| 1 |
On 02/22/2013 02:55:07 PM, Mick wrote: |
| 2 |
> On Friday 22 Feb 2013 08:04:38 Helmut Jarausch wrote: |
| 3 |
> > Hi, |
| 4 |
> > |
| 5 |
> > I have no experiences in debugging a slow network. |
| 6 |
> > |
| 7 |
> > At home, I have a router which is connected to my telephone line via |
| 8 |
> > VDSL2. |
| 9 |
> > I have 2 PCs one of which is connected by an ethernet cable (i.e. |
| 10 |
> > wired) while the other |
| 11 |
> > one uses a wireless connection which is specified as 56 Mbit/s. |
| 12 |
> > |
| 13 |
> > When copying data from one machine to the other one I see varying |
| 14 |
> > speeds from only 0.5 Mbits/s up to |
| 15 |
> > more than 20 Mbits/s. I have no idea why it is so slow some times. |
| 16 |
> > There are no other wireless devices nearby. |
| 17 |
> |
| 18 |
> Where do you see these transfer speeds? On the wired machine, or on |
| 19 |
> the |
| 20 |
> wireless machine? |
| 21 |
> |
| 22 |
> What do you use to check the transfer speed? |
| 23 |
|
| 24 |
One tool is ttcp which is very similar to netcat and this concerns me |
| 25 |
most. |
| 26 |
The other tool is a remote webserver for measure internet speed |
| 27 |
|
| 28 |
> Do the speeds go up when you used a Cat5e cable at full duplex? |
| 29 |
|
| 30 |
The PC which is wired doesn't have those problems since I can transmit |
| 31 |
data from |
| 32 |
my remote office machine at the highest possble rate which is specified |
| 33 |
by my |
| 34 |
internet provider. |
| 35 |
|
| 36 |
|
| 37 |
> |
| 38 |
> Have you scanned for access points to be absolutely sure that there |
| 39 |
> are no |
| 40 |
> other wireless devices, or APs? |
| 41 |
> |
| 42 |
> Do you use encryption? |
| 43 |
|
| 44 |
Yes, but since I do get optimal performance "some time" I don't think |
| 45 |
this |
| 46 |
matters. |
| 47 |
|
| 48 |
> |
| 49 |
> What frequency and channel are you on and have you tried to change |
| 50 |
> channel/frequency? (some domestic devices like cordless phones, |
| 51 |
> wireless |
| 52 |
> earphones, bluetooth, microwaves, perimeter sensors, etc. can cause |
| 53 |
> co-channel |
| 54 |
> and adjacent channel interference and/or force sharing of the |
| 55 |
> bandwidth). |
| 56 |
> |
| 57 |
> Are both machines and the router's LAN on the same MTU? |
| 58 |
|
| 59 |
Sorry, what's "MTU" ? |
| 60 |
> |
| 61 |
> |
| 62 |
> > What tools and techniques can I use to debug this situation? |
| 63 |
> > |
| 64 |
> > Many thanks for a hint, |
| 65 |
> |
| 66 |
> First use ifconfig to see how many dropped packets you get on both |
| 67 |
> machines. |
| 68 |
> Then use iwconfig, or cat /proc/net/wireless on the PC running the |
| 69 |
> wireless to |
| 70 |
> see how many discarded packets you get, fragmentation, etc. as well |
| 71 |
> as nwid |
| 72 |
> packets from different networks on the same frequency. |
| 73 |
> |
| 74 |
> Then check with e.g. iptraf-ng what network connections are running, |
| 75 |
> their |
| 76 |
> rate(s) and if any of these should not be there. |
| 77 |
> |
| 78 |
> The router stats may also be revealing in this respect, as well as |
| 79 |
> dropped |
| 80 |
> packets on the WiFi. |
| 81 |
> |
| 82 |
> I hope the above will give something to work with, others may have |
| 83 |
> better |
| 84 |
> ideas and more specialised tools for troubleshooting this. |
| 85 |
|
| 86 |
Yes, that'll keep me busy a while. |
| 87 |
> |
| 88 |
> PS. I haven't mentioned spectrum analysers to keep this thread |
| 89 |
> within the |
| 90 |
> realm of quick fixes, but if you do have EM interference and the |
| 91 |
> source of it |
| 92 |
> is obscure, you may need more professional equipment to nail this |
| 93 |
> problem. |
| 94 |
> -- |
| 95 |
> Regards, |
| 96 |
> Mick |
| 97 |
|
| 98 |
Many, many thanks Mick, |
| 99 |
regards, |
| 100 |
Helmut. |
Archives