Design of Decision making module in NIC for PNG packet acceleration

This paper presents a novel approach to decision-making module for Network Interface Cards (NICs) to enhance packet processing efficiency in Field Programmable Gate Array (FPGA) based networking environments. Our design, which is based on the Corundum platform, uses Deep Packet Inspection (DPI) to find Portable Network Graphics (PNG) file signatures in the first eight bytes of incoming packets. While other packets take the usual processing paths, those recognized as PNG files are offloaded to a specialized hardware accelerator for decompression. By drastically lowering the host CPU’s processing cost, this offloading improves network throughput and lowers latency. Functional simulations and synthesis results show a 24.87% performance boost with only a slight increase in area usage (23.9%) and power (69.6%). By integrating decision-making intelligence within the NIC, our approach offers an efficient method to accelerate network traffic processing.