Testing 802.11n

I have done some reading, into the subject of 802.11n speeds, over the WiFi chipset of my laptop, which its Realtek driver, supposedly supports.

The main fundamental fact to know about this, is that whether 802.11n speeds are in fact enabled, is determined from the router, and not as much from the local driver, where for all I know, the support could be without error.

This feature requires, that all the access to the router is set for 802.11n enabled, including in my case, the Guest Network, which was not so previously. More specifically, the form of security must be WPA2, not WPA1 or WEP, even for any Guest Network.

Where previously my client was set to using Channel 11, it now registers as using Channel 1, which means that although the used frequencies are in the same band as before, they can now span a larger part of this band, to support a single download or upload.

Further, in this mode, the slowest device on the LAN, also determines the fastest speed at which 802.11n will work, even when active. Hence, if I had an older smartphone and a tablet with poor or no 802.11n support, too bad, that tablet would set the limit, on how fast my WiFi speeds are going to be. In fact, both my phone and tablet are relatively recent, and have not interfered much in this trial.

When I run the ‘iwconfig‘ command on the laptop ‘Klystron’ as root, it tells me that “802.11bgn” is enabled, but that the maximum speed available will be ~72mbps. What this states, is the hardware limit of my WiFi chipset. This number already falls short of what 802.11n should allow, but is nevertheless the limit of my chipset.

Because of the DSL package I subscribe to, downloads to my LAN from the internet are limited to a much slower speed. Thus, I cannot use the Internet to test this. But I can use transfers from one computer on my LAN to the other, to test what my WiFi speed has become.

It used to be, that if I synced files between ‘Phoenix’ and ‘Klystron’, this local transfer was limited to about 1.5 Megabytes per second, while now that I have changed my configuration as described, I can get this transfer to take place at 3-4 Megabytes per second.

This low bitrate should not be alarming, because it is partially determined by how fast ‘Phoenix’, the older of the two machines, can encrypt the data, and the decryption of the data at that rate already consumes about 25% of the CPU time, on all 4 cores, belonging to the faster computer ‘Klystron’. This is not a speed test.

I have been wondering, whether I could have been suffering from an additional problem on ‘Klystron’, to whatever its behavior was when I simply closed the lid. And 2 additional things which may have been going wrong, could have been repeated requests from ‘Klystron’, to my router, of IPv6 addresses, while another could have been some half-formed use of 802.11n.

The way I now have it, any use of global IPv6 addresses from my laptop have been disabled, in ways that are safer for the configuration than what I had before. And support for 802.11n is currently enabled. If there is going to be some sort of malfunction, associated with this latter, enabled feature, I would like it to happen now, so that I can determine with some certainty, that there was ever an 802.11n problem with this chipset and driver. Personally, I am skeptical.



I have retained the advantage of the better touch-pad on ‘Klystron’.

When I received my Hewlett-Packard laptop in 2013, it came with Windows 8.1 installed, and I had named it ‘Maverick’. One feature it had, and which I was afraid to lose, should I take Windows off it and replace that with Linux, was the very fine quality of its touch-pad, which on many laptops replaces a mouse. This one touch-pad is so subtle, that it forms an effective replacement for a mouse, while on most earlier laptops, I found that I needed to replace the internal pointing device, with a Bluetooth mouse. The way this works can easily be a software detail.

Well now that this laptop is running Linux, and is renamed ‘Klystron’, I find that again, the functioning of its touch-pad is smooth and satisfactory enough, not to require replacement with an external, Bluetooth mouse.

I think though that this fact is not so much due to Debian / Linux being able to support a “Synaptics Touch-Pad” in general, but rather due to how Kanotix / Spitfire has been set up out-of-the-box, to recognize a lot of different hardware options from the Live USB Key, by Kanotix Developers.

In a similar way, I have never been able to set up a graphics chip-set or card, just using straight Debian. I have always needed to rely on the special scripts and abilities that Kanotix comes bundled with, to install my graphics card. This does not really mean that I could never learn. I have just never had to do so manually.