Warakorn - Fotolia
An important part of any network administrator's job is to monitor the network for performance, traffic usage, faults and availability, and to respond quickly to issues. A network-monitoring tool is a software-based or a software-hardware combination that watches the network from end to end, collecting data on hundreds of performance metrics, among them bandwidth, latency, responsiveness and CPU use of hosts. A comprehensive network monitoring foundation tracks the network's behavior and issues alerts when it exceeds a performance threshold -- whether that means dipping below an acceptable level or when network traffic spikes.
Most network-monitoring products work well for typical small to midsize networks, whether wired or wireless. But more complex networks -- enterprise networks and distributed environments -- need a comprehensive platform that offers visibility into physical and virtual servers, wide area network (WAN) links, a software-defined network (SDN) architecture, cloud-based services, network-based applications and the increasing number of mobile devices that connect to the network.
How network monitoring works
In addition to keeping tabs on the overall health of a network, careful network monitoring can facilitate proactive strategies, such as justifying the cost of hardware or infrastructure upgrades required to eliminate chronic network bottlenecks. This capability can also save money over the long term by enabling an enterprise to reduce repetitive troubleshooting issues.
Organizations that rely on a service provider to oversee portions of their network connectivity can use network monitoring to enforce service-level agreements (SLAs), which state the level of service a provider agrees to meet. Without clear insight into how a network is truly performing, customer enforcement of an SLA is difficult to impossible.
Finally, a properly functioning network provides higher user satisfaction that, in turn, reduces calls to the helpdesk and subsequent follow-ups.
All kinds of built-in and open source network-monitoring tools are freely available. The Ping utility, for example, is adequate for diagnosing simple host connectivity issues. Microsoft Network Monitor provides network packet capture and analysis, which a network admin might use to troubleshoot network problems. Nagios is a popular open source network-monitoring tool; it isn't easy to set up but it provides accurate data on many metrics. However, free tools also tend to be narrow in focus or require add-ons or plug-ins for expected or additional functionality, such as auto alerts.
The free or patchwork approach may be affordable, but it's often inefficient. Businesses need a robust network-monitoring product to ensure uptime, uninterrupted access and ease of use.
Networking monitoring features
A comprehensive network-monitoring product should be able to detect, monitor and analyze a network and its devices in real time, enabling an admin to respond based on warnings and alerts. At a minimum, your tool should be at least moderately easy to implement and configure, and it should support multiple vendor devices. The product you choose -- at minimum -- should include autodiscovery, node and device inventorying, automatic and configurable trouble alerts and warnings, and they all should be tied together through a Web-based centralized management interface. The interface should include a dashboard with easy-to-read graphs and tables that provide different views of network status; it should also have a network topology map (or the ability to generate one), as well as commands for modifying network settings and troubleshooting issues. Some interfaces provide viewing and reporting only.
With the transition to IPv6 well underway, your choice of network-monitoring tools must also be able to detect and analyze Internet Protocol version 4 (IPv4) as well as IPv6 traffic and protocols.
Another feature to consider is whether to pursue an agent-based or agentless monitoring approach. A network-monitoring tool typically uses agents or sensors that collect and transmit data to a management console for analysis. Agents require some resources to run, which can affect performance. An agentless product, on the other hand, is designed to have little to no impact on current processes.
Finally, many but not all network-monitoring approaches are application-aware, which means they can detect and monitor all applications and services operating across a network. This is an important benefit, because it enables administrators to understand if a performance problem stems from either the network or the application itself. An application-aware network monitoring tool lets IT staff track an application's response time in detail, including server processing, network request and network response.
Nice-to-have network monitoring features
The ability to spot trends -- both historical and what's expected -- can be a highly useful resource for technical staff. Trend data lets an admin look back at network performance data, whereas trend predictions provide a glimpse into the most likely future of network performance, based on past behavior.
Automatic capacity planning is a handy feature but not needed by every type of environment. It facilitates network-infrastructure management and optimization by sending an alert to a network administrator whenever available memory, bandwidth or network capacity is about to be exceeded.
The bottom line
Look for a comprehensive network-monitoring product that meets your needs today and provides scalability for future network expansion. You have plenty of choices, both open source and commercial. Open source is freely available and accessible, but the commercial versions tend to provide all required features in an integrated package.
Arm yourself with app performance monitoring tools
Future of monitoring makes the networking engineer role a critical one