Next generation Wi-Fi - 802.11ac in brief

At the end of last year, 802.11n finally crushed under itself (at least in terms of chipset shipments) all previous Wi-Fi standards - some experts voiced figures of about 70%. Having tried the high speeds, the public wants more, and the heirs are tipped to 802.11ac, promising up to gigabit speeds. Let's see what we really get.

What with speed?

802.11n used the following to increase throughput:

• Optimized modulation and packet transfer mechanisms allowed for “breaking through” from 54Mbps to ~ 75Mbps
• Then Channel Bonding was turned on - channels 40 Mhz wide (twice as wide as the traditional 22 Mhz) provided a doubling of speed - up to 150Mbps.
• Then the Multiple Spatial Streams mechanism was activated, of which the standard can be up to four, which allows reaching in theory 150 * 4 = 600Mbps

802.11ac is going to "catch up and overtake" in the following ways:

• The channels are 80Mhz and 160Mhz wide, which allows you to instantly double / quadruple the 802.11n results.
• The maximum number of Spatial Streams increased to 8, which allows you to double the speed.
• Optimization of modulation and packet transmission methods allows you to squeeze a little more resource and to ensure that high speeds will be available not only within a radius of 4m from the access point.

Total, adding up all the factors, we can get the speed in theory more than eight times the 802.11n performance - about 5Gbps. Marketers are jubilant, People havala. In practice, this speed is practically unattainable:

• The new channels no longer fit into the 2.4Ghz range, so 802.11ac will only work in 5Ghz. But in 5Ghz, it's not that simple. In the same Europe, it is possible to work seamlessly only on the first four channels: 36/40/44/48 - the rest should include DFS / TPC (coexistence with radars), which excludes the possibility of building a more or less reliable network. And in these 4 channels only 1 channel 802.11ac will fit, and that “only something” at 80Mhz. As an exercise, try to calculate how many 160Mhz channels fit into the entire 5Ghz range in different regions . Some hopes are placed on the 802.11ad version of the standard, which operates at 60Ghz, where megahertz is available more, but this frequency range is still clearly defined only in the US, and its communication range is measured in tens of meters (which is very good to reset the video from the phone on a TV or using a wireless mouse, but not for serious networks). So, we divide the speed in half.
• 8 Spatial Streams require a radio module with 8 antennas and a suitable environment (so that all 8 streams end up on the client). How it will look like a Dual-Radio access point with 8x8: 8 MIMO (16 antennas with Dual Radio!) Can only guess. :) Mobile devices will most likely not go further than 4x4: 4 because of the need to save electricity and space in the case. So, cut the speed still in half.

So we get a maximum doubling of 802.11n speeds - which in itself is still not bad.

Why do we still go to 802.11ac?

In addition to speeds, 802.11ac offers two key improvements:

• Beamforming - the ability to dynamically change the antenna pattern (which is realistic for an antenna array of 8 elements). Ideally, this means that the coverage area of ​​the access point optimally adjusts to the current location of customers. Beamforming is not new to Wi-Fi, it has even been made part of the 802.11n standard. But part of the optional! In 802.11ac it will become part of the mandatory. It is unclear how exactly Beamforming will work in 802.11ac and whether it will have at least some benefit in the end , but it is clear that it is being introduced to maximize the effect of the next (and main) improvement.
• In 802.11ac MU-MIMO will finally appear! Wi-Fi networks are half duplex: while one is transmitting, the rest are listening. Packets are transmitted sequentially — one packet is transmitted at a time. If the “pipe” at 450Mbps (802.11n 3x3: 3 MIMO) is streaming at 1Mbps - 1/450 bandwidth is used. If data for another client arrives at the same time, it will not be possible to use unused bandwidth. As a result, the use of 802.11n ultrahigh speeds in networks with a large number of slow clients (i.e., corporate) is very small. MU-MIMO allows you to split the "pipe" into several "tubes of smaller diameter" and transmit data in parallel. This technology is well known to telecom operators. So far, they are talking about two variants of the implementation of MU-MIMO in 802.11ac: SDMA (Space Division Multiple Access) allows you to transfer data to different clients via different Spatial Streams (that's where Beamforming is needed!), Downlink MIMO allows you to split OFDM subcarriers into groups, and dynamically (sort of) allocate to each client the required number of subcarriers. Thus, even if there are 2x2: 2 MIMO clients sitting on the access point, you can still use the full potential of the “pipe”.

Total

As you can see, even if you limit the maximum speed to one Gbps, the 802.11ac standard promises significant benefits for both domestic (high speed) and corporate networks (efficient utilization of these very high speeds in networks with a large number of clients). In addition, to support the new radio modes, you need to completely change all the equipment, which promises substantial benefits to vendors and integrators - everyone is happy, in general :) Currently, the standard is expected to be ratified at the end of 2012 / early 2013, but more optimistically customized vendors have already presented chipsets and products based on them at CES-2012 . Whether the fate of 802.11n, adopted a few years later than expected, is prepared, time will tell. We are waiting, sir.
')
UPD: Appeared the first iron and tests