Re: Security statement

Jason Williams (streak@ccwf.cc.utexas.edu)
Wed, 3 Dec 1997 14:16:04 -0600 (CST)


On Tue, 2 Dec 1997, ike relucio wrote:
> in order for anyone to be able to intercept any of your transmission from
> point A to point B, they would need to know before hand the path that your
> packets would take to reach your destination. But the path is only
> determined when the packet it sent, doesn't it ? The protocols do not state
> or guarantee that each packet will take the same route -- only that the
> packet will reach the destination. This was one of the main considerations
> when the internet was designed.

It's my understanding that the Internet uses hierarchical routing...which
is pretty much static I believe. TCP/IP doesn't use dynamic or random
routing so the path packets take is pretty much the same over time. As an
experiment..try running traceroute to a known up host several times in a
row. It's been my experience that the path doesn't change. It can be
quite annoying at times too. When a router goes down and prevents you
from going from point A to point B, packets aren't rerouted around that
router. Consequently, everything comes to a stand still and prevents you
from getting to your destination host.

> So the only points where your packets could conceivably be intercepted in
> it's entirety would be that portion of the path that is fixed --- possibly
> the first few legs of the journey as the packet leaves your computer and
> goes to your ISP as well as the last few legs of the journey just before
> reaching the destination.

See above.. :)
Most people won't have access to the backbone routers anyway...but they
might be much more likely to have access to their own gateway. As an
example of this...a recent posting in the university newsgroup
(utexas.dorms.general) showed the network administrators sniffing packets
to determine what was eating up their bandwidth. Come to find out it was
an enormous amount of Quake traffic (about 3.3Mbps for over an hour).
It's not too hard for administrators to sniff packets though doing
something useful with them is another matter.

> Example, if, everytime a router received a packet, the router can choose to
> pass it to one of two possible paths and the router does a good job at load
> balancing, then the first router in the path gets 100% of the packets, the
> second router gets only 50% because the packets are split between two
> routers, the third gets only 25% because there are 4 possible routers that
> could receive the packet, and so on .....Nice theory if it actually works
> that way :-)

Nice theory...but I don't believe it does work that way. If there was
more load balancing, router congestion wouldn't be as big of a problem as
it is today. Routing tables may change a lot, but they don't change
everytime a packet is received I don't believe.

--
streak@ccwf.cc.utexas.edu    * Jason Williams -- Austin, Tx.  |     |
streak@mail.utexas.edu       * University of Texas at Austin  | ___ |
streak@cs.utexas.edu         * BS Computer Science             \_|_/
*************** http://ccwf.cc.utexas.edu/~streak/ **************|