Towards Robust Multi-Layer Traffic Engineering: Optimization of Congestion Control and Routing

Abstract

Abstract-In the Internet today, traffic engineering is performed assuming that the offered traffic is inelastic. In reality, end hosts adapt their sending rates to network congestion, and network operators adapt the routing to the measured traffic. This raises the question of whether the joint system of congestion control (transport layer) and routing (network layer) is stable and optimal. Using the established optimization model for TCP and that for traffic engineering as a basis, we find the joint system is stable and typically maximizes aggregate user utility, especially under more homogeneous link capacities. We prove that both stability and optimality of the joint system can be guaranteed for sufficiently elastic traffic simply by tuning the cost function used for traffic engineering. Then, we present a new algorithm that adapts on a faster timescale to changes in traffic distribution and is more robust to large traffic bursts. Uniting the network and transport layers in a multi-layer approach, this algorithm, Distributed Adaptive Traffic Engineering (DATE), jointly optimizes the goals of end users and network operators and reacts quickly to avoid bottlenecks. Simulations demonstrate that DATE converges quickly.

Design for Optimizability: Traffic Management of a Future Internet

“…Although congestion control and routing operate independently, their decisions are coupled. The joint system is stable, but often suboptimal [7]. Furthermore, traffic engineering does not necessarily adapt on a small enough timescale to respond to shifts in user demands.…”

Network Resource Allocation for Competing Multiple Description Transmissions

“…In each iteration t, by solving the following problem (12) over (E si , r si , m si ), for each description i, each source s determines its distortion exponent E si (t), rate r si (t) and packet loss exponent m si (t) (13), the source adjusts its offered prices per unit traffic load reduction for each description and each link on its path for the next iteration.…”

“…Along another thread, some researchers have studied the joint network congestion control and multipath routing, see [12], [14], [15], [26] and the references therein. Joint optimization of the routing and congestion control was considered, such that the network can be stable, robust and the users can have better QoS for the applications.…”

Interactions of Intelligent Route Control with TCP Congestion Control

“…The interactions of TCP congestion control with adaptive routing have been studied from a more theoretical perspective in [2]. The most related paper to our work is [5], which focused on the stability of intradomain traffic engineering when the traffic is generated from persistent TCP connections.…”