http://www.smallnetbuilder.com/wireless/wireless-features/30938-dont-mess-with-wmm enabling WMM should increase wireless performance.
Although this article is rather old, how is the RV220W expected to behave?
In my case, speed measurements show a clear decrease in wireless performance when WMM is enabled.
WMM enabled - max 15MBits/s (ticked WMM checkbox - default settings)
WMM disabled - max 40MBits/s
Any ideas? Thanks in advance.
The packets are going to be "arranged" so that Priority of traffic type is gained and would i guess be a hit on performance since more CPU time to analyse the packet structure and arrange it in WMM priority..
I cannot see why WMM would ever increase throughput. it just helps prioritize.
I guess its like turning on IPS, this slows throughput but increases security.
Frame Burst does in most cases increase throughput.
Does it really take so much processing power to prioritise the wireless packets that results in an over 50% drop in throughput rate?
I guess this depends on the amount of wireless devices as well. do forget each device added to a wireless cell shares the bandwitdh. the cpu is only 400Mhz and has many features to process all at the same time.
I dont understand why Network devices dont have Dual cores now since nearly all new Mobile phones have dual cores and at least 1Ghz processors! its like the manufacturers live in the past with technology of SMB devices.
I have the impression that the wireless card/subsystem in a router has its own separate CPU which does the processing of the wireless packets. The main CPU for this router (Cavium CN5010), although it is running at only 400MHz, it is supposed to be very fast for networking applications which is verified by the tests conducted here.
I found his review very flawed since it does not replicate day to day usage of WebGui lag, random reboots, dropped WAN connections to say the least.
if you can confirm this thesis with other makes of wireless adaptors then raise an issue with SBSC.
I have found no difference in CPU usage (25% - 30% on average in both cases).
I did more tests:
It seems that only WAN <-> Wireless is affected. The speed LAN <-> Wireless is the same regardless whether WMM is enabled or not.
...this seems to be a general problem. Enabling WMM leads to a decreased download speeds under certain conditions. So I keep it disabled.
Apple devices seems to be affected (probably they works as they are supposed to). Here are a few sources:
Or it can be a bad driver:
Hi, just to throw some food for thought. WMM is WI-FI MULTIMEDIA. WMM is like a QoS but not a bandwidth throttle. Like QoS it identifies traffic based off the same information background, best effort, video, and voice (802.1d). It is very much expected to see a decrease in background and best effort traffic, especially in conjunction with using voice and video applications. This will directly correlate in to a decrease of MBPS for such classified traffic. Also keep in mind, that it is not WMM's job to set priority policies, it is application or device which is sending the data.
The thing is, wireless is not the same as an ethernet wire. Comparing a Cat 6, full duplex 1000 meg link is definitely apple and oranges compared to almost any wireless connection which is why you do not see a dramatic or any decline at all on a LAN with a basic QoS for wired devices. I'm also not sure how the RV220W works inside, but if it's like a switch, there should be a TCAM where the WMM resides, as that is true for QoS on the small business switches. I don't know this answer.
Please rate helpful posts
I still have the concern that WMM is somehow broken.
the IP packets are classified by their Differentiated Services Code Point (DSCP) value (0x00 - 0x3F, 6bit).
In general there are only a few classes (you named it) to simplify things:
The classification takes place inside wireless client (the application or the operating system has to set the DSCP value) or inside the router ("CoS to DSCP" in conjunction with QoS).
Depending on the DSCP value and wireless load, wireless packets are delayed by the receiver (e.g. the router in case of uplink traffic) by either not sending or delaying ACKs. The result is that other stations in the network get a higher probability to access the medium. But dropping ACKs will and delaying ACKs can result in a complete loss of the transmitted data. This is bad because already transmitted data needs to be sent again.
So, again my question: Why should enabling WMM slow down my wireless downlink speed if no other is accessing the same medium?
The settings on my laptop are the same, only the router config has changed.
The reason is affects downlink is because the traffic from the sender (uplink) is already classified when it is received by the receiver (the router). The inbound traffic, lets say over the WAN, this traffic may or may not be classified and needs to be reclassified based off the headers which is not the job of the router's WAN port.
I would strongly suspect that if you do a LAN to LAN file transfer there is little or no difference.
Please rate helpful posts
Just enabled WMM again and everything gets incredible slow. Doesn't matter what and where I download from (vimeo, LAN server, speedtest, ...) and which wireless client I use (Windows Laptop with 11g, Macbook with 11n network card).
In a former posting I said that only WAN<->Wireles is affected, which I cannot hold true anymore.