System Design and Fault Tolerance Analysis of the Distributed Storage System with Solid State Recorder on Spacecraft

Abstract

Many aerospace missions have a large number of scientific payloads which are widely dispersed. Such situations bring the storage system great challenges in capacity, data reliability and payloads connection scheming complexity. Aiming at these problems, a Distributed Storage System (DSS) on spacecraft is proposed in this paper. Firstly, Solid State Recorders (SSRs) based on NAND Flash are introduced as the local storage nodes to build a DSS on spacecraft. Secondly, a series of typical system operations, such as node capacity balance and backup out-of-place, are designed to execute the system tasks. Thirdly, techniques of Forward Error Correction (FEC) and Automatic Repeat Request (ARQ) are used to make up a fault tolerance mechanism to improve the system reliability. The FEC schedule including Hamming code, RS code and Low Density Parity-Check Code (LDPC) is introduced in detail. Hamming code is used against Single Event Upset (SEU) inside the FPGA, RS code is used against SEU and retention problems in Flash medium, and LDPC is used to get a high channel gain when data is sent to earth. The FEC methods are analyzed and proved to be able to decrease the error rate efficiently