Windowed ping, paper and software.
mathis at pele.psc.edu
Mon May 2 20:57:58 UTC 1994
I am in the process of prepairing a network performance diagnostic for
publication. The paper is in anon.psc.edu:pub/net_tools/WPing.ps, and the
abstract is attached at the end of this message. Some of you may have seen
early drafts at the IETF.
It has always been my intent to release this code to the network operations
community first. Windowed ping has had as much impact on our (PSC's) ability
to diagnose performance problems as traceroute has had on our ability to
diagnose routing problems. We are publishing it now because we are concerned
about the overall performance of the U.S. Internet during the transition from
the T3 NSFnet.
There will come the day when *your* *worst* *customer* has a copy, and wants
to know what your are going to do about some swamp. It can also be used for
some truly nasty attacks by your competitors. Windowed ping is sufficiently
dangerous to cause anxiety.
I had planned to limit its dangers by individually labeling the binarys such
that a packet sniffer could identify the original authorized copy of the
program. The labels had simple encrypted seals designed to take a wizard with
adb longer to break than to rewrite it from scratch. All of this labeling code
is already working.
There are two problems with this scheme:
I had planned to distribute the binarys by e-mailing uuencoded tar files
directly to authorized people. However, the messages are about 250KBytes,
which is larger than most mailers can safely inhale. Some other mechanism is
need to distribute the labeled binarys.
Second, my available pool of compile platforms for generating binarys doesn't
have nearly adequate coverage. The source is more or less posix so it will
run on almost anything. (raw sockets are beyond the spec)
I am soliciting suggestions from this community as to how to proceed.
I will consider sending sources to appropriate people, or even to this
After the paper is published, people who have no constructive use for the
software will be requesting it. I would like a clean, fair way to protect the
community from them.
Windowed Ping: An IP Layer Performance Diagnostic
The Internet is suffering from multiple effects of its own rapid growth.
Network providers are in the uncomfortable position of deploying new products
and technologies into already congested environments without adequate tools to
easily assess their performance. In this paper we present a diagnostic that
provides direct measurement of IP performance, including queue dynamics at or
beyond the onset of congestion. It uses a transport style sliding window
algorithm combined with either ping or traceroute to sustain packet queues in
the network. It can directly measure such parameters as throughput, packet
loss rates and queue size as functions of packet and window sizes. Other
parameters, such as switching time per packet or per byte, can also be
derived. Many of the measurements can be performed either in a test bed
environment (yielding the most accurate results), on single routers in
situ in the Internet, or along specific paths in the production Internet.
We will illustrate several measurement techniques.
More information about the NANOG