StreamFusion connects a Rust-based Arrow DataFusion execution engine to SQL workloads on Flink 2.2.0 and 2.2.1. It runs an entire supported query natively and falls back to the existing Flink path when it cannot do so. Developer benchmarks report throughput improvements ranging from about 1.31x to 3.41x across tested state-backend configurations. The operator documentation confirms whole-query fallback, so these results do not imply that every operator is accelerated or that production workloads will achieve the same gains.