It will be my first post on this blog 🙂
Today I want to show my tests of two wireless access points. The short specification of this two devices presents as follows:
802.11N – Cisco Aironet 2600 series AIR-CAP2602I-E-K9 with autonomous ap IOS software 15.3.3-JC
“802.11n with 3×4 multiple-input multiple-output (MIMO) technology with three spatial streams, which sustains 450-Mbps rates”
Full spec: Link
802.11AC – Cisco Aironet 2700 series AIR-CAP2702I-E-K9 with autonomous ap IOS software 15.3.3-JC
“802.11ac with 3×4 multiple-input multiple-output (MIMO) technology supporting three spatial streams. This architecture offers a sustained 1.3-Gbps rates”
Full spec: Link
So both devices have 3×4 spatial stream. The main purpose of this test is to show performance diffrence between 802.11N and 802.11AC wireless standard. The differences between the two standards in the brief are:
- – the width of a channel, 802.11N supports up to 40Mhz channel while 802.11AC (phase 1) support up to 80Mhz, and for phase 2 of this standard it is 160Mhz. Cisco Aironet 2700 is a 802.11AC phase 1 device.
 
- – the modulation, 802.11N supports up to 64-QAM, for 802.11AC it is 256-QAM
 
As wireless client I will use notebook Dell Latitude e7240 with Intel 7260AC wireless card which has 2×2 spatial stremas. The notebook is running under Opensuse Linux 13.1 with 4.2.5-1 kernel version and iwlwifi-7260-15 firmware for wireless card. Unfortunately I do not have device supporting 3×3 spatial stream.
Let’s get to the test…
I set radio interface on both devices like shown below.
Cisco Aironet 2600:
| 
					 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22  | 
						interface Dot11Radio1  description 802.11N 5GHz Radio  no ip address  no ip route-cache  !  encryption mode ciphers aes-ccm   !  ssid JAN  !  antenna gain 0  peakdetect  dfs band 1 2 3 block  beamform ofdm  channel width 40-below  channel 5240  station-role root access-point  world-mode dot11d country-code PL both  no cdp enable  bridge-group 1  bridge-group 1 subscriber-loop-control  bridge-group 1 spanning-disabled !  | 
					
Cisco Aironet 2700:
| 
					 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22  | 
						interface Dot11Radio1  interface Dot11Radio1  no ip address  !  encryption mode ciphers aes-ccm   !  ssid JAN  !  antenna gain 0  peakdetect  no dfs band block  stbc  channel width 80  channel 5240  station-role root access-point  bridge-group 1  bridge-group 1 subscriber-loop-control  bridge-group 1 spanning-disabled  bridge-group 1 block-unknown-source  no bridge-group 1 source-learning  no bridge-group 1 unicast-flooding !  | 
					
For performance testing I use iperf3 between two computers. One of them is connected by gigabit ethernet to the access point, and the second (Dell Latitide) is connected to the wireless network. It looks like below.

Let’s see how the connection looks, when I connect notebook to Aironet 2600. The distance between devices is about 2 meters. I’ll check the statistics of wireless interface in notebook.
| 
					 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25  | 
						interface Dot11Radio1 JAN-DELL:/home/jan # ethtool -S wlp2s0 NIC statistics:      rx_packets: 605591      rx_bytes: 695567571      rx_duplicates: 1      rx_fragments: 605486      rx_dropped: 248      tx_packets: 443918      tx_bytes: 608352571      tx_filtered: 0      tx_retry_failed: 0      tx_retries: 568      beacon_loss: 0      sta_state: 4      txrate: 300000000      rxrate: 300000000      signal: 230      channel: 0      noise: 18446744073709551615      ch_time: 18446744073709551615      ch_time_busy: 18446744073709551615      ch_time_ext_busy: 18446744073709551615      ch_time_rx: 18446744073709551615      ch_time_tx: 18446744073709551615  | 
					
I am especially interested in two parameters – txrate, rxrate. It look’s very good, the Intel 7260 wireless card achieved maximum connection bitrate for 2×2 spatial stream device in 802.11N standard – 300Mb/s Rxrate and 300Mb/s Txrate.
Now I launch iperf to see how the performance looks, when device passes real TCP traffic.
On the server I started iperf which is listening for incomming connections on 8888 port.
| 
					 1  | 
						iperf3 -s -p 8888  | 
					
I’m going to perform download test. On the client I launched iperf in 10 processes to simulate 10 tcp sessions:
| 
					 1 2 3 4 5 6 7 8 9 10 11 12  | 
						iperf3 -c 192.168.50.222 -p 8888 -P 10 -t 1000 -R [  4]  24.00-24.87  sec  3.50 MBytes  33.6 Mbits/sec                   [  6]  24.00-24.87  sec  1.85 MBytes  17.8 Mbits/sec                   [  8]  24.00-24.87  sec  3.42 MBytes  32.9 Mbits/sec                   [ 10]  24.00-24.87  sec  1.28 MBytes  12.3 Mbits/sec                   [ 12]  24.00-24.87  sec  2.52 MBytes  24.2 Mbits/sec                   [ 14]  24.00-24.87  sec  1.81 MBytes  17.4 Mbits/sec                   [ 16]  24.00-24.87  sec  3.11 MBytes  29.8 Mbits/sec                   [ 18]  24.00-24.87  sec  1.89 MBytes  18.1 Mbits/sec                   [ 20]  24.00-24.87  sec  1.73 MBytes  16.6 Mbits/sec                   [ 22]  24.00-24.87  sec  1.85 MBytes  17.8 Mbits/sec                   [SUM]  24.00-24.87  sec  23.0 MBytes   220 Mbits/sec   | 
					
and upload:
| 
					 1 2 3 4 5 6 7 8 9 10 11 12  | 
						iperf3 -c 192.168.50.222 -p 8888 -P 10 -t 1000 [  4]  19.00-20.00  sec  2.13 MBytes  17.9 Mbits/sec    0   29.7 KBytes        [  6]  19.00-20.00  sec  2.18 MBytes  18.3 Mbits/sec    0   31.1 KBytes        [  8]  19.00-20.00  sec  2.10 MBytes  17.7 Mbits/sec    0   29.7 KBytes        [ 10]  19.00-20.00  sec  1.97 MBytes  16.5 Mbits/sec    0   26.9 KBytes        [ 12]  19.00-20.00  sec  2.02 MBytes  17.0 Mbits/sec    0   26.9 KBytes        [ 14]  19.00-20.00  sec  1.97 MBytes  16.5 Mbits/sec    0   26.9 KBytes        [ 16]  19.00-20.00  sec  2.05 MBytes  17.2 Mbits/sec    0   28.3 KBytes        [ 18]  19.00-20.00  sec  1.96 MBytes  16.5 Mbits/sec    0   26.9 KBytes        [ 20]  19.00-20.00  sec  2.03 MBytes  17.0 Mbits/sec    0   28.3 KBytes        [ 22]  19.00-20.00  sec  2.10 MBytes  17.6 Mbits/sec    0   29.7 KBytes        [SUM]  19.00-20.00  sec  20.5 MBytes   172 Mbits/sec    0   | 
					
Wow! 🙂 220 Mb/s download and 172 Mb/s upload speed. In my opinion these are very good results for 802.11n standard.
Let’s check the 802.11AC with Aironet 2700. The connection status on Dell Latitude is showing:
| 
					 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25  | 
						interface Dot11Radio1 JAN-DELL:/home/jan # ethtool -S wlp2s0 NIC statistics:      rx_packets: 2458256      rx_bytes: 3658398341      rx_duplicates: 1      rx_fragments: 2458183      rx_dropped: 181      tx_packets: 472888      tx_bytes: 442820819      tx_filtered: 0      tx_retry_failed: 0      tx_retries: 3045      beacon_loss: 0      sta_state: 4      txrate: 866700000      rxrate: 650000000      signal: 222      channel: 0      noise: 18446744073709551615      ch_time: 18446744073709551615      ch_time_busy: 18446744073709551615      ch_time_ext_busy: 18446744073709551615      ch_time_rx: 18446744073709551615      ch_time_tx: 18446744073709551615  | 
					
Despite the short distance connection parameters were not as stable as in the case Aironet 2600. I do not know why Intel 7260 did not achieve full rate, which is 866Mb/s Rx/Tx Rate for 2×2 spatial stream client on 80 Mhz channel width. I saw 866Mb/s rates only temporarily. In Addition, there was a problem with the client connection which was hanging up every minute. I solved this problem by remove 802.11d configuration.
| 
					 1 2  | 
						interface Dot11Radio1 no world-mode dot11d country-code PL both  | 
					
I do not understand why it had such an impact. These settings works well in 802.11N.
But, how real speeds I can achieve in iperf test with Aironet 2700?
Download speed:
| 
					 1 2 3 4 5 6 7 8 9 10 11 12  | 
						iperf3 -c 192.168.50.222 -p 8888 -P 10 -t 1000 -R [  4]  72.00-73.00  sec  3.56 MBytes  29.9 Mbits/sec                   [  6]  72.00-73.00  sec  4.18 MBytes  35.1 Mbits/sec                   [  8]  72.00-73.00  sec  8.95 MBytes  75.1 Mbits/sec                   [ 10]  72.00-73.00  sec  7.93 MBytes  66.5 Mbits/sec                   [ 12]  72.00-73.00  sec  3.97 MBytes  33.3 Mbits/sec                   [ 14]  72.00-73.00  sec  6.93 MBytes  58.2 Mbits/sec                   [ 16]  72.00-73.00  sec  3.30 MBytes  27.7 Mbits/sec                   [ 18]  72.00-73.00  sec  4.19 MBytes  35.2 Mbits/sec                   [ 20]  72.00-73.00  sec  5.80 MBytes  48.7 Mbits/sec                   [ 22]  72.00-73.00  sec  3.86 MBytes  32.3 Mbits/sec                   [SUM]  72.00-73.00  sec  52.7 MBytes   442 Mbits/sec  | 
					
Upload speed:
| 
					 1 2 3 4 5 6 7 8 9 10 11 12  | 
						iperf3 -c 192.168.50.222 -p 8888 -P 10 -t 1000 [  4]  10.00-11.00  sec  3.55 MBytes  29.8 Mbits/sec   16   35.4 KBytes        [  6]  10.00-11.00  sec  3.40 MBytes  28.5 Mbits/sec   19   36.8 KBytes        [  8]  10.00-11.00  sec  3.37 MBytes  28.3 Mbits/sec   18   35.4 KBytes        [ 10]  10.00-11.00  sec  3.28 MBytes  27.5 Mbits/sec   26   31.1 KBytes        [ 12]  10.00-11.00  sec  3.79 MBytes  31.8 Mbits/sec   19   36.8 KBytes        [ 14]  10.00-11.00  sec  3.48 MBytes  29.2 Mbits/sec   24   32.5 KBytes        [ 16]  10.00-11.00  sec  3.52 MBytes  29.5 Mbits/sec   18   32.5 KBytes        [ 18]  10.00-11.00  sec  3.71 MBytes  31.1 Mbits/sec   22   38.2 KBytes        [ 20]  10.00-11.00  sec  4.15 MBytes  34.8 Mbits/sec   30   36.8 KBytes        [ 22]  10.00-11.00  sec  3.26 MBytes  27.3 Mbits/sec   17   33.9 KBytes        [SUM]  10.00-11.00  sec  35.5 MBytes   298 Mbits/sec  209  | 
					
442 Mb/s download and 298 Mb/s upload speed.
Nice, almost half gigabit per second. But I was expecting a better result in spite of twice the width of the channel and denser modulation in 802.11ac
That sums up.

Is it worth to upgrade wireless infrastructure from 802.11N to 802.11AC?
It depends on your needs. 802.11N is still fast enough for most applications, and it is stable. But if you work with tasks which are generating heavy network traffic, such as copying big files over local network, you should consider switch to 802.11AC.