RE: Re: [OT] Bandwidth software

["Ian Davidson" <ian@xxxxxxxxxxxxxxx>]

Thanks Andy/Patrick

I appreciate that other factors can impact the network performance,
although they will be looked at separately. My only concern at the
moment is 2 wireless links we are installing this week. One should be
easy as it is only 1.2km away but the other although within the
manufacturers spec is getting close to the limit. The units self
calibrate over the path so once the link is up I wanted to see what
throughput we can achieve. We are having a very important demonstration
next week so I really have to get this one right, once the crane and
riggers leave site there would be no way to get either back before the
demo date!

Ian D
(Hoping for good weather this week!)


-----Original Message-----
From: Patrick Lidstone [mailto:patrick@xxxxxxx]
Sent: 20 September 2004 15:30
To: Ian
Subject: [ukha_d] Re: [OT] Bandwidth software

--- In ukha_d@xxxxxxx, "Andy Laurence" <andy@a...> wrote:
> From: Ian Davidson [mailto:ian@r...]
> > a) Test the maximum sustainable bandwidth between to points on a
LAN.
>
> FTP?  See how quick a file of size X transfers, and work it out from
that?

Bear in mind that network performance depends on a multitude of factors,
all the way up the stack - from physical network interface, through
operating system, chosen protocol and the efficiency of the particular
application on top. Performance across the various permutations can vary
wildly, and may not even saturate your link.

In short, if you choose a real world test like ftp, ensure the usage
scenario mirrors the final "end user" configuration as closely as
possible if you need "meaningful" figures.

Patrick