NetMeter graph looks like a pulse train... is this normal?

[thiggins]

I did misunderstand. That's much clearer.

Your experiment pretty clearly points to a bug in the client card, or at least unoptimized code. It looks like when you enable the 5 GHz band in the client, it causes the card to periodically scan in both bands, which is understandable. But the scan is causing throughput in the currently associated connection to drop, which it shouldn't.

If there isn't a new driver available from Dell, there's not much you can do besides disabling 5 GHz band on the client.

 

[Jdragon]

Thanks Tim. I have the most recent driver, so SOL I guess.
Best regards.

 

[Jdragon]

No! Not S.O.L.! Three words, "Advanced Networking Service". I was hell bent on finding a way to disable the "scanning" that we thought was stealing the bandwidth periodically and this was the culprit. Disabled it, and now I have NO PULSE TRAIN.

I did a computer restart and a dialog popped up claiming some Dell remote service could not run without "Advanced Networking Service" running... so my guess is some embedded Dell software was doing something screwy.

Learned my lesson though: ALWAYS DO A CLEAN OS INSTALL IF AT ALL POSSIBLE (not just a factory OS install like I did).

Thanks again for your help Tim... you spurred me think in different ways that I would not have come up with on my own.

 

[thiggins]

Thanks for reporting back. Looks like "Advanced Networking Service" is some sort of Dell remote support agent. http://www.file.net/process/hnm_svc.exe.html

 

[iliketofish]

Similar issue perhaps

I'm working through a problem (I've been at it for weeks now) that may be similar to this thread's discussion. I'm trying everything mentioned here in this thread.

Here's my setup:

Router: WNDR3700
Desktop: WNDA3100v2
Laptop: Intel 4965agn
Synology NAS: DS107e
Vonage Router (disconnected for testing)

Windows 7 OS on desktop, Vista x64 on laptop

Attached here is a netmeter image of the WNDA3100v2 with a file transfer to the NAS. The comparison is a file transfer between the desktop/NAS and the desktop/laptop, desktop wired to router, laptop wireless to router.

When I force 54Mbps at 2.4ghz, the graph is rather steady, nice looking. When I set the connection to 300Mbps at 2.4ghz, then it looks like the attached picture.

Would be grateful for any help.

 

[thiggins]

What is the large gap in the NetMeter picture?

How much distance between laptop and router? Are there other networks in the area? Jumbo frames will make no difference with a wireless connection.

 

[iliketofish]

What is the large gap in the NetMeter picture?

How much distance between laptop and router? Are there other networks in the area? Jumbo frames will make no difference with a wireless connection.

The gap: I stopped a transfer and started another later. Between the router and the adapter are 3 meters or so in the same room. (Jumbo frames was a test for the wired connection...just rounded it out with the wireless one!)

More info:
54g at 5ghz looks fine, too.
WPA-PSK2 (AES) security
Channels on both radios set to avoid collision with other units in the vicinity.
Cordless phone unplugged and batteries removed.
Microwave 5 meters away.

 

[thiggins]

As long as the microwave isn't running you'll be fine. When it is running, it will disrupt your 2.4 GHz communications.

I don't think you have the same issue. If I read your LAN Speed Test results correctly, you're getting ~ 50 Mbps write and ~80 Mbps read. These are typical speeds with N gear, even with channel bonding (40 MHz channel) enabled.

802.11n has much higher throughput variation than 11g due to the MIMO techniques used. But average throughput still ends up being higher.

 

[iliketofish]

As long as the microwave isn't running you'll be fine. When it is running, it will disrupt your 2.4 GHz communications.

I don't think you have the same issue. If I read your LAN Speed Test results correctly, you're getting ~ 50 Mbps write and ~80 Mbps read. These are typical speeds with N gear, even with channel bonding (40 MHz channel) enabled.

802.11n has much higher throughput variation than 11g due to the MIMO techniques used. But average throughput still ends up being higher.

The LAN connection desktop to NAS with jumbo frames was 97 write, 147 read. The most I get with the wireless N is around 30, or 10% of the theoretical max. Where should I look for the problem? What I'm really working on is the Desktop WNDA3100v2 to Synology NAS: DS107e connection.

 

[thiggins]

Go look at the wireless charts and you'll see that the speeds you're getting are typical of N products.
Draft 11n Does Not Equal 100 Mbps Ethernet

 

[iliketofish]

Go look at the wireless charts and you'll see that the speeds you're getting are typical of N products.
Draft 11n Does Not Equal 100 Mbps Ethernet

Okay. I think I read that a while back. In other words, Wireless N doesn't do any better than G, at least in what I'm experiencing. (Attached is a screenshot of my desktop to NAS over 5ghz N.) I guess I'll just have to live with it, if you say so! Is that what you'd say? What about these wild fluctuations in throughput?