Re: RE:Website published on ADSL

["Paul Gordon" <paul_gordon@xxxxxxx>]

>Paul Gordon wrote:
> >
>
> > 12   741 ms   701 ms   701 ms  213.120.207.222
> > 13   975 ms  1139 ms  1099 ms  172.16.93.126
> > 14  1126 ms   962 ms   865 ms  172.16.93.53
> > 15   961 ms   837 ms   810 ms  172.16.100.57
> > 16   783 ms   988 ms   879 ms  172.16.109.5
> > 17   920 ms   426 ms   412 ms  62.7.250.131
> > 18     *     1016 ms  1139 ms  host213-120-145-98.btopenworld.com
>
> > Now, what concerns me is the hops numbered 13 through 16 - the
>172.16.X.X
> > ones. These are in the non-routable private address space as
defined by
> > RFC1597 and CANNOT be on the internet.
>
>My guess is that these are transit routers within BT's infrastructure.
>The 213.120.207.222 is the border router which connects the internet to
>BT's network.
>

Hmmm, I guess that could be so, - the 62.7.250.131 at hop 17 is also an
address in BT's provider block, I believe, so it could be within BT's
network right up to that point. I find it quite bizzarre that BT would
route
traffic in that way though - doesn't it seem a bit "odd" to you?

>You're right, they're private IPs, and as such you can't ping them
>directly, because they're non-routable. However, there's nothing to
stop
>them seeing you. When you traceroute to 213.120.145.98 and the packet's
>TTL expires at one of these 172.16 routers, it throws your packet away
>and replies with an ICMP TIME_EXCEEDED message, where the *source* IP
>address is the router's 172.16.* address and the destination address is
>your IP address, which is routable, hence the packet gets back to you.
>

Surely the TTL should not expire in this case, - isn't the default TTL
value
30?

>If one of these routers were attempting to send any packets to you for
>which a reply might be reasonably expected (for example a TCP SYN), you
>wouldn't see a 172.16 address, because the packet would have to be
NAT'd
>in order that the router might see a reply - for example, you'd
probably
>see the packet as having a source IP address 213.120.207.222 - NAT'd by
>that border router.
>
>However, since an ICMP TIME_EXCEEDED requires no reply, it would be a
>waste of time to NAT the packet, and besides, that would lead to a
>misleading picture of the network (although I've seen plenty of NAT
code
>that performs NAT on all ICMP packets).
>
>If you try traceroute to another 213.120.145.* address, which
presumably
>are other people's DSL IP addresses assigned by BT, you'll see a
similar
>trace.
>

And you didn't spot my (of course) deliberate mistake in my original
hypothesis, which is that if I were taking a route through his company's
internal LAN, why would I hit the external (IE ADSL) interface of the host
"emperor-zurg" on 213.X.X.X rather than the internal (IE
Ethernet) interface
on 172.16.X.X ??

Interesting one, - it does "seem" to suggest that it's working,
but that
then is at odds with the earlier testing we did the other week. So, that
makes it a 1-all draw at the moment, and the jury is still out (awaiting
the
results of a manual count?)

What we need is someone to tip the balance, and do it from a BT OpenWoe
line....

Paul G.

BTW, sorry about the plaster dust in that PC - I left the loft hatch open
when I was having the hall/stairs/landing replastered!! - everything up
there is COVERED. (still, at least I got plenty of cables in the wall
before
the plaster went up  :-)


_________________________________________________________________________
Get Your Private, Free E-mail from MSN Hotmail at http://www.hotmail.com.

Share information about yourself, create your own public profile at
http://profiles.msn.com.


-------------------------- eGroups Sponsor -------------------------~-~>
eLerts
It's Easy. It's Fun. Best of All, it's Free!
http://click.egroups.com/1/9699/3/_/2065/_/974576426/
---------------------------------------------------------------------_->