Massive data explosion, data storage is the key.
At present, the data shows exponential growth, and the data scale has risen from GB, TB and PB to EB and even ZB. According to Hyperion's prediction, by 2025, the global data space will grow to 163ZB, which is 10 times of the data generated by HPC in 201year. The explosive growth of data gave birth to the development and application of digital technology, but it also put forward higher requirements for calculation and storage.
In high performance computing (HPDA), computing, storage and network are indispensable. In the past, the focus of industrial innovation was to pursue higher computing power. With the rapid development of big data, diversified computing power and other related technologies, the focus of high-performance computing began to evolve from computing-centric to data-centric computing; Traditional HPC began to develop into high performance data analysis (HPDA). According to IDC statistics, 67% of high-performance computing centers (HPC) around the world are already using AI and big data-related technologies, and HPC is accelerating its integration with AI and big data, moving towards the era of high-performance data analysis with data-intensive characteristics.
In the HPDA era, the data volume of various industries ushered in a blowout growth. In the evolution of seismic exploration from two-dimensional to three-dimensional, the data volume has increased by 10-20 times; In the revolution of upgrading movie rendering from 2K to 8K, the amount of data increased by 16 times; In the field of satellite mapping, the detection accuracy has been reduced from 20 meters to 2 meters, and the data volume has increased by nearly 70 times year-on-year.
In addition to the surge in data scale, the challenges of complex business models and low analysis efficiency are also calling for more efficient storage.
As the carrier of data, storage has gradually become a new kinetic energy to promote the development of HPC industry. However, the traditional HPC storage has barriers in mixed load performance, cost, cross-protocol access and so on, which can not match the needs of HPDA scenarios. How to break the limitations of storage performance, cost and efficiency, and fully release the data potential, has become a restrictive factor for the upgrading of HPC industry.
High-performance data analysis and storage accelerate the development of HPC industry
At present, as a supporting platform for data application and data analysis, as well as the key infrastructure of a powerful country in science and technology, data storage has become a national weapon, and it plays an important role in the fields of financial core transactions, new oil and gas exploration, gene sequencing, autonomous driving, meteorological prediction and space exploration. The ability of data storage and processing has become the key to improve the level of government management, improve the operational efficiency of enterprises and enhance the development resilience of enterprises. Data storage is becoming a solid foundation for accelerating digital transformation.
The new industry changes and the important position of data storage have put forward new challenges to high-end storage, and at the same time, they are accelerating the innovation of storage technology-expanding from some scenes of HPC to the whole scene of HPC/HPDA, and storage has begun to undertake the heavy responsibility of accelerating the transformation of the industry to "data-intensive". According to the data of the HPC market space in 2020 by Hyperion Research, an international authoritative analysis organization, the growth rate of data storage is the first, which is much higher than the average growth rate of the overall market.
High-performance data analysis (HPDA) storage can match the high-end storage of various HPDA scenarios, and can be used for data-intensive HPDA application scenarios such as gene sequencing, meteorological ocean, supercomputing center, energy exploration, scientific research and industrial innovation, intelligent medical care, deep learning, and face recognition. To achieve a leap-forward improvement in efficiency, quality and cost performance.
It is worth noting that Huawei OceanStor Pacific series of next-generation high-performance data analysis (HPDA) storage can efficiently cope with key challenges such as ultra-high-density design, mixed load design and multi-protocol interoperability, and promote the HPC industry to upgrade to data-intensive. At present, it has been successfully applied to many domestic foreign enterprises and institutions in high-performance computing scenarios such as autonomous driving, gene sequencing, weather forecasting and satellite remote sensing.
As an important engine of high-performance data analysis, storage is fully releasing the application value of HPC, promoting the continuous progress of HPC industry, crossing the gap from "calculation-intensive" to "data-intensive", and constantly promoting the prosperity and healthy development of human society.