Wireless speeds are a big lie?

[techs]

I see a LOT of wireless routers and wireless network adapters as a computer technician. I also live in an area where many people have 80 Mbps internet connections.

I have been keeping track of the actual throughput and it is pretty awful for wireless-N. In fact the highest I have seen repeatable is 52Mbps for a card that says its connected at 300 Mbps.

I don't know how much overhead goes into error correction and encryption protocols but thats still pretty low. Does the USB eat up bandwith?

Am I the only one seeing this? Is this actually the maximum that wireless N is capable of real world?

 

[BrightCandle]

Wireless rarely gets close to its theoretical speeds for any one individual user. An N user is kind of lucky to get 100Mbit/s on a 300Mbit/s connection and then only really in a fairly uncontested airspace.

Its not a lie but they certainly struggle to get near peak performance unlike wired connections which very reliably get within single percentage points of peak.

 

[MontyAC]

Actual speed is generally less than 1/2 of the stated speed. From PCMag:

You however, will never see those speeds. Issues such as channel overlap and interference means a router in a typical home or office environment will never reach these theoretical speeds touted by vendors. When testing at PCMag, which has a real-world testing environment, if we see speeds close to half of what a vendor says a router is capable of, that's excellent bandwidth!

http://www.pcmag.com/article2/0,2817,2347539,00.asp

 

[JackMDS]

Most of the numbers are statement of the inner work of the Chipset.

As an example, telling One the size of the car's engine does not necessarily correspond to the speed that the vehicle can achieve.

The statement about a distance of 300 feet indoor is like stating how much a car can do on a special straight specially paved down hill road. I.e., the condition of the indoor are not disclosed.

If you put an Access Point in a space like a Boeing 787 hangar, you can call it indoor and might do 300 feet.

In other words, you can Not be successful in life without understanding what is behind the Marketing and how numbers and words are used to confuse the customers.

 

[Red Squirrel]

My wireless AP says it can do 300Mbps, which makes me laugh, because it has a 100mbps uplink. Most of them simply blatantly lie. Well, it's more indirect lying. The AP itself may support 300mbps, but only from the client to the AP. it's kinda like when people pay for 5mbps DSL in rural areas. They get 5mbps right up to the AFC, and then it's a T1 uplink to the CO. :awe:

 

[techs]

This has become a problem for me. Despite my router being 3 feet from my wireless adapter I max out at about 38Mbps. Yet my wired nic gets 77 Mbps.

 

[spidey07]

It's not a lie. You are indeed connected at a 300 Mbs data rate. This is the real data rate and it's clocked as such.

Wireless is half duplex, throw in management frames and retransmissions and your actual thruput is less than half of data rate.

Wireless is nowhere near the performance of wired and never will be. It's just limits of physics.

 

[spidey07]

This has become a problem for me. Despite my router being 3 feet from my wireless adapter I max out at about 38Mbps. Yet my wired nic gets 77 Mbps.

3 feet is too close. Back them apart some and orient antennas per the sticky. Use 5 ghz with 40 meg channel for optimal performance.

 

[cmetz]

It's not a lie. You are indeed connected at a 300 Mbs data rate. This is the real data rate and it's clocked as such.

The equipment is (might be) capable of selecting and using a 300Mb/s basic data rate. In nearly any real-world scenario, however, the equipment will select a lower data rate based on signal quality feedback, because a reliable link at the 300Mb/s data rate can't be achieved under your real-world conditions. So your "300Mb/s" N router will quickly jump down to a 50-100Mb/s actual selected data rate. (might be - both ends need to be N, and for "300Mb/s" you need HT40 support as well, and then we're also assuming the device can bridge/route/NAT at that throughput, and so on...)

Then all the other issues like duplex, management frames, ACK clocking, TDM among stations in STA mode, and gap times come into play and deliver even less real-world performance.

techs, in networking, it is very common to see "marketing numbers" - that is, numbers that are not completely outright LIES, but have very significant gotchas to them that require interpretation - and a very cynical eye - to understand what those numbers mean in the real world. This problem is worse in the consumer end of the segment, where there is less serious testing by trustworthy authorities, and the buyers nearly universally don't have the skills, gear, knowledge, and time to do real lab testing.

You're just at your favorite big box store, and one vendor's box says N150, one says N300, and one says N600 - so which one is better? The bigger number, right? Sadly, it's just not that simple. The N150 device might have superior RF performance that does better in the real world than the others.

As far as your specific scenario goes, if you're using HT40 on an RF channel pair that's otherwise clear (change channels and check), and you're getting 38Mb/s throughput, that does sound lower than you should realistically expect. I would suggest trying another adapter on the client end and trying another wireless router, and seeing if changing those out gives you a better result.

What devices do you have, and how are you testing performance?

 

[JackMDS]

I can see three options.

1. If it is Really important, One should shave some time of TV watching and instead make him/her knowledgeable about the issue at hand.

2. One can concentrate and do the best in his carrier of choice, make a lot of money and then will have the flexibility to keep spending until the best solution can be found.

3. Useless Drama Queening/Kinging on on FaceBook, Tweeter, On line forums and the like (which seems to be the preferred solution of most people).

I can help with option 2.

Get two of these - http://www.newegg.com/Product/Produc...82E16833320091

Flash them with DD-WRT make One your main Router.

http://www.dd-wrt.com/wiki/index.php/Asus_RT-N66U

Use one as your main Router.

Then Place the second Router next to the Desktop, and configure it as a Wireless Bridge feeding the computer via the Ethernet jack.

http://www.dd-wrt.com/wiki/index.php/Wireless_Bridge

 

[techs]

As far as your specific scenario goes, if you're using HT40 on an RF channel pair that's otherwise clear (change channels and check), and you're getting 38Mb/s throughput, that does sound lower than you should realistically expect. I would suggest trying another adapter on the client end and trying another wireless router, and seeing if changing those out gives you a better result.

What devices do you have, and how are you testing performance?

I currently have a Netgear WNDR3700 router. I have tried it with at least 8 different wireless card. Four from my shop and four notebooks in for service.

I previously had a Buffalo, model number I don't remember, maybe N300, since it was in the shop to be set up for a customer. Both claimed 300Mbps speeds.

I have tried numerous channels.
Strangely the highest speed of about 38 Mbps was attainable with five of the wireless cards, the rest getting less. And this highest speed was the same with both routers.

I actually have a TP-Link card on the way for another customer, which I will test.

 

[JackMDS]

About 40Mb/sec. (yielding LAN transfer of about 4MB/sec. B=Byte b=bit) is the most common Bandwidth that regular 802.11n using the 2.4GHz band would provide.

The way to improve on it is to use Dual band Wireless source and connect the Wireless N to the 5GHz band.

A good full Dual Band (Wireless Router and client card) can go up to about 30MB/sec. LAN Transfer which is the functional output of 300Mb/sec. and above Wireless Entry Level Network.

Any type of transmission is highly dependent on a Good impedance matching Antenna. Most of the consumer Wireless (especially Laptops) include really crappy" pieces of wires (aka Antenna).

Rest assure that when an OEM manufacturers test its Wireless chipset they are using the best Antenna for the test bench assembly. I doubt the any of the Brands that box the chipset into a market Wireless product really bothers to thoroughly test the device's performance., they just coping the specs from the OEM website.

 

[dawks]

Wireless is more an art than a science. There are so many things that can affect its performance, from drivers/settings, cordless phones, cell phones, bluetooth, cellular saturation, to the shape/size of a room and its materials, and the electrical wiring in the wall, all the way to outter space 'weather' and radiation.

Also check out Wireless antennas Azimuth charts if you're curious. Each antenna transmits and receives in a different manner. And yes, being to close can have an effect on the signal. Even having something thats not blocking line of site, but is near line of site can cause signal loss.
http://www.ubiquitiworks.com/airMAX-Omni.asp click on antenna in the middle of the page, just below pricing to see some Azimuth charts.

Also note, if you are looking at hitting high speeds from the internet, it might in fact be your router that's unintentionally limiting speeds. Some routers don't have fast enough processors to do Stateful Packet Inspection SPI at high speeds.
I know some of the popular/older Linksys 54G series routers would have issues with internet connections beyond 65mbps.

 

[Harrod]

it's kinda like when people pay for 5mbps DSL in rural areas. They get 5mbps right up to the AFC, and then it's a T1 uplink to the CO. :awe:

I hate working on AFC equipement, it's the stuff that was origonally intended just for voice and somehow they were able to adapt them to do data as well. Do you ever have them lock up and stop passing traffic to where you have to switch protection or toggle the software in them?

 

[suncare]

It's not a lie. You are indeed connected at a 300 Mbs data rate. This is the real data rate and it's clocked as such.

Wireless is half duplex, throw in management frames and retransmissions and your actual thruput is less than half of data rate.

Wireless is nowhere near the performance of wired and never will be. It's just limits of physics.

Pretty agree.

 

[techs]

Well, I am hoping 802.11ac can actually provide a performance bump.
This review is somewhat encouraging:
http://www.tomshardware.com/reviews/wi-fi-802.11ac-router,3386.html

Not bad for draft.

 

[Red Squirrel]

I hate working on AFC equipement, it's the stuff that was origonally intended just for voice and somehow they were able to adapt them to do data as well. Do you ever have them lock up and stop passing traffic to where you have to switch protection or toggle the software in them?

I don't work on them myself, I just monitor them among lot of other stuff and can log in to them, but I can't recall having them do that but I've only been there a year. I just know that the parts are really hard to get for them, such as the rectifiers. had one sitting in alarm on our DMS for months while they tried to source out a rectifier.

Got to hand it to them though, to be able to operate in these extreme temps they're built fairly robust. Most of em are in cabinets on poles or on the side of a street etc. Same with some DMS equipment like Star hubs. Most of the issues we see are usually related to the cable plant and not the equipment itself. Especially in older areas that use the paper cables. A bit of water gets into those and all hell breaks loose. So much stuff the average person does not realize that goes into making the phone/internet possible.