Re: Router speeds...

[Chris Buechler <funsec () chrisbuechler com>]

Dr. Neal Krawetz wrote:

> The tests:
>  - Benchmark using http://www.speakeasy.net/speedtest/
>    Select the two cities that give the best and consistent performance.
>  - Only connect one router at a time.
>  - Limit router usage to the benchmark (don't go web surfing or nmapping
>    at the same time.)
>  

that's a bad way to benchmark firewalls/routers.  The Internet is too 
unpredictable for that to be a reliable, repeatable indication of 
anything, when using a single source like that. 

I'd suggest setting up the routers with a machine on each side, and run 
something like iperf between the LAN and WAN sides for an indication of 
best-case single TCP connection maximum throughput. 

A more common area where these home boxes fall apart is under high pps 
load.  BitTorrent is probably the best way I know of to get this kind of 
real world load.  I'd get as many BitTorrent downloads going as you can, 
and see what kind of sustained max throughput you can get.  Under this 
situation, I would test directly connected to the Internet, and would 
make sure to be using enough torrents with sufficiently fast seeds to 
max out the available Internet bandwidth that you have at your 
disposal.  I'd recommend the FreeSBIE torrent (it always seems to have 
enough fast seeds to peg my 6 Mb cable modem, last I tried it at least), 
and a couple Linux distros or something similarly large for good measure. 

Then repeat both tests for machines on the wireless interfaces. 

-Chris
_______________________________________________
Fun and Misc security discussion for OT posts.
https://linuxbox.org/cgi-bin/mailman/listinfo/funsec
Note: funsec is a public and open mailing list.