Introduction
Since its invention in 1973 and subsequent adoption as the network standard IEEE 802.3, Ethernet has grown to become the most widely used technology in networking business systems such as back-office mainframe computers, printers and data servers. Over time, factors such as improved speeds, ease of installation and maintenance, decreasing hardware costs and scalability have sustained the fast growth and adoption of Ethernet as the preferred networking technology. In recent years, there have been growing efforts to transfer these improvements to industrial networks which traditionally have been non-Ethernet-based (Cisco 2010, p.3). This integration has created what has come to be known as the industrial Ethernet. This essay briefly examines various aspects of an industrial Ethernet and the reasons why it is increasingly gaining adoption in the distributed systems of industries.
Legacy industrial networks
Industrial Ethernet refers to the integration of Ethernet technologies into traditional industrial networks (Cisco 2010, p.6). Industrial networks have traditionally been characterized by separate information, device and control networks. The information network is normally made up of corporate applications for running administrative functions. The control network is usually linked to the corporate network and connects control and monitoring devices such as automation controllers, minicomputer controllers, I/O nodes, servers and human workstations (Cisco 2010, p.2). The device network is linked to the control network made up of I/O devices such as sensors and motion equipment such as robotics, nano controller, frequency drives and motors (Cisco, 2010 p.2). This networks use a combination of different technologies for interconnectivity. These technologies include variants of fieldbuses technologies such as DeviceNet, Profibus, Modbus, CANopen, and SERCOS I / II to link the devices in addition to gateways or routers to link to the corporate network which mostly runs on Ethernet standards. Each Fieldbus solution normally operates at a different proprietary protocol. Although such networks can meet the resilience and real-time requirement of automation and control systems of a typical manufacturing environment, they have a number of significant drawbacks. First, owing to different vendors of Fieldbus solutions coupled with different specifications, it is common to witness many networks intertwined so as to establish complete connectivity. This also means more manpower and frequent training. Second, it was not easy to upgrade the network say in functionality due to the limited bandwidth of such networks. Third, it is quite involving to maintain and keep the network running. In an industrial Ethernet, the three separate networks are fused into a single network whereby the “fieldbus-specific information used to control I/O devices and other manufacturing components are integrated into Ethernet frames” (Cisco 2010, p.3). The type of Ethernet deployed, however, differs significantly from that of traditional Ethernet architecture in the corporate data networks.
Requirements and Implementation
The harsh environments of industries mean that the Ethernet should be specially designed to withstand extremes of temperature, shock and unique power voltages common in these places. (Cisco 2010, p.7) This calls for industrial-grade Ethernet hardware and software. The core difference however is the consistent and fast data transfer requirement of the automation and control systems. A downtime, however minor can greatly affect production in an industry. The data flow should be consistent and stable. This calls for the implementation of special fault tolerance and redundancy measures in addition to other technologies to support such networks in terms of intelligence capabilities, high availability, security, and management functions (Cisco 2010, p.7).
A managed switched Ethernet architecture has been cited as the most appropriate for an industrial network (Cisco 2010, p.8). This network is best suited to offer better traffic control, monitoring, security, and diagnostics of an industrial environment. Redundancy is a critical factor in designing industrial network topologies. The two major designs of industrial Ethernet are the ring and redundant star topology (Cisco 2010, p.11). In redundant star topology, data-layer switches already connected to higher level network switches provide connection to station-based devices such as PLC, robots and PC-nodes (Cisco 2010, p.11). In the ring topology, the nodes and layer two (data layer) switches are connected together (Cisco 2010, p.12). The layer two switches are connected to higher-level switches. Ring topology is preferred over redundant star because it offers higher availability, lower costs and ease of cabling (Cisco 2010, p.11).
There have been fears that Ethernet opens the traditionally “closed” industrial network exposing them to outside attacks. However, improvements in Ethernet have resulted in better measures to maintain availability, integrity and confidentiality of the automation and control systems (Cisco 2010, p. 9).With Ethernet it is possible to create security zones and implement security policies through VLAN configuration, firewall, access control, demilitarized zones, encryption technologies among other up-coming superior security technologies.
A critical requirement of industrial networks is that they must be highly reliable. To enhance reliability, a number of techniques can be applied at each layer of the OSI model. At the physical layer, redundancy can be enhanced by use of redundant components (power supplies, supervisors, etc) and redundant devices (switches and routers). At the data link and network layer we can have redundant links/paths and IP routing respectively (Cisco 2010, p.10).This measures will go a long way in ensuring a continuous network with little or no downtime critical for factory-floor applications (Cisco 2010, p.10). Real-time performance can be achieved by using technologies that prioritize and filter traffic such as Internet Group Management Protocol (IGMP) and QoS in addition to those that segment the network (Cisco 2010, p.15).A managed switched Ethernet will also feature intelligent switches with integrated management capabilities that ensure ease of managing such a network(Cisco 2010, p. 18).Such capabilities include DHCP option 82 and DHCP Port-Based Allocation ( only in Cisco devices)which enable quick and smooth resolution of IP-related issues to maintain network uptime.
Acceptance and adoption
Despite great improvements in Ethernet technologies tailored for industrial purposes, there are those who contend it is still lacking in terms of real-time capabilities and security. They argue that traditional fieldbus systems offer superior real-time capabilities critical to industrial environments compared to industrial Ethernet which is inherently non-deterministic and unreliable (Kuhnle 2009, p.102). They also point out that adopting Ethernet in industrial network ‘opens up’ such networks and expose them to outside attacks. Other critics argue that industrial network are best secured in their traditional “closed” form. However, industrial Ethernet has continued to gain acceptance and its growth and adoption is predicted to remain high (Cisco 2010, p.1).
Conclusion
Industrial Ethernet provides manufacturers with a cheap, improved connectivity, secure, easy to manage and upgrading option to complement their traditional network systems. Incorporating this new technology in their systems will enable them enjoy new innovative technologies already in traditional Ethernet networks. This will enable them realize improved efficiency, low costs and enhanced ability to meet the demands of their markets.
References
Cisco Systems 2010. Industrial Ethernet: a control engineer guide, Cisco Systems. Web.
Kuhnle, H 2009. Distributed manufacturing: paradigm, concepts, solutions and examples. Springer, New York.