Network load of idle systems
in [Wireless Internet]
|
Prev: NEWS: First SMS Trojan for Android is in the wild
Next: Introducting Poken Take Your Social Networking Offline
From: Wim J on 10 Aug 2010 17:34 John Navas wrote: > On Tue, 10 Aug 2010 15:36:19 -0500, in > <5jd3669b7jpatgl614h6ogeicutt2nei65(a)4ax.com>, Char Jackson > <none(a)none.invalid> wrote: > >> On Tue, 10 Aug 2010 08:27:28 -0700, John Navas >> <spamfilter1(a)navasgroup.com> wrote: >> >>> On Tue, 10 Aug 2010 17:05:34 +0200, in >>> <4c616abf$0$22945$e4fe514c(a)news.xs4all.nl>, Wim J <nospam(a)pc004.local> >>> wrote: >>> >>>> In my neighbourhood, I see about 10 access-points being >>>> active. About 40% on channel 1, 20 % percent on channel 6 >>>> and 40 % on channel 11. So I take channel 4 and 13 for >>>> my two access-points. I thinks thats the best I can do. >> To the OP, hopefully you live in an area where using channel 13 is >> allowed, i.e., not in the US. Thanks for the warning, but channel 13 is allowed here (Netherlands). >> >> Be careful when saying, with a broad stroke, that wireless is faster >> than broadband Internet. > > It's nonetheless rare for wi-Fi to be an actual bottleneck: > * Super-speed broadband is still relatively rare. > * Wi-Fi usually runs fast enough for even super-speed broadband. > * Speed tends to be limited by remote servers to less than > super-speed (something ISPs probably count on). > I agree with that. My line is 10Mbit ADSL, but there are some cable companies offering 80Mbit here. And VDSL2 is starting its roll-out (but slowly). Wim
From: Jeff Liebermann on 10 Aug 2010 18:32 On Tue, 10 Aug 2010 23:34:42 +0200, Wim J <nospam(a)pc004.local> wrote: >I agree with that. My line is 10Mbit ADSL, but there >are some cable companies offering 80Mbit here. And VDSL2 >is starting its roll-out (but slowly). Watch your router speeds. Most routers are not able to pass 80Mbits/sec wired or wireless. The best you can do via wireless 802.11g is about 25Mbits/sec. See table at: <http://www.smallnetbuilder.com/component/option,com_wireless/Itemid,200> When a wireless router is idle, you'll have 10 tiny SSID broadcasts every second. The load is negligible. If you're tied to large number of machines, via a switched network, you'll see a substantial number of broadcast packets from the machines, such as ARP requests. Finding a blank channel is going to be difficult. Landing between two channels is a bad idea because you'll get some interference from both channels. Also, the strongest signal is not the best criteria. A stong nearby signal, but which doesn't pass much traffic, doesn't create anywhere near as much interference as a weaker signal, that's on all the time. The problem is that few utilities show how much traffic is moving by SSID. Fortunately, Kismet does (Pkts column). Run it all day and the channel with the fewest packets would be a good start. -- Jeff Liebermann jeffl(a)cruzio.com 150 Felker St #D http://www.LearnByDestroying.com Santa Cruz CA 95060 http://802.11junk.com Skype: JeffLiebermann AE6KS 831-336-2558
From: Char Jackson on 10 Aug 2010 18:57 On Tue, 10 Aug 2010 14:03:22 -0700, John Navas <spamfilter1(a)navasgroup.com> wrote: >On Tue, 10 Aug 2010 15:36:19 -0500, in ><5jd3669b7jpatgl614h6ogeicutt2nei65(a)4ax.com>, Char Jackson ><none(a)none.invalid> wrote: > >>On Tue, 10 Aug 2010 08:27:28 -0700, John Navas >><spamfilter1(a)navasgroup.com> wrote: >> >>>On Tue, 10 Aug 2010 17:05:34 +0200, in >>><4c616abf$0$22945$e4fe514c(a)news.xs4all.nl>, Wim J <nospam(a)pc004.local> >>>wrote: >>> >>>>In my neighbourhood, I see about 10 access-points being >>>>active. About 40% on channel 1, 20 % percent on channel 6 >>>>and 40 % on channel 11. So I take channel 4 and 13 for >>>>my two access-points. I thinks thats the best I can do. >> >>To the OP, hopefully you live in an area where using channel 13 is >>allowed, i.e., not in the US. >> >>>Your best bet is probably to avoid the channels with the strongest >>>signals (for both access point and client radios), not necessarily the >>>most access points. >>> >>>>In general, performance is very good, so I think most >>>>of my neighbours are not the kind of heavy users. >>> >>>Wi-Fi is designed to coexist, and has enough capacity that even degraded >>>performance tends to be faster than broadband Internet and thus not a >>>bottleneck. >> >>802.11g is probably still dominant, with a max throughput of about 24 >>Mbps. It doesn't take much interference or congestion to cut that in >>half or even less. Meanwhile, many ISP's are selling service at the >>8-16 Mbps level, with some plans being higher, (such as 22 Mbps with >>bursting to 30 Mbps in my case), not to mention the new DOCSIS 3 >>service tiers that are beginning to roll out with 50 Mbps service, and >>announcements/rumors of doubling and quadrupling that. >> >>Be careful when saying, with a broad stroke, that wireless is faster >>than broadband Internet. > >It's nonetheless rare for wi-Fi to be an actual bottleneck: Only if you stretch the definition of rare so that it has a bit of overlap with common. > * Super-speed broadband is still relatively rare. > * Wi-Fi usually runs fast enough for even super-speed broadband. I don't know what Super-speed broadband means to you, but Comcast's basic service, to use a large ISP as an example, is 12 Mbps. Time Warner's Roadrunner is similar, if not identical. It's not at all rare for a WiFi connection to drop below that rate in the face of congestion and interference. The ironic thing is that you'd be arguing my side, and have in the past, if you had brought it up instead of me. > * Speed tends to be limited by remote servers to less than > super-speed (something ISPs probably count on). I run into that less and less as the weeks go by. These days, it seems to be limited to a few of the open source repositories. All of the more mainstream sites allow full speed downloads, at least in my experience.
From: John Navas on 11 Aug 2010 15:01 On Tue, 10 Aug 2010 15:32:26 -0700, in <fvj366tvkcu7pe4484vpcuk6790g3ig4nh(a)4ax.com>, Jeff Liebermann <jeffl(a)cruzio.com> wrote: >Finding a blank channel is going to be difficult. Landing between two >channels is a bad idea because you'll get some interference from both >channels. In general, but I have seen high interference cases where it produced better results than being on the standard channels (1, 6, 11). >Also, the strongest signal is not the best criteria. A >stong nearby signal, but which doesn't pass much traffic, doesn't >create anywhere near as much interference as a weaker signal, that's >on all the time. The problem is that few utilities show how much >traffic is moving by SSID. Fortunately, Kismet does (Pkts column). >Run it all day and the channel with the fewest packets would be a good >start. The important part there is "all day" (including all night) because you might otherwise miss periods of heavy traffic, and even that might not be enough if (say) someone is simply out of town, which is why I generally recommend going for the standard channel with the weakest signals, and only investigating further if performance proves to be a problem. -- John "Assumption is the mother of all screw ups." [Wethern�s Law of Suspended Judgement]
From: DanS on 11 Aug 2010 17:29
> I don't know what Super-speed broadband means to you, but > Comcast's basic service, to use a large ISP as an example, > is 12 Mbps. Time Warner's Roadrunner is similar, if not > identical. It's not at all rare for a WiFi connection to > drop below that rate in the face of congestion and > interference. The ironic thing is that you'd be arguing my > side, and have in the past, if you had brought it up > instead of me. > >> * Speed tends to be limited by remote servers to less than >> super-speed (something ISPs probably count on). > > I run into that less and less as the weeks go by. These > days, it seems to be limited to a few of the open source > repositories. All of the more mainstream sites allow full > speed downloads, at least in my experience. I've got T/W in WNY, and typically saturate my full 10mbps d/l speed using only 2 d/l threads from GigaNews. |