Networking help

[Bob Bell]

On Tue, Dec 17, 2002 at 09:08:18AM -0500, pll at lanminds.com <pll at lanminds.com> wrote:
> In a message dated: Mon, 16 Dec 2002 20:50:31 EST
> bscott at ntisys.com said:
> >  This is just a guess, but the timeout on ARP is 20 seconds, IIRC.
> 
> Interesting.  I didn't know that!

    Actually, I think the traditional timeout is something like 20
minutes, not 20 seconds.  On Linux, I think this has been replaced by
some complex calculation on a number of variables, but at least seems to
be greater than 1 minute.  See `man 7 arp`.

> However, let's assume 'T' did exist.  So what?  IMO, the icmp packets
> should never have gone to 'T' anyway.  Shouldn't have gone back out
> the interface they came in on?

    I was going to reply, but I think Ben already answered this adequately:

  No!

  ****** IP routing is a stateless operation. ******

  There is no such thing as "return traffic" in IP.  Higher level protocols,
like TCP, might implement virtual circuits, but the IP layer itself is
stateless.  Routing decisions are based entirely on the destination address
of each packet.  Period.

  (Again, load balancing, policy-based routing, and other things are
increasingly common exceptions to this rule, but they are rarely active
unless you explicitly configure them.)

-- 
Bob Bell <bobbell at zk3.dec.com>
-------------------------------------------------------------------------
 "We trained hard to meet our challenges but it seemed as if every time
  we were beginning to form into teams we would be reorganized. I was to
  learn later in life that we tend to meet any new situation by
  reorganizing; and a wonderful method it can be for creating the
  illusion of progress while producing confusion, ineffectiveness, and
  demoralization."
   -- Gaius ("Petronius Arbiter"), who died 66 A.D.