Building a Spaceborne Integrated High-performance Processing and Computing Platform Based on SpaceVPX

Abstract

An efficient way to reduce the pressure on communication bandwidth between satellite and ground is to process information and images autonomously, remove incorrect picture data, and communicate useful and usable information. Therefore, this study proposes and develops an open and distributed high-performance integrated processing and computing platform based on the SpaceVPX architecture to support efficient and flexible information processing of satellites in space. The high-performance processing and computing platform is built on the principles of software hierarchy, hardware modularization, and resource heterogeneity, which refines the granularity of parallel computing from devices to hardware modules, improving parallelism and resource utilization, and extending the data plane and control plane of the SpaceVPX architecture from within to between devices, realizing information transfer between halves. At the same time, it acts as a “virtual backplane” bus, enabling information transmission between all hardware modules in the processing and computing platform, and realizing the interconnection of hardware module resources throughout the satellite. A single module has a computational power of 15TOPS and a transmission bandwidth of 200Gbps. Meanwhile, with the design ideas of software-defined network and software-defined hardware, the hardware modules in different physical locations are formed into a logical and organic whole, which effectively supports the goal of multi-use and parallel reuse of the integrated processing and computing platform with computing tasks allocated on demand and hardware modules enabled on demand, and can further enhance the processing and computing capacity of the entire satellite, laying the foundation for deep integration and linkage between the satellite platforms.