From chuck.atkins at kitware.com Fri Mar 5 04:20:26 2010 From: chuck.atkins at kitware.com (Chuck Atkins) Date: Fri, 5 Mar 2010 04:20:26 -0500 Subject: [CDBUG-talk] Abysmal wireless network performance. Message-ID: I'm seeing some very poor performance when using WiFi and I don't really know where to start looking. For example: A ping to my gateway shows reasonably fast responses when no other network traffic is happening: PING 192.168.1.1 (192.168.1.1): 56 data bytes 64 bytes from 192.168.1.1: icmp_seq=0 ttl=64 time=6.938 ms 64 bytes from 192.168.1.1: icmp_seq=1 ttl=64 time=8.561 ms 64 bytes from 192.168.1.1: icmp_seq=2 ttl=64 time=9.516 ms 64 bytes from 192.168.1.1: icmp_seq=3 ttl=64 time=5.939 ms ... But if I start any other network traffic, say an ftp transfer, then the response time skyrockets and the transfer rate reflects this: ... 64 bytes from 192.168.1.1: icmp_seq=63 ttl=64 time=6.849 ms 64 bytes from 192.168.1.1: icmp_seq=64 ttl=64 time=4.978 ms 64 bytes from 192.168.1.1: icmp_seq=65 ttl=64 time=251.623 ms 64 bytes from 192.168.1.1: icmp_seq=68 ttl=64 time=291.099 ms 64 bytes from 192.168.1.1: icmp_seq=69 ttl=64 time=387.732 ms 64 bytes from 192.168.1.1: icmp_seq=70 ttl=64 time=485.485 ms 64 bytes from 192.168.1.1: icmp_seq=71 ttl=64 time=457.686 ms 64 bytes from 192.168.1.1: icmp_seq=72 ttl=64 time=507.699 ms 64 bytes from 192.168.1.1: icmp_seq=73 ttl=64 time=117.663 ms I'm running 8-STABLE and using the iwn driver with an Intel 4965agn card. This problem only happens over wireless and is not an issue when using a hard wired connection. Your thoughts? - Chuck -------------- next part -------------- An HTML attachment was scrubbed... URL: From jaime at snowmoon.com Fri Mar 5 08:18:00 2010 From: jaime at snowmoon.com (Jaime) Date: Fri, 5 Mar 2010 08:18:00 -0500 Subject: [CDBUG-talk] Abysmal wireless network performance. In-Reply-To: References: Message-ID: On Mar 5, 2010, at 4:20 AM, Chuck Atkins wrote: > I'm seeing some very poor performance when using WiFi and I don't > really know where to start looking. Any problem will become more obvious as more traffic is placed on the link. This is typically an exponential growth, too. Have you checked the space near your computer and near your base to see what other wifi networks are detectable? Lately, I've been finding a number of people with bases on the same channel as a neighbor. This might contribute to your problem. If you find this, move to another channel. Ideally, follow the 3- channel-separation rule. For example, if you see a lot of bases on channels 6 and 11 and one on 8, then only 1 and 2 are viable for you. This is because 3, 4, and 5 are your three channels of separation from those bases on channel 6. If there are no "good" channels, find the one with the fewest and weakest competition for a signal near the base and your most common work spaces. Also, Apple's bases have a feature called "Interference Robustness." it trades speed for stability (e.g. compensates for microwave ovens.) Lastly, look at wifi noise levels. Any decent wifi analysis tool will tell you the noise level, the channel, and the signal strength. The "real" or "useable" signal is the signal minus the noise. So reducing noise can make a big difference. Noise can be microwave ovens, some cordless phones, and other kinds of wireless equiment. Good luck, Jaime From chuck.atkins at kitware.com Fri Mar 5 10:00:58 2010 From: chuck.atkins at kitware.com (Chuck Atkins) Date: Fri, 5 Mar 2010 10:00:58 -0500 Subject: [CDBUG-talk] Abysmal wireless network performance. In-Reply-To: References: Message-ID: I'm fairly certain it's not an issue with the access point, proximity to, interference, channel, etc., since I only see the problem in FreeBSD. Linux WiFi on the same machine shows no degradation. Chuck Atkins R&D Engineer Kitware, Inc. (518) 371-3971 x603 -- "Mathematicians are tools for turning coffee grounds into formulas.", Paul Erdos On Fri, Mar 5, 2010 at 8:18 AM, Jaime wrote: > On Mar 5, 2010, at 4:20 AM, Chuck Atkins wrote: > > I'm seeing some very poor performance when using WiFi and I don't really >> know where to start looking. >> > > Any problem will become more obvious as more traffic is placed on the link. > This is typically an exponential growth, too. > > Have you checked the space near your computer and near your base to see > what other wifi networks are detectable? Lately, I've been finding a number > of people with bases on the same channel as a neighbor. This might > contribute to your problem. > > If you find this, move to another channel. Ideally, follow the > 3-channel-separation rule. For example, if you see a lot of bases on > channels 6 and 11 and one on 8, then only 1 and 2 are viable for you. This > is because 3, 4, and 5 are your three channels of separation from those > bases on channel 6. > > If there are no "good" channels, find the one with the fewest and weakest > competition for a signal near the base and your most common work spaces. > > Also, Apple's bases have a feature called "Interference Robustness." it > trades speed for stability (e.g. compensates for microwave ovens.) > > Lastly, look at wifi noise levels. Any decent wifi analysis tool will tell > you the noise level, the channel, and the signal strength. The "real" or > "useable" signal is the signal minus the noise. So reducing noise can make > a big difference. Noise can be microwave ovens, some cordless phones, and > other kinds of wireless equiment. > > Good luck, > Jaime > -------------- next part -------------- An HTML attachment was scrubbed... URL: From jaime at snowmoon.com Fri Mar 5 11:39:59 2010 From: jaime at snowmoon.com (Jaime) Date: Fri, 5 Mar 2010 11:39:59 -0500 Subject: [CDBUG-talk] Abysmal wireless network performance. In-Reply-To: References: Message-ID: <98CFDEAE-FCC1-4E2C-949F-B31DFA146D41@snowmoon.com> On Mar 5, 2010, at 10:00 AM, Chuck Atkins wrote: > I'm fairly certain it's not an issue with the access point, > proximity to, interference, channel, etc., since I only see the > problem in FreeBSD. Linux WiFi on the same machine shows no > degradation. Sorry, then. I honestly don't use wifi on FreeBSD. I wish I knew more and could help. Jaime From kevinmca1 at verizon.net Fri Mar 5 20:30:15 2010 From: kevinmca1 at verizon.net (Kevin McAleavey) Date: Fri, 05 Mar 2010 20:30:15 -0500 Subject: [CDBUG-talk] Abysmal wireless network performance. In-Reply-To: References: Message-ID: <6.1.2.0.1.20100305202730.01cc1260@incoming.verizon.net> A thought ... I've seen some genuine squirrelies myself ... when I setup wifi on our machines, I make it a point (Verizoff is our DSL here, and they're AWFUL) to add mtu 1468 in my ifconfig settings. I forget what the default for wifi is, something on the order of 2290 or some such, and it's not a valid MTU at all. Give that a shot, that's what it sounds like to me. Wired ETH does the proper 1500, but the wifi default in BSD 8 is just not right ... At 04:20 AM 3/5/2010, Chuck Atkins wrote: >I'm seeing some very poor performance when using WiFi and I don't really know where to start looking. For example: > >A ping to my gateway shows reasonably fast responses when no other network traffic is happening: > >PING 192.168.1.1 (192.168.1.1): 56 data bytes >64 bytes from 192.168.1.1: icmp_seq=0 ttl=64 time=6.938 ms >64 bytes from 192.168.1.1: icmp_seq=1 ttl=64 time=8.561 ms >64 bytes from 192.168.1.1: icmp_seq=2 ttl=64 time=9.516 ms >64 bytes from 192.168.1.1: icmp_seq=3 ttl=64 time=5.939 ms >... > >But if I start any other network traffic, say an ftp transfer, then the response time skyrockets and the transfer rate reflects this: > >... >64 bytes from 192.168.1.1: icmp_seq=63 ttl=64 time=6.849 ms >64 bytes from 192.168.1.1: icmp_seq=64 ttl=64 time=4.978 ms >64 bytes from 192.168.1.1: icmp_seq=65 ttl=64 time=251.623 ms >64 bytes from 192.168.1.1: icmp_seq=68 ttl=64 time=291.099 ms >64 bytes from 192.168.1.1: icmp_seq=69 ttl=64 time=387.732 ms >64 bytes from 192.168.1.1: icmp_seq=70 ttl=64 time=485.485 ms >64 bytes from 192.168.1.1: icmp_seq=71 ttl=64 time=457.686 ms >64 bytes from 192.168.1.1: icmp_seq=72 ttl=64 time=507.699 ms >64 bytes from 192.168.1.1: icmp_seq=73 ttl=64 time=117.663 ms > >I'm running 8-STABLE and using the iwn driver with an Intel 4965agn card. This problem only happens over wireless and is not an issue when using a hard wired connection. > >Your thoughts? > >- Chuck > >_______________________________________________ >CDBUG-talk mailing list >CDBUG-talk at lists.nycbug.org >http://lists.nycbug.org/mailman/listinfo/cdbug-talk ------------------------------------------- Kevin McAleavey KNOS Project http://knos.yolasite.com/ -------------- next part -------------- An HTML attachment was scrubbed... URL: