as a con-trol level backbone. HSE maps the existing Foundation Fieldbus H1 protocol to UDP

over IP and uses conventional 100BaseTX Ethernet cable. PID and other field control options are already built into Foundation

Fieldbus via the user layer Fieldbus Func-tion Blocks that are common to both H1 and HSE. Foundation Fieldbus networks are

standardized as Type 1 and Type 5, respec-tively, of the multi-headed IEC 61158 Field-bus protocol standard. Unlike the H1

network, HSE is not rated for intrinsically safe operation.

The contrast between the battle to establish the original Foundation Fieldbus

protocol stack and the relative lack of competition for HSE at the control level is an irony not lost on many fieldbus watchers. The protracted path to speci-fication of the H1 standard was littered with turf wars involving virtually

every leading member of the process automation supplier community. The HSE experience is shaping up to be much different, however, with the net-work emerging as a common Ethernet-based control level network already

included in the new system designs of many process automation suppliers. Emerson Process Management was among the first to publicly announce their intentions, but similar announcements are expected shortly. At this

time only HSE Linking Devices from ABB and SMAR have been certified by the Foundation.

)��)��(��.��#+��#��#��)��*��%�� +��%��/++� '��

�� .0�)��

3K\VLFDO

'DWD�/LQN

��

,(&��

# ��.�� #��.9�� :��&:��)

# ��.�� #��.9�� :6


$SSOLFDWLRQ

�8VHU�/D\HU�

)LHOGEXV 'DWD�/LQN�/D\HU

)LHOGEXV 0HVVDJH�6SHFLILFDWLRQ

))�)XQFWLRQ�%ORFNV��'HYLFH�'HVFULSWLRQV

,(((��(WKHUQHW

,(((��(WKHUQHW

8'3��7&3�,3

)LHOGEXV 0HVVDJH�6SHFLILFDWLRQ

))�)XQFWLRQ�%ORFNV��'HYLFH�'HVFULSWLRQV



Convergence of the majority of the process control community around the HSE standard results in process applications drawing largely from a single

technology base while discrete machine control remains a contended field. Siemens, who serves both process and machine control customers, remains the primary process automation supplier not in the Foundation Fieldbus

camp as the company continues to market their Profibus architecture, includ-ing Ethernet-based PROFInet and the field level Profibus PA network.

��.��7��(�� )��While HSE is designed mostly for use at the control network layer, some PAS/DCS suppliers are likely to extend the network into the field for use with their own I/O multiplexers and smart devices. Like other Ethernet pro-

tocols vying for use at the device level, Foundation Fieldbus HSE will be used to connect field devices when the environmental, power supply, and safety requirements of Ethernet are all met at a lower cost than H1. The in-

centive lies in eliminating the need for HSE Linking Devices by eliminating the two-tier H1/HSE architecture. Substituting an Ethernet switch for each Linking Device simplifies the network architecture and reduces the end cost.

Ability to use HSE at the field level will also be advantageous on a perform-ance basis since the network was not designed to have other protocols running underneath it.

� 9�,� � +��-�; �� +�� ,�

��

��&!��,6�< = ��.��.�� 4� 07*�%��')8�� .��+��

�+��*��:�2<;��.��')8�� .23��+��

'��.��

��/3< �. -��.��.��<>92,�� +�� #��.��.��

��.��

#�� -�� +��.��(-��*?�� '��.��/3< ��.��+��

�(��(��/��)��)��(��.��)��

��"��7��,��"� ��.��9�� Recent additions to the Foundation Fieldbus function block library expand

the scope of potential applications for the network. In September of this year the Fieldbus Foundation released specifications for fully configurable Flexi-ble Function Blocks (FFBs) that allow the system developer to specify both

the number and type of I/O parameters plus the algorithm to be configured. These fully configurable FFBs complement the previously released pre-configured FFBs, designed largely for use with remote I/O, where the num-



��1�-��0��'��2��,��,��0� �!!��'��**��

�� $��'��*$�� !��&�'��

3&��"��4��,��0��'$��+5�

ber and type of I/O were predefined but the algorithm is configurable. The fully configurable FFBs, on the other hand, can be used for complex applica-

tions such as batch control, coordinated drive control, PLC sequencing, and other tasks not accounted for in the existing library of standard or pre-configured Fieldbus function blocks. With the addition of this configurable

functionality, the Fieldbus Foundation is positioning 100 MB HSE as the one network for all types of automation applications.

Although the Fieldbus Foundation organization is targeting applications be-

yond process control, where the potential competitive field widens, the organization has so far not joined the IAONA organization that seeks to promote commonality between the disparate industrial Ethernet protocols.

This is also true of the Profibus User Organization (PNO), which is Founda-tion Fieldbus’ primary competition.

��$+��,�&��1��

The mission of IAONA, or the Industrial Automation Open Networking Al-liance, has evolved since its original inception in November of 1999. Initially

formulated as still another group targeting development and promotion of an Ethernet specification in the industrial automation space, IAONA first pursued a merger with the IDA group and then

ultimately re-positioned itself as a neutral umbrella organization for the disparate industrial Ethernet factions.

The first step towards this umbrella role was cemented in a

Memorandum of Understanding signed with competing indus-trial Ethernet groups ODVA and IDA in November 2000. This agreement represents a long-term vision to eliminate the differences between associated

industrial Ethernet protocols and guarantee a minimum level of interopera-bility. Potentially common areas not currently specified in existing protocols are to be pursued jointly under the IAONA umbrella and implemented in

member protocols. In addition to its affiliation with ODVA and IDA, IAONA has extended membership invitations to other Ethernet groups such as the Profibus User Group and the Fieldbus Foundation who have yet to

join the organization.



��$+��,��1��In its original incarnation as an industrial Ethernet protocol propo-

nent/developer, IAONA formed numerous working groups charged with developing various aspects of the technology. This organization was aban-doned with migration towards its new role as an umbrella organization, but

under the current mandate new working groups have again formed to ad-dress potential areas of commonality within specific topics. The chairmen of

these Joint Technical Working Groups (JTWGs), along

with representatives from IAONA Europe, IAONA US, and partner associations such as ODVA and IDA form the Technical Steering Committee that governs the

IAONA organization. A separate marketing group is re-sponsible for the seminars, trade fairs, and other educational and promotional activities the group is un-

dertaking to advance the cause of industrial Ethernet.

While the IDA effort is broad in scope, it largely stays away from trying to resolve the real-time protocol issues

associated with industrial Ethernet and concentrates on higher-level concerns. IAONA’s stated first priority is to ensure that devices supporting competing industrial

Ethernet protocols don’t disturb each other when they are on the same network. Ultimately, development of a common API that will allow communication between the

incompatible industrial Ethernet object models is IAONA’s long-term Holy Grail.

+�� *�� ' �,��IAONA sees itself as the neutral repository for common industrial technol-ogy, a position they’ve signaled should extend to industrial XML schemas as well. Many of IAONA’s activities will be focused on sanctioning various ex-

isting technologies from other standards organizations and submissions of member companies. In the area of connectors, for example, the EtherNet/IP SIG within ODVA has proposed its new IP 67-rated industrial Ethernet con-

nector for consideration by IAONA while IAONA itself has expressed support for CENELEC’s industrial Ethernet standardization effort.

In other areas, however, such as their ultimate goal of developing a common

industrial Ethernet API as a means of rationalizing important application-layer components such as object models and device descriptions, the Joint

� +��-�; �� 9<��

0��*�� 9��!�'&�'��$�9��4��

,�+!�� *��9��*�� 9��$ ��9��*�*��9��

,��.�� 9��+��9��9��95��

��.�� .��9��9��+��*��9��

-�.�� -�.��9��*��4��!�>@5��

'��.�� "�.�� 4� ��.��

��1��2��"�� 3 �+��##��#(� ��



��1�-��6�'(��'�� *� ��*�� !!��"�''��''�"�� $��0� �!��0�&7� ��!��'��+

Technical Working Groups will need to issue deliverables in the form of specifications or guidelines. This will also be true for

intermediate phases such as the management of duplicate IP addresses. As a result, there is a perception building in the marketplace that IAONA could ultimately emerge as still an-

other higher-level industrial Ethernet protocol.

IAONA hopes to plow much of its output back into member protocols, par-ticularly in the uncharted areas of potential commonality, but the reality is

that it may be difficult to convince the participating Ethernet camps to amend or change their existing specifications. The strategy to develop a common API for object-level communications rather than strive for a com-

mon object model addresses this reality. In general, however, IAONA could face an uphill battle convincing entrenched competitors to incorporate the group’s recommendations and specifications.

$�&�:�� !�

At the year 2001 ISA trade show the OPC Foundation announced its inten-

tion to extend the existing OPC DA (Data Access) specification to allow run-time interoperability across disparate industrial Ethernet networks. Func-tioning independently of any specific industrial Ethernet real-time protocol,

the new OPC DX (Data Exchange) will operate at a high level in the automa-tion architecture and enable peer-to-peer communication between controllers, HMI, and other intelligent devices.

OPC DX will be beneficial for customers faced with the need to exchange data within heterogeneous, non-interoperable industrial Ethernet environments:

for example, a packaging house with packaging lines supplied by different machine builders that support competing Ethernet protocols. We know that OPC

DX will include the previously announced OPC XML architecture, and if DX lives up to its promises then each of the compet-ing Ethernet networks may access objects using this common protocol. OPC

DX will function as a middle ground, allowing run-time read/write commu-nication between the controllers, HMI, and other high-level devices that reside at the upper levels of the architecture.

��4�� #��#'�� +� �(�� .�� #��

+6(

';'; ';'; ';'; ';';

2WKHU

3URWRFROV

';'; ';';

��

352),QHW+6(



�

�

1(7:25.�

�

�

6321625�6��

�

7$5*(7�$3�

3/,&$7,216�

1(7:25.��

75$163257�

/$<(56�

�

$33/,&$7,21�

/$<(5�

�

�

2%-(&7�02'(/��

�

�

&211(&725�

�

�

67$786�

(WKHU1HW�,3� 2'9$��&RQWURO1HW�,QWHUQDWLRQDO��DQG�,QGXVWULDO�(WKHUQHW�$VVQ��

&RQWURO�WR�HQWHUSULVH��FRQWURO�OHYHO�QHWZRUN��,�2�QHWZRUN��WLPH�FULWLFDO�DSSOL�FDWLRQV�

7&3�,3�DQG�8'3�WKURXJK�SURWRFRO��HQFDSVXODWLRQ�

3URGXFHU��FRQVXPHU�&,3�SURWRFRO�XVHG�LQ�&RQWURO1HW�DQG�'HYLFH1HW��KDV�DJUHHG�WR�FR�RSHUDWH�ZLWK�,$21$�

%DVLF�&,3�OLEUDU\�SOXV�(WKHU1HW�,3�VSHFLILF�REMHFWV�DQG�GHYLFH�SURILOHV�

%D\RQHW�VW\OH�,3��HQFDS�VXODWHG�5-��

6KLSSLQJ��

)RXQGDWLRQ�)LHOGEXV�+6(�

)LHOGEXV��)RXQGDWLRQ�

&RQWURO��,�2�QHWZRUN�IRU�SURFHVV�FRQWURO��SODQW�FHQWULF�HQWHU�SULVH�LQWHJUDWLRQ�

8'3� 3XEOLVK��VXE�VFULEH�IURP�RULJLQDO��)LHOGEXV��VSHFLILFDWLRQ�

)LHOGEXV�IXQFWLRQ�EORFNV��GHYLFH�GHVFULSWLRQV�

5-�� &HUWLI\LQJ��

,$21$� )RXQGLQJ�ERDUG�LQFOXGHG�+LUVFKPDQQ��-HWWHU��(6(��1RZ��PHPEHUV�LQ�,$21$�(XURSH��

+DV�DJUHHG�WR�IXQFWLRQ�DV�QHXWUDO��XPEUHOOD�RU�JDQL]DWLRQ��FXUUHQWO\�IRU�(WKHU1(7�,3�DQG�,'$�

/RRNV�WR�H[�WHQG�FRPPRQ�DOLWLHV�RI�H[LVW�LQJ�HIIRUWV��VXFK�DV�(WKHU1HW�,3�DQG�,'$�

/RRNV�WR�H[WHQG�FRPPRQDOLWLHV�RI�H[LVWLQJ�HI�IRUWV��VXFK�DV�(WKHU1HW�,3�DQG�,'$�

:RUNLQJ�WRZDUG�&RPPRQ�$3,�YHUVXV�FRPPRQ�REMHFW�OLEUDU\�

+DV�HQGRUVHG�&(1(/(&��HIIRUWV��(WKHU1HW�,3�KDV�RIIHUHG�WKHLU�LQGXVWULDO��FRQQHFWRU�

6SHFLI\LQJ�

,'$�� $*�(��-HWWHU��.XND��/HQ]H��3KRHQL[�&RQ�WDFW��6LFN��6FKQHLGHU��(OHFWULF�

'LVWULEXWHG�PRWLRQ�FRQWURO�ZLWK�PXOWL�D[LV�FRRUGLQDWLRQ�

8'3��7&3�,3�� 3XEOLVK��VXE�VFULEH�IURP�57,��KDV�DJUHHG�WR�FRRSHUDWH�ZLWK�,$21$�

'HYHORSLQJ�WKHLU�RZQ�REMHFW�OD\HU��

� 6SHFLI\LQJ��VKRZLQJ�SDUWLDO�SURWRFRO�SURGXFWV�

352),QHW� 3URILEXV��,QWHUQDWLRQDO�

&RPSOH[��PDFKLQHU\�

8'3��7&3�,3��352),EXV�RU�RWKHU�GHYLFH�QHWZRUNV�IRU�UHDO�WLPH�

'&20�:LUH�SURWRFRO��53&�

(PSKDVL]HV�'&20�EDVHG��REMHFW�LQWHUIDFHV��DXWRPDWLRQ�REMHFWV�GHYHORSHG�LQ�FRQ�WUROOHU�6:�RU�YLD�352),EXV�IRU�UHDO�WLPH�GHYLFH�OHYHO�

5-�� 6SHFLI\LQJ��VW�SURG�XFWV�GXH�<(��

��

The initial OPC DX implementation will be based on the existing OPC DA specification plus its core Microsoft technologies COM and DCOM. The ul-

timate intent is to migrate the specification to the Web Service Architecture and accompanying Internet, SOAP, XML, and Microsoft .NET technologies.

The OPC Foundation is targeting December 2001 for availability of the speci-

fication and sample code with prototype products expected at the Hannover Fair in April 2002. Most of the leading industrial Ethernet network consorti-ums, with the exception of IDA and IAONA, came out in support of OPC DX

when it was first announced. The DX specification will not impact the speci-fications of the respective industrial Ethernet protocols.



�� !� ��'�� 9"��$(��9"��$��*��-��

��1��.�� 9��.��+�� + ��&��+&��&��&��

�1 ��#��-�� '��.��.� ��-��.�� 5��*��*5�� 5��*��*��1 ��#��$�� @��.��'��88��4� �� -��8 ��+A�� ..*!��*-��.��8 ��0��8"��+�� +A�� 9�� &!��* �� ::� 1��+��1��5��$:�� 1��!��.��'��68 (��.��6� (��-�� 6�� (��!��!��'��

��1�� #��$�� .��.��"��+��

�� .��@�� '��;� +A��$��@��+� �� .��'��"��#��>�� -��;� '��+�� 0��'�� -��6��18��-��#��$�� '�� !��'�� "��'��<8; �B��+��$��

1�� 2=/)9�0�� 6�� *�� .�� 9��9�� +�� 9�� .�� *�� 1�� 6��+�� 2333 �� *�� .��9 �0�� $�� C� ��

��0��+�� +��0� ��.��*�� +� �0� #� �� . �� +� �� .��0�

D�� $�� .�0�E� �� .�� .. ��+�� -�� 0�C� � �� -�� .�� .�� +�� .�� .��4�� 1��+��.��9��9.��9��F

�0�� 6��9!�� "��9"� ��9��3<3<)�-�!��F8/2*�82*233391��F8/2*�82*22339��F��.�G�0��+ ��

>��+��0��+ ��

!�� "��H"� ��9��3<3<)�-�H8/2*�82*2333H1��8/2*�82*2233

��!&��0�%� +=+

�>��'��*%��!��(

8$�� %��!��(

6�!&$�0%��!��(

��#(�%�?��

��0�'��%��

��%�8�

��&$�0�%��

��:�� %��

��(��*9=� ,�*�� !�'�� *��!��

Documents

The industrial ethernet protocol wars fieldbus revisited