A Comparative Synthesis Approach to Optimal Network Designs with Indeterminate Objectives

When managing wide-area networks, network architects must decide how to balance multiple conflicting metrics, and ensure fair allocations to competing traffic while prioritizing critical traffic. The state of practice poses challenges since architects must precisely encode their (somewhat fuzzy) intent into formal optimization models using abstract notions such as utility functions, and ad-hoc manually tuned knobs. In this paper, we present the first effort to synthesize network designs with indeterminate objectives using an interactive program-synthesis-based approach. We make three contributions. First, we present a novel framework in which a user's design objective, and the synthesis of a program (network design) that optimizes that objective are done in tandem. Second, we develop a novel algorithm for our framework in which a voting-guided learner makes two kinds of queries (Propose and Compare) to the user, with the aim of minimizing the number of queries. We present theoretical analysis of the convergence rate of the algorithm. Third, we implemented Net10Q, a system based on our approach, and demonstrate its effectiveness on four real-world network case studies using black-box oracles and simulation experiments, as well as a pilot user study comprising network researchers and practitioners. Both theoretical and experimental results show the promise of our approach.