Unlike NETGEAR, Linksys chose to enable the dynamic band steering that Broadcom built into its XStream technology. This means that clients can be moved between the two radios as signal levels change. Here is Linksys' description of how this works from their Reviewer's Guide:
Linksys Smart Connect band steering is dynamic. Devices are constantly monitored and moved to the 5 GHz band that will yield the best performance for all devices.
Devices with fast connection speeds are placed in the low 5 GHz band (5 GHz-1 radio). Devices with slow connection speeds are presumed to be legacy devices or located farther away from the router and are placed in the high 5 GHz band (5 GHz-2 radio) because it has slightly higher transmit power. If one of the 5 GHz bands experiences a high utilization rate, it will steer one of its devices to another band in order to prevent oversubscription.
A "bounce detection" algorithm ensures a device is not steered too often within a certain timeframe to prevent frequently moving devices between 5 GHz bands. Devices are steered only when they are idle so their activity (i.e. gaming, streaming video) will not be interrupted by the steering process.
Like the R8000, the EA9200 does not support 2.4 / 5 GHz band steering. Clients are only steered between the two 5 GHz radios.
I used the same devices that I used for the R8000 Smart Connect testing:
|Device||Type||Network Map ID|
|Moto X smartphone||1x1 AC||android-d4c097|
|NETGEAR R7000 in client bridge mode||3x3 AC||WLANTEST-STA
|Laptop with NETGEAR A6200 USB adapter||2x2 AC||x220i|
|iPad 2nd gen||1x1 N||SNB - iPad 2|
|iPod Touch 5th gen||1x1 N||Kyra's iPod|
Table 5: Smart Connect Test devices
The EA9200 was located in the wireless testbed upper test chamber with the door open. The bridge mode R7000 was in the lower chamber with the door closed. This allowed me to use the testbed programmable attenuators to control the signal and therefore the link rate of the R7000. I set 20 dB of attenuation so that the R7000 throughput wouldn't dominate the other AC devices. All other devices were located within 6 feet of the EA9200, outside the test chamber and all received a nice, strong signal.
I found the EA9200's band-steering was indeed dynamic. In my first set of tests, the 3x3 AC client insisted on connecting to the 5 GHz-2 radio along with the two N devices, despite a few power cycles. When I moved the two portable AC devices to another room to reduce their link rates, they nicely moved to the 5 GHz-2 radio. However, when I moved them back to their original locations, they stayed attached to the 5 GHz-2 radio. Note that all connections were idle when I moved devices around.
When I went to write up this review, I found I had neglected to set the attenuation for the R7000 3x3 bridge client. When I set everything up for a retest, all clients connected to the 5 GHz-1 radio as shown in the Network Map below. The R7000 bridge is shown in the map as a generic Network Device and the computer wired to it running the IxChariot endpoint shown as WLANTEST-STA.
Network Map - All clients on 5GHz - 1 radio
So I ran an IxChariot test and got 86 Mbps total downlink throughput. The plot's similar colors make it difficult to tell, but using the legend we see that the highest throughput is achieved by the 1x1 N 5th iPod.
Throughput - All clients on 5 GHz-1 radio - downlink
Running an uplink test yielded higher total throughput (107 Mbps) with the most going to the 1x1 AC MotoX this time.
Throughput - All clients on 5 GHz-1 radio - uplink
After a power cycle, clients came up attached as they "should" be, i.e. the "slow" N devices on 5 GHz-2 and "fast" AC devices on 5 GHz-1. Here's the Network Map to prove it.
Network Map - All clients on 5GHz - 1 radio
The downlink test in this configuration yielded 237 Mbps total throughput with the 3x3 AC device getting the lion's share. This is a 2.8X improvement (86 to 237 Mbps).
Throughput - N clients on 5 GHz-2, AC on 5 GHz-1 - downlink
Uplink throughput improvement was more modest, only 1.6X (110 to 169 Mbps).
Throughput - N clients on 5 GHz-2, AC on 5 GHz-1 - uplink
Note to Linksys: A tablular presentation of device connections like NETGEAR has in the R8000 (shown below) would have made it a hell of a lot easier to track device comings and goings while testing Smart Connect. The Network Map is pretty, but the font size is waaay too tiny on the wireless device icons.
NETGEAR R8000 device connection table
I've said it before: $300 is a lot to spend on a router. And judging from comments in the Forums, people are starting to push back. For that money, people expect to see huge performance increases and, frankly, AC3200 products don't deliver. Yes, they do provide higher total throughput if you have more than a couple of 5 GHz devices in constant simultaneous use. So you should see fewer problems if you have a lot of video streams going, but only if devices are using 5 GHz.
I give Linksys credit for enabling the dynamic band-steering capabilities of Broadcom's XStream technology. But even in my limited experiments, I found "Smart Connect" wasn't very. Even with strong signal levels, devices with very dissimilar link rates ended up on the same radio. Smart Connect is going to have to get a lot more intelligent for people to cough up the extra dough.
Between the R8000 and EA9200, I'm giving the nod to NETGEAR. Yes, the R8000's Smart Connect implementation uses static device assignments that aren't changed until the router is power cycled or devices disconnect and reconnect, but it provides similar total throughput gains. It also has better 5 GHz performance, range in particular, a much better feature set including OpenVPN and full outbound port filtering to better control internet service access. And street price is already headed down, making it currently $25 cheaper than the EA9200.
The most compelling reason, at least for router geeks, is that DD-WRT is available for the R8000, albeit in early form. At least Linksys hasn't advertised OpenWRT support for the EA9200; they messed that up enough with the WRT1900AC.