Transmission Control Protocol/Internet Protocol (TCP/IP) is to Ethernet as spices are to food. You rarely get the best experience from one without using the other. That's because the TCP/IP suite is the key, behind-the-scenes technology that brings the Internet and intranets to life -- just as spices bring out the flavors in our food. As a result, TCP/IP is inextricably linked with Ethernet in the business and consumer worlds.

But what many people don't understand is that using Ethernet and TCP/IP only guarantees that Ethernet devices can coexist within a network; it does not guarantee that these devices can communicate effectively. For devices to be able to understand and interoperate with one another, a common application-layer protocol is needed.

The TCP/IP architects understood this problem and designed these protocols to work with multiple application layers. To date, many are used with the TCP/IP suite -- e.g., protocols for file transfer (File Transfer Protocol [FTP]), e-mail (Simple Mail Transfer Protocol [SMTP]), and World Wide Web (Hyper Text Transfer Protocol [HTTP]).

Now that Ethernet has infiltrated the plant floor, networking organizations and automation vendors are promoting several new application-layer protocols. These protocols are intended to make Ethernet suitable for the rigors of industrial automation. However, not all technology is created equal.

Introduced in early 2000, EtherNet/IP™ is the front-runner of all the industrial Ethernet solutions. Part of the reason is that it is based on open technology, using the same application layer -- the Common Industrial Protocol (CIP) -- as DeviceNet and ControlNet. This approach offers myriad benefits to automation users and vendors alike: low product development costs, ease of use, simple product and network integration, and multi-vendor interoperability.

Further, more than one vendor or organization backs EtherNet/IP. ControlNet International (CI), the Industrial Automation Open Networking Alliance (IAONA) and the Open DeviceNet Vendor Association (ODVA) recognize and/or support this networking standard. As a result, EtherNet/IP delivers interoperable Ethernet products from a large community of vendors.

In fact, there is a full catalog of EtherNet/IP products available today, which means that manufacturers can currently deploy complete solutions with a flat (or single network) EtherNet/IP architecture. Simply put, it is the Ethernet-for-automation vision realized.

What Is EtherNet/IP?

It's important to understand the basic technology behind the name before delving too far into the specifics. EtherNet/IP is best defined by deconstructing its name. The "EtherNet" portion is fairly straightforward. It refers to commercial off-the-shelf Ethernet (IEEE 802.3) -- not a modified, proprietary version -- and the TCP/IP suite.

"IP," on the other hand, stands for Industrial Protocol and is what distinguishes this network. Unlike many industrial Ethernet options, EtherNet/IP uses an already proven, open protocol at the application layer (i.e., CIP). So EtherNet/IP is CIP deployed on an Ethernet TCP/IP network, just as DeviceNet is CIP deployed on a Controller Area Network (CAN).

In terms of how it works, EtherNet/IP uses TCP/IP to send explicit messages -- those in which each packet not only has application data, but also carries the meaning of the data and the service to be performed on the data. With explicit messaging, nodes must interpret each message, execute the requested task, and generate responses. These types of messages are used for device configuration and diagnostics, and are highly variable in both size and frequency.

EtherNet/IP also leverages standard User Datagram Protocol/Internet Protocol (UDP/IP -- part of the TCP/IP suite) transport services, which provide higher performance and multicast functionality for real-time -- also known as implicit -- messaging. With implicit messages, the application data field contains only real-time I/O data. The meaning of the data is linked to an identifier that is defined at the time the connection is initially established, reducing the processing time in the node during runtime. Such messages are low overhead, short, and provide the required, real-time performance needed for control.

So because it leverages both TCP/IP and UDP/IP protocols to encapsulate networked messages, companies can use EtherNet/IP for both information and control applications.

EtherNet/IP -- Here and Now

EtherNet/IP is almost three and a half years old. In that relatively short time, this network has seen a significant amount of interest and activity. A year after CI and ODVA unveiled the specification, Rockwell Automation launched the first wave of EtherNet/IP products, which centered on the Allen-Bradley ControlLogix™, ProcessLogix™, PLC® 5, and SLC™ 500 controller families. Since then, many other vendors have either developed EtherNet/IP products or announced plans to do so.

Today, EtherNet/IP-enabled hardware and software abounds. This includes vision systems from Cognex and DVT, weigh systems from Mettler Toledo, ID systems from Datalogic, Balogh, and SICK, motion controllers from AMCI and Bosch Rexroth, and communication boards from Hilscher and HMS Industrial Networking. In addition, Wind River has added an EtherNet/IP stack to its Tornado development environment, and companies like HMS, Prosoft, and Woodhead offer a wide variety of bridging devices to allow companies to upgrade almost any existing network to EtherNet/IP.

And the list goes on. All of these products are available and operating in plants around the globe. In other words, EtherNet/IP is a field-proven technology.

With the introduction of the Allen-Bradley CompactLogix™ L35E controller, EtherNet/IP solutions now span the full range of control-system requirements. As a result, engineers have the option of using a flat EtherNet/IP architecture to network many of their applications -- from a complex brewery line to a simple motion control system. However, there are pros and cons to this (and any) approach.

Not Everything to Everyone

Most people don't lug stereo components and speakers everywhere they want to listen to music. In the same regard, portable CD players and radios aren't capable of providing the high fidelity and clarity audiophiles dream about. The point is that no piece of technology is capable of addressing every need. That holds true for industrial networks as well. EtherNet/IP is bound by the inherent limitations of Ethernet, which makes it less than ideal for every application. For example:

• It's not as robust as ControlNet. ControlNet is a real-time, control-layer network providing high-speed transport of control data and information, like EtherNet/IP. However, ControlNet does not require active infrastructure components; it uses passive taps and network media. It also offers physical media redundancy as a standard option. EtherNet/IP, on the other hand, uses active components like switches. If a component goes down, the entire network is affected. Likewise, it requires two separate EtherNet/IP networks, infrastructure components, etc. to provide media redundancy similar to ControlNet.


• It doesn't scale as far as DeviceNet. The DeviceNet communication link is based on a broadcast-oriented protocol, the Controller Area Network (CAN). Consumer and commercial demand for CAN has been the driver in lowering the price and increasing the performance of CAN chips, which are currently available for less than 20 euro. This makes it cost effective and easy for vendors to add DeviceNet connectivity and mass-produce their products. The same is not yet true for EtherNet/IP. More components are required -- a microprocessor, TCP/IP stack, etc. -- and the cost of an EtherNet/IP chip can run more than 90 euro.


• It's not as widely adopted as Profibus. More than 1,000 Profibus products are currently available, and it is one of the most widely used networks in Europe. The number of companies building any type of industrial Ethernet hardware is in the hundreds. However, it's interesting to note that the same dynamic once existed in the office environment. Ethernet was not always the de facto, universal standard. But once it gained momentum, it was impossible to hold back. In similar fashion, Ethernet is now proliferating on the plant floor.

Many Things to Most Companies

So why would a manufacturer choose EtherNet/IP? To start, there isn't one network that is as robust as ControlNet, as scalable as DeviceNet, etc; each network has its own strengths and weaknesses. However, EtherNet/IP has enough features and functionality to make it work for a wide variety of applications. As a result, it's more versatile than most networks.

Because EtherNet/IP provides both control and information capabilities, it also gives companies the ability to collapse their communications architecture (if they choose) -- which brings numerous benefits. Namely, a single-network design saves time and money. That's because it is accompanied by one configuration software package. In addition, plant electricians are required to learn only one installation method (even if it is a bit more complex). In this case, familiarity does not breed contempt. It breeds a manageable learning curve, and moderate training and manpower costs.

Another plus: IS and IT departments worldwide have been using Ethernet for years. Such long-term exposure to the Ethernet technology has produced an expansive knowledge base and unparalleled resources, most of which translates directly to EtherNet/IP. For instance, a secondary benefit of converged technologies has been the realization that IP telephony and I/O control have very different challenges, but share many of the same solutions.

The primary reason manufacturers turn to EtherNet/IP is not its similarity to fieldbus, but that it offers several new capabilities -- attributes exclusive to an Ethernet-based network. These include:

• New services like voice and video


• Variable topologies


• IT integration


• Remote management


• Internet connectivity


• Flexible security.

Rather than provide mounds of technical detail, the best way to describe these capabilities is to illustrate each through real-world examples.

The Before and After

A traditional application would look something like the following. An information network links a programmable logic controller to an upper-level, MES-type system. The controller also is connected to a second network -- a dedicated I/O bus -- to link all the peripheral devices on the factory floor. The end user would use the I/O network to modularize the construction and wiring of the system, gather remote data, and manage the speed of production.

In comparison, the controller could now be linked to an EtherNet/IP network for both information and I/O using a flat network architecture. Depending on the application, this setup can offer not only better performance than the traditional application, but additional functionality as well.

• New services like voice and video -- e.g., installing a camera to watch the plant floor and product flow remotely.


• Variable topologies -- e.g., setting up all the programming terminals on a virtual LAN, which would allow engineers to connect the terminals on a single network, yet segregate and isolate multiple control systems at the same time.


• IT integration -- e.g., accessing the MES system from the plant floor to review and refresh, in real time, maintenance scheduling information.


• Remote management -- e.g., programming the control system to send a message to a maintenance pager when alarms occur.


• Internet connectivity -- e.g., accessing an online troubleshooting guide to adjust an inoperative sensor.


• Flexible security -- e.g., restricting access to a port on a switch, which results in restricted access to a specific I/O block.

These additional capabilities provided by Ethernet give manufacturers a significant advantage, assuming EtherNet/IP can meet all of their industrial network requirements. These capabilities are not, and cannot be, available on a fixed-purpose industrial network. Moving to this type of architecture marks a philosophical and operational shift for most companies -- one that opens a realm of new possibilities to an innovative user.

And After the After?

Like all thriving technologies, EtherNet/IP is not static. So the organizations and their member companies responsible for its growth, including Rockwell Automation, are sure to add features that will help companies evolve with changing business requirements, such as safety, synchronization, and wireless utilities. But no one knows what lies on the horizon for sure. Regardless, Rockwell Automation has a reputation for Ethernet innovation that will carry forward into its work with EtherNet/IP:

• First Ethernet-connected PLC (Pyramid Integrator -- 1990)


• First PLC with embedded Ethernet port (PLC-5 -- 1994)


• First small logic controller with embedded Ethernet port (SLC 500 -- 1995)


• First routable industrial Ethernet architecture (NetLinx -- 1999)


• First PLC with multiple Ethernet ports (ControlLogix -- 1999)


• First PLC with real-time I/O on Ethernet (ControlLogix -- 2001)


• First PLC with embedded I/O on an Ethernet port (CompactLogix -- 2003).

And Rockwell Automation is striving to deliver many more of these firsts in the coming months and years. Just as Michelin three-star restaurants deliver the best-prepared food, EtherNet/IP delivers the finest in industrial Ethernet networking solutions.