It’s not uncommon for networks to suffer from performance problems. This is especially true in large organizations, where traffic is typically high and can often overwhelm the network infrastructure.
A way to troubleshoot such problems is to use traffic measurements. With traffic measurements, network administrators can identify which parts of a network are underperforming and where the bottleneck is. By focusing on these slow points in a network, admins can determine whether one or more components need to be added or replaced.
In this article, we will look into how and why we should measure network performance as well as some free and commercial products tools.
What is Network Performance
Network performance is measured by throughput, latency, or bandwidth. Latency has become the most often reported metric, but this blog will focus on throughput. Throughput is defined as the amount of information that can be sent over a connection at a specific time. For example, if your wireless router advertises it offers “up to 12mbps”, this means it can send 12 megabits of data every second. Bandwidth is defined as the maximum capacity for an internet connection while latency is the amount of time it takes to get a response from a remote server or device.
What does that mean?
Network performance on paper may look good but in the real world, the results may be disappointing. Here are a few examples:
- A user clicks a link on a website and it takes 5 seconds to load the page.
- A VoIP phone call has choppy audio and one-way video.
In both of these cases, the latency is too high for what was expected. In other words, it takes too long for the response to come back. To solve this problem, designers can reduce latency or increase throughput.
Why is Network Performance Important?
A slow-performing network can have a huge impact on your business. An example of how bad it can be is provided by Akamai, who reports that online shoppers abandoned their carts at a rate of 70% on websites experiencing periods of degraded performance. As you can see in the illustration below, customers will go to another website if they are not getting what they expect.
There are many other examples where poor network performance has caused significant losses for companies. Here are a few:
– Airlines lose revenue due to poor network performance. For example, when United Airlines tried to use electronic passenger boarding passes for connecting flights in 2008, the company was forced to stop doing so after 8 hours because of problems with the Gogo wireless service used by the airline. The cost to United Airlines was roughly $1M in 2008 dollars. United Airlines has also experienced problems with the Gogo service about 50 times in 2012 and 2013, which costs them roughly $50k per day.
– In 2006, Comcast lost a lawsuit that cost them $16M for network performance issues. Comcast was sued for $875M in 2011 by a new customer who said they were promised internet speeds of 6Mbps and yet he only got 0.5Mbps. Comcast won the lawsuit on appeal in 2012, but this case demonstrates that customers will take companies to court after experiencing poor network performance. Some companies have chosen to implement unlimited data plans (e.g., T-Mobile) in an attempt to avoid these problems.
Why is Network Performance Important for Telecom Professionals?
Telecom professionals need to monitor network performance because they are responsible for making sure all subscribers get the best possible experience. In other words, any time a customer calls with a complaint about their service, the telecom professional’s goal is to resolve the issue by either …
– Fixing a problem on their end or
– Recommending actions for the customer to take (e.g., install a WiFi router, update computer software or hardware)
Primary Goals of Network Performance Monitoring
The primary goals of Network Performance Monitoring are:
- Identify performance issues before they become problems: Working closely with various departments to ensure that downtime doesn’t hurt the business.
- Optimize your infrastructure for cost-effectiveness and efficiency: Identify ways to improve the current infrastructure so it delivers better results at a lower price.
- Prevent downtime and service interruptions (ex: an email server going down)
- Perform continuous service improvements (ex: reduce authentication ticket response time from 20 seconds to 2 seconds)
How to Measure Network Performance?
There are many methods for measuring network performance. For example, you can use the software on your computer, an app on your mobile device, or specialized hardware. The best way to measure network performance is to look at the average delay of packets across the session.
The internet is made up of many networks. Some of these networks are shared by many users (e.g., cellular data providers) while others are private (e.g., corporate or university networks). This means that some services will not perform well if they do not have enough bandwidth allocated to them. The following are some of the most common techniques used to measure network performance.
- One way to measure network performance is to use the software on your computer that can take measurements when you are using a service (e.g., Skype). There are several programs like this available (e.g., Ping Plotter, Wireshark). The downside of this approach is that you have to run the software every time you are about to use the service. This makes it difficult, if not impossible, to take measurements in real-time while using the service.
- An app on your mobile device can also be used to measure network performance. For example, the Ookla Speedtest app allows mobile device users to test how fast their internet is and compare it with other users in their area. This approach has the same downside as the first one: you have to run an app every time you want data.
- Large enterprise networks use entire systems called Network Management Software (NMS) to measure performance. This type of system is typically hosted on servers inside the enterprise network and can be used to monitor lots of things like security, inventory, etc. What does this have to do with measuring network performance? NMS systems usually collect data about many different factors that affect network performance (e.g., links, server response times, network configuration). The advantage of this approach is that the data is collected in real-time and does not require you to run software every time you want to measure performance.
Tools for measuring network performance
There are many tools for measuring network performance. The following are some of the most common tools. Their cost and complexity vary depending on your needs.
- ping – free and simple, but it only works for measuring the speed of your computer’s connection to another computer.
- traceroute – free and simple, but it only works for measuring the path between two computers.
- Iperf – free and more powerful, but it is a command-line tool. Also, you have to install software on your computer before you can use it.
- Speedtest – a very popular app for measuring the speed of your mobile data connection with any internet service provider (e.g., 3G or 4G).
- Netflix’s Fast.com – this is an easy-to-use web app that can be used for measuring both mobile and wired connections.
However, the best way to measure network performance is to look at the average delay of packets across a session. The following section will explain why using tools like Ping or traceroute are not good for this purpose.
Ping is one of the most basic tools for measuring network performance. It simply sends echo requests between two computers and then measures how long it takes for the echoes to be returned.
The problem with Ping is that it only measures the time it takes for a packet to travel from one computer to another. This is not good because if a packet goes through a congested link, then there may be a significant delay even though the throughput is quite high.
Traceroute works by sending packets with incrementally increasing time to live (TTL) numbers to the destination. The first packet sent should have a TTL of 1, which means that if it doesn’t make it to the destination, then one router along the path will send an error message back. The next packet sent has a TTL of 2 and so on until the destination is reached. Traceroute then measures how long it takes for each echo to come back. This is good because it shows us the path that packets take from source to destination, but it does not give any information about the average delay across a session.
What we want to measure is the time delay between when a packet enters a network and when it leaves. To do this, we use Iperf .
Iperf is just like Ping because it sends packets with time stamps. The difference is that, instead of measuring the time it takes for an echo to return, Iperf measures the time till the packet reaches its destination. If you want to learn more about Iperf, you can check out my YouTube video on Iperf . The important thing to know is that it can measure the average delay of data across an entire session.
Netflix’s Fast.com is just like having Iperf installed on your computer because it runs inside your browser.
Netflix’s Fast.com measures the effective speed of your broadband connection, which is a combination of processing time, waiting time at routers along the path from source to destination, and actual throughput. In other words, it reflects end-to-end network performance. There is no difference between its measurement and what Iperf or Fast.com measure because they are all capturing the same information.
However, there are two potential weaknesses with Fast.com. One is that it only works if you are using Netflix because it needs to send packets across your connection to determine the speed of your connection. The second weakness is that its results can be manipulated. For example, if your computer is behind a router that performs network address translation (NAT), then the router will change the source IP addresses of packets going to Netflix’s Fast.com, which means they cannot be traced back to their source.
Fortunately, these weaknesses can be mitigated by using a website like Speedtest .net. Unlike Fast.com, this website does not require any special code to be installed on your computer and it allows you to run the speed test from a variety of servers around the world.
For large enterprises, there are several tools of choice when it comes down to network performance monitoring tools. Here are a few main ones:
Solarwinds Orion Network Performance Monitor, which has a free 30-day trial
Paessler PRTG Network Monitor, a cost-effective option for small and medium businesses.
In addition to the tools listed above, there are also open-source alternatives such as OpenNMS and LibreNMS.
Advantages of measuring network performance
1. Increase awareness of baseline performance
2. Improve customer experience by identifying problems and quickly resolving them
3. Identify trends such as an increase in traffic, packets dropped, or latency that can be used to plan future upgrades or changes
4. Use baselines to monitor for anomalies and identify when things go wrong
5. Measure where the bottlenecks are, e.g., if it is your internet connection or your local network
6. Measure for specific services that you have implemented to ensure they are performing as expected
7. Increase productivity by identifying areas of improvement and testing new scenarios
8. Set SLAs based on measurable metrics instead of using subjective opinion
9. Show the value of the network to management to get a budget for upgrades or new equipment
10. Increase knowledge about networking in general by understanding performance characteristics and pitfalls
Some Challenges While Measuring Network Performance
1. Sometimes the test results are not consistent, i.e., they are too slow or sporadic to be of use
2. Using tools for testing requires specific configuration that not all organizations have
3. Some tools cannot accommodate changes in network architecture without reconfiguration
4. Some tools do not support mobile devices making them less useful in the field
5. Some tools require intermediate knowledge to set up and use correctly
6. Programmers may not be able to access some of these tools, so it is more difficult for them to test their code
7. These tools can be too complicated or redundant when all you need is a simple speed test
8. Some tools are not free for all users, e.g., only the organization or a specific department can access it
9. Inaccurate results due to network congestion
10. Slow load time when a user runs tests, making them less useful if you need quick results
11. Methods of configuring these tools vary from one to another, leading to confusion when tested between different users
12. Some tools give an average of all the tests run while others may show each test individually or in separate tables
13. Cannot be used for legacy network equipment due to lack of compatibility
14. Not enough functionality to suit large enterprises with complex networks
15. Additional performance degradation caused when sending results to a testing server
How to improve your networks’ speed and reliability with the right tools and knowledge
Network performance monitoring tools are designed to measure the reliability and speed of your network. To get the most out of them, you need to know how they work and what they can do for you.
1) Increase awareness of baseline performance: Tools like Solarwinds Orion Network Performance Monitor, Paessler PRTG Network Monitor, and LibreNMS will allow you to see exactly how well or poorly your network is performing at any given moment in time. This information tells IT, staff, when there’s a problem before users start complaining about slow connections or other issues that may not even be related to the network itself. It also gives management insight into whether their service-level agreements (SLAs) are being met by providing clear data on performance indicators.
2) Improve customer experience by identifying problems quickly and resolving them: With Solarwinds Orion, you can set up alerts to notify you when your network performance drops below a certain level for any reason. This will let you address issues before they cause serious problems – including downtime – to maintain the best possible service.
3) Create standard network metrics: Most of the tools mentioned will record different types of data to give you a clear picture of how your network is performing for both internal and external hosts. This data can be used to produce reports that show not only where performance issues are, but also help identify areas in need of improvement – e.g. which devices are slowing down the network or generating the most traffic. This type of analysis also makes it easier to predict future bottlenecks and work proactively to prevent them from occurring.
4) Configuration: To use these tools effectively, you must first understand your network configuration – how it is built, as well as where problems are likely to happen before they do. This can be done manually or through automated discovery, depending on the capabilities of your system(s).
5) Interoperable testing: Tools like Solarwinds Orion and Nagios will give you detailed information about how network performance changes as you add more devices and users into the mix. This is especially important for ISPs who want to add new customers while making sure existing ones won’t suffer because of increased competition.
6) Accessibility: Solarwinds Orion and Nagios have clearly-defined interfaces that are easy for beginners to understand and use effectively, even if they don’t have a degree in computer science. LibreNMS is designed for Linux users, so Windows-based managers may need assistance from tech-savvy employees or outside consultants.
7) Scalability: Solarwinds Orion and Nagios are both scalable; the dashboards can be shared across multiple computers to provide remote access. LibreNMS is designed for small networks with no more than 100 devices.
8) Extensibility: Solarwinds Orion is highly extensible, so third-party tools can be integrated with its dashboard to provide additional functionality. LibreNMS supports plugins written for the PHP scripting language, but they aren’t included in the software by default and must either be installed manually or through a package manager like YUM.
9) Cost: Solarwinds Orion and Nagios are free for personal use, but commercial licenses must be purchased if you plan to use them in a business setting. LibreNMS, on the other hand, is free for both personal and commercial use.
10) Support: Solarwinds Orion is backed by a large support team, and offers various levels of paid subscription to fit your needs. Nagios offers both free and premium tiers of support, but they aren’t official partnerships like the one with Solarwinds.
11) Community: Nagios is maintained by volunteers who work together to develop new features and address potential issues. There’s little outside support available for LibreNMS aside from community forums, but it is open-source software that can be modified by anyone with the know-how.
Network Performance Monitoring, or NPM for short, is the process of gathering information about your network’s health and using it to identify issues related to performance or downtime before they become serious problems. Tools like Solarwinds Orion and Nagios can provide you with hundreds of different metrics that will help you keep track of how your network is performing in real time and which devices or applications are causing problems that could lead to downtime.
As seen, network performance can make or break a company. With the right tools and methods, you can measure your system’s performance and identify any potential issues.y any potential issues.