1: \begin{abstract}

2: 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.

3: 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.

4: %

5: We make three contributions. 

6: %

7: First, we present a novel framework in which a user's design objective, and the synthesis of

8: a program (network design) that optimizes that objective are done in tandem. 

9: %The framework generalizes and formalizes \textit{comparative synthesis}, a new approach for learning a user's quantitative intent through queries. 

10: %

11: Second, we develop 

12: %the first algorithm for comparative synthesis, 

13: a novel algorithm for our framework

14: in which a voting-guided learner makes two kinds of queries (\Validate{} 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. 

15: %and show attractive guaranteed convergence rates, 

16: %both for general case and for an important and a widely used class of objective functions.

17: %

18: Third, we implemented \name{}, a system based on our approach, and demonstrate its effectiveness 

19: on four real-world network case studies using black-box oracles and simulation experiments, as well

20: as a pilot user study comprising network researchers and practitioners.

21: %

22: Both theoretical and experimental results show the promise of our approach.

23: 

24: \end{abstract}

25: