EtherCAT vs. EtherNet/IP

EtherCAT and EtherNet/IP are both Ethernet application layer protocols designed to move inputs and outputs between controllers and endpoints on a manufacturing machine. That sounds like they’re pretty similar, and they are. However, though the results may be similar, how they accomplish moving I/O data couldn’t be more different.

EtherCAT is a highly flexible Ethernet network that uses something called “processing on the fly.” EtherCAT messages are passed to the next node in the ring before being processed by that node, providing the network with blazing speed and efficiency. It also means that EtherCAT can offer flexibility in topology and incredible synchronization. Like the competitive offerings, it’s a complete solution. EtherCAT includes a safety protocol and multiple device profiles, among other things. EtherCAT also benefits from a strong user group, the EtherCAT Technology Group.

EtherNet/IP, on the other hand, is an implementation of the Common Industrial Protocol (CIP). CIP is a well-defined data representation, connection management and messaging protocol that operates over many Transport and Physical Layers. It defines how devices connect to one another, how they represent data to external entities and how data moves from device to device. CIP is an object-based technology that exposes data over a CIP network as a collection of attributes (the data values) grouped in categories called objects. EtherNet/IP is simply the implementation of CIP over Ethernet TCP/IP.

EtherCAT has some significant advantages over EtherNet/IP.
  • It’s fast – blazingly fast.
  • It supports integrated black channel safety. Black channel in this context means that the safety network itself doesn’t appear to be there. The bus system in a black channel network does not perform any safety-related tasks but only serves as a transmission medium.
  • You get full redundant operation with an extremely fast switch over.
  • You get “processing on the fly” – all the I/Os are processed as the message moves through the node from the input port to the output port. Outputs are taken out of the message and inputs are added to it.
  • Telegrams in an EtherCAT message can be up to 60KB, meaning that you can move a lot of data from a node to the master.
  • Master devices simply issue the message and receive the response. Single message in – single message out. Much simpler than an EtherNet/IP Scanner that must keep track of messages to all of its slave devices.
  • There is no practical limit to the number of slave nodes on an EtherCAT network.
  • EtherCAT can serve as the transport protocol for some other protocol like BACnet, EtherNet/IP, PROFINET IO, or Modbus.
  • Slave devices can support any TCP/IP application program like HTTP, HTTPS, or HTML.
  • Slave devices can be designed to process TCP data without having to implement a TCP/IP stack. TCP/IP processing can happen in the EtherCAT master device and the UDP or TCP packet data passed to the slave device.
  • It requires no underlying bus system to manage nodes and connections.

The main disadvantage of EtherCAT is that slave devices need to embed a specific hardware ASIC to implement EtherCAT. Its data model is very different and can be very difficult to understand.

EtherNet/IP also has its advantages.
  • EtherNet/IP is the way that all Rockwell Logix Controllers connect field devices. It is the most widely used application layer manufacturing protocol in North America.
  • Ethernet/IP uses all the transport and control protocols used in traditional Ethernet including Transport Control Protocol (TCP), Internet Protocol (IP), and the media access and signaling technologies found in off-the-shelf Ethernet interface cards.
  • The core of EtherNet/IP is the CIP which is also used in ControlNet, DeviceNet, and CompoNet. This standard organizes networked devices as a collection of objects. It defines the access, object behavior and extensions that allow widely disparate devices to be accessed using a common mechanism.
  • EtherNet/IP is a standard application layer on top of Ethernet and TCP/IP. No special hardware is required. Cyclic messages are transferred using the User Datagram Protocol, or UDP. Acyclic messages are transferred using TCP.
  • There’s a multitude of support tools, training, and documentation about how to design and integrate an EtherNet/IP system with Rockwell controllers.
  • Rockwell provides the Add On Profile program which improves how a vendor’s device integrates with a Logix PLC.

Just like it’s cousin, EtherNet/IP has its disadvantages, too. For one, it has limited bandwidth and speed. Rockwell PLCs can only move packet sizes of around 500 bytes. A significant issue for embedded devices is that EtherNet/IP TCP and UDP messages must be processed, and that consumes significant processor bandwidth. And finally, there’s no real information modeling. You can create objects with attributes, but you can’t link objects together into any sort of hierarchy.

Is one better than the other? No. If you need to integrate into a Rockwell architecture, EtherNet/IP is the best choice. If you need the massive speed of EtherCAT (and most architectures don’t), then it’s the better choice. EtherNet/IP is less complex than EtherCAT, which is so flexible and functional that it drives up the complexity.

EtherNet/IP and EtherCAT are different animals with distinctive characteristics and different ways of moving data. One is not better than the other. Each can be the right choice depending on the specific application considerations.