IP networks are dynamic, complex and expanding.

Efficient, proactive visibility is absolutely necessary when it comes to troubleshooting performance issues.

Pinpointing the issues that degrade application performance, such as latency, jitter, loss, changes in QoS or alterations in bandwidth require real-time active performance metrics. As noted in  Network Troubleshooting in Today’s Distributed Data Center, network performance fires that are not immediately found and put out can infect multiple IP network services.

Many organizations utilize Simple Network Management Protocol (SNMP) to look at individual device and terminal performance.  SNMP provides statistics on specific devices or elements on the network. You gather details about every component of the network, but not how those components are working together to meet the performance requirements.

In some ways, data gathered using SNMP is like a detailed car inspection. You will see how every component of the car is functioning, from the fuel pumps to the brake pads. But, once the car gets on the road, you have no true understanding of the speed of the car or how long it is going to take to get to your destination.

Unfortunately, SNMP only provides partial data on what you need to know to predictably and consistently manage your IP Network. SNMP tools require devices to be owned and configured, as well as a server to run the SNMP software and receive the data. These tools give important device-based metrics such as CPU, utilization, free memory and temperature readings;  the statistics will report how many bytes a given interface received. These tools do not offer  comparison data to see actual achievable bandwidth or end-to-end performance. This can provide a false sense of security– without real-time statistics detailing the available, utilized and total capacity of the network, it is very difficult to actually measure and understand your  network and application performance.

With SNMP, devices are only polled every 5-15 minutes, often  missing critical performance details. The overall health of the network may appear “fine” since transient problems such as 30 seconds of latency or a major drop in bandwidth will be visible on an average network health report. When SNMP tools detect an issue, it may have occurred 10 minutes ago and network engineers have to back track to troubleshoot.

SNMP is only the first step to monitoring network performance by ensuring device health.

Beyond this, there are other critical areas that must be monitored and managed to assure overall performance and to quickly pinpoint problems between the devices. Every network engineer needs end-to-end visibility into the paths between those devices, especially over the WAN or third-party network which is not owned or controlled by the organization.

To ensure network performance, network engineers need complete, unbroken visibility through every hop along a path from the source of the application traffic to its destination, even through segments that pass through service provider and carrier networks. Performance characteristics such as bandwidth, jitter, loss and latency, are critical to complete the full picture of network health.

Want to learn more about why you need visibility beyond your device health, and into the networks you may or may not own? Visit our website or start your own free network assessment and monitoring trial!