Even before many consumers, enterprises and technology vendors have upgraded to WiFi 6/6E, the currently emerging version of the technology, there is growing consensus that the next generation, WiFi 7, will be needed to compete squarely with 5G for many emerging use cases beyond the traditional, largely indoor domains.
There is even fear of 5G invading some indoor enterprise sectors currently dominated by WiFi, while there is a lot at stake in the big public venue market where both technologies are contenders.
In this context it was interesting to observe recent comments from US testing group National Instruments (NI), which conceded that significant advances must be made in that area to ensure that WiFi delivers on its promises for lower latency, higher capacity, and particularly the greater robustness required by some of the new industrial use cases. It is those use cases where 5G promises to bring wireless communication much closer to the quality and resilience of wired networks, which obviously cannot serve mobile processes such as robotics and drone operation.
The game was given away somewhat by NI’s go-to-market director and head of semiconductor marketing, David Hall, who was quoted by RCR Wireless News suggesting that WiFi 6 does not necessarily suffer from “severe technology limitations”. This sounds rather like a politician’s denial, especially as Hall added that WiFi 7 offers more spectrum to serve all of the emerging use cases and devices that rely on wireless connectivity.
It is other aspects around latency and interference mitigation where more rigorous testing will be required to reassure enterprises hovering between WiFi 7 and 5G that the former would meet all their requirements. Yet NI has identified support in WiFi 7 for wider 320 MHz transmissions, double the 160 MHz of WiFi 6, and for 4096-QAM, four times the 1024-QAM of WiFi 6, as the two improvements that will require new testing equipment.
In the case of 320 MHz transmission it is fairly clear that testing equipment must advance in scalability and linearity, ensuring consistent measurement across a wider spectral range. It is also worth noting that although not part of WiFi 7, WiGig operating in 60 GHz spectrum is also coming along and will impose similar testing challenges to 5G mmWave. Such networks are currently being deployed mostly on a line-of-sight basis, but testing must take accurate account of non-line-of-sight reflections. That is still work in progress but at least WiFi, or WiGig, will be able to take advantage of R&D in the 5G arena.
Another point emerging from the expansion of 5G and now WiFi across such diverse spectral bands is the demand for testing matched to overall use cases rather than just networks or infrastructures themselves.
German electronics and testing group Rohde & Schwarz is among those developing “use case classes” for test parameters that can be invoked to ensure that 5G QoE meets the needs of given application classes, which will avoid having to evaluate every relevant app separately for a field such as industrial IoT, or for that matter consumer gaming.
Previously, under LTE for example, user experience testing was confined to well-known consumer apps such as Facebook or YouTube to test performance. This becomes wasteful when extended to multiple lower profile apps. If WiFi is to address use cases similarly, it may need to adopt comparable dedicated testing.
Testing for more advanced QAM under WiFi 7 is also connected to developments across the whole wireless world and is concerned with greater sensitivity to measuring impact of noise on signal degradation. QAM emerged in the era of analog TV and was then revamped for digital as audio and video migrated towards IP transmission in alignment with alphanumeric data.
The objective then became to modulate as many digital bits as possible on a single cycle or Hz of a wave. Under QAM, this is done by mapping what are known as symbols comprising multiple input bits to variations in phase and amplitude of two waves superimposed at one quarter of a cycle offset, or 90 angular degrees.
In the beginning there was 2-QAM where a single bit of information was encoded as a 1 or 0 represented by two different states of the wave. This was then extended to 4-QAM or two bits, altering the phase and amplitude accordingly at two places within a single cycle. Since two bits can encode four different values, 00, 01, 10 and 11, this enabled four modulation states or values, hence the name 4-QAM. Then under 8-QAM, three bits are encoded, yielding eight modulation states, while 16-QAM encodes four bits and 16 modulation states, and so on.
WiFi 6 supports 1024-QAM, which modulates 10 bits within each Hz cycle, allowing 1024 different modulation combinations. This will be increased to 4096-QAM under WiFi 7, which means there will then be 12 bits modulated per cycle, increasing the density of the constellation further. With each bit’s encoding closer spatially, susceptibility to noise increases, imposing ever greater processing demands on the signal processing equipment.
It also raises the bar for testing, because as NI’s Hall pointed out: “The more states you have, the harder it is to tell them apart and the higher performance the hardware has to have in order to transmit and receive that signal.”
So to assess whether a WiFi component meets requirements for transmitting and receiving the 4096 discrete phases of magnitude supported by WiFi 7 4096-QAM, test equipment also has to reach those standards.
QAM incidentally is fundamental to the OFDM multiplexing employed by WiFi as well as other fixed and wireless transmission schemes, including 5G New Radio. OFDM is a data transmission scheme based on frequency division multiplexing (FDM) catering for harsh channel conditions without need for complex filters to eliminate noise.
OFDM divides a channel into subcarriers at different frequencies across a band separated by guard intervals to protect against cross-channel interference. QAM is then the digital modulation scheme for each of those sub-carriers. There is no underlying mathematical limit on the density of QAM and therefore the amount of data that can be carried in each Hz of bandwidth, but that is not accounting for noise, which does in practice always impose a maximum data rate defined by Shannon’s Law.