Determining the Cause of a High Retry Percentage |

Determining the Cause of a High Retry Percentage

Advances in Wi-Fi technology have made Wi-Fi the preferred access method for everything from social media to business-critical applications. This has raised expectations for high performance to new levels, and any negative impact needs to be resolved immediately. Retry rates are the most reliable indicator of performance issues, and quickly diagnosing high retry rates and isolating them to a root cause is an essential activity for maintaining high performance of a Wi-Fi network.


Wi-Fi has come a long way over the years. From devices to access points (APs) to antennas to management software, there are numerous aspects of the technology that are just plain better than they were before. It is not even a trade-off. We pay less, we get more and it works...usually.

The advances in Wi-Fi have led to one unfortunate side effect for network engineers: people expect things to work. No longer is a shaky wireless connection greeted with a shake of the head and an eye roll. Today, a trouble ticket gets sent whence the Wi-Fi came and users wonder how it is that their home Wi-Fi can work better.

In some ways, it is the normal maturity of technology. Few people purchase an automobile -- a mature technology, I think we all can agree -- if the car can’t be reliably maintained. Fast speeds, heated seats and pinpoint handling don’t matter if you can’t tell when the oil needs changing. Maintenance is a requirement, not a feature.

With so many links in the Wi-Fi chain getting better -- the devices, the APs, etc. -- it has become a situation where the trouble spots have moved to the air. Engineers need to know when that most sensitive part of the network is troubled, and why.

The good news is that the most reliable symptom of trouble in the air is known: retry percentages. In Wi-Fi, a retry is a retransmitted frame. (“Frame” refers to what is commonly called a “packet”. In Wi-Fi, the term “frame” is preferred because “packet” implies that an IP header is present. Many Wi-Fi frames do not include IP headers.) A retry occurs following a collision.

The presence of high percentages of collisions -- identified in Wi-Fi frames as retries — is bad news for Wi-Fi performance. A collision means that a Wi-Fi frame was unsuccessful and unsuccessful Wi-Fi frames waste channel time. Collisions also increase the Wi-Fi contention window, which is another waste of channel time. The loss of channel time due to failed frames and larger contention windows means less available channel time for real, useful data. That's bad for Wi-Fi performance.

Identifying whether retries are happening is relatively simple, especially when using a tool like NETSCOUT AirMagnet WiFi Analyzer PRO. Whether looking big or small, AirMagnet has a screen for it. Want statistics on a whole group of APs and stations? AirMagnet’s Channel screen shows retry counts and percentages for every device that has a Wi-Fi radio, one channel at a time. Want retry statistics targeted for one AP or station? AirMagnet’s Infrastructure screen can handle that, with integrated one-click device filters to boot.

The problem is that seeing Wi-Fi retries is like seeing leaking fluid coming from a car. You know something's wrong, but you may not know exactly what's wrong. It is in that spirit that this paper goes beyond just getting the raw retry numbers. The goal here is to understand the root causes of high retry percentages, and to understand how AirMagnet WiFi Analyzer PRO can be used to determine which cause (or causes) is the source of Wi-Fi performance issues.

Common Causes

First, the causes. There are several common causes of Wi-Fi performance issues, all of which commonly lead to high Retry Percentages. They are:

Resumo de conformidade PCI

Activity from wireless devices that do not use 802.11/Wi-Fi protocols, such as wireless headsets, automated lighting systems, security cameras and a host of others, can cause Wi-Fi collisions.

Hidden Nodes

A “hidden node” is a Wi-Fi AP or station that is unable to hear nearby Wi-Fi traffic, but is capable of interfering with that same Wi-Fi traffic. It sounds like an odd situation because it can’t happen in wired networking. If a wired device can reach something, then it can hear it.

In wireless networking, the area that a transmitted frame reaches may be very different than the area that a receiving device can hear. See the diagram below:

hidden nodes Stations and APs located in hidden node areas cause collisions, especially when Wi-Fi networks get busy.

Sticky Data Rates

The IEEE 802.11 standard may govern the regulations and protocols of Wi-Fi, but many aspects of device behavior are left open to vendor interpretation. One of those aspects is data rate selection. Each Wi-Fi station and AP has the right to choose which data rate is used for each transmitted frame.

If radio frequency (RF) channel conditions change due to mobility, congestion or other common enterprise Wi-Fi occurrences, APs and stations must change data rates accordingly or run the risk of transmitted frames not reaching intended receivers, which will result in more retries.

Mismatched Transmit Power

As with data rate selection, radio transmit power is ungoverned by the IEEE 802.11 standard. Both are left to APs and stations for the same reason — because different environments and devices have different Wi-Fi requirements — and both can cause the same result: Wi-Fi collisions.

Radio transmit power often causes collisions if APs and stations use transmit power levels that differ significantly. When AP transmit power goes too high above or below station transmit power, the lower powered party may mistakenly believe that high data rates are stable, thus causing failed transmissions and, ultimately, retries.

Disabling of Low Data Rates

Modern Wi-Fi best practices sometimes call for low OFDM data rates, including 6, 12 and 18 Mbps, to be disabled. The goal is to inflate throughput test results, as 802.11 management traffic occupies less Wi-Fi channel time when higher rates are required.

Disabling low data rates often has the unintended consequence of popping retry percentages once actual users arrive. Natural changes to the physical environment (doors opening and closing, people moving around, etc.) may compromise the stability of RF links between APs and stations. The IEEE 802.11 standard provides low data rates as a fail-safe for such occasions. If those rates are forbidden, the fail-safe is absent and excessive Retries may be the result.

Effective Diagnoses

The latter two causes of high retry percentages can be corrected by making simple changes to Wi-Fi infrastructure configurations. Network engineers can enable all OFDM data rates and configure AP transmit power to fall within a range that matches ranges commonly used by modern Wi-Fi stations. An AP transmit power range of 14 to 17 dBm works well with mid-2010’s era smartphones, laptops and tablets.

NETSCOUT AirMagnet WiFi Analyzer PRO allows data rate and transmit power configurations to be viewed for many AP models. Network engineers can simply navigate to the Infrastructure screen of AirMagnet WiFi Analyzer PRO, click the desired AP from a list on the left panel of the screen, then select the AP Details tab from the lower right panel of the screen.

Information gathered from AP Beacon frames is displayed in the AP Details panel of AirMagnet WiFi Analyzer PRO’s Infrastructure screen. The IEEE 802.11 standard makes a listing of enabled data rates mandatory in Beacon frames, while the inclusion of AP transmit power information is optional.

Potência de transmissão

Determining whether one of the former three causes (interference, hidden nodes or sticky data rates) is one of the reasons for high retry percentages is more complicated. It can require knowledge of radio frequency communication and the behavior of modern Wi-Fi stations, along with how to us an enterprise-class Wi-Fi analysis tool, like AirMagnet WiFi Analyzer PRO.

The most fundamental radio frequency concept to understand is that Wi-Fi communication is ultimately in the hands of the receiving party. If RF interference or competing Wi-Fi traffic appears on a channel near the receiving AP or station, then a “collision” (failed frame transmission) is likely. If RF interference or competing Wi-Fi traffic appears on a channel near the transmitting AP or station (or halfway in between), then successful frame transmissions are quite possible.

Wi-Fi is very much a location-specific networking technology. Neighbors and interference and all of the other things that affect Wi-Fi performance really only matter if the problem makes its way to the location of the receiver.

Portability is essential when attempting to perform location-specific analysis, and AirMagnet WiFi Analyzer PRO offers just that. WiFi Analyzer PRO can be run on a Windows laptop or NETSCOUT OptiView XG analysis tablet. Either option gives network engineers the ability to rove to areas where users have reported Wi-Fi issues and investigate the cause of high retry percentages from there.

At this point we would be remiss if we didn’t mention that the AirMagnet product line also includes NETSCOUT’S best-in-class Wi-Fi spectrum analyzer, AirMagnet Spectrum XT. Spectrum XT with WiFi Analyzer PRO, whether running on a laptop or integrated in the handy tablet form factor of the Optiview XG, makes for a killer combo.

The combination of AirMagnet Spectrum XT with WiFi Analyzer PRO makes an especially effective solution for identifying whether Interference is causing wireless retries. Spectrum XT is a spectrum analyzer, and spectrum analyzers show RF activity from all sources, whether the source is Wi-Fi or interference. Spectrum XT also displays the received signal strength of interference sources, often making tracking a simple task.

Interference can still be identified using AirMagnet WiFi Analyzer PRO, but it may take some work. The primary Wi-Fi indicator of interference is excessive retries, but interference-based retries will occur mostly in the area surrounding the interference source. Therefore, a Wi-Fi device must be moved to locations suspected of having interference issues, all while keeping an eye on the retry percentages of Wi-Fi traffic that is received by the device. Received retry percentages for Wi-Fi devices can be viewed in the WiFi Analyzer PRO by navigating to the Infrastructure screen, changing the stats pane (lower right corner) to display “Rx Total/% Total” and then clicking on the device in the left side pane.

Retry Rates

Identifying another one of the common causes of high retry percentages — hidden nodes — is done almost in the opposite way to identifying performance. Hidden node problems can be diagnosed in AirMagnet WiFi Analyzer PRO by viewing retry percentages for transmitted data, not received. If a device or AP has a hidden node problem, some of its transmitted data will fail (thus causing retries) if a hidden node becomes active.

A tell-tale sign of a hidden node is when received retry percentages are significantly lower than transmitted retry percentages, but that alone doesn’t prove the presence of a hidden node. Transmitted retry percentages can exceed received retry percentages for a number of reasons, most commonly mismatched transmit power. However, if the AP transmit power is configured to match Wi-Fi device transmit power, then an asymmetrically high transmitted retry percentage might be indicative of a hidden node problem.

Retry Rates

If a hidden node problem is suspected, at that point it is recommended to consult a floorplan of the facility. Hidden nodes exist because some AP or Wi-Fi device is on the same channel, close enough to cause collisions at the location of the receiver and is out of range from the transmitter. It is actually another case where the AirMagnet product suite, including that integrated with the OptiView XG, can help, because this suite includes AirMagnet Survey PRO, which allows for floorplans and channel patterns to be viewed.

The final common cause of high retry percentages, sticky data rates, may be the most simple to diagnose using AirMagnet WiFi Analyzer PRO. One must simply look at which data rates are being transmitted by an AP or Wi-Fi device, then check to see if too many high rate frames are being used. APs and Wi-Fi devices are supposed to use lower data rates when transmitting if retries are occurring. Some APs and Wi-Fi devices don’t, and the result is lots of high rate traffic that isn’t received, has to be re-sent and ends up just wasting every other AP and Wi-Fi device’s channel time in the process.

AP and Wi-Fi device data rates can be viewed in the Infrastructure screen of AirMagnet WiFi Analyzer PRO. Data rates are tallied in the same place where retry percentages are shown, which is the stats pane. First, an AP or Wi-Fi device must be selected in the left pane of the Infrastructure screen. Then the rates area, followed by the bytes area, must be expanded in the stats pane. Once the stats pane has been set to display “Tx Total/% Total”, the work can begin. The percentage of traffic transmitted at each rate can be seen and cross-referenced to the retry percentage of the AP’s or Wi-Fi device’s transmitted data. High retry percentages coupled with high data rates likely means that the selected AP or Wi-Fi device is sticky with its data rates.

Sticky data rates is the one cause of high retry percentages that can’t be solved. You can configure APs to use similar power as stations. You can configure APs so that all OFDM rates are enabled. If an interference source is identified, you can remove the interferer or change the channel of the Wi-Fi network. And if a hidden node problem is happening, there are a number of potential solutions: moving APs, changing AP channels, disabling AP radios, adding APs and lowering the RTS threshold on APs or Wi-Fi devices (if the RTS Threshold setting is configurable) are all potential solutions.

APs and Wi-Fi devices that are sticky with data rates are configured that way by their manufacturer. As network engineers, we are not AP and Wi-Fi device manufacturers, so we just don’t have that level of control.

The good news is that Wi-Fi markets are as competitive as ever. If an AP, smartphone, laptop or any other Wi-Fi device has problems with sticky data rates, tell the manufacturer. No company wants to make a low performing Wi-Fi device.

The topic of Wi-Fi retry percentages introduces a lot of variables, and that can be frustrating when trying to troubleshoot problems or just perform general analyses on Wi-Fi deployments. Knowing the common causes of high retry percentages can be helpful. Using NETSCOUT AirMagnet WiFi Analyzer PRO to identify the cause of chronic high retry percentages can also be helpful, and when that problem is eliminated, the feeling can be downright joyful.

Powered By OneLink