The goal of this work is to design a comprehensive transport layer model to improve application performance.
Our work ECON is a cross-layer modelthat accurately and scalably predicts application layer performance. ECON is built over a highly accurate model to predict transport layer performance, which is then used to predict the performance of higher layers.
Our initial ECON work focussed on applications that run on loss-based transport protocols. We are currently focusing our attention on BBR (Bottleneck Bandwidth and RTT), a new congestion control algorithm that does not use loss as a signal for congestion. Instead, BBR attempts to send BDP (Bandwidth Delay Product) amount of packets irrespective of losses.