The Benefits of Ethernet for Distributed Measurement and Control Systems

Just as the measurement and control industry adopted PC technology standards, such as operating systems and internal bus architectures over the last two decades, engineers who need distributed intelligence and I/O today are applying communication technologies standardized by the PC industry. Traditionally, vendors created their own proprietary communications networks to connect to sensors, pass data between control and data acquisition devices, and communicate to the enterprise. As a result, interoperability was poor, equipment was expensive, and improvements were protracted. The PC industry has standardized on general-purpose communications buses, making it easy to design a distributed measurement and control system. By using flexible software, virtual instrumentation abstracts much of the communication details and helps engineers and scientists select the data bus that best fits their needs -- whether it is PCI, PXI, GPIB, USB, FireWire, Ethernet, or future communications buses. Ethernet has become the de facto standard for highly distributed systems. There are four main benefits of using Ethernet that have driven its adoption for distributed measurement and automation.

Four Benefits of Ethernet for Distributed Measurement and Control

1. Ubiquitous Technology
A prime benefit of Ethernet in distributed measurement and control systems is the standardization of equipment and tools by the PC industry, which drives rapid improvements in performance, features, and ease of use while decreasing costs. Today, for simple test systems, the equipment cost of an Ethernet communication system for local data acquisition or control can be less than $50, and you can purchase the components at a neighborhood electronics store. For high-end systems where reliability is a prime concern, commercial off-the-shelf (COTS) technology designed to provide high levels of uptime also is available. Vendors such as Cisco Systems offer components that incorporate redundancy and automated switchover to meet the needs of companies that rely on Ethernet for Web stores, back-end ordering systems, and manufacturing systems. With these products, you can detect network problems and fix them within seconds using technologies such as spanning tree protocol, which provides path redundancy. For environments with long cable runs or with high electromagnetic noise, you can use fiber optics -- a standard Ethernet offering -- or for mobile communications, you easily can configure your system to use wireless Ethernet via IEEE 802.11, another standard Ethernet option. Companies such as Hirschmann and Woodhead even manufacture a line of switches and cables with an IP67 rating designed for operation directly on machines.

The IT industry also has tools to manage connectivity and security for data acquisition or control systems that use Web-based communications, like the Web server built into every National Instruments LabVIEW Real-Time target. For instance, with off-the-shelf tools such as managed switches, you can configure bandwidth priorities and set up logical groupings of network devices through VLANs and for security, you can use standard tools to set up virtual private networks (VPNs), enable IP filters, limit port access, and monitor for unauthorized access.

2. Simplified Communication between Machines
In the past, engineers building distributed systems were often forced to standardize on one vendor because of difficulties implementing machine-to-machine (M2M) communication using hardware from multiple vendors. The problem was that each vendor offered a specific communications bus that was not supported by other vendors' equipment. Now that companies are standardizing on Ethernet, it is possible to connect multiple devices on one physical bus. The seven layers of the OSI architecture allow Ethernet to support different protocols simultaneously on a common physical connection. On a single Ethernet bus, you may have FTP, HTTP, and Modbus TCP communications happening simultaneously. For instance, you can use NI LabVIEW software to create a distributed system that communicates between LabVIEW nodes using the shared variable feature, collects data from a Modbus TCP PLC, and reads configuration data from the built-in FTP server on a LabVIEW Real-Time target. With the communications abstractions built into LabVIEW 8, it is now simple to support numerous protocols using the new shared variable architecture. For instance, today you can add the new National Instruments cFP-1808 Ethernet backplane to any system based on LabVIEW and automatically map the I/O directly into a shared variable that can be read from and written to via a simple LabVIEW icon. Additionally, you automatically can republish every I/O point added via a shared variable on a Windows PC using a built-in OPC server, allowing most major data acquisition and control vendors to read them. There are even gateways available that map I/O data from legacy networks such as DeviceNet and PROFIBUS directly onto an Ethernet network, making M2M communications easier. National Instruments has references on how to work with industrial networks, such as Modbus, PROFIBUS, CANopen, CCLink, and DeviceNet, using gateways and LabVIEW at ni.com/comm.

[+] Enlarge Image


Figure 2. Compact FieldPoint-1808 Distributed Network Variables






Figure 3. LabVIEW Project Manager with Shared Variable to cFP-1808 Thermocouple Module

3. Communications to the Enterprise
One of the main benefits of Ethernet is the ability to easily communicate between machines and enterprise systems (M2E). Most enterprises have an existing Ethernet network in place, typically in the form of a local area network. Users share a variety of data through such networks, ranging from management-level reports and supply-chain management data to corporate databases accessed by individual workstations. In applications such as end-of-line test in which the quality of a product is verified, it often is necessary to store the test data in a database for traceability. An Ethernet-enabled data acquisition system can access corporate databases through standard interfaces such as ODBC, SQL, and ADO. With this ability, the Ethernet I/O system can deposit acquired data onto corporate databases where you can use it to create reports and analyze the performance and condition of the plant. The same holds true for connectivity with enterprise resource planning (ERP) systems such as those made available by SAP and Oracle. These systems can directly connect to the data acquisition system via Ethernet and can use the acquired data to improve the understanding and visibility of manufacturing performance and the condition of resources. Ethernet measurement and control systems may also have the ability to connect to the Web, send e-mails, or pass information using file transfer protocols. With LabVIEW Real-Time controllers, users on the network directly can access the embedded Web server through a Web browser and monitor the state of the device and the data being acquired and can even modify the acquisition parameters through a custom interface.

4. Bandwidth
Ethernet’s bandwidth makes it suitable for a wide range of measurement and control applications. The standard today is fast Ethernet running at 100 Mb/s, and some measurement systems today also are incorporating Gigabit Ethernet. However, the key to Ethernet's speed is the network design. Traditionally Ethernet networks used hubs to connect individual nodes. A hub physically connects all of the devices and transparently passes the Ethernet packet to all devices connected to the hub. Because a hub retransmits data from one node to all of the nodes, the network bandwidth is shared by all of the devices connected to the hub. This is why the network becomes congested when a person using a computer on a hub downloads a large file or group of files from another computer. On a network with hubs, two devices transmitting at the same time causes a collision and both sides back off and wait for a random time interval to resend the data. This process is known as Carrier Sense Multiple Access with Collision Detection (CSMA/CD). When a node is on a network in which collisions can occur, the device must concurrently transmit and listen for a collision. Because the device always listens for a collision on a network with a hub, it cannot simultaneously send and receive data. This type of network is known as a half-duplex network.

One of the most important advancements in contemporary Ethernet networks is the use of switched Ethernet. An Ethernet switch maintains a lookup table of the MAC addresses of the devices connected to each port. When a switch receives an Ethernet packet, it compares the destination MAC address to the MAC addresses stored in its internal lookup table and then internally connects the two ports. In a network consisting completely of switches, every node has a dedicated communications port and the switch manages the connection between nodes. If two devices try to communicate to the same node, the switch buffers one of the messages to avoid collisions. A switched network eliminates the possibility for collisions and enables the individual nodes to operate in full duplex mode where they both transmit and receive at the same time, effectively doubling the total network bandwidth. Additionally, because a switch can make multiple connections simultaneously, the network bandwidth is not shared among all the devices.

When compared with traditional serial networks, the speed improvements of Ethernet become apparent. In one wastewater application, a traditional 485-based serial network was considered for a citywide data acquisition and control system. Because RS485 systems have slow data rates, the system was specified to achieve an update from each station once every 10 minutes. When the city examined the proposals, they decided to install an Ethernet system instead of a serial-based system. This not only gave city officials greater bandwidth but also helped them implement an event-driven protocol where each station updated the main control station whenever a data value changed outside of the deadband. The resulting system achieves almost instantaneous updates from every station so the operator knows the real-time status of every node on the network.

Adoption of Ethernet for Industrial Networks
Ethernet is continuing to gain adoption in distributed measurement and control applications due to lower costs, device interoperability, communication to the enterprise, and high bandwidth. Looking forward, the largest trend affecting the future of Ethernet is deterministic Ethernet for applications that require tight synchronization and data transfer between nodes. In standard Ethernet timing, collisions on hub-based networks introduce variations, network switches introduce propagation delay, and individual nodes introduce delay. These delays can cause problems for any time-based system including control applications. There are a number of new Ethernet standards being introduced to provide varying levels of determinism including EtherNET/IP (ODVA), PROFINET (Siemens), EtherCAT (EtherCAT Technology Group), ETHERNET Powerlink, and SERCOS-III (IGS). Each protocol offers varying levels of performance and cost but generally require all nodes on the Ethernet segment to support the protocol. Fortunately, LabVIEW software abstractions, such as the shared variable, make it possible for you to select the best communications bus for your measurement and control applications.

Learn how to work with industrial networks and use gateways with LabVIEW.

Todd Walter
Industrial Measurement and Control Product Manager
todd.walter@ni.com

23 ratings | 4.17 out of 5

Read in English Chinese (Traditional) Spanish | Print

Reader Comments | Submit a comment »