``Large-Scale TCP Models Using Optimistic Parallel Simulation''
G. Yaun, C. D. Carothers, and S. Kalyanaraman
In Proceedings of the 17th Workshop on Parallel and Distributed
Simulation (PADS '03), June 2003. BEST PAPER AWARD.
ABSTRACT
Internet data traffic is doubling each year, yet bandwidth does not
appear to be growing as fast as expected and thus short falls in
available bandwidth, particularly at the ``last mile'' may result. To
address these bandwidth allocation and congestion problems,
researchers are proposing new overlay networks that provide a high
quality of service and a near lossless guarantee. However, the central
question raised by these new services is what impact will they have in
the large? To address these and other network engineering research
questions, high-performance simulation tools are required. However, to
date, optimistic techniques have been viewed as operating outside of
the performance envelope for Internet protocols, such as TCP, OSPF and
BGP.
In this paper, we dispel those views and demonstrate that optimistic
protocols are able to efficiently simulate large-scale TCP scenarios
for realistic, network topologies using a single Hyper-Threaded
computing system costing less than $7,000 USD. For our real-world
topology, we use the core AT&T US network. Our optimistic simulator
yields extremely high efficiency and many of our performance runs
produce zero rollbacks. Our compact modeling framework reduces
the amount of memory required per TCP connection and thus our memory
overhead per connection for one of our largest experimental network
topologies was 2.6 KB. That value was comprised of all events used to
model TCP packets, TCP connection state and routing information.