Slow WiFi speed… Who do you call?

July 30, 2018
Even when home WiFi bottlenecks are completely outside of the providers’ responsibility or control, they have to handle subscriber complaints. To assist in the troubleshooting task, this article will discuss the usual suspects when it comes to slow WiFi: signal strength, interference and equipment performance. It also will provide some suggestions on how to avert those service calls or truck rolls in the first place, by qualifying the customer’s WiFi with a site survey and educating them about their WiFi performance.

Service provider support centers are inundated with complaints about “slow WiFi.” They spend billions in capex and opex to build and maintain the best possible access and transport networks, while subscribers will judge them on the perceived quality of the last segment, the one that the carrier has the least control over: the WiFi network.

Yet even when bottlenecks are completely outside of the providers’ responsibility or control, they have to handle subscriber complaints. To assist in the troubleshooting task, this article will discuss the usual suspects when it comes to slow WiFi: signal strength, interference and equipment performance. It also will provide some suggestions on how to avert those service calls or truck rolls in the first place, by qualifying the customer’s WiFi with a site survey and educating them about their WiFi performance.

Signal strength

WiFi technology is designed to be resilient and allow connectivity even in the most adverse conditions. But, getting a successful WiFi connection doesn’t guarantee good performance (see Figure 1). To understand the underlying mechanisms behind WiFi performance, it is useful to go back to the principles of radio frequency (RF) transmission.

Figure 1. Example signal, noise and SNR level measurements.

WiFi uses RF waves to transmit information. As the waves travel through air, the signal gets attenuated. This phenomenon is called free space path loss, which is a function of distance and frequency. For a 2.4 GHz WiFi signal with a clear line of sight to the WiFi router, the signal’s amplitude will be attenuated by roughly 70 dB over a distance of 100 feet. The decibel scale is logarithmic; therefore, a 3 dB loss corresponds to dividing the transmit power by 2, and a 10 dB loss corresponds to dividing it by 10.

As the signal travels within homes or offices, the quality is degraded even further by obstacles. Big obstacles, like walls, attenuate the signal the most. A drywall (gypsum) type of material attenuates the signal by only 3 dB, whereas a concrete wall attenuates it by more than 10 dB.

When the WiFi receiver detects a low signal level, it will negotiate with the transmitter to use a lower data rate. This mechanism ensures that data can be properly decoded without requiring retransmissions. As an analogy, this technique is similar to adapting your car’s speed to the road conditions. Reasonable people will lower their speed in heavy rain, but if they don’t, they may never arrive in one piece to their destination. Therefore, the farther you are from the WiFi access point, or the more signals get attenuated by obstacles, the lower the data transfer speed.

In the outer rings of the signal coverage, even when WiFi connection is still possible, the negotiated speed may be just a few megabits per second, even though the equipment may be capable of gigabit speed. In practice, the top data speed can only be achieved if the receiver and transmitter are 10 to 20 feet away and with a clear line of sight.

Interference

In an ideal world, just ensuring a good signal level would be sufficient to guarantee good WiFi service. But in reality, the WiFi signal suffers from noise interference.

There are two types of noise interference sources that commonly plague WiFi networks: interference coming from other WiFi networks and interference coming from non-WiFi sources. The disruption from interference can result in lowered data speeds or complete signal loss, depending on the nature and severity of the interference source.

WiFi’s RF channels are a shared medium, and all devices connected on the same frequency (channel) share the same “air time.” The 802.11 protocol rules require that each device only transmits packets when there is no other device transmitting. Devices achieve this by sensing the RF frequencies to detect if there is any energy. If energy is detected, transmission will be postponed until the medium is free.

Therefore, the available bandwidth is effectively divided by the number of devices attempting to transmit on the same frequency. A channel with a high number of transmitting devices will not leave much free air time and won’t allow much data to be transmitted. The take away is that even if the signal level is high enough to ensure optimum transmission speed, the RF frequency channel’s capacity may not allow this transmission to take place. To continue our car analogy, even a fancy Italian sports car will get stuck in a traffic jam, but when the traffic jam clears, it will be able to reach its destination faster than the other cars.

How “busy” the RF channel is, is measured by the utilization level. For example, a utilization level of 50% indicates that RF activity is detected half of the time on the channel. An important fact frequently overlooked by users is that not only do the devices connected on the local network contribute to this RF activity, but all the neighboring WiFi devices configured on the same channel contribute as well. This is particularly a problem in dense urban areas, where the signal from tens or even hundreds of access points can be detected (see Figure 2).

Figure 2. Utilization levels in dense urban environments.

RF activity is not limited to WiFi devices. A lot of other devices, such as microwave ovens, cordless phones, Bluetooth devices and many more, generate RF signals in the 2.4 GHz frequency band (see Figure 3). These devices constitute sources of non-WiFi interference. They are difficult to track down because conventional WiFi equipment does not scan or list non-WiFi equipment. The explanation for this is that non-WiFi equipment does not follow the 802.11 protocol to advertise their presence. The use of dedicated test equipment is required to detect and address them.

Figure 3. Interference from a microwave oven on the 2.4 GHz frequency band.

Equipment performance

Over the years, IEEE 802.11 standards have evolved to bring faster and faster speeds - from the early days of 802.11 in 1999, with a meager speed of 54 Mbps, to the latest 802.11ac standards, which can achieve a maximum theoretical speed over 3 Gbps. The 802.11 standards were designed with backwards compatibility in mind, to ensure that upgrading WiFi routers to the latest generation would not prevent older client equipment from connecting.

However, consumers often overlook the fact that the highest speed advertised in the standard can only be achieved if both the transmitter and receiver devices support the same generation. Therefore, if a subscriber upgrades their WiFi router to a model supporting 802.11ac, but their client device (PC, tablet, phone, etc.) is not upgraded, the advanced capabilities offered by 802.11ac will not be used. Conversely, if a subscriber upgrades their client device to a newer generation, but not their WiFi router, they won’t be able to take full advantage of the upgrade. Very often customers overlook this fact and simply get frustrated by the lack of speed improvements after purchasing a new device.

The 802.11n and 802.11ac standards also support a method called Multiple Input Multiple Output (MIMO) that requires multiple antennas. In a device with four MIMO antennas (4x4), user data is broken down into four streams that are transmitted in parallel, one stream per antenna. On the receiving end, the four data streams are received by four antennas and the data is reassembled. This method can effectively transmit four times “faster” than devices with a single antenna. However, it requires both the transmitter and the receiver to have the same number of MIMO streams capability to take full advantage of the feature.

The reality is that even though WiFi routers often have four-MIMO streams, there are no clients at this time that can match it. Mobile devices like phones or tablets often have one or two antennas, while only high-end laptops will have three antennas. Unfortunately, the number of antennas in a client device is often difficult to figure out, as they are not externally visible in a phone, tablet or laptop, and manufacturers don’t make this information easily accessible.

Preventing unnecessary support calls

All these considerations combined make it difficult to predict the actual data rates on a given WiFi network. Basic knowledge of RF propagation rules can help determine optimal placement for the access points, but accounting for all obstacles to establish the equipment’s exact coverage range is close to impossible, unless a technician can run a passive or an active site survey (coverage test) during the service turnup.

During a passive site survey, the customer’s WiFi access point(s) is in place, and the technician simply walks around the facility with test equipment to measure the received signal’s strength to ensure that the quality is sufficient in the locations where WiFi will be used. If dead zones are detected, the access point can be moved or a wireless repeater (or mesh router) can be added to improve the range. During the site survey, the test equipment can also detect all the neighboring access points and measure channel utilization levels. Channels with high utilization levels should be avoided, and technicians can also reconfigure the customer’s access point to use a different channel. Some test tools will automate this process and provide channel recommendations to the technician.

An active site survey goes beyond the aforementioned measurements by having the test equipment connect to the subscriber’s access point and run data traffic to perform upload and download tests to a server (see Figure 4). The active survey establishes the actual data throughput as experienced by the subscriber, in the locations where they will be using the service. Technicians can quickly determine whether the achieved upload and download rates would meet service-level agreement (SLA) requirements and readiness for high-bandwidth traffic like audio and video streaming.

By educating customers and setting the right expectations during the installation process, technicians can share information and tips, making sure their customers understand the basic nature of WiFi and that the environmental conditions may change over time.

Figure 4. Dual ended active site survey with upload and download tests

Eve Danel is a senior product manager at VeEX Inc. responsible for WiFi, Carrier Ethernet and mobile fronthaul/backhaul test solutions. Prior to joining VeEX in 2011, she was a product manager at Ditech Networks responsible for VoIP voice quality. From 2000 to 2005 she served as a product manager for Sunrise Telecom's Ethernet test and measurement products.