What is a Network KPI?
A network key performance indicator (KPI) is a measurement and a benchmark to achieve optimal network performance goals. To support these goals, measuring actual performance against the KPI goals helps the network team make decisions to improve and sustain network performance and service levels and meet the KPI objective. These decisions include:
What is the Difference Between a KPI and a Metric?
Many people use the terms KPI and metrics interchangeably, as if they are the exact same thing. Sometimes, they are.
Think of KPIs as a special kind of metric. As a measurement, a KPI is a metric. In fact, all KPIs are metrics. But metrics are a broader term. Metrics are quantitative measurements and some metrics track things that quite frankly aren't all that important.
Smart networking professionals identify the most important metrics and treat them as KPIs, creating goals and objectives and measuring actual performance against their set bars. What distinguishes a KPI from a run-of-the-mill metric is the word key. This is how an ordinary metric gets elevated into a KPI – by linking the metric to key objectives and goals.
Common Network Metrics and Key Performance Indicators (KPIs)
Each enterprise has its own idea as to what to measure and what goals to set. That means each organization will have their own set of KPIs. At the same time, they will continue to collect other metrics. But these are not tied to objectives and they are not treated in the same elevated way as a KPI.
For many, Service Level Agreements (SLAs) are not just goals, but requirements. These SLAs can themselves be a KPI or can be based on several separate KPIs. Quality of Service (QoS), application performance and end user experience (sometimes called application experience) are similarly often considered as KPIs.
Here are 16 of the most commonly used network metrics and KPIs:
- Application Layer Latency: This metric tracks application latency by monitoring and measuring application layer request/response pairs.
- Bandwidth Usage: This is perhaps the most easily understood metric, one that non-IT pros regularly grasp. “Monitoring bandwidth is one of the main reasons to perform network management, as it shows which applications, traffic types and network segments gobble the most bandwidth. With this knowledge, IT can make sure there is enough availability for business-critical applications and services—with as little latency as possible,” the Optimize Bandwidth for Fast, Smooth Traffic Flows blog explained.
- Connection Time: The amount of time needed to create a connection between two devices. This metric can indicate problems with server response, connection issues or other network problems.
- Connectivity: It’s important to check the connection between devices and nodes on your network. If there’s a malfunctioning device or improper connection on a network, various services can face downtime or performance issues. Connectivity issues commonly occur due to malware that targets specific nodes and can impact the performance of a specific area of a network.
- CPU Utilization: The percentage of CPU utilization for a specific component within a device, such as a server.
- Error Rates: This is as it sounds: the amount of errors that happen as data is transmitted, which can occur from signal interference or collisions. A high error rate can stem from faulty hardware or other network issues.
- Network Availability: Network availability measures your network's uptime over set periods of time and shows how well the network is supporting key services.
- Network Jitter: Jitter occurs when network latency becomes variable and inconsistent. End users complain about it when video gets jittery or voices sound glitchy, though jitter can degrade all areas of network performance.
- Network Latency: Network delay, often called network latency, is a core metric/KPI that tracks network performance by tracking the time it takes for data transfer from one place to another.
- Packet Loss: This is the percentage of packets that don’t make it to their intended destination. Packet loss creates network latency and reduces network performance.
- Quality of Service (QoS) Metrics: These metrics include a number of discrete measurements such as priority levels and bandwidth, which help define the quality of specific services such as VoIP.
- Response Time: This is simply how fast a network request reaches and gets a response from the targeted device.
- Round-trip time: The amount of time in milliseconds it takes for a ping request to be returned from the remote device.
- Server response time: How much time it takes for a sever to receive, complete and respond to a request.
- Throughput: Throughput is the rate of data transmission across the network. More specifically, throughput is how many data packets actually reach a network destination. Low throughput could result from packet loss, which network managers should attend to.
- Uptime: Tone of the most basic measures, uptime tracks what network resources are up, which are down and which may have subpar performance.
View All of The ABCs of Infrastructure Monitoring
Looking to start on the basics of IT infrastructure monitoring? Our alphabetized index is an excellent place to begin or extend your education. View all of our current topics.