Boston Linux & Unix (BLU) Home | Calendar | Mail Lists | List Archives | Desktop SIG | Hardware Hacking SIG
Wiki | Flickr | PicasaWeb | Video | Maps & Directions | Installfests | Keysignings
Linux Cafe | Meeting Notes | Blog | Linux Links | Bling | About BLU

BLU Discuss list archive


[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: Byte throughput of an interface?



 Casey Callendrello wrote: 
> I've always liked iperf.  It has lots of configurable options, like TCP 
> window size, packet length, and whether or not to do a bi-directional 
> test.  Plus, it's free in every sense of the word. 
> 
> http://dast.nlanr.net/Projects/Iperf/

Ah, but that's fundamentally different, in that it's adding traffic to see 
what capacity is, whereas the things I was suggesting (and I think what the 
OP was asking about) measures what's already going over the interface. 


> On Nov 7, 2007, at 10:11 PM, Matthew Gillen wrote: 
> 
>> John Chambers wrote: 
>>> Something that turned out to be surprisingly difficult to  find:  How 
>>> does  one  learn  the  actual data rate (in Bytes) of a linux network 
>>> interface? 
>>> 
>>> There are lots of tools that give the packet rate, starting with good 
>>> old "netstat -i".  It's interesting and useful info.   But  what  I'm 
>>> trying  to  find  is  the total byte (or byte/sec) throughput of some 
>>> interfaces.  I know it's gotta be there somewhere, probably in one of 
>>> those  directories  under  /proc, but I don't seem to be guessing the 
>>> right keywords to find it. 
>>> 
>>> Anyone know off the top of their head? 
>> 
>> I like iptraf (choose "General interface stats") from 
>> http://iptraf.seul.org/ as a text-mode network monitor.  I think the 
>> issue 
>> is that the number isn't directly in /proc, but rather there are some 
>> things 
>> in /proc that you could use to calculate it (ie the Tx/Rx 'bytes' 
>> fields of 
>> /proc/net/dev).  Have fun with the parsing ;-) 
>> 
>> You could also run 'ifconfig' and parse the output of that 
>> periodically (the 
>> last line for each interface has tx/rx bytes) 
>> 
>>> I've considered using tcpdump, but that seems like a huge cpu load to 
>>> get data that's probably just lying about somewhere ... 
>> 
>> Yea, that's way overkill. 
>> 
>> Matt 
>> 
>> -- 
>> This message has been scanned for viruses and 
>> dangerous content by MailScanner, and is 
>> believed to be clean. 
>> 
>> _______________________________________________ 
>> Discuss mailing list 
>> [hidden email] 
>> http://lists.blu.org/mailman/listinfo/discuss
> 
> 


BLU is a member of BostonUserGroups
BLU is a member of BostonUserGroups
We also thank MIT for the use of their facilities.

Valid HTML 4.01! Valid CSS!



Boston Linux & Unix / webmaster@blu.org