Why 802.11ax is not just higher data rate

Article By : John Lukez

802.11ax is first and foremost designed to tackle the problem of network capacity, which has become a larger issue in dense environments.

As the next-generation WLAN standard, 802.11ax promises greater capacity and more robust data transmission than previous Wi-Fi standards. It represents the most fundamental change in Wi-Fi operation since 802.11n, which first emerged in products in 2004 with the promise of 100Mbits/s data speeds. 802.11ax is first and foremost designed to tackle the problem of network capacity, which has become a larger issue in dense environments such as airports, sporting events, and campuses.

802.11ax borrows heavily from LTE, utilising technology proven in cellular networks to increase system capacity by as much as ten times more users for the same spectrum. Thus, there are many new system aspects in 802.11ax requiring validation that have been unknown in the Wi-Fi community.

OFDM evolves to OFDMA

The biggest change with 802.11ax over earlier WLAN systems is the adoption of orthogonal frequency division multiple access (OFDMA). OFDMA systems allocate blocks of time and frequency—called resource units (RUs)—based on a central resource (the access point). In OFDM systems, the user occupies the entire channel. As more users are added, each user's requests for data competes for media access via carrier sense multiple access with collision detection (CSMA/CA). Too many users can create a bottleneck, resulting in poor quality of service (QoS) as these users request data, especially in high bandwidth applications such as streaming video.

Figure 1 compares OFDM to OFDMA spectrum usage. In OFDM, the users occupy all the subcarriers (RF spectrum) all the time for a given data request, whereas in OFDMA the users occupy only a subset of subcarriers for a prescribed amount of time. OFDMA requires that all users transmit at the same time, so each user is required to buffer its packets to a defined number of bits so that all users stay aligned in time regardless of the amount of data to transmit. Additionally, an OFDMA access point (AP) can dynamically change the amount of spectrum a user occupies based on how much bandwidth that user needs. For example, a streaming video user would need more subcarriers (spectrum) as opposed to an e-mail transmission that doesn't need real time performance.

802.11ax_testing_01 (cr)
Figure 1: With ODFM, a user occupies an entire channel, but not in OFDMA.

The dynamic allocation of spectrum required to support OFDMA adds considerably Wi-Fi's spectrum complexity and stresses the radio hardware in ways not previously seen. Timing synchronisation, frequency alignment and response time characteristics matter. New test methods are needed to ensure the system performs correctly. Poor or mediocre response timing may not result in total system failure, but could lead to reduced system capacity impairing the key performance advancement in 802.11ax, and leading to poor product reviews.

Power control

OFDM systems allocate an entire portion of spectrum (all subcarriers) for an amount of time for one user (left side of Fig.1). Subsequent users then get all the subcarriers for the next amount of time based on CSMA/CA. OFDMA systems share the spectrum with multiple users at the same time (right side of Fig. 1).

As a result, power control is needed to ensure that a user very close to the access point (AP) does not drown out another user farther from the AP. Widely different power levels between users at the AP will result in increased inter-carrier interference (ICI), receiver compression, leakage and carrier frequency offset (CFO) due to timing misalignments between Wi-Fi user/station (STA) devices. In 802.11ax, the AP commands the STAs to adjust their power up or down based on the target received signal strength indicator (RSSI) at the AP side. The STA first estimates path loss by subtracting measured RSSI (at STA) from AP transmit power (encoded inside the packet). The STA then transmits a signal with power equal to the target RSSI plus estimated path loss. Devices closer to the AP transmit less power while devices farther away transmit more power to overcome the greater path loss. Figure 2 illustrates the process.

802.11ax_testing_02 (cr)
Figure 2: With multiple users sharing the same spectrum gives rise to a need for power control.

The 802.11ax standard defines device classes based on how accurately they control their power. Class "A" devices control their transmit power within ±3dB while Class "B" devices control their power within ±9dB. In earlier 802.11 systems (pre-802.11ax), STA devices simply transmitted their maximum designed/allowed power. Of course, a class A device will help improve system capacity and have a higher selling price, so manufacturers will have an incentive to ensure class A performance, which requires more stringent testing and calibration. The new requirements of dynamic power control in 802.11ax require testing to verify that the system operates correctly and capacity isn't degraded due to one STA improperly controlling its transmit power.

To test the power control functionality of a STA device, the test system needs to mimic an AP's real-time responses (Figure 3). In a power-control test, the device under test (DUT) will adjust its transmit power based on information encoded in the downlink signal from the tester (acting like an access point). The DUT (STA) will transmit the corresponding power back to the tester, where the tester measures the transmitted power from the DUT, and then sends back a corresponding command to increase or decrease the power. This entire process requires real-time behavior with latencies on the order of a few hundred microseconds. The diagram below shows how this process takes place.

802.11ax_testing_03 (cr)
Figure 3: In a power-control test, an 802.11AX tester emulates an access point, adjusting the power of the DUT.

Dynamic power control can introduce settling time issues in an 802.11ax system. As the power changes, delays in power settling due to power amplifier response times or power control loops may become an issue. This behaviour needs to be checked to ensure settling is fast enough and there are not undesirable overshoot conditions (Figure 4).

802.11ax_testing_04 (cr)
Figure 4: Power control settling response time is one of several measurements needed for testing 802.11ax STA devices.

To perform these tests, you need to measure power on a packet-by-packet basis and accurately report the power per packet as a function of time to validate proper response of the STA to power control commands. Because these measurements are made as part of the power control testing, no additional test time is incurred.
 
Next: Timing matters for 802.11ax users »

Leave a comment