Industrial Networking: Layers of Connectivity

Jan. 31, 2012
Ethernet has captured the bulk of networking implementation over the past several years. One reason for its success lies in its flexibility.

Properly speaking, Ethernet only defines the physical layer—cables, connectors and firmware that enables communication over the media—but there is also the Open Systems Interconnection (OSI) stack. This seven-layer model encompasses the set of protocols that enable all the cool things.

The OSI stack is what most people refer to when they say Ethernet. The Physical layer includes such standards as IEEE 802.11 or 802.15 that are often referenced within Automation World articles. Layer 2 is the Data link layer, which is not commonly discussed. The Network layer, layer 3, includes Internet Protocol (IP), which finds things on the network. The Transport layer includes TCP and UDP. These are commonly combined as in TCP/IP or UDP/IP. The Session layer, Presentation layer and Application layer complete the seven layers of the model.

All of these allow for much flexibility. But Ethernet’s foundation is in probability theory. If you want determinism, that is, the assurance that a data packet will be received when you expect it to be received, you need to work with the stack. There are many ways to accomplish this feat: CC-Link, EtherCat, Ethernet Powerlink, EtherNet/IP, ProfiNet and Sercos III are some of the methods devised to enable determinism.

Automation World will be providing a service to the engineering community this May 22-23 in the Chicago area during The Automation Conference. My six-year-old dream is about to come true—all the fieldbus proponents in one room on one day, so engineers can see all of them and ask the tough questions. Join me there as we learn the differences and similarities Ethernet, and how to choose the one that’s right for you.