You are not logged in.

#1 2008-08-13 22:24:54

ljshap
Member
From: Ossining, NY
Registered: 2008-01-23
Posts: 160

Wireless speed haircut using wpa vs wep vs no security.

I'm just curious what type of speed haircuts people experience when using wpa vs wep vs no security.  I would have expected something like a 10 to 20% decrease in download and file transfer speeds instead of what I'm seeing.

  My non scientific tests are as follows:

                                    d/l                  u/l
wpa               
                                 1.280            692             
                                 1,189             338
                                    909          1,602

wep                         8.089           1,043


no security
                                11,485          1,658
                                  4,813           1.072
                                   3,047          1.842

  Please ignore the obvious anomilies in the test speeds other than the slow wpa transfers.


   The tests were taken at dslreports, but reflect my "real world" observations including transferring files.  I used a linksys 54GS router, a linksys54??? wireless desktop card running on a 15MB cable connection.  The connection was establish using netcfg2.

    I was just curious if other people are experiencing the same type of speed slowdowns using wpa.  Before I started using netcfg, I assumed I screwed up, but the information needed for netcfg seems pretty simple and basic. 

    If this is typical, I might turn the access point off, and use wep when I need wireless access.


Live Free or Die !

Offline

#2 2008-08-14 09:56:01

iphitus
Forum Fellow
From: Melbourne, Australia
Registered: 2004-10-09
Posts: 4,927

Re: Wireless speed haircut using wpa vs wep vs no security.

I'd expect a minor hit from wpa but not on that scale. With the spread of those results, there's not much you can accurately draw from it. The test probably isn't the best one either.

They dont match with my results either. I can near max my laptops' 802.11b  (theoretical max ~ 1.38MB) wireless through WPA. I've got about 1.0MB/s reproducable consistent download speed from a local aussie ftp mirror.

It could be a configuration issue, crappy router, hardware or driver issue.

Offline

#3 2008-08-14 11:38:38

dhave
Arch Linux f@h Team Member
From: Outside the matrix.
Registered: 2005-05-15
Posts: 1,112

Re: Wireless speed haircut using wpa vs wep vs no security.

This is a usage of the word "haircut" that I didn't know. What does it mean, exactly?


Donate to Arch!

Tired? There's a nap for that. --anonymous

Offline

#4 2008-08-14 17:04:34

ljshap
Member
From: Ossining, NY
Registered: 2008-01-23
Posts: 160

Re: Wireless speed haircut using wpa vs wep vs no security.

dhave wrote:

This is a usage of the word "haircut" that I didn't know. What does it mean, exactly?

From Wikipedia

In finance, a haircut is a percentage that is subtracted from the par value of the assets that are being used as collateral. The size of the haircut reflects the perceived risk associated with holding the assets.

For example, Treasury bills (which are seen as fairly safe) might have a haircut of 1%, while for a stock option (which are seen as less safe) the haircut might be as high as 30%.


------

   Sorry about that, I just should have said the reduction in speed for using WPA vs WEP over no encryption.  Sometimes the word gets used very broadly for any reduction in value of an asset such as  bandwidth due to risk or other reasons.  In this case it would be a speed penalty for using encryption.  I assume it comes from a barber shop haircut where something is taken off the top.

   Hope that helps some.


Live Free or Die !

Offline

#5 2008-08-14 19:49:03

cactus
Taco Eater
From: t͈̫̹ͨa͖͕͎̱͈ͨ͆ć̥̖̝o̫̫̼s͈̭̱̞͍̃!̰
Registered: 2004-05-25
Posts: 4,622
Website

Re: Wireless speed haircut using wpa vs wep vs no security.

upstream results are not the best test scenario, as they are entirely too variant.
For testing, I would connect a box to one of the ethernet ports on the linksys, and then use iperf to test between a node on the wireless to the wired node.

note: wireless is more suseptible to packet loss, so udp throughput over wireless will most likely be crappy. Make sure to test with tcp instead of udp.


"Be conservative in what you send; be liberal in what you accept." -- Postel's Law
"tacos" -- Cactus' Law
"t̥͍͎̪̪͗a̴̻̩͈͚ͨc̠o̩̙͈ͫͅs͙͎̙͊ ͔͇̫̜t͎̳̀a̜̞̗ͩc̗͍͚o̲̯̿s̖̣̤̙͌ ̖̜̈ț̰̫͓ạ̪͖̳c̲͎͕̰̯̃̈o͉ͅs̪ͪ ̜̻̖̜͕" -- -̖͚̫̙̓-̺̠͇ͤ̃ ̜̪̜ͯZ͔̗̭̞ͪA̝͈̙͖̩L͉̠̺͓G̙̞̦͖O̳̗͍

Offline

#6 2008-08-15 00:35:33

ljshap
Member
From: Ossining, NY
Registered: 2008-01-23
Posts: 160

Re: Wireless speed haircut using wpa vs wep vs no security.

cactus wrote:

upstream results are not the best test scenario, as they are entirely too variant.
For testing, I would connect a box to one of the ethernet ports on the linksys, and then use iperf to test between a node on the wireless to the wired node.

note: wireless is more suseptible to packet loss, so udp throughput over wireless will most likely be crappy. Make sure to test with tcp instead of udp.

It was my "real world" experience that caused me to make the speed tests and to try wep and no encryption.  The tests only confirmed what I was already seeing and was drastic enough to not be concerned about speed variations or anomalies.   WPA browsing was sluggish with d/l speeds of 1 to 1.5 MBs but both WEP and no security was responsive with good downloads including pacman upgrades.   I am usually able to close to my rated bandwidth of 15MBs using WEP with 128 bit encyrption.   
 
The wireless card is in a desktop and would be a pain to connect with a cable.  The fact that I'm getting such good speeds with WEP indicates that the router and card are working OK.  At this point I'm just going to go with 128 bit WEP encryption and try to turn off SSID broadcasting when I'm not using the wireless computer.

  I pretty much assumed it was my fault when I was using ifconfig, iwconfig and iwpriv to make the connection, but netcfg2 seems pretty easy and hard to mess up.  I'm really not that concerned about the security with WEP, but I was curious if anybody else had the same type of experiences.

Thanks


Live Free or Die !

Offline

#7 2008-08-15 20:43:06

ljshap
Member
From: Ossining, NY
Registered: 2008-01-23
Posts: 160

Re: Wireless speed haircut using wpa vs wep vs no security.

Not exactly sure what I was doing wrong but I'm getting good connectivity now using wpa.
  I set the router to use wpa2 personal mode and the encryption type to TKIP and AES.  I set the netcfg2 config file to wpa since I couldn't find any setting corresponding to wpa2 personal. 

  If there is a speed penalty for using wpa, it seems to be pretty minor.  What a difference.


Live Free or Die !

Offline

Board footer

Powered by FluxBB