[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]
[PCN] Paper on Measured Rate Termination
Dear colleagues,
we finally managed to revise our paper on "PCN-Based Measured Rate
Termination (MRT)".
http://www3.informatik.uni-wuerzburg.de/~menth/Publications/papers/Menth08-Sub-9.pdf
It is important for the WG for several reasons:
1) It documents pitfalls with different MRT versions. We should take
them into account in the edge behavior drafts. E.g., inter-measurment
times (IMT) are required at the egress node also for MRT-ITR (MRT with
indirectly calculated termination rates) to avoid overtermination in the
presence of IEAs with significantly different RTTs. That's the issue
Daisuke brought up on the list.
2) It shows how many termination steps are required with MRT-DTR (MRT
with directly calculated termination rates) in case of packet loss.
3) It introduces proportional flow termination policies that allow CL to
terminate flows without overtermination even if there are only a very
few flows per ingress-egress-aggregate (IEA) and multiple IEAs on a
bottleneck link.
4) It shows that SM poorly performs for small IEAs even if flow
termination policies respect safety margins for termination. This is
probably one of the most important reasons why SM is not good enough and
CL is standardized as a second system, but the issue was nowhere
documented so far.
These are the major issues. A summary of all findings is provided in
Section 4.9.
Thanks a lot for all your comments so far, they greatly improved the
contents and the readability of the paper. Further comments are welcome.
Kind regards,
Michael
--
Dr. Michael Menth, Assistant Professor
University of Wuerzburg, Institute of Computer Science
Am Hubland, D-97074 Wuerzburg, Germany, room B206
phone: (+49)-931/31-86644 (new), fax: (+49)-931/888-6632
mailto:menth at informatik.uni-wuerzburg.de
http://www3.informatik.uni-wuerzburg.de/research/ngn