The Analysis of SECDED in Data Storage Transfer and Algorithm Optimization

Abstract

The SECDED can correct up to 2-bit errors and detect up to 3-bit errors. To further improve the algorithm, this paper proposes solutions. For example, Using High- Dimensional Sphere Packing to increase the rate of our encoding method. The problem is that Assuming the data with an M, how can set the parity bit length of K meet the requirements of correcting a mistake? K checksum bits can have a value. One of these values indicates the data is accurate. The remaining 1-value means that the errors in the data can meet: $-1 > \boldsymbol{m}+\boldsymbol{K} (>\boldsymbol{M}+\boldsymbol{K}$ for the total length of the encoding). In theory, a K check code can determine which one (including the information code problems and check code). In the future, decentralized network architecture and native artificial intelligence (AI) capability are two significant trends of 6G networks. The existing centralized AI models that rely on cloud servers or terminals will be challenging to sustain the distributed intelligent cooperation requirements of multi- terminals and multi-nodes in 6G networks. Data collection and processing, AI in model training, model deployment, and reasoning get some new challenges through this new decentralized network environment. Aiming at the characteristics of heterogeneous mass terminal equipment, the significant difference in computing capacity, and dynamic change of communication network conditions in the 6G network decentralized computing environment, this paper analyses the development trend of decentralized artificial intelligence and relevant technologies and theories. It puts forward relevant forward-looking technical challenges and research directions.