Media bridge AC speed identical to N speed, can't break 155mbps

[vixro]

I am having trouble increasing the speed of my home network. I recently purchased 2 RT-AC66U's to increase the speed to my MyBookLive in the main living room. My previous real time transfer speeds were about 11.5-12.5MB/s to my mybook live over the wireless-n 5ghz connection I had created with my netgear router and e4200 bridge. After setting up the 2 AC66U's, the transfer speed increased to about a 16MB/s write and a 20MB/s read. That's it! The drive can transfer at 205mbps (25.6MB/s) write and 525mbps (65.6MB/s) read when connected via cat 6 ethernet cable.

If I set the router at a lower channel on the 5ghz line I will see a decrease in write speed by HALF in comparison to the upper channels but the read speed does not change. Changing the transmit power from 40-200 in small increments also has absolutely no change in the speed of the transfer. Using inSSider shows only 2 other close by 5ghz signals on channels I don't use, but have also tried to use just in case.

I have tried a 3rd AC66U and replaced the router as well as the bridge without any change in transfer speeds.

I have moved both routers to different locations with about a 1MB/s change in speeds both up and down.

I have tried the .270 official firmware, as well as the .270.26b merlin firmware, as well as the .354 official and .354.29b merlin firmwares. The .270 firmwares increase the transfer speed by about 10mbps on the read/write.

I have tried a 20mhz channel, 40 mhz channel, 80 mhz channel, and auto channel.

20mhz 122mbps read / 88mbps write.
40mhz 135mbps read / 90mbps write.
80mhz 155mbps read / 105mbps write.

I have tried wireless mode auto as well as wireless mode N only. The transfer speeds were about 30mbps different.

I have tried no encryption as well as WPA2-Personal encryption, speeds were around the same.

I did also notice a thread where a user suggested using the router as an AP instead where they were seeing the same throughput that I was before the change. I cannot change my main router to an AP because my internet access is authorized via PPPoE. I tried this anyway just to test transfer speeds across my network and they were IDENTICAL to my previous speeds.

I do not understand how people are getting these 500+mbps transfer rates whatsoever. I have not even come close to breaking 160mbps once, let alone 300 or 500 or higher. I have moved the routers anywhere from 1feet apart, to 3, to 10, to 50. The speeds never change more than 30mbps.

The absolute max speed based on firmware changes, channel settings, mode changes, and prime location that I have achieved is 112mbps write and 156mbps read. This is in relation to a 4MB size packet. A 1MB size packet gets me almost 200mbps read, but it seems to just be a burst speed.

Why did I spend $380 on this media bridge for a 33% increase in speed over my previous setup? What am I doing wrong?

My CCNA certification that I earned 10 years ago obviously didn't help me whatsoever in real life.

 

[thiggins]

Where are you seeing reports of 500 Mbps throughput in real world use?

In my experience, the speeds you are reporting are reasonable for draft 11ac.

 

[vixro]

What speed test did every benchmark I read across the net use so I can replicate their results?
Why is my n performance almost identical to my ac performance on the same router when all of these benchmarks show a huge increase over wireless-n on the same router? (my speeds look in line for wireless-n but not ac).



http://www.pcworld.com/article/2621...est_802_11ac_router_on_the_market_so_far.html

http://reviews.cnet.com/routers/asus-rt-ac66u-802/4505-3319_7-35406080-2.html

http://www.smallnetbuilder.com/wireless/wireless-reviews/32090-asus-rt-ac66u-wireless-retest

http://forums.smallnetbuilder.com/showthread.php?t=11166

 

[thiggins]

I can speak primarily to the SmallNetBuilder benchmarks.

First, you need to consider that Windows filecopy has a higher overhead than the IxChariot test tool we use. So filecopy is going to tend to show lower throughput.

The iperf/jperf tool used in the PC World review is very dependent on Window Size, Buffer Length and MSS settings. I don't think the reviewer specified the settings he used. The CNET review doesn't state the performance measurement method.

The new chamber-based test method we use is showing much higher throughput than the previous over-the-air method. There are two factors here. One is the ASUS PCE-AC66 adapter, the other is the Octobox Multipath Emulator (MPE).

The adapter connects via PCI-e slot, which has higher bandwidth than Gigabit Ethernet. Its driver also is different than the driver embedded in the ASUS router.

MIMO depends upon decorrelation between the streams to increase throughput. The higher the decorrelation, the higher the throughput. Decorrelation in an over the air environment primarily depends on multipath, which is highly dependent on your physical space.

Because the MPE is an RF tight controlled environment, it produces very high decorrelation, which will tend to push throughput higher.

The easiest thing you can try is a different test tool. You can try iperf/jperf, which can be a pain to set up. But there are easier to use tools you can try.

And just to state the obvious, you should make sure that the two computers you are using to run the tests are capable of running at the speeds you are hoping to measure using an Ethernet connection.

 

[vixro]

Could not achieve anything over a consistent 19MB/s transfer rate so I returned the routers to Amazon and purchased powerline networking. I don't get the 19MB/s I had with the wireless AC but I get a consistent 17MB/s and video streaming works great now. All for $70 instead of $380.

 

[roscolo]

Could not achieve anything over a consistent 19MB/s transfer rate so I returned the routers to Amazon and purchased powerline networking. I don't get the 19MB/s I had with the wireless AC but I get a consistent 17MB/s and video streaming works great now. All for $70 instead of $380.

If all you were trying to achieve is video streaming, you didn't need AC speeds anyway. N speed is more than adequate.

 

[vixro]

Wireless had glitches in the stream. Even though I was the only 5ghz stream within 500 feet. Different channels and faster speeds did not matter, 7MB-9MB/s rate videos would not stream properly. I thought AC speeds would help, especially the "low latency" 66u, but wired was the only thing that worked. Even if is through the power line.

 

[thiggins]

Wireless has neither low nor consistent latency. The throughput speeds you see quoted are average values. N and AC wireless frequently has large (50% or more) and long (1-2 second) "dropouts" / glitches.

If you want reliable video streaming over wireless, you must drop the bitrates. There are some damned good encoders today that can provide very decent HD streams. But if uncompressed Blu-ray rips are what you want, go Ethernet.