How to Measure Network Performance: 5 Network Metrics
As more companies continue to rely on SaaS and cloud applications to run their businesses, it becomes important for them to ensure their network infrastructures can withstand the demand, and that they’re able to offer their services quickly and reliably. Continuous network performance monitoring can help you ensure that your network is always performing at its highest level. So, we’re running you through exactly how to measure network performance, and what network metrics you should be looking at.
With network demands increasing, it’s important for monitoring to be continuous and detailed. When it comes to how to measure network performance, there are several different metrics that you can analyze.
Using a network performance monitoring software, your business can easily track these metrics in real-time, and understand how they’re hurting your network’s performance. Not only will you be able to quickly find network problems, but you’ll also be able to collect historical data, and compare good performance, with below par.
Let’s get into it!
Just so we’re on the same page, we’re going to give you a quick definition of network performance, at least from our point of view.
A more technical definition of network performance is "the analysis and review of collective network statistics, to define the quality of services offered by the underlying computer network [that is] primarily measured from an end-user perspective."
More simply, network performance refers to measures of service quality of a network as seen by the customer.
There are three important concepts in that definition:
Measuring Performance: Network performance is something that can be measured. It’s important to measure good performance versus bad performance using network metrics.
Network Quality: Performance refers to the quality of the network connection. Is the connection stable and fast? Or is it slow and constantly cutting out?
The Customer: Performance is always experienced by the user of the network, not just by the tools measuring performance. This is why it’s important to have a network performance monitoring solution that can measure end-to-end network performance from the user’s perspective.
When using a network performance monitoring software, like Obkio, users can install network monitoring Agents at different points within the network architecture to perform a complete network assessment of network performance and health. In some cases the Agent is installed next to the firewall to monitor ISP performance (WAN) and in other cases at the far-end of the LAN network to have a complete end-to-end performance monitoring solution.
Essentially, a network’s performance is its ability to meet the user’s expectations. Which means that network performance is user-dependent.
Some networks have small requirements while others need to perform at high levels to sustain multiple systems or services. A well-performing network is one where the actual performance exceeds usage, in the sense that the performance should always be greater than the need in order to avoid congestion or a struggle in the network
Check out our blog post on Fault Monitoring vs Performance Monitoring to make sure you understand the differences between the two!
In this article, we’re going to give you a rundown of the top network metrics. Stay tuned for more detailed articles about the cause, impacts, and troubleshooting process of each metric. Follow us on LinkedIn, Facebook, Twitter to find out when these posts arrive!
Now, let’s get how to measure network performance!
When it comes to how to measure network performance, it’s important to know which network metrics you need to examine. Depending on the specific issues that affect your network, not every metric is going to be important for you to look at. But there are some metrics that are essential for any businesses to consider.
Let’s go through the most essential network metrics that your company should absolutely be monitoring.
In a network, latency refers to the measure of time it takes for data to reach its destination across a network. It is usually measured as a round trip delay, in milliseconds (ms), taking into account the time it takes for the data to get to its destination and then back again to its source.
Measuring the round trip delay for latency is important when knowing how to measure network performance because a computer that uses a TCP/IP network sends a limited amount of data to its destination and then waits for an acknowledgement that the data has reached its destination before sending any more. Therefore, this round trip delay has a big impact on network performance.
Consistent delays or odd spikes in delay time are signs of a major performance issue that can happen for a variety of reasons. Most delays are actually undetectable from a user’s perspective and can therefore go unnoticed but can have a huge impact when using VoIP, or unified communication systems suchs as Zoom, Skype, Microsoft Teams and so on.
A network performance monitoring (NPM) solution is pivotal for tracking latency because it can track and log these delays, to find the source of the problem.
Latency also affects maximum throughput of a data transmission, so how much data can be transmitted from point A to point B in a given time. We’ll be covering throughput in point 4.
But the reason that latency affects throughput is because of TCP (Transmission Control Protocol). TCP makes sure all data packets reach their destination successfully and in the right order. It also requires that only a certain amount of data is transmitted before waiting for an acknowledgement.
A common analogy of the relationship is to imagine a network path like pipe filling a bucket with water. TCP requires that once the bucket is full, the sender has to wait for an acknowledgement to come back along the pipe before any more water can be sent.
If it takes half a second for water to get down the pipe, and another half a second for the acknowledgement to come back, this equals a latency of 1 second. Therefore, TCP would prevent you from sending any more than the amount of data, or water in this example, that can travel in any one second period.
Essentially, latency can affect throughput, which is why it’s so important to monitor.
To put it not so lightly, jitter is your network transmission’s biggest enemy when using real-time apps such as unified communications, including IP telephony, video conferencing, and virtual desktop infrastructure. Simply put, jitter is a variation in delay. Otherwise known as a disruption that occurs while data packets travel across the network.
There are many factors that can cause jitter, and many of these factors are the same as those that cause delay. One difficult thing about jitter is that it doesn’t affect all network traffic in the same way.
Jitter can be caused by network congestion. Network congestion occurs when network devices are unable to send the equivalent amount of traffic they receive, so their packet buffer fills up and they start dropping packets. If there is no disturbance on the network at an endpoint, every packet arrives. However, if the endpoint buffer becomes full, packets arrive later and later.
If you’ve ever been talking to someone on a video call or other unified communication system, and suddenly their voice speeds up significantly, then slows down to catch up, or keeps fluctuating between the two - you have a jitter problem. Jitter can also be caused by the type of connection you use. A connection on a shared medium, such as a cable, is more likely to have higher jitter than a dedicated connection. So that’s something to keep in mind when choosing a connection medium.
Looking for more info? Check out our blog post on the top 3 network monitoring use cases to find out when you should be monitoring network performance!
Packet loss refers to the number of data packets that were successfully sent out from one point in a network, but were dropped during data transmission and never reached their destination.
It’s important for your IT team to know how many packets are being dropped across your network to be able to take steps to ensure that data can be transmitted as it should be. This provides a metric for determining good or poor network performance.
A network performance monitoring software, like Obkio, uses a tactic which involves generating and measuring synthetic traffic in order to count the number of packets sent and the number of packets received.
Packet loss is usually expressed as a percentage of the total number of sent packets. Often, more than 3% packet loss implies that the network is performing below optimal levels, but even just 1% packet loss might be enough to affect VoIP quality.
Packet loss is something that is determined over a period of time. If you record 1% packet loss over 10 minutes, it can suggest that you have 1% during the whole 10 minutes, but it can also be that you have 10% packet loss over 1 min and then 0% over the remaining 9 minutes. How can you figure that out? Well that’s why Obkio calculates packet loss every minute, so you always get an up-to-date and precise measure of packet loss.
Throughput refers to the amount of data passing through the network from point A to point B in a determined amount of time. When referring to communication networks, throughput is the rate of data that was successfully delivered over a communication channel.
Throughput is usually measured in bits per second (bit/s or bps),
"Internet Connection Speed" or "Internet Connection Bandwidth" is a general term used by internet companies to sell you high-speed internet, but is used by default to mean throughput, which is the actual rate of packet delivery over a specific medium.
That’s why the best way to measure throughput is using Speed Tests.
A Speed Test is the best solution for giving you an idea of how fast your internet connection is right now. Essentially, a speed test measures speed by sending out the most amount of information possible throughout your network, and monitoring how long it takes to get delivered to its destination.
A network performance monitoring solution like Obkio allows you to run speed tests manually, or schedule speed tests between monitoring Agents, or a group of multiple Agents to ensure your speed or throughput is constantly being monitored.
Obkio also allows you to perform speed tests with multiple TCPs at the same time, which makes for the most accurate speed test results.
The last two network metrics are fairly self-explanatory, just by their names. Both have very noticeable and substantial effects on real-time applications and unified communication systems like Zoom, Teams, Skype etc. because they make it very difficult for people to speak to each other in a clear and coherent manner.
In a simplified way, packet duplication refers to when data packets are duplicated somewhere in the network, and are then received twice at their destination. Many times, if the source of the data believes that a data packet was not transmitted correctly because of a packet loss, it may retransmit that packet. The destination source may have already gotten the first packet, and will receive a second duplicate packet.
Once again, in the example of a video chat, packet duplication may cause you to hear as though someone is repeating words or sentences as they’re speaking to you - which isn’t a very pleasant experience.
Packet reordering is also pretty self explanatory and occurs when data packets arrive at their destination in the wrong order. This can happen for various reasons, such as multipath routing, route fluttering, and wrong QoS queue configuration.
Packet reordering is also very simple to spot. If you’re talking to someone over video call and all of a sudden words in their sentences sound scrabbled and out of order, it may be because the data arrived in the wrong sequence.
Once again, a network performance monitoring solution will be able to catch these problems, right as they happen. Having continuous monitoring of your network, whether from your head office, data center, or home office, means that you’ll catch these network issues long before you’re on an important video call with a client who can’t understand a word you’re saying because of packet loss or packet reordering.
If you want to learn more about why you should be monitoring your network, you can check out our blog post on the Top 7 Reasons Why You Should Monitor Network Performance.
Now you may be wondering how all these metrics could possibly play a part in how to measure network performance. All the metrics we mentioned, in addition to user requirements and user perceptions, play a role in determining the perceived performance of your network.
Each metric on its own gives you an idea of how your infrastructure is performing, but you need to look at all of these factors to give a true measurement of network performance.
The best way to measure and quantify user experience is by measuring User Quality of Experience (QoE). Quality of Experience (QoE) allows you to measure performance from the end-user perspective and is essentially the perception of the user of the effectiveness and quality of the system or service. In fact, users base their opinions about the network exclusively on their perception of QoE.
Measuring QoE is a culmination of all these network metrics we discussed, as well as the ability of the network to meet the user’s expectations. That’s basically what network performance is all about.
Other metrics you can use to measure QOE include:
The MOS score was created by the ITU, a United Nations agency that sought to facilitate international connectivity in communications networks, and created a metric that could be measured and understood by all.
The MOS has been originally developed for traditional voice calls but has been adapted to Voice over IP (VoIP) in the ITU-T PESQ P.862. The standard defines how to calculate a MOS score for VoIP calls based on multiple factors such as the specific codec used for the VoIP call.
The MOS score is a rating from 1 to 5 of the perceived quality of a voice call, 1 being the lowest score and 5 the highest for excellent quality.
You can learn more about MOS score in our article on Measuring VoIP Call Quality with MOS.
VoIP quality refers to the quality of voice communications and multimedia sessions over Internet Protocol (IP) networks, such as the Internet
Obkio’s network performance monitoring software calculates the VoIP Quality for each network performance monitoring session every minute. Obkio measures VoIP Quality with MOS score even if there is no ongoing call, to provide a proactive monitoring over packet capture solution.
Don't wait for bad user experience complaints to start troubleshooting! This Quality of Experience (QoE) metric helps IT pros understand the complex impact of network performance on VoIP.
Nowadays, the best performance monitoring tools can synthesize a rather precise evaluation of the QoE score based on their measurement of all the previously discussed performance-affecting metrics!
And that brings us to…
As you can see, there are a lot of factors to keep in mind when measuring performance, and all them need to be monitored simultaneously to really provide a concrete conclusion.
Lucky for you, the number one key to learning how to measure network performance is finding a solution that measures throughput, latency, packet loss, jitter, and more, to give you a simple and quick overview of your network.
As soon as a problem occurs, with anyone of the metrics being measured, you’ll be notified - even before it reaches the end user.
This isn’t just a shameless plug, really, it’s a way to make your life easier, make your business run smoother, and help you ensure that your network is always performing at its highest level. It'll also be a key tool in helping you identify network performance issues. Your users and employees will thank you.
Try the solution for yourself and see what it tells you about the way your network is performing.
Get a free trial of Obkio’s Network Performance Monitoring Solution by signing up here. You can also learn more about our products below:
- The Network Performance Monitoring Agent
- Network Device Monitoring
- Network Performance Monitoring
- Remote Network Monitoring