huawei releases the data storage chapter of the "towards an intelligent world 2024 white paper"
2024-09-19
한어Русский языкEnglishFrançaisIndonesianSanskrit日本語DeutschPortuguêsΕλληνικάespañolItalianoSuomalainenLatina
[shanghai, china, september 19, 2024] during huawei connect 2024, huawei released the "towards an intelligent world 2024" series of white papers (hereinafter referred to as "white papers"). the white paper covers 9 major industries and sectors, including data storage. in the data storage chapter, huawei combines digitalization, intelligence, and analysis of data elements to elaborate on the outlook and action suggestions for the future evolution of data storage, helping customers build the best data infrastructure in the era of ai big models.
huawei releases the data storage chapter of the "towards an intelligent world 2024" series of white papers
in the era of the surging digital economy, thousands of industries such as finance, electricity, manufacturing, education, medical care, and operators are moving from digitalization to digital intelligence at an unprecedented speed. data has become a key factor in their successful transformation: massive historical data has been awakened, a large amount of new data has been generated by ai, and diversified original data needs to be retained for a long time. at the same time, data processing efficiency has become a prerequisite for the full release and value of ai productivity.
data storage is a carrier of data. huawei has put forward five major prospects for its future development in the white paper:
outlook 1: deploy ai-ready data infrastructure with storage-computing separation architecture to accelerate the emergence of intelligence: in the era of ai big models, the coordinated evolution of computing power and storage power and on-demand expansion have become basic requirements for deploying data infrastructure. the enhancement of core capabilities such as horizontal expansion of capacity, linear growth of performance, and multi-protocol interoperability will effectively help thousands of industries become ai-ready.
outlook 2: improving data processing efficiency with all-flash storage and releasing data value: facing the data processing pressure brought by large model training requirements and intelligent upgrade trends, all-flash storage, with its high performance, large capacity, and low power consumption, combined with emerging data paradigms such as vector rag (retrieval enhanced generation) and long memory storage, can greatly improve the overall performance of the system and achieve storage-based computing.
outlook three: building a data security system that combines prevention and treatment: intrinsic security becomes a basic requirement for storage. by building all-flash backup storage and multi-layer anti-ransomware, we can move from passive response to attacks to active and comprehensive protection, enhance corporate immunity, and achieve strong security through storage.
outlook 4: build an ai data lake foundation to make data visible, manageable and usable: data has become the key competitive advantage of enterprise ai applications. the data weaving capability will break the data silo dilemma and allow enterprises to store, circulate and use massive multi-source heterogeneous data.
outlook 5: full-stack pre-integrated training/pushing machines support the rapid implementation of ai big models in the industry: by pre-integrating infrastructure, tool software and other components, and coordinating the system integration capabilities of ai big model suppliers, we build a one-stack delivered training/pushing machine to enable digital upgrades in thousands of industries.
based on the above outlook, the white paper also provides detailed action suggestions to help build a data infrastructure that meets the requirements of the digital era. in the future, huawei data storage will continue to work with all sectors of society to provide advanced data storage for more new technologies and new applications, and jointly create a better future for the digital era.
