At present, all parts of the country are vigorously promoting the construction of new infrastructure, and this trend will continue to be maintained. With the continuous advancement of new infrastructure represented by 5G, industrial Internet, artificial intelligence and other technologies, it will inevitably bring more ubiquitous connectivity, more massive data, more diversified data types, and more common data analysis and processing needs, which bring great challenges to storage systems.
At the "2020 China New Infrastructure Leadership Summit" on December 18, which was guided by Xinhuanet and the Academic Department of the Chinese Association for Science and Technology Association, hosted by Xinhuanet Client and China Institute of Electronics, hosted by Zhiding Media, and co-organized by Technology Walker and global mobile communication system association (GSMA), Zhang Xinfeng, deputy general manager of Shuguang Storage Division, shared in his keynote speech the best practices of how Shuguang can help customers digitally transform through its massive storage system in the current "new infrastructure" environment. Bring new thinking to the industry's solution to storage challenges.
Zhang Xinfeng, deputy general manager of Shuguang Storage Division
New changes in the storage industry
Today we have entered an era of big data, the scale of data is showing explosive growth, at the same time, data is becoming the core production factor of many enterprises, in the digital transformation of enterprises, the basic supporting role of data is becoming increasingly prominent, more and more enterprises through data analysis and processing to gain competitive advantage.
In his speech, Zhang Xinfeng cited a set of IDC research results: the global data circle will grow more than 5 times from 2018 to 2025, when the global interconnected devices will exceed 150 billion, most of which will generate data in real time.
Another significant feature of the big data era is that the amount of data generated by the edge and terminal is increasing, and this momentum will continue unabated with the popularity of 5G. Self-driving cars, for example, generate 3TB of data per hour, not including entertainment information and GPS data on in-vehicle devices. As car-to-vehicle connections become more commonplace, and as more and more pattern-specific machine learning and AI algorithms are implanted, the amount of data will increase significantly.
In response to the latest changes in the industry, the storage industry has made a lot of efforts. In Zhang Xinfeng's view, there are four major trends in the storage field that deserve attention:
The first is innovation in storage media. Innovation focuses on two directions: one is the pursuit of faster storage media, the pursuit of real-time, a large number of new technology applications to create better performance, such as SSD, NVMe; the other is a larger capacity, and is a high capacity at low cost, such as quantum storage technology, the purpose is to provide a medium for the cheap storage of massive data.
Second, different industries present completely different demands on storage. For example, streaming media applications generally pay more attention to performance jitter; financial securities generally pay more attention to latency issues. In order to meet different needs, it is necessary to provide a software-hardware, customized storage system.
Third, the intelligent management requirements of storage. The growth of storage volume has brought about the complexity of operation and maintenance, making automatic early warning and automatic operation and maintenance of storage systems more and more critical.
Finally, there's the larger amount of data, which is one of the biggest challenges for storage systems today. New technologies such as big data, the Internet of Things, the Industrial Internet, and 5G have brought about an explosive growth of diversified data, especially the exponential growth of unstructured data stored by enterprises in the local, edge, or public cloud as file or object storage.
"In the past, our storage reached the petabyte level is a very large-scale application, and in the past two years, we have contacted a lot of single-user data scale of hundreds of petabytes, which is not suitable for this kind of massive data traditional storage system method." Zhang Xinfeng said.
Immerse yourself in polishing and produce good mass storage products
Shuguang is a local enterprise born from the Chinese Academy of Sciences, mainly for government and enterprise customers to build IT infrastructure and various solutions and services. After more than 20 years of development, Sugon currently has seven product lines, of which storage is one of its two core businesses (the other core business is server). As the core business of Sugon, Sugon has continued to invest in storage technology for many years and has made many technological breakthroughs, especially in massive storage. At present, a first-class team has been established in China that has mastered the core technology of distributed storage, and has more than 200 related technology invention patents. According to IDC data, Shuguang's self-developed massive distributed storage system ranked in the top two domestic IDC products for 7 consecutive years.
According to Zhang Xinfeng, Sugon's massive storage system architecture is divided into three layers: the lowest layer is the hardware platform, based on the fully standardized x86 platform, adapted to Various processors and high-speed network products such as Intel, such as 25GB, 100GB, 200GB network; the middle is the most core distributed software layer, the most critical of which is Sugon's self-developed local disk management system, which is very scalable, performance and reliability compared with similar software; the top layer is the application, including NFS, FTP, etc to meet the needs of users to store different data.
"The system is fully optimized for massive unstructured data storage scenarios, and is widely applicable to a variety of application areas with data sharing needs, such as the Internet, media asset management, video editing and processing and other industries or scenarios." Zhang Xinfeng said.
Zhang Xinfeng said that Shuguang massive storage system has the following three core advantages:
The first is high performance. High performance depends primarily on the architecture itself, which is a fully integrated architecture. At the same time, the single-node performance has also been optimized, which can provide better performance for users who pursue extremes. In addition, the system leverages state-of-the-art networking technology and extensive optimizations on network protocols to further improve storage performance.
Secondly, the high reliability of the storage system. The reliability of the entire system first comes from the reliability of a single hardware, and then combined with the software to achieve the reliability of the entire storage system, the cluster, all firmware, all nodes and networks are redundantly designed, there is no single point of failure.
Finally, in terms of ease of management, the storage system is implanted with intelligent algorithms, and new nodes are added to identify automatic processing at the first time, which never requires personnel to do too much intervention management. At the same time, multiple systems can be unified management, and one management system can manage 64 sets of clusters, which means that dozens of EP-level data can be managed through a management interface.
Zhang Xinfeng introduced that thanks to Sugon's continuous innovation in technology, Shuguang storage system has now played a role in more than 30,000 enterprises across the country, and the capacity of managing data has exceeded 10 EB, which has played an important role in the digital transformation of many politician customers.
"Today's data storage needs have reached a new stage, in order to cope with massive data we need to make changes in technology, solutions, products, Sugon distributed storage system is what we have made." Zhang Xinfeng concluded.