Full Text: Download PDF
With the proliferation of Ethernet-based networks in industrial settings, choosing the right infrastructure is vital for successful network management. Alongside simple switches, gateways, and routers, managed switches have emerged, offering integrated functions that make network configuration and maintenance easier to accomplish. Greg examines the features of a managed switch platform and presents the concept of a lean managed switch.
A managed switch is a powerful tool that gives users control over an industrial network. By implementing a managed switch, users can remotely monitor and administer networks as well as switch functions.
Today’s managed switches offer a variety of abilities. While some complicated networks might require higher-end capabilities, traditional managed switches are very expensive. Many smaller networks require more control than an unmanaged switch provides but can get by with just a few key management functions. Some of the most frequently requested management functions include:
- Bootstrap Protocol (BootP)
- Simple Network Management Protocol (SNMP)
- Rapid Spanning Tree Protocol (RSTP)
- Internet Group Management Protocol (IGMP) snooping and query functions, which are especially important for EtherNet/IP applications
A closer look at these features demonstrates how managed switches provide network configuration capability, remote monitoring and diagnostics, and integral IT-compatible network redundancy.
What is BootP?
BootP allows a network device to obtain an IP address over the network. Each device on the network has a unique MAC address, a six-octet ID number assigned by the manufacturer (for example, 00:A0:45:08:CD:8D). When this device is added to a network, it broadcasts a request for an IP address. A BootP server on the network sees the request and sends a BootP reply, assigning an IP address (for example, 192.168.1.10) to the device. This makes the device accessible to higher-level network communications using that IP address.
For Industrial Ethernet, MAC addresses serve as the foundation for networking to establish communication and direct data traffic. This is the Layer 2 level of communications in the Open System Interconnection (OSI) model. IP addresses are assigned to devices and switches to support the higher-layer protocols used to produce complex, functioning networks. Once a managed switch has an assigned IP address, users can easily access, configure, and monitor it via a standard Web browser. In addition, the switch can then respond to standard networking diagnostic tests, such as pinging. A switch without an IP address cannot provide this simple yet powerful network diagnostic capability.
What is SNMP and why does it matter?
Network management systems use SNMP to monitor devices on the network for conditions that warrant administrative action. SNMP managed devices describe configuration and management information in the form of variables called Object Identifiers (OIDs). Management applications can query and sometimes set OID variables.
OIDs are natively arranged in a numerical hierarchy, such as 1.3.2.11.11.4. A management information base translates the numeric OIDs into a more human-friendly format, such as SysName. A software component called an agent runs on the managed system and reports these variables to the managing system via SNMP through IP. Widely used and multivendor supported, SNMP is the de facto standard and most popular protocol for managing diverse networks.
SNMP management capability is useful and important for all Ethernet networks, including EtherNet/IP applications.
ODVA EtherNet/IP infrastructure guidelines require switches with both Web- and SNMP-accessible port status and diagnostics functions for large-scale control enterprise or networking. These functions are also recommended for all general-use applications.
Why should I care about RSTP?
Unless an Ethernet network has a method for providing redundancy, loops can lead to network failure. Multiple active paths or loops in topology between network devices create several problems. First, the MAC address table the switch uses can fail because the same MAC addresses are seen on multiple ports. Second, broadcast packets forwarded in an endless loop between switches can result in a broadcast storm. Broadcast storms can consume all available CPU resources and bandwidth, overwhelm network devices, and cause those devices to fail, requiring a reboot to recover network operation.
RSTP is an OSI Layer 2 protocol defined in the IEEE Standard 802.1D. As its name suggests, RSTP creates a spanning tree within a mesh network containing connected Ethernet switches. RSTP disables the links that are not part of that tree, leaving a single active path between any two network devices. The protocol allows a network design to include spare (redundant) links, providing automatic backup paths if an active link fails without risking the danger of loops or requiring backup links to be manually enabled/disabled.
More vendors support RSTP than any other redundancy method. A switch with RSTP can integrate with existing Ethernet systems and IT practices. Unlike proprietary redundancy mechanisms, RSTP’s development as an open standard allows users to integrate RSTP supporting switches from multiple vendors into a single redundancy system. The protocol allows flexible redundancy for any topology: ring, tree, mesh, or a combination of topologies.
What’s the big deal about EtherNet/IP and IGMP snooping?
EtherNet/IP is a multivendor Industrial Ethernet technology managed by ODVA. The EtherNet/IP infrastructure guidelines require switches with IGMP snooping and IGMP query functions in all EtherNet/IP applications for general use or large-scale control enterprise and networking.
Under the ODVA standard, a very small, isolated system with a low device count can use an unmanaged switch. However, the standard also clearly specifies that designers must either precalculate the total multicast traffic to which each unmanaged switch will be exposed or test the configuration in advance.
EtherNet/IP devices can generate a great deal of multicast traffic. A multicast packet is a message addressed to a group of nodes. It is necessary to limit which end devices receive the traffic to avoid overloading them and causing them to fail.
When a switch without IGMP snooping receives multicast messages, it floods all ports, potentially overloading end devices and other network switches. A switch with IGMP snooping, however, forwards multicast messages to only the devices that request the traffic.
When an EtherNet/IP device wants to consume multicast data, it will transmit an IGMP join message. All IGMP snooping switches receive these join messages. The switch then snoops on the join messages as they pass to determine which ports will receive the multicast data. This restricts the multicast data to only the ports and connected end devices that expect and can handle the traffic, as shown in Figure 1.
EtherNet/IP requires IGMP query support on at least one switch or router in the network. For IGMP snooping to work properly, the network must have at least one switch or router that supports IGMP query. Periodically, this device will query the end devices in the network regarding which multicasts they wish to receive. The end devices send an IGMP join report, which updates the multicast/port associations.
When multiple IGMP queriers are in the network, the IGMP querier with the lowest IP address acts as the network querier. If the original device fails or is removed, the device with the next lowest IP address becomes the IGMP querier.
If the system does not have a switch or router that supports IGMP queries, multicast traffic problems are likely. For example, IGMP snooping switches can act erroneously by forgetting the learned multicast/port associations and then flooding all ports and devices or neglecting to forward multicast traffic to any devices at all, including those that should receive it.
A leaner switch
Many of today’s networks require capabilities beyond what an unmanaged switch can provide but do not use managed switches with full functionality because of their high costs. Newer switches such as Phoenix Contact’s Lean Managed Switch (Figure 2) provide commonly requested management functions at a lower price.
The Lean Managed Switch targets the specific features described earlier – SNMP, RSTP, IGMP, and Web-based management capability. These abilities allow industrial network managers to add resources, monitor and diagnose networks, and provide redundancy as needed. IES
Phoenix Contact
717-944-1300
info@phoenixcon.com
www.phoenixcontact.com
