Arlington, Virginia
December 1998
Abstracts and presentation slides
Target
Goals for RM Congestion Control Algorithms
<pdf
version>
Brian
Whetten,GlobalCast Communications
Strawman
Congestion Control Specifications
Mark
Handley, ACIRI Sally
Floyd, ACIRI
Congestion
Control Using Dynamic Rate and Window
Dah
Ming Chiu, Sun Labs
Abstract: Examines the advantages
of using window and rate as the control parameter in a transport's congestion
control. Proposes a way to combine the use of both window and rate as
controls to get the best of both worlds - smooth transmission
with bounded losses (the STBL algorithm). The algorithm
has two phases. In the first, a congestion window is dynamically selected
at the sender. Subsequently, congestion control is
based on rate. Finally, various suggestions for applying STBL to multicast
congestion control were discussed.
Congestion Control
Jamal
Golestani, Bell Laboratories
Update
on Scalable Session Message
Sally
Floyd, ACIRI
Rate-Based Fair Multicast
Congestion Control
Injong
Rhee, NCSU
Round
Trip Estimation for Multicast Congestion Control
Anindya
Basu, Bell Laboratories
Abstract: In multicast communication,
the estimation of receiver round trip times is necessary for congestion
control. For example, in rate based regulation schemes, round trip times
may be required for the proper adjustment of the transmission rate. This
scheme estimates round trip times in a way that avoids the problem of ack
implosion at the source by using a hierarchical tree organization of receivers
in a multicast group for congestion feedback. At the same time,
it does not require measurement of one-way delays, so that
clock synchronization between the source and the receivers
is unnecessary. Furthermore, this scheme works even if the network layer
topology for multicast and the transport layer topology for
congestion feedback are different and do not add significantly
to the overheads.
Simulation
Study of Monitor-Based Flow Control for One-to-Many Bulk RM in NS2 with
Regard to TCP-Friendliness
Tetsuo
Sano, NTT
Abstract: Presents the simulation
study of the evaluation of MBFC in regard to TCP-friendliness. The
simulation was done in NS2. Researchers were able to confirm the effectiveness
of MBFC by adjusting system parameters properly. Moreover, they discovered
that: (1) setting the transmission rate to the worst receiver leads to
an extremely unfair bandwidth sharing occupied by TCP flow, (2) if RM only
experiences a relatively small loss, it should not give up its bandwidth,
otherwise TCP monopolizes it; and, (3) introducing RED gateways leads
to more desirable results for MBFC than using normal droptail gateways.
These findings suggest that routers should introduce a mechanism to punish
self-complacent traffic employing no or poor adaptive CC mechanisms.
The
MINC (Multicast Inference of Network Characteristics) Project
<pdf
version>
Don
Towsley, UMass
Generic
Multicast Transport Service (GMTS) "An Architecture Update"
<pdf
version>
Brad
Cain, Bay Networks and Don
Towsley, University of Massachusetts
LMS
Performance and Applications
Christos
Papadoulos, WUSTL
Multicast
Security: Towards a Standardized Solution
<pdf
version>
Ran
Canetti, IBM Research
Roger
Kermode,
Motorola
Updated 3/3/99