Paper Title: Computing Storage in the AI IoT Era

Paper Abstract: Within the IoT and AI technology development, data processing, sharing, transporting now becoming more and more cost and time sensitive. The edge computing and data decentralization need more and more computing power to solve complex issue. And the heterogeneous computing and storage resource requires more efficient and proper ways on scheduling and arrangement. We analyze the evolution history about the computing and storage, then give out the idea that the AI-IoT need a unified computing-storage combine solution. Then for the storage side, we proposed one AI based data sharing way to let controller to do the Key-value data mapping and searching speedup which directly reflect the IDC data mining result. This technology will let the data storage structured and data searching more power efficient. The goal is let the COT for big data processing more efficient and substantial.

Paper Author: Tang Sun, Director, Design Verification, Beijing Starblaze Technology Co. Ltd.

Author Bio: Profound experience on ASIC verification and SOC integration. Over 10+ chips success and Marvell/Apple cooperation ASIC project handling standalone. Now focusing on the AI + Storage technology study as the AI Storage Chip architect.