Re: [Phys-l] Missing universities

[Jason Alferness <alf@u.washington.edu>]

At 12:26 PM 4/13/2006, you wrote:
> Jason Alferness wrote:
> > Umm... no, not this time, but have used it before... worked fine then...
> >
> > So trying it... it traced just fine... showed hops, RTT's, and whatnot...
> > It'd certainly show if it was getting plugged at your campus borders or
> > something... Isn't that what you're implying they might see... or else 
> what
> > specifically are you expecting to see?  Am I missing something?
> >
> > Here's from my windows box:
> > ------------
> > Tracing route to carnot.physics.buffalo.edu [128.205.17.137]
> > over a maximum of 30 hops:
>
> OK, now try it on something that doesn't respond to ICMP ECHO
> packets, such as:
>
>      www.kutztown.edu
>      www.esu.edu
>      www.bloomu.edu
>      www.cup.edu
>      www.lhup.edu
>      www.wcupa.edu



OK, if we have firewall issues, I'll gladly concede the standard windows 
version will potentially fail miserably as a tool and be far too 
limiting...  I had understood that suddenly he was having connectivity 
problems with a large block of the country... I suppose it's possible they 
all started blocking ICMP Echo and port 80 all at the same time, but 
thought that less likely than it getting clogged up very close to 
home.  Windows traceroute may well have helped on that front... And of 
course, there are very legitimate reasons to block things that you don't 
absolutely need...  Maybe I'm wrong, I don't know...


> I recommended tcptraceroute for a reason.
>
> ICMP is IP protocol 1
> TCP  is IP protocol 6
> UDP  is IP protocol 17
>
> If you think hosts (and firewalls!) treat ICMP exactly the same as TCP,
> you're going to be very disappointed.


I certainly don't think all protocols are necessarily treated equally... 
but it was sounding to me as if suddenly he lost connectivity with much of 
the world...  I've set up iptables firewalls, too.. have some concept of 
how they work, though just enough to make it do what I wanted...  If we're 
dealing with these sorts of issues, then I'd say you're absolutely correct 
that he'll need a more sophisticated tool.



> > OK, the Linux trace gives more meaningless digits of precision for RTT's,
>
> Linux isn't the issue.  I didn't mention Linux.  Precision isn't
> the issue either.  The issue is TCP versus ICMP.  You can install
> tcptraceroute tools for microsoft boxen ... but the microsoft
> built-in tracert is ICMP only -- unless there has been a major
> upgrade while I wasn't looking.


Nope, you're absolutely correct in this regard as far as I know...


> > The linux version also has
> > more options for type of service, ports...
>
> Ports?  Ports???  Unless I am greatly mistaken, ICMP doesn't even have ports.
>    http://www.ietf.org/rfc/rfc792.txt
> That's one of the basic facts underlying the point I've been making.

I understand this... and I think you're correct... that's why I listed type 
of service first ( the more correct "protocol #" apparently slipped my mind 
in my scribbling... sorry) and with respect to the linux version... the 
linux version does support different protocols, including TCP/IP, I believe...



> > blah, blah... but do I need them
> > for this and if so for what?  I can think of scenarios that it would be
> > nice to know more, but usually problems are more simple than I first think
> > they may be.
>
> Yeah, but sometimes they're not ultra-simple.  I'm not suggesting this is
> rocket surgery ... but it doesn't have to be rocket surgery to be beyond
> the reach of the microsoft built-in tracert.


True enough... I've found a number of problems with it, though... and one 
doesn't need extra $$ (I don't know a darn thing about tcptraceroute, if 
it's free/share/$$-ware or anything... It may well be a great tool)

> I recommended tcptraceroute for a reason.  Many web servers (and the 
> associated
> firewalls) are narrowly tuned to accept TCP to port 80 plus a few related 
> things
> ... and not much else.

Fair enough... but with so many dropping off the map at once, my first 
thought was a more local problem... It certainly can be one of many things.

Cheers,


> _______________________________________________
> Forum for Physics Educators
> Phys-l@carnot.physics.buffalo.edu
> https://carnot.physics.buffalo.edu/mailman/listinfo/phys-l



-------------------------------------------------------------------------------------------

"A person, who is nice to you, but rude to the waiter/waitress, is not a
nice person. (This is very important. Pay attention.  It never fails.) "

Dave Barry from "FIFTEEN THINGS THAT IT TOOK ME OVER 50 YEARS TO LEARN "


Jason Alferness
University of Washington
Room B256B, Physics & Astronomy Building
Campus Box 351560
Seattle, Washington.  98195-1560

Phone: (206) 221-2974
FAX : (206) 685-0635
email: alf@phys.washington.edu