Home > Articles

  • Print
  • + Share This
From the author of

The Role of Network Management

Network management is the means by which a network is operated and maintained throughout its lifecycle. This includes day-to-day maintenance and operation as well as large reconfigurations. Figure 1 illustrates the typical network management approach used in large service providers and some enterprises.

Three network devices are shown (a tiny fraction of a real network): an IP router, a PBX, and a copier/scanner. We assume in this case that the PBX and copier devices don't host element management system (EMS) entities—this capability is provided by the external EMS application. Why might devices exist with no management infrastructure? Often, this is the case for legacy devices or for very simple devices. The devices may date from a time when management wasn't such an issue, or there simply might not be adequate resources on the device.

Figure 1Figure 1 Network management hierarchy.

In Figure 1, the IP router does have an embedded EMS, so it can be directly accessed by the network management system (NMS).

In passing, why might you need to manage a copier? Surprisingly, copiers are now advanced network components in their own right, with hard disks, network interfaces, and so on. There have been reports of individuals using copiers for illegal file sharing over the Internet, so these devices have to be carefully and securely managed. This points to another important networking concept: Never assume it's safe to network any device and then forget about it!

The level of network management abstraction increases as we move up the chain in Figure 1 from the device level (routers, switches, hubs) to the EMS, through the NMS, and on to the operational support system (OSS). Each layer in this hierarchy has a unique network management role.

  • The focus of a device typically has little regard for the overall network; for example, an IP router has its configuration tables and protocols set up and it then proceeds to route incoming packets to its neighbors. Each such device adds to the overall intelligence in the network. It's axiomatic to say that the network IQ is the sum of the IQ of all devices in the network.

  • As we've seen, an EMS may reside on a device or it may exist as a standalone application. In either case, the EMS tends to focus on one or more devices rather than the overall network. The EMS can be used to view and update the configuration of one or more devices. Typically, the EMS implements the management interface, composed of at least the CLI, one or more SNMP agents, and possibly other network management technologies—based, for example, on XML, HTML, etc.

  • An NMS is arbitrarily complex, generally resides on one or more external servers, and may support remote clients. It usually provides a networkwide perspective, allowing the creation of entities that involve numerous devices. Such multi-device objects are often called services and include layer 2/3 VPNs and virtual circuits such as MPLS LSPs, ATM SPVCs, etc. The NMS uses the EMS to execute required operations on network devices. This division of labor between the NMS and EMS layers provides a useful simplification to the NMS software. The NMS asks the EMS to do task X and the EMS goes away, does X, and tells the NMS when it's finished. Another important area often handled by the NMS is network device software upgrade. In this area, the operator might decide to select a number of devices for an automatic firmware upgrade using FTP.

  • The OSS provides support for operational issues such as billing, processing customer orders, creating and managing workflows, trouble ticketing, scheduling truck rolls, managing inventory, and so on. In other words, the OSS provides a more business-centric way of looking at the overall network management function. As with the NMS-EMS case, the OSS can leverage the NMS features. The OSS does this using "calls" into software interfaces in the NMS (for instance, using CORBA, as shown in Figure 1) when it needs to do things such as retrieve fault data, create or update services, and so forth. (For more details on the OSS, NMS, and EMS, see [1] in the References section at the end of this article.)

Each layer of the hierarchy provides a unique view of the network management function. As we've now seen, each layer can leverage the services offered by the layer immediately below it.

  • + Share This
  • 🔖 Save To Your Account