From Multiple Decompositions to TRUMP: Traffic-Management Using Multipath Protocol

Traffic management is the adaptation of source rates and routing to efficiently utilize network resources. Traffic management today includes congestion control, routing and traffic engineering. In this paper, we perform a top-down redesign of traffic management using recent innovations in optimization theory. First, we propose a new objective function that captures the goals of both end users and network operators. Second, using various optimization decomposition techniques, we generate four distributed algorithms that divide traffic over multiple paths based on feedback from the network links. These distributed algorithms are provably stable and optimal. Third, combining the best features of these distributed algorithms, we construct TRUMP: a new traffic management protocol that is distributed, adaptive, robust, flexible and easy to manage. Packet-level simulations show TRUMP behaves well with realistic topologies, feedback delays, capacities, and traffic loads. Overall, we show that using optimization decomposition as a foundation, simulations as a building block, and engineering intuition as a guide can be a principled approach to protocol design.