Network Management
Keeping the Network Alive from Afar Network management is the process of documenting, monitoring, troubleshooting, and configuring network devices. Network management gives visibility to the networking staff. The routers and switches in a network have the same components as a regular PC. There are a CPU (or two), memory, storage, and network interfaces. The primary difference from a PC is that network equipment is highly optimized to perform certain functions, such as passing packets quickly. The nature of today s networks is that network equipment is distributed everywhere. Remote branches, data centers, locations around the world all these places on the network have routers, switches, servers, and other networking gear. Network administrators are responsible for the health and well-being of a company s network. Their goal generally is to provide uninterrupted network services. If a network outage or fault occurs, they want to know about it as soon as possible and isolate the problem quickly. When a problem occurs in the network, it can bring business to a grinding halt. The network administrators must find, isolate, and fix the fault as quickly as possible to restore business operations. The longer the fault persists, depending on the number of people and processes affected, the greater the monetary losses the company might experience. For example, a manufacturing line might stop operating. Or a stock-trading firm might stop trading. A hospital might lose the ability to track patients. A theme park might be unable to collect tickets. When a network fault occurs, a network administrator needs to gain visibility to the network, taking into close consideration any changes that might have occurred recently, to determine the source of the problem. Aside from dealing with faults, network management facilitates the orderly upgrade or maintenance of network devices. Remember, a company might have network devices installed all over the world, so it is impractical for a network administrator to physically travel to all the various locations to upgrade code. Several process models can be adopted to form the basis of a network management process framework, including FCAPS and ITIL. FCAPS stands for Fault, Configuration, Accounting, Performance, and Security. FCAPS represents the thinking about network operations going back to the 1980s. In short, the areas of network operations and management under FCAPS are broken down into these areas: Fault: Finding and correcting network problems. Configuration: Monitoring and controlling network devices and configuration. Accounting: Ensuring that devices are distributed appropriately and providing the ability to account (and bill) for usage. Performance: Measuring the throughput in the network and looking for potential bottlenecks. This area is especially important when determining capacity for applications and users. Security: Protecting the network from inadvertent mistakes and intentional sabotage. Another more recent network management process framework is the Information Technology Infrastructure Library (ITIL). ITIL takes a more serviceoriented viewpoint, focusing on the end services provided by the network, instead of the network itself. The key operational areas of ITIL are as follows: Service strategy identifies the IT services that could be developed as strategic assets for use by internal or external customers of a business. Service design develops a service strategy into an IT service design that addresses aspects of the proposed service, including availability, capacity, continuity, and security. Service transition focuses on implementing and creating a production service, including change management, release management, configuration management, and service knowledge management. Service operation focuses on the operations aspects to maintain services at defined service levels, including problem management and incident management. Continual service improvement is the ability to continually improve the quality and service levels of the services the IT department delivers.
148 Network Availability FCAPS and ITIL have similarities. ITIL takes a broader perspective, looking at the total lifecycle of network services. ITIL also takes a viewpoint of the services the network is providing rather than the network itself. Whether you employ FCAPS, ITIL, or another methodology to ensure network availability, management, and operations, it s important that you adopt and use a methodology that meets the company s goals. The elements that make up an effective network-management strategy are documentation, network-management protocols, and troubleshooting tools. Network Documentation: A Must-Have A relatively simple yet often overlooked process of network management is documenting how all the network devices connect to each other. Network documentation becomes indispensable during network problems. Documentation takes many forms: Logical network diagrams show how all the devices are connected. Cabling charts document where cable runs in a building. These charts are vital when you want to figure out where the other end of a cable terminates. (Is it an office or another floor?) A database correlates IP addresses and other network information with individual users and locations. Where more than a couple of people are responsible for a network, you need change-control documentation. Change control is the process of documenting any changes to the network before you make them so that a group can review the changes and ensure that one set of changes doesn t overwrite another. You can reference logs from network devices when you want to determine the source of a problem. Cisco devices send console messages and Simple Network Management Protocol (SNMP) traps (discussed next) to a remote server for this purpose. By studying the log of messages, a network administrator might be able to determine the source of a problem. Network-Management Protocols Many tools and protocols help you effectively manage network devices. These tools and protocols help you configure, back up, monitor, and measure network devices. Network-management software makes efficient use of public domain protocols to discover and manage networks. The first protocol is simply the Transmission Control Protocol/Internet Protocol (TCP/IP) ping tool. Network-management software, in its simplest form, uses ping as a heartbeat monitor. Ping sends a single request to a device, and the device is expected to send back a single response when it receives the request. Thus, if your system does not receive a response for a period of time (remember, packets might drop in the network occasionally, so ping needs to retry three or more times), you can assume that the device is unreachable for some reason. An unreachable device indicates that somewhere between the networkmanagement workstation and the destination device, packets couldn t pass. Commercial network-management software indicates this fault by coloring a graphical representation of a device red on a map of the network. (Green indicates that everything is okay.) Whereas ping provides a heartbeat, SNMP lets you get and set information between a network-management station and a managed device such as a router or switch. SNMP is a simple protocol that is similar to a database-retrieval program. The managed device maintains a database of information such as the health of its components (CPU, network interface, buffers) and throughput (how many packets are passing through an interface). Using SNMP, a network administrator can send a new configuration file to a device, upgrade its software, check
Network Management 149 its health, and measure how many packets are flowing through the device. You can store virtually anything about the device in the database in such a way that a network-management workstation can learn about or configure any aspect of the network device. SNMP also provides traps. Rather than a command-and-response, like other SNMP commands, traps are unsolicited responses. Managed devices, such as a router, send a trap when something needs attention. For example, if an interface goes down, a router sends a trap, notifying the network-management workstation. The workstation then indicates that there is a fault on the device. Depending on the vendor, you might use other protocols to manage a device. Troubleshooting Tools The final component of network management is troubleshooting tools. These tools help the network administrator isolate and correct a network problem when it occurs. These tools include cable testers, packet analyzers, and regular computers used to query information on location. Cable testers and other physical-level devices let you determine whether a cable has a physical problem. For example, a cable might be too long and out of spec, or a device nearby might be causing interference and disrupting the flow of traffic. Packet analyzers allow a network administrator to monitor traffic on part of the network. The analyzer not only captures the traffic, but also decodes the contents of each packet into human-readable form. These tools are indispensable when you want to determine the source of a problem or an application s behavior. Because packet analyzers capture all traffic, a hacker can use the tool not only to observe data as it flows through the network, but also to introduce lethal and disruptive traffic into the network. The final tool is a regular laptop computer. A network administrator can install network-management, packet-analyzer, and database software and carry all the software tools he needs to troubleshoot and correct a problem on location.
150 Network Availability At-a-Glance: Network Management Why Should I Care About Network Management? If you use a computer attached to a network at work, school, or elsewhere, chances are that the network manager knows about you and the types of programs you use. You are being watched! Although user supervision is a part of network management (and the part most users tend to worry about), there is much more to it than that. As networks get more complex and more intelligent, it becomes necessary to put tools in place to help manage the network and ensure that it is operating efficiently and effectively. What Problems Need to Be Solved? An ideal network management system is everywhere, all the time. With increased network intelligence and the addition of smart end devices, it is possible to have every point in the network accounted for and part of the overall management reporting system. This can make the task of active network management very difficult. To handle network management, tasks are usually assigned to one of the five FCAPS categories fault, configuration, accounting, performance, and security. Each category has its own methods, strategies, and protocols: Fault detection and correction. A good network management scheme will quickly find and isolate problems. Configuration. As the network grows, manual configuration of devices becomes prohibitively difficult. Configuration also includes monitoring functions. Accounting tracks usage, distribution, and billing. In many cases billing is used to justify departmental budgets. Servers account for and report usage, and track billing. Employee unknowingly leaves a backdoor on a web server. Security procedures find it and block it. Are you there? Ping 10.23.172.66 Performance measures are used to ensure that the network is operating efficiently. If any bottlenecks are found, the network manager can open alternative paths. Security measures protect the network. Most people assume that hackers are the biggest threat, but many attacks come from inside the network edge. Yes, I am. IP address replied at 10:30:12 a.m. SNMP communicates with, and configures and receives information from, the network. Everything Gets Documented Hi Paully. Paully, Hi. Hi. Paully. Are you there, Paully? Faulty devices report failures and new paths are established.
Network Management 151 At-a-Glance: Network Management Nightmare of Mobility Keeping track of thousands of devices was hard enough for network managers when everything basically stayed put. The huge surge in wireless networking has added hundreds or thousands more mobile devices, including laptops and PDAs, making network managers jobs harder. Some strategies for keeping track of mobile devices are in covered in Part VIII, Mobility.
152 Network Availability Network Management Monitor Network Management Functions Advantages of Network Management Backhoe= Line Break The network manager configures the network and monitors its realtime performance and security. Network Manager Server San Francisco NY L.A. Atlanta Dallas (H.Q.) Performance L.A. Router 3 Packets Lost 1587 Packets Sent Total Time: 3:11 Security IDS Reports Possible Spoofing Attempt on NY Router Shut Down Monitor Network Management Benefits The entire network can be designed, configured, and monitored from a single location. This reduces the need for travel. I need to get a line repair on the north/south route between NY and Atlanta.
This page intentionally left blank