$O(K)$-Approximation Coflow Scheduling in $K$-Core Optical Circuit Switching Networks

This paper tackles the challenge of scheduling coflows—a key abstraction for managing related data flows in distributed systems—in modern data center networks (DCNs) that use multiple independent optical circuit switching (OCS) cores. Optical switches offer high bandwidth and energy efficiency, but multi-core setups introduce complexity: inter-core traffic allocation must overcome cross-core coupling, and intra-core scheduling must respect port exclusivity and reconfiguration overhead. The authors propose an algorithm that combines linear programming (LP)-guided global coflow ordering, inter-core flow allocation, and intra-core circuit scheduling to minimize total weighted coflow completion time (CCT).

For K-core OCS networks under an asynchronous reconfiguration model, the algorithm achieves approximation ratios of 8K and 8K+1 for zero and arbitrary release times, respectively. The framework is also applicable to H-core electrical packet switching (EPS) networks, offering ratios of 4H and 4H+1. This work fills a gap in coflow scheduling research for multi-core OCS networks, providing the first theoretical performance guarantees for this setting. The results are backed by a 15-page paper with 6 figures, submitted to arXiv on April 24, 2026.