802.11ac Draft and 802.11n Performance Comparison

    I recently wrote a small 802.11ac forecast . The first hardware with support for 802.11ac Draft and the first reviews have already appeared in the USA .
    In general, of course, one should not judge the entire technology by one product (in which, by the way, far from all 802.11ac features are implemented), but you can make a first impression.

    For those interested, I recommend reading the full review, but I will consider one picture that interests me very much. Here we compare the bundles “Buffalo router (802.11ac, 3x3: 3 MIMO, channel 80mHz) paired with another one” and “Linksys (802.11n, 3x3: 3 MIMO, channel 40mHz) with Intel Wireless-N6300”.

    image

    When transmitting a single data stream, the implementation of 802.11ac in this router is not of interest: 802.11n reaches 110Mbps at 40Mhz (and this, in fact, is not enough), 802.11ac - 170Mbps at 80Mhz. In the same frequency range, you can put two 802.11n devices and get 220Mbps +. So, real improvements in Mbps / Mhz are not visible. It is difficult to say why this happens: a detailed research methodology has not been published, and it is quite possible - just the features of the firmware. Theoretically, we should have seen at least a three-fold increase (2x channel + 256-QAM versus 64-QAM).

    But look at what happens when you increase the number of threads in a cell! 802.11n “blown away” already on three streams, resting on the ceiling of 165Mbps (which is more like the truth and very good for a consumer device, but 3x3: 3 MIMO can be more - 165Mbps in Chariot also show points with 2x2 MIMO), while 802.11ac "accelerates" to seven! It is clear that the specific speed indicator per stream drops, but it still remains within reasonable limits: 438/7 = ~ 62Mbps effective bandwidth (not datarate!) Is even higher than 165/3 = 55Mbps, for 802.11n. Here you can clearly see the potential of the new technology, which squeezed out of the cell almost three times the speed (as expected), and provides greater capacity and scalability of the cell.

    Again, it is not yet possible to judge why this happens (the advantages of the new technology are manifested only in multithreaded mode). I hope that soon the base of tested devices will expand and it will be possible to draw any reasonable conclusions. In the meantime, suggest the versions in the comments!

    Also popular now: