netdev
[Top] [All Lists]

Re: missing icmp errors for udp packets

To: Chris Wedgwood <cw@xxxxxxxx>
Subject: Re: missing icmp errors for udp packets
From: Pekka Savola <pekkas@xxxxxxxxxx>
Date: Tue, 31 Jul 2001 22:48:37 +0300 (EEST)
Cc: <kuznet@xxxxxxxxxxxxx>, <therapy@xxxxxxxxxxxxx>, <netdev@xxxxxxxxxxx>
In-reply-to: <20010801074547.H8228@weta.f00f.org>
Sender: owner-netdev@xxxxxxxxxxx
On Wed, 1 Aug 2001, Chris Wedgwood wrote:
> On Tue, Jul 31, 2001 at 10:12:12PM +0300, Pekka Savola wrote:
>
>     Me neither.  It would make 'ping -f' testing of your ISP's
>     connections rather inconvenient ;-) ...
>
> As someone who until recently was involved in architecture and
> planning for a large ISP/carrier who's network spanned 3 continents (I
> just like saying that, it sounds better than it really is!) I can
> tell you plenty of people use similar tests.
>
> They are bogus.  As is traceroute.
>
> ping & traceroute are very useful, but there results can often be
> misleading.
>
> For example, cisco routers, of which sadly there are a few still in
> use, do no respond to ICMP packets terribly reliably when they are
> busy, which is pretty reasonable (the route packets instead).

Who said I was pinging Cisco routers?  If I have two servers 100 ms off
each other, I make them 'ping -f' each other.  This does test the
infrastructure and forwarding capabilities a bit.

Traceroute isn't optimal as you noted, as the routers have to pull
the packet with expiring TTL off the "fast path", and this is often
subject to the rate-limiting considerations also.

-- 
Pekka Savola                 "Tell me of difficulties surmounted,
Netcore Oy                   not those you stumble over and fall"
Systems. Networks. Security.  -- Robert Jordan: A Crown of Swords


<Prev in Thread] Current Thread [Next in Thread>