Towards accurate congestion control models : validation and stability analysis
Krister Jacobsson, Håkan Hjalmarsson, Karl Henrik Johansson · KTH Publication Database DiVA (KTH Royal Institute of Technology) · 2006
Fluid flow models have turned out to be instrumental for analysis and synthesis of primal/dual congestion control algorithms which rely on aggregated information from a network path. In particular stability has been analyzed using such models. Departing from the theory of modeling for control, we refine the fluid flow model by augmenting the customary model of transport latencies, link price and source control with estimator dynamics and sampling properties. The impact of cross traffic and changes in network configuration is incorporated as well. As a demonstration, the modeling framework is applied to FAST TCP for which a fluid flow model is derived which through packet simulations is shown to provide accurate quantitative and qualitative information such as prediction of stability regions, behavior to cross traffic and which dynamics that influence the closed loop behavior. This model is also compared with pre-existing FAST TCP models and it is illustrated that using appropriate sampling rates and previously neglected estimation dynamics may have a large impact on the closed loop properties. We also introduce a novel link model that is validated towards packet/level data.