Implementing 802.11ac/ax Does Not Guarantee High Throughput

 
After upgrading a Wi-Fi network to new technologies that boast speeds in excess of 1Gpbs, network engineers are often frustrated to find that the performance of network applications and services still lags. After all, if the data rate promises 10x the speed, shouldn’t we see a similar improvement in end user experience? Well, not necessarily. By now, many of us have experienced that especially in the Wi-Fi space, 1Gbps does not really mean 1Gbps. There are many variables that go into delivering data over radio waves, and as network engineers we are tasked with delivering the best performance possible.

When testing and troubleshooting a Wi-Fi network, it is critical that we understand the difference between PHY data rate, bandwidth, and true throughput – as these values can vary dramatically depending on the environment. These measurements can also help to determine if a given area is performing as it should, or if there is room for improvement.

Wait – PHY data rate, bandwidth, and throughput – aren’t those basically the same thing? Not exactly…

Let’s define these metrics first, then look at how to measure them.

PHY Data Rate is the maximum rate at which the physical layer can transmit. It represents the rate that data is transferred over the channel, which includes protocol headers, control and management frames. Usually, this is the number advertised by the access point or controller.

Bandwidth is often used synonymously with the data rate. It usually represents the maximum rate at which data can be transmitted but does not include any physical or data link layer overhead.

Throughput is the actual rate achieved by the data in flight. Only the data frames are taken into consideration when measuring throughput, this measurement does not include control frames, management frames, retries, or any other protocol overhead.

To troubleshoot a performance problem, or to test the rate of transfer in each area, it is best to focus on actual throughput measurements. This is commonly done using tools like iPerf, which can create data streams to measure the throughput between two devices. It is not uncommon for the throughput to be less than 50% of the PHY data rate. If it falls to a much lower value, we can start to focus on other metrics such as noise, channel interference, retries and utilization to determine why it is low.

Example: Application Comparison/Baselining

 There are many different terms I use to cover ‘Application Baselines”, such as ‘Profile’, ‘snapshot’ and many other synonyms.

The Most Common Reasons for Wi-Fi Roaming Problems

You are standing in one area of the building, wirelessly connected, and application performance works great. But as you walk to another area, you notice your radio icon loses a signal strength line or two and things start to lag, or even drop completely.  

Everything Else (Paul Smith)

 

Multi-Tasking is a fantasy. There are plenty of high-energy people out there who say they can Multi-Task, and many a job-seeker has laid claim to this skill in an interview. The truth is, they are lying. Humans can only do one thing at a time.

Anatomy of a Connection

When I analyze a trace file with a client, I always look for the beginning of the TCP connection or application.

The reason is quite simple, there are many items that you can only see at the beginning of the capture. Sure you can make assumptions if you don’t have the beginning captured, but life is so much easier if you had it.