I did not use the old extender test procedure, but instead created a new test suite using the Revision 10 wireless testbed. The new set of tests presents a much more comprehensive view of extender performance. Version V126.96.36.199 firmware was used for all the following tests.
First, let's get storage performance out of the way. I used the router storage performance test setup, but ran only NTFS tests. I measured a respectable 15.7 MB/s write and 30 MB/s read with the testbed computer connected directly to an EX8000 gigabit Ethernet port.
Throughput vs. Attenuation
Since the EX8000 can function as a tri-band AP, I first tested throughput vs. attenuation performance (or rate vs. range, aka rvr), using the Revision 10 Throughput vs. Attenuation test with the EX8000 in the large octoScope test chamber with turntable.
The EX8000's 2.4 GHz radio was set to channel 6 and "Up to 400 Mbps" (40 MHz) bandwidth. Keep in mind the octoScope Pal test client is limited to 20 MHz link rate, however. The EX8000's low band 5 GHz radio (Ch 36-48) was set to Channel 40 and "Up to 866.7 Mbps" (80 MHz) bandwidth. The 5 GHz high-band radio (Ch 149-161) was set to Channel 149.
NETGEAR EX8000 in chamber
I'm not including the results in the Wireless Charts because that's not the EX8000's primary function. But I grabbed the results from the current top-ranked Ubiquiti AC Pro 3x3 AC1750 class access point for comparison, which is valid because it used the same Revision 10 process. Remember, all the following tests were done with a 2x2 802.11ac client.
2.4 GHz downlink shows the difference between the 2x2 NETGEAR and 3x3 Ubiquiti. The EX8000 has both higher average throughput and maintains connection longer. Both point to superior range/coverage for the NETGEAR.
NETGEAR EX8000 throughput vs. attenuation - 2.4 GHz down
2.4 GHz uplink shows mixed results with the Ubiquiti having a slight lead with stronger signals (lower attenuation), but the NETGEAR taking the high road as signal levels drop.
NETGEAR EX8000 throughput vs. attenuation - 2.4 GHz up
5 GHz downlink again shows the two-stream EX8000 with a distinct advantage over the three-stream Ubiquiti. Also included are the results from a run with the same octoScope 2x2 client associated with the EX8000's 4x4 5 GHz high-band radio. The extra two streams really make a difference, even though the client link rate is the same.
The additional link gain provided by the extra two streams makes the difference. This is why you see four-stream products at the top of the Router Ranker. The low 0 dB attenuation result is likely a glitch; I ran the Channel 149 test only once.
NETGEAR EX8000 throughput vs. attenuation - 5 GHz down
5 GHz uplink results are surprisingly low. I've seen this effect in a few recent reviews and found it's due to my use of only a single connection in the Revision 10 process. I ran a quick experiment and found if I increased the number of iperf3 connections, throughput would increase to look more like downlink. (The iperf3 command line sets a 32 KByte buffer size with 2 MByte TCP window.) The four-stream Channel 149 connection still produces the best result in the group.
NETGEAR EX8000 throughput vs. attenuation - 5 GHz up
So it looks like the EX8000's radios make it a pretty good AP. For extender performance, I paired the EX8000 with a NETGEAR R7800 Nighthawk X4S, a Qualcomm-based AC2600 class 4 stream MU-MIMO enabled router. Since the R7800 is both top-ranked and Qualcomm-based, it should provide "best evidence" of the EX8000's capabilities.
A Wi-Fi extender is only as good as its wireless backhaul, so that's what this portion of our test focused on. The basic approach was to place the R7800 in one octoScope RF chamber and the EX8000 in another. The two are connected via a programmable attenuator that was used to ramp down the signal between the two, simulating performance at different "distances".
The octoScope Pal test client was also connected to the EX8000 RF chamber's antennas via a splitter and attenuator. But the test did not change the Pal attenuator, leaving it at 0 dB attenuation. The setup is shown below.
Extender test configuration
Here's the R7800 parked in one of the octoBox 18 chambers...
NETGEAR R7800 in chamber
...and the EX8000 in a second 18" chamber.
NETGEAR EX8000 in chamber
Throughput between router and extender was measured with different connections to the EX8000: 2.4 GHz wireless—5 GHz wireless and gigabit Ethernet—with uplink and downlink traffic run in each case. One set of tests was run with the R7800's 5 GHz radio set to Channel 40, causing a two-stream backhaul connection; the other was with the R7800 set to Channel 149, resulting in a four-stream link. Each connection type was tested separately.
The first composite plot compares how downlink throughput with a 5 GHz two-stream connection changed for each connection type, as the "distance" between R7800 and EX8000 increased, i.e. as attenuation increased. (Downlink is R7800 to EX8000; uplink is vice-versa.). The difference between the Ethernet and 5 GHz backhaul lines is fairly small, meaning that the EX8000 is pretty efficient in passing backhaul throughput along. 2.4 GHz throughput is obviously limited by the bandwidth of the client connection.
NETGEAR EX8000 backhaul vs. attenuation - downlink - Ch 40
The next plot compares throughput with the same setup, but now uplink traffic changed for each connection type. 2.4 GHz throughput is again limited by the lower bandwidth of that connection. Ethernet and 5 GHz throughput again track pretty closely.
NETGEAR EX8000 backhaul vs. attenuation - uplink - Ch 40
Next, the backhaul link switched to four-stream using Channel 149 and downlink throughput measured. Note the Ethernet-connected bandwidth increases even more at low attenuation values (higher signal levels), but the 5 GHz wireless connection can't follow because it's only a two-stream connection. 2.4 GHz is again limited.
NETGEAR EX8000 backhaul vs. attenuation - downlink - Ch 149
Finally uplink with four-stream backhaul shows Ethernet bandwith staying about the same as with two-stream backhaul and both 5 GHz and 2.4 GHz throughput about the same as downlink.
NETGEAR EX8000 backhaul vs. attenuation - uplink - Ch 149
During testing, the attenuation range eventually caused the R7800 and EX8000 to disconnect. When that happened, I never saw the connection quickly switch over to using 2.4 GHz backhaul. I was able to force the EX8000 to use a 2.4 GHz backhaul link by setting 51 dB of attenuation between it and the R7800 and waiting a minute or so for backhaul to come back up on 2.4 GHz. When that happened, the EX8000 GUI showed a 120 Mbps 2.4 GHz link rate and I measured around 35 Mbps throughput via a 5 GHz wireless connection.