ISP question [ was traceroute , pls ]

[Vu Pham]

David A. Bandel wrote:
> On Fri, Oct 17, 2008 at 11:15 AM, vu pham <vu at sivell.com> wrote:
>> David A. Bandel wrote:
>>> On Fri, Oct 17, 2008 at 9:10 AM, vu pham <vu at sivell.com> wrote:
>>>> Are the ISPs supposed to let the traceroute packets get thru ?
>>>>
>>>> I am on a server, which cannot traceroute all the way to the destination,
>>>> and I still be able to make the smtp connection to the destination.
>>>>
>>>> So I guess I cannot conclude about connections based only on the
>>>> traceroute
>>>> path, is that correct ?
>>> Correct.  There are a _lot_ of clueless admins out there.  Most are
>>> Windoze admins, but some find their way into other networking
>>> positions.  They are under the mistaken illusion that if they block
>>> things like ping (ICMP) and traceroute (in UNIX, udp, in Windoze
>>> ICMP),
>> You make me remember the first time I learned about how traceroute works by
>> using the TTL field. That's a nice invention, isn't it ?
>> It tooks almost 20 years since the Internet came into being to have such a
>> simple but very useful tool like that.
> 
> ???  Um, I know I used traceroute (and tcpdump) and others some 30
> years ago (we're talking UNIX systems now, not DOS -- Windows didn't
> even exist then).  That is less than 20 years after the arpanet (now
> known as the Internet) came into being.  And ttl's are as old as the
> IPv4 protocol (actually older).
> 

traceroute was written by Van Jacobson in 1988, and Arpanet started in 
1969, which is 19 years apart. I just checked again, and I was wrong in 
that calculation because  tcpip protocol only started in 1974, and 
traceroute is based on tcpip.

Even that, it still took 14 years for such a nice invention, and it was 
5 years after Ping was written (http://ftp.arl.mil/~mike/ping.html)

By the way, I found this note is interesting, too:

http://www.kohala.com/start/papers.others/vanj.99feb08.txt


Vu