In-Network Splitter Function Enabling Highly-Parallel Event Stream Processing

Cloud data center workloads are comprised with a<br/>substantial volume of data-analytics tasks which rely heavily on<br/>data parallel stream processing. Splitting the data streams into<br/>independent data partitions and their distribution across multiple<br/>operators is a heavy operation than can become a performance<br/>bottleneck. Emerging in-network hardware accelerators in cloud<br/>infrastructures can enhance the performance of the splitter function and improve the parallelism degree of data-intensive cloud applications. Building on Programming Protocol-independent<br/>Packet Processors (P4), we devised a P4-based splitter fully in the<br/>data plane to benefit from line-rate performance of P4 packet<br/>processors, e.g., Tofino Switch. We present a summary of the<br/>splitter switch capabilities and complement our proposed design<br/>with new discussion points for future consideration.