The purpose of this paper is to study three properties of Adaptive Virtual Queue (AVQ) schemes. These properties include: (1) stability in the presence of feedback delays, (2) its ability to maintain small queue lengths, and (3) its robustness in the presence of extremely short flows.
The main contribution of this paper is to propose a simple and easy to implement AVQ algorithm that only relies on keeping track of the virtual queue size without actually enqueuing or dequeuing packets.
(1) Active Queue Management (AQM) schemes are used by routers to
intelligently select packets to mark (or drop) in a manner that conveys
information about the current state of the network to the users.
(2) (1) The implementation complexity of AVQ is comparable to RED, (2) AVQ
is primarily a rate-based marking, (3) AVQ regulates utilization, (4) AVQ
adapts the capacity of the virtual queue, and (5) the deired utilization
and damping variables can be adjusted to determine the stability of the
system.
(3) The authors ran five simulation experiments to test their AVQ
algorithm under various conditions.
I would rate this paper as a 4 because this paper does present a deployable AQM algorithm known as AVQ. This algorithm, which is an extension of previous work in the field of AQM, performs better than a number of other well-known AQM schemes.
To justify their results, the authors run five simulation experiments under various conditions to show that their algorithm performed better than exisitng AQM algorithms. In section 4 of this paper, the authors derive a lengthy proof of their underlying assumption. I found the authors' approach to be a solid and concise one.
One limitation of their approach is that this algorithm, I believe, would be succeptable to attacks by malicious users (as would all AQM algorithms).
One lesson researchers should take away from this work is that good research often comes in the form of studying current solutions to open problems and extending them in a simple or novel way.