Technical Summary
Unmanaged large buffers in today's networks have given rise to a slew
of performance issues. These performance issues can be addressed by
some form of Active Queue Management (AQM) mechanism, optionally in
combination with a packet scheduling scheme such as fair queuing.
This document describes various criteria for performing precautionary
characterizations of AQM schemes.
Working Group Summary
Over it's lifetime, this document changed scope a couple of times.
It currently contains guidance for characterizing AQM algorithm
performance, though it had started out to prescribe evaluation
methods for AQM algorithms being considered in the working group. As
time went on, the focus shifted, and the WG did not use the
guidelines in a prescriptive way. The current guidelines are not
fully parameterized, and it is assumed that someone performing a
characterization would consider how to parameterize configurations
relevant to their own networks or scenarios of interest.
Ideally there would be running code (as mentioned by Dave, among
others) for performing the characterizations, e.g. implemented using
some tool such as ns-3, mininet, etc, however, this has not come
about yet.
Where it sits now, the document is suitable as Informational
guidance, but it is not a prescriptive and fully defined spec, like
for example, an IPPM metric would be. It simply collects guidance
and description of the set of things to look at and the measurements
of interest in common AQM performance testing situations.
Document Quality
The document is of sufficient quality that someone (e.g. a graduate
student) could read it and implement proper tests.
Personnel
Wesley Eddy (wes@mti-systems.com) is the shepherd, and Mirja
Kühlewind (ietf@kuehlewind.net) is the AD.