Last time we checked, it was reasonably easy to get a bus transfer in New York, but do not expect that level of ease when doing transfers from one industrial bus to another. More than likely, you'll get more attitude than a rush hour bus driver.
What you need is a good map to avoid the roadblocks along the way and a clear and logical itinerary that avoids the potholes. Help is here.
The interconnection and integration of existing heterogeneous sub-networks into a homogeneous network has always presented challenges. Evolving and converging protocol standards along with emerging multi-protocol components suggest a clear understanding of the problems and solutions has never been more important.
These concerns include reliability, redundancy, robustness, and security. Most importantly, the multi-bus points of integration should be as seamless as possible, and a consistently high level of security should persevere throughout.
Where legacy equipment is still performing as required, an efficient way to integrate these components into the overall scheme while preserving security and reliability is vitally important. As new equipment and processes evolve, a consistent strategy of deployment of Ethernet-supported interfaces insures proper integration with a minimum of downtime and re-engineering delays.
The wide range and availability of Ethernet solutions and its support from standards groups, vendors, OEMs, and industry provides the assurance that it will continue to evolve. From the early days of 10Mb coaxial cable products, Ethernet has moved into the gigabit range and beyond. Bandwidth and media support outstrip any other transmission and access control protocol set.
The range and types of industrial networks and buses are very broad, as many evolved to handle specific types of industries and related applications. Some lay claim to more universality and interoperability.
Our focus will be on these network architectures, since a prime prerequisite for choosing a common network platform must include its ability to work with other networks and be standards-based. For convenience, the term “network” will mean network or bus.
There are a number of ways to characterize the various networks, none that provide clear dividing lines; for our purposes, let us group them as proprietary and open standard.
Even proprietary systems may provide gateway solutions for interconnectivity, but they require more configuration, modification, and testing than an open standard system.
In addition, some networks attempt to standardize on the application and message syntax, or upper layers in Open System Interconnect (OSI) terminology. Let's look at the lower layers responsible for getting the data to a destination device, user, or application.
If you've never seen the OSI model, avert your eyes from the diagram you see here. Too late. It's not possible to read any book on data communications and networking and avoid seeing this conceptual view of generic network architecture. Fortunately, we will spend most of our time in layers one and two, which is where Ethernet operates.
The following networks provide their own proprietary layer one and two implementations, but some offer alternate access via an Ethernet interface. In some cases, the media may also be unique to the network. Some define an architecture using layer concepts similar to the OSI model above.
Lower speeds and a variety of topologies characterize these networks, but Profi-bus, Modbus, and Foundation fieldbus have also joined the Ethernet bandwagon. While each network is important in its own right, none can claim its physical and data link layer protocols are good interconnect strategies, and thus need some type of gateway to communicate with other industrial networks.
The following networks came about with Ethernet in mind or evolved to support Ethernet and some TCP/IP functionality.
The open standard networks support multiple topologies over Ethernet at speeds of 100Mbit/s or better and provide TCP and UDP interfaces. This should make clear the evolution to high speed Ethernet over a variety of topologies.
Industrial Ethernet supports the integration of existing sub-networks into a homogeneous network that includes effective routing, redundant links, and beefed up security. Many Ethernet vendors offer products designed for industrial environments. Ethernet hubs, managed and unmanaged switches, media converters, and edge switches, hardened for hostile environments, are available at speeds up to the Gigabit range and operational over a variety of topologies using coax, copper, and fiber. Since many industrial networks already support Ethernet, integration is often straightforward.
Software is usually GUI-based to more effectively manage, configure, and monitor industrial networks. Ethernet managed and unmanaged switches, including edge switches, coupled with the right topology, provide the best solution to the control and support of remote sites such as power substations and unattended sites.
Industrial Ethernet networking has inherent advantages. By utilizing a standards-based solution that supports multi-vendor implementations, industrial Ethernet users enjoy highly reliable systems with rapid recovery, reduced costs of deployment, and a guaranteed upgrade strategy as needs evolve. Redundancy and self-healing Ethernet networks provide the desired 24/7 uptime.
One other advantage of Ethernet is Power over Ethernet (PoE), which can greatly simplify the wiring of the many sensors, monitors, and input devices found in industrial Ethernet environments.
This makes possible arrangements like an integrated substation network, which interconnects substations and central operations systems. Numerous Intelligent Electronic Devices such as relays, sensors, meters and Remote Terminal Units, as well as surveillance cameras, VOIP phones, and other devices, connect in substation Local Area Network (LAN).
Serial protocol devices connect via routers or terminal servers, and Ethernet devices, including Power-over-Ethernet enabled video cameras, connect directly to Ethernet switches. The substation LAN connects to a Wide Area Network (WAN) router to transmit data to central operations systems and centers for processing and storage.
The hard choices of integration include decisions of how to connect to proprietary systems. This will vary from industry to industry. Most legacy systems that continue to perform well are candidates for some form of gateway interface unless local management elements are fully effective or isolation from other networks is desirable. The security features of many Ethernet switches can block intrusions, and most routers offer firewalls and filtering options to keep systems secure.
When an older proprietary system is not performing, migration to Ethernet will permit a number of enhancements.
Gateways make communications possible between dissimilar systems. The range and types of gateway devices are broad, and the configuration and proper matching of two different interfaces can be a daunting task.
The difficulty arises from the number of layers that the two architectures must match and integrate. The ideal is to make the gateway transparent to both systems. System A thinks it is just talking to another member of its network and is not aware system B is different.
When a gateway only has to deal with the routing, addressing, and transmission of data, the configuration is relatively simple. Referring back to the OSI model, this covers layers one to three.
When it involves applications and message syntax, things get a lot stickier and more time consuming. The more layers that must be converted, the more processing overhead is involved. This can be unfortunate for critical real-time systems.
When the conversion involves the interconnection of systems and the transport of data, several Ethernet vendors offer components that efficiently handle the task.
The amount of effort this takes, of course, is based on how dissimilar the systems are and how much control of the subnets is required.
There are many ways Ethernet components and standards can manage and provide for redundancy, robustness, security, and flexibility of design for many industrial networks. Ethernet is also the best integration strategy available to network planners and architects.
Topology and redundancy: Ethernet works with many topologies insuring the right topology for the job. At the edges of a network with geographically separated devices, the ring topology supported by Ethernet-managed switches provides several advantages.
Ethernet switches that support IEEE 802.1w, the Rapid Spanning Tree Protocol (RSTP), provide redundant links that can quickly recover from topology changes and add to the reliability of the ring. Because RSTP is designed to work with all topologies, some vendors offer proprietary and/or standards-based redundancy protocols that can significantly reduce recovery time down to as little as 50ms in the simple rings that are often used at the edge of a network.
Since the ring is comprised of devices with point-to-point links, signal reshaping and retransmission of the sending leg reduce the possibility of transmission errors. Cabling costs are also significantly less than installing a separate link to each remote device as in a star or mesh topology. Where all devices are located together, a simple star or bus topology works just fine.
Some Ethernet switches support dual homing. In Ethernet LANs, dual homing is a network topology that adds reliability by allowing a device to connect to the network by way of two independent connection points (points of attachment). One connection point is the operating connection, and the other is a standby or back-up connection that cuts on in the event of a failure of the operating connection.
Security: The 2003 Slammer worm attack on portions of the Northeast U.S. power grid confirmed the need for better security than currently implemented. The Energy Policy Act of 2005, which goes into effect the summer of 2006, provided a further push for a higher level of security in power systems. Both Ethernet and TCP/IP provide several sophisticated security features honed in IT departments and equally available to industrial Ethernet users.
Several TCP/IP-based and IEEE-based standards have emerged to handle intrusions over Internet-like connections. These include various forms of user authentication, password protection, and encryption. Managing a remote Ethernet component (switch, router, and hub) is most effective using standard GUI-based protocols.
The User-based Security Model of the SNMPv3 standard specifies the use of the Data Encryption Standard, or DES-CBC, using a 56-bit key. Each manager must know the privacy key of each agent with which it communicates. Any Ethernet switches employed should provide remote access security for Telnet CLI communication, SNMP management, and Web-interface access.
Ethernet, because of its high bandwidth, is the best protocol for deploying physical security devices at remote and peripheral sites. PoE adds ease of supplying power to remote security devices.
Virtual LAN support: VLANs are widely used today for reducing broadcast traffic by limiting the size of a shared-traffic domain. Since crossing a domain involves a routing decision, the security of a given domain is certain. A VLAN creates separate network segments that can span multiple Ethernet switches. A VLAN is a group of ports designated by the switch as belonging to the same broadcast domain. The IEEE 802.1Q specification establishes a standard method for inserting VLAN membership information into Ethernet frames.
VLANs, as the name suggests, create virtual LANs administratively. Instead of going to the wiring closet to move a cable to a different LAN segment, the same task can transpire remotely by configuring a port on an 802.1Q-compliant switch to belong to a different VLAN. The ability to move end stations to different broadcast domains by setting membership profiles for each port on centrally managed switches is one of the main advantages of 802.1Q VLANs.
There are two reasons to maintain proprietary buses: legacy systems that are still providing satisfactory service, and highly tuned and specific applications.
However, in a world where costs, high availability, and future proofing are key operational objectives, industrial Ethernet is the clear winner for new deployments.
Industrial Ethernet provides the best support, redundancy, security, integration, and migration for industrial sites as they continue to evolve in the post-9/11 era.
Jim Hammond (jhammond59@comcast.net) is a technical trainer and writer in New Mexico who has specialized in networking and data communications for 30+ years. He has worked for Memorex, Compaq, Tandem, and HP. Larry Castelli (lcastelli@garrettcom.com) is Industrial Product Manager at GarrettCom, Inc. in California where he leads the company's control and industrial automation product marketing. He has two degrees in physics and an MBA.
View the online version at www.isa.org/intech/20070201.
Fast ForwardThere are two reasons to maintain proprietary buses: legacy systems that are still providing satisfactory service, and highly tuned and specific applications. Industrial Ethernet provides the best support, redundancy, security, integration, and migration for industrial sites as they continue to evolve in the post-9/11 era. VLANs, SNMPv3 support, encryption, and SSL connections provide a secure environment as networks grow and open. |
TerminologyPower over Ethernet, or PoE, technology describes a system to transmit electrical power, along with data, to remote devices over standard twisted-pair cable in an Ethernet network. It is useful for powering IP telephones, wireless LAN access points, Webcams, Ethernet hubs, computers, and other appliances where it would be inconvenient or not possible to supply power separately. Per international standard, it provides up to 48 volts DC with a maximum current of 400 milliamps. |
RESOURCESOpen bus integration Ethernet strategies in practice Innovative integration best medicine |