Print

Please login or register.
1 Hour 1 Day 1 Week 1 Month Forever
Login with username, password and session length

[Lucid]

When I transfer files from my DND-323 to the PC (DWA-556) the speed in which it takes to transfer a 500MB vid is around 7.5-8MB/sec. I have placed the router on first floor in an unobstructed area. The PC is nestled between my desk and fileing cabinet on the second floor on the opsotie side of the house (22ft Away in a straight line frlom the the router to the PC. I have moved the PC, changed the wire positions, diabled the firewall SPI (which increased the speed from 4-5MB/sec to 7.5-8MB/sec.

The router is running ona mixed mode b,g,n. I have tried N only and get EXACTLY the same results with bandwidth set to Auto 20/40mghz and only at 40 or 20mghz. I have also locked the channel to the least used channel (10). I do notice that there is interference but I was getting approximately 9MB/sec when I have my pc wired at 1000GB with and without jumbo frames at 4000.

I don't belive that 8 MB/sec is fast enough. I swaer I used to get 20MB/sec on my old wored network. This is why I upgraded to the DWA-556. I thouhg mayye the wire on my lan (which are new CAT6) may be the issue.

I have followed the stick not on achieving 300 MB/s connections and windws does report this but I find that the router shows a very different picture. With current router settings I am setting this:
802.11n (2.4GHz) 240MB/s Signal strength:66

Currently transfering a file at 7.5-8MB/s is approximately 25% usage of bandwidth. I believe 20MB/s should be acceptable.

I do notice the following in my wireless stats (clear before transfer of file):
Sent: 387607
TX Packets Dropped : 0
Received : 173268
RX Packets Dropped : 802
Errors : 128

Is anyone else finding similar results? Are there any methods of increasing bandwidth usage?

TIA

Lucid
 

[funchords]

1 MB/s throughput is 8+ Mbps.  8 MB/s is ~65+ Mbps.  That's 3x the performance you can expect from 802.11g.  I'm impressed, that would rival any 100 Mbps wired network. 

If 20 MB/s file transfers are the goal over wireless, forget about it until you go three-stream dual band (maybe http://www.dlink.com/press/pr/?prid=460 for example).  Your old wired network must have been a gigabit network and it probably  had some really fast SATA drives.

Some things to try -- in general --

1. Ensure transmission rate is set to Auto on both the NIC card and the access point.  Do not lock a rate to avoid fallback.  Allow it to find the most efficient rate for throughput (your real goal), not data rate.
 
2.  On the computer end, orient the antennas are oriented similar to http://www.gadgetpeople.co.nz/images/products/DLink_DWA-556_135184.gif (That's probably good enough. I'd prefer the middle antenna to be angled 45 degrees directly right or left, then the other two pointed so that all 3 antenna tips are about equa-distant like the legs of a tri-pod).

3.  No obstruction within 5 inches (one antenna length) of an antenna tip

4.  No metal obstruction between the card's antennas and the AP's antennas.

[Lucid]

After defaulting the DNS-323 and the Router my Gigabit lan knocks up to 23-24MB/sec solid. It was 12 or so not too long ago. The reset definately helped. I moved the router further frlm the wall using the settings for 300mb/s link. I turned my computer 45 degrees to allow a clear line of sigth (so to speak) and now achived the 300mb rate on the DIR-655 wireless info page. However, transfers are still 7MB/s.

I am going to try you advice but in respect to acheiving the 20MB/s over wireless I was assured that this was possible.

Cheers,
Lucid.

[Lucid]

I started a new thread becuase I am unsure if the question here is relavent to this one:

http://forums.dlink.com/index.php?topic=3629.0

[funchords]

I am going to try you advice but in respect to acheiving the 20MB/s over wireless I was assured that this was possible.

Sorry -- 20 MB/s would be 160 Mbps ... that will not happen with a DIR-655 which only operates on the 2.4 GHz band.  You can probably get around a third of that -- around 7 MB/s -- which you are.

I've not recently seen a wireless anything in the field do significantly more error-free throughput than a third of its datarate on access-divided spectrum (where nodes have to listen for and hold off for other nearby devices). 

In the 2.4 GHz band, where an actually clean 40 MHz signal and actually using ShortGI is "optimistic" (practically impossible), you're looking at a practical sustained maximum of a third of 130 Mbps.  That falls right in there with the results that you're getting.  Anything above that is bonus

Your speeds with the DIR-655 (rates are best case):
802.11g + 4 more subcarriers = 58 Mbps
Forward error correction = 65 Mbps
2 streams of MIMO* (multiplies above) = 130 Mbps

What you really won't access very often (rates are best case):
ShortGI -- (too many incompatible stations on 2.4 GHz) = add 7.2 Mbps before multiplying in MIMO
40 MHz -- (too many incompatible stations on 2.4 GHz) =  multiply by 2 before MIMO

MIMO is extremely sensitive to the environment and uses reflections to its advantage. Slight movements of a laptop (thus moving the antennas) can negate it. People walking in the rooms of the transmitter or receiver, even if they are not walking directly between the path of the antennas, can provoke a retransmission and reduce throughput.

[Lucid]

 If I knew this I wouldn't have bought the adapter. Man....7MB/s is crap. I like my 25MB/s on my gigabit. I'm going ot have to return this product.

Thanks for the detailed info.

Lucid

[funchords]

You're welcome. 

I actually think the capabilities are pretty great.  But even as wireless advances will continue, so will physically-connected (wired) speeds.  I don't know that we'll ever see the day where wireless equals or passes wired throughput speeds.

[Lucid]

The following is advice given from Lycan when I first noticed my wired speed problems. He advised to used WirelessN given my situation and that wireless N achieved fast results over wired....

First putting the router in the basement is a horrible idea. The router should be centrally located in the house.
N speeds tout 300mbps. Thats mega BITS per second. Average data throughput on a decent N network is about 32MBps. A wired connection can achieve 80 mbps, so around 10-15MBps. Give or take.

MB = Megagbyte
Mb = Megabit

Read this quote again....

a decent N network is about 32MBps.

I think I should get my money back after all this.

Cheers,

Lucid

[funchords]

32 MB = 32 * 8 bits/byte = 256 Mbps (without protocol overhead, without sharing the channel, without errors)

Look, I can't blame anyone -- except the industry -- who believes this [exploitive deleted].  Go to the DIR-655 web page right now, I read this as the FIRST product feature:  "Draft 802.11n Speeds Up to 14x Faster than Wireless G*"

Note that asterisk -- it's the symbol of the generation for "I'm crossing my fingers." It refers to the mousetype at the very LAST that says,

* Maximum wireless signal rate derived from IEEE Standard 802.11g and draft 802.11n specifications. Actual data throughput will vary. Network conditions and environmental factors, including volume of network traffic, buildings materials and construction, and network overhead, lower actual data throughput rate. Environmental factors will adversely affect wireless signal range. Wireless range and speed rates are D-Link RELATIVE performance measurements based on the wireless range and speed rates of a standard Wireless G product from D-Link. Maximum throughput based on D-Link draft 802.11n devices

Well the specifications talk about stuff that doesn't apply to you.  Something more than 9 out of 10 of us will not be able to get 40 MHz channels, necessary for half of that claim.  Some large number of us won't be able to do spatially-differentiated MIMO. Some large number of us won't be able to do 400microsecond GI.

As for 2.4 GHz N-routers go, the DIR-655 is an awfully good one -- the capability of which 90%+ of us can never realize because we live too close to other signals that continuously prevent its use.

Now the industry will correctly point out that network equipment is always sold describing its best-case environment.  1000 Gbit equipment seldom results in 125 Mbps throughput owing to environmental, protocol, bus-speeds, CPU engagement, etc. etc. etc.  We professionals understand why the industry does this because we take it all into account. 

Drivers of cars are similarly savvy.  They buy a car capable of 0-60 in no time and speeds up to 125 MPH and braking that stops on the 10th of a cent, but they also know they'll probably never do those things where they drive.  That works for cars because there are a lot more savvy buyers of cars then there are savvy buyers of networking gear.

The industry, not just D-Link, but especially the members of the 802.11n certified-to-a-nonratified-evolving-standard-so-that-nothing-really-ever-gets-finished-nor-can-be-trusted cartel, owe the public a duty of setting more realistic expectations in their marketing.

I don't blame you for being angry.

[Lucid]

I hear you brohter. Raise the flag and you'll see me marching along. Since the industry knows realistic throuput that is what should be advertised (unlike the car - which can if you push it do what they claim...you can travel to Germany/Italy/Greec for that one!)

As for duty to the public? If you lve in Canada there is a cusomer protection site that states quite plainly as in this case that a letter stating that you want your money back for nonperformance of product as advertised is enough. Of course you have to return the product!

I don't blame Dlink in general but that quote from Lycan really bothered me. I based my purchasing decsion on his claim which is 3x's faster that realistic values. I am contacting Dlink regarding the return tomorrow.

Cheers,

Lucid

[funchords]

I don't blame Dlink in general but that quote from Lycan really bothered me. I based my purchasing decsion on his claim which is 3x's faster that realistic values. I am contacting Dlink regarding the return tomorrow.

Lycan helps a lot of folks on here and multiple times I've seen him be realistic about expectations.  I don't know if context or accident or generalities explain the stuff that you quoted, but I doubt he intended to mislead anyone (his job is support and he'd have to help someone make that work).  I've read plenty of his posts on various topics where he tells it just like it is.

These days, "decent N network" in a home probably means 5 GHz-based, all N, with 40 MHz bonding and with two spatial MIMO streams.  And I'm kinda guessing that would be the case, because I haven't invested in the new gear to set that up.  It would be a several-hundred dollar investment for me to upgrade all my G gear in order to test that theory.  But as long as 2.4 GHz is as crowded as it is, the N technologies that create the biggest boosts will simply not work.

Best wishes.

[Lucid]

No doubt Lycan has been a great help here but that misinformation wasted my money. In any case I am sure that Dlink can help out.

Cheers,

Lucid.

[Lycan]

I DID make a typo.
32Mbps. Comon read the rest of the post I say a realistic throughput of 10-15MBps. Which we CAN achieve. In fact I've seen it go higher. Also trying to base a request to return the product based on Mod's  technical support thread is silly.

The following is advice given from Lycan when I first noticed my wired speed problems. He advised to used WirelessN given my situation and that wireless N achieved fast results over wired....

"First putting the router in the basement is a horrible idea. The router should be centrally located in the house.
N speeds tout 300mbps. Thats mega BITS per second. Average data throughput on a decent N network is about 32MBps. A wired connection can achieve 80 mbps, so around 10-15MBps. Give or take"

MB = Megagbyte
Mb = Megabit

Read this quote again....
I think I should get my money back after all this.

Cheers,

Lucid

You claim that I gave you misinformation, but then state that this quote is in response to wireless performance issues you were having with the hardware.

No doubt Lycan has been a great help here but that misinformation wasted my money. In any case I am sure that Dlink can help out.

Cheers,

Lucid.

So I'm confused, did I give you incorrect information that encouraged you to purchase the product or was I trying to help you with hardware you were already having trouble with?

[Lucid]

Hey Lycan...don't get me wrong I am not trying to make a flame ware out of this. Typo or no typo I purchased something based on inaccurate information (regardless of how it was communicated). You have helped me a lot in the past and I do account for that. Please don't take this personally since I am the one who purchased the device and may have lost money.

In terms of being clear on the issues at hand (this is getting way off topic now) you have been involved in
1) NAS DNS-323 performace
2) Wireless N Performace
3) Wireless vs. Wired vs. Powerline advice and implementation

As I said you have helped with a variety of issues. I NEVER beat on the MOD's (you included). I bought a product that was communicated to me to be faster than a wired network. If I knew you made a typo I would have not even had to ask for the advice in the first place. Yes I did read the rest of the post as well. The 10-15MBps reads to me as if you are describing the wired network.

In any case my issues are solved and I will ask for a refund on the product. There is no need anyone to take the issue to heart. If anything I should because I wasted the money on a technology that isn't what advertisments make it out to be. As funchords similarly wrote...it is an industry wide misconception based on misleading marketing.

I appologise for the inconvenience.

Lucid.

[santoniuk]

Lucid do you not know that 100 Mb LAN only has a theoretical max of roughly 12 MB/s transfer speeds. Now I know some of these switches on the Dlinks say giga speed but take in consideration the wifi is a max of 300 Mbps at optimal distance and without "noise"

In reality 8 MB/s is ideal over N wifi. You have overhead, conjestion, drive latency, packet loss, etc to contend with.

If you had a large single file over copper at giga speeds with jumbo packets turned on and no mixed mode connections on the switch or lan for that matter (100 and giga on the same segment) then and only then could you get upwards of 120+ MB/s. You would also need PCI express or internal NICs that are not limited by the 133 MB/s bus of PCI.

In short 8 MB/s second is great transferring form a NAS with sata over the air to a desktop. There is no way you will ever get 20 MB/s over wifi for long durations. You are also wrong about false advertisements. No where are they using any claims outside what is already clearly defines and is industry standards.

If you believe standards are false, please contact http://www.iso.org/iso/home.htm and complain that they are wrong.