This study explores the application of Software Defined Networking (SDN) to optimize data center network performance by employing a series of innovative techniques tested within a Mininet-emulated dumbbell network topology. We precisely assess network behavior under different protocols and configurations, including IPerf3 in both Cubic and Bottleneck Bandwidth and Round-trip propagation time (BBR) modes, as well as UDPST (User Datagram Protocol Speed Test). Results highlight IPerf3 in BBR mode as superior in stability and performance. Additionally, we investigate various caching architectures—singular, split, and multi-cache—with multi-cache configurations demonstrating the best performance in terms of capacity and access time. Furthermore, network coding strategies like XOR coding and Random Packet Spraying (RPS) are analyzed for their effectiveness in reducing packet loss and retransmissions, with RPS showing a notable decrease in retransmission rates. Collectively, these insights contribute significantly to SDN implementations aimed at enhancing the operational efficiency of data centers.

Understanding discourse structures of news articles is vital to effectively contextualize the occurrence of a news event.To enable computational modeling of news structures, we apply an existing theory of functional discourse structure for news articles that revolves around the main event and create a human-annotated corpus of 802 documents spanning over four domains and three media sources.Next, we propose several documentlevel neural-network models to automatically construct news content structures.Finally, we demonstrate that incorporating system predicted news structures yields new state-of-theart performance for event coreference resolution.The news documents we annotated are openly available and the annotations are publicly released for future research 1 .