802.11ac has seen a tremendous uptake in 2014 and the industry expects this technology to be adopted much more quickly than the previous standards.
IEEE 802.11ac has proven popularity now. That's because service providers have a critical need to offload their traffic and because the industry doesn't foresee any other major standards emerging in the near future. Capacity management came about as a big challenge as the industry sees a tremendous increase in demand for it, especially because many new devices don't have wired internet access. Furthermore, many users like to bring all of their devices to work – the BYOD (bring-your-own-device) phenomena -- sometimes up to a total of five connected devices at the same time. This demand for capacity is partially driving the demand for 802.11ac as well.
Although many 802.11ac products have appeared on the market, there is still hesitation on when to schedule 802.11ac deployments. Some of the challenges associated with this are cost of deployment, identifying the right product, and a steep learning curve. The good thing, though, is that 802.11ac is now an approved IEEE standard and Wi-Fi Alliance has a Wi-Fi Certified ac certification program available. 802.11ac opens up new possibilities for throughput, capacity, improved price-to-performance ratio, as well as the ability to use 5-GHz bands.
Several testing challenges emerged while testing 802.11ac standard, though.
Bandwidth: The 802.11ac standard functions at the 5 GHz license-free bands in the 80 MHz or 160 MHz bandwidth, which is much greater than earlier WLAN standards. According to Aeroflex, a single capture of the full signal bandwidth is required for conducting transmitter tests to measure EVM (error vector magnitude), frequency, power and spectrum flatness. The spectrum mask measurement provides the analysis of a significantly wide bandwidth; this operation is executed via spectrum stitching by recording and capturing snapshots of the signal and stitching, them in the frequency domain. For receiver tests, generating a full bandwidth signal waveform is necessary to stimulate the device under test (DUT). This allows receiver sensitivity tests during various modes of operation.
The spectrum specification of an 802.11ac channel is 160 MHz wide.
Multiple spatial streams: The 802.11ac standard increases MIMO (multiple input, multiple output) support up to eight spatial streams, allowing a terminal to transmit and receive a signal to and from several users on the same frequency band simultaneously. In the R&D application, MIMO requires test equipment to perform encoding or decoding of composite signals with multi-path channel emulation. At the manufacturing stage, the WiFi testing focus shifts to RF component calibration and quality assurance.
High-density modulation: The 802.11ac release highlights modulation schemes at up to 256 QAM (quadrature amplitude modulation), that is four times greater than 64 QAM used in previous WLAN standards. As a result, obtaining the required signal quality for high-rate transmission became more challenging. To accurately measure 802.11ac signals, the remaining EVM of the test equipment must be better than the lowest EVM requirement or negative 32 dB at 256 QAM, to avoid jeopardizing production numbers, according to Aeroflex.
Challenges in WiFi offload testing: Apart from the challenges mentioned, WiFi offload testing is expected to extend from performance tests in the laboratory to OTA (over-the-air) field tests for network assessment. Network equipment vendors do most of the testing in the lab using cabled RF to simulate network loads. With WiFi offload, mobile operators will ensure that site assessment will be an integral aspect of testing, and this will require expertise in OTA tests.
I have been tracking 802.11ac for a while now, and have even upgraded my home router to the standard. I haven't run real-world throughput tests on it, but it seems like it is getting close to the Gigabit Ethernet speed that is used as the backhaul for it in most cases. This, if nothing else, would seem to be a real limitation on further development (not to mention a little embarassing). Even some lucky cable customers with 1 Gb fiber could potentially be in the situation of having that be the bandwidth limitation.
Do you think that 802.11ac could shame the wired Ethernet guys into upping their game?
I don't know if "shame" is the right term. I assume you're talking about the service providers here, who have been steadily updating their core networks and last mile connections, over the years. It's labor-intensive, and therefore expensive work. But for example, DOCSIS 3.1 theroretically can deliver 10 Gb/s downstream and 1 Gb/s upstream, and it was approved as a standard earlier this year.
I am curious to see 802.11ac in the real world. I have 802.11n at home, and find that in the 5 GHz band, the bit rate will vary from the high 100 Mb/s or low 200 Mb/s ranges, to the AP's maximum of 270 Mb/s. My assumption is that the 2 X 2 or 4 X 4 MIMO used is not all that dependable, maybe even affected by people walking around the house. A system that depends on 8 X 8 MIMO will most likely exhibit this same behavior.
I see the shame as being directed more towards the Ethernet standard people rather than the ISP's, although there is certainly a portion to go in that direction as well. While 802.11 has gone through several upgrades Ethernet has remained stuck at GigE for quite a while now. There are standards for faster Ethernet, but they are nowhere near mainstream. It's pretty bad when wireless throughput starts to pass wired.
The marketing material around 802.11ac is starting to talk about 1300 Mbps. Yes, I know that that number is heavily asterisked, but so is wired Ethernet. Even if you allow for the larger fudge factor common in wireless, it is still starting to look like wired Ethernet is about due for an upgrade.
Actually, better antenna diversity should significantly increase throughput in the conditions that you describe.
IEEE 802.3-2012 includes 40G and 100Gb/s versions, in Section 6. These aren't commonly available in home or even office PCs, but I don't think that's what they are aimed at. These higher rates are primarily used in telco trunks, replacing SONET links, for instance.
Why, are you feeling limited by the 1000BASE-T commonly available in PC motherboards these days? I don't see that as any problem, for some time to come.
As to antenna diversity, that was my point. MIMO only works as well as the lack of correlation between the different propagation paths. That lack of correlation is not guaranteed, by any means. It all depends on conditions at the time. So if people walk about, or if your AP and your STA are not positioned just right, there's no guarantee you can make effective use of 4 X 4, much less 8 X 8.
At home, from an upstairs AP to a downstairs STA, the correlation between the propagation paths seems to vary constantly for me, with 802.11n. So I was only saying that I wouldn't be a bit surprised if the same occurs with 802.11ac.
Why did you upgrade to ac? Did it give you a boost in performance? I just upgraded from 802.11g to 802.11n, but only because the g router needed a power cycle once a week. The n router has better range, the ability ot add guest networks, and USB sockets where I keep some network storage.
The Linksys g router that I replaced is still in production and is fine for many homes.
@MeasurementBlues, my previous router went to a relative whose router died on them. I ended up getting a Netgear Nighthawk. I have not done performance testing on it, but it does seem to perform nicely. I have not seen any reduction in range, and it might be a bit better. I also liked the USB ports on this device and the more-comprehensive setup options.
For home use. the high speeds of 802.11ac - 500 MB/s or more - are not going to help you if your incoming DSL line is only around 20 MB/s. In a corporate environment where there is more data shunted around internally there is more benefit to it. Ditto if you have a fibre incoming line.
I had an 802.11g router (11 MB/s) and when it failed upgraded to an 802.11n (54 MB/s) and did notice a very slight difference, but probably most noticeable on printing speed - which of course is internal.