FORC: High-Throughput FPGA Accelerator for ORC Files

This research article presents FORC, a high-throughput FPGA accelerator designed to optimize the decoding of Apache ORC files in big data analytics. The study highlights the challenges of data decompression and decoding on CPUs, demonstrating that FORC achieves up to 12.9GB/s decoding throughput on AMD/Xilinx Alveo U280 FPGA, significantly outperforming traditional CPU implementations. The paper details the architecture, integration, and performance results of FORC, which addresses the IO and co