Inspur KaiwuDB: 'One step faster” – creating a database that better understands the Internet of Things

From June 11 to 13, 2023, the OAGS2023 Open Atom Global Open Source Summit will be held in Beijing. As an annual event in the open source industry, this summit, with the theme of "Open Source Empowerment, Inclusive Future", gathers advantages from various fields such as government, industry, academia, and research, and brings together top celebrities to discuss the future of open source. Wei Kewei, CTO of Inspur KaiwuDB, was invited to attend the summit and share the theme of "Building a "One Step Faster" AIoT Database", combining key technology trends in database technology such as distributed, HTAP, multi-mode, AI4DB and innovative data scenarios in the Internet of Things era. , and discuss the development trends of databases in the new wave of industry with industry experts and technology enthusiasts present.

Picture: KaiwuDB CTO Wei Kewei delivers a keynote speech

According to data, as of 2022, there will be a total of 13 billion Internet of Things (IoT) devices in the world, and this number is still growing rapidly, and is expected to double within four years. By 2030, three-quarters of the world's "dumb" devices are expected to be connected to the Internet. The huge by-product of the Internet of Things is data, which means that all connected devices will become the singular point of data explosion. The IDC report shows that the data generated by IoT devices increased from 0.1ZB in 2013 to 4.4ZB in 2020, a 50-fold increase in 7 years; this data will reach 79.4ZB in 2025, an astonishing growth rate. Wei Kewei said that companies that create and deploy IoT devices are facing unprecedented challenges in data management capabilities: they not only need to think about how to use IoT devices efficiently, but also how to handle the explosive growth of IoT data, including massive time series data collection. Integration of time series data and transactional data, efficient data management for easy access and call, data integration analysis and visual presentation, etc. However, this kind of data storage and processing is often beyond the capacity of traditional data management methods.

KaiwuDB found in practice that the data management scenarios in the field of Internet of Things not only put forward massive time series data processing requirements, horizontal expansion requirements, and high cardinality requirements for the database, but also required deep integration with high-value relational data. Cloud The need for edge-end collaboration, the need for quick response for real-time control, the need for trend analysis and assisted decision-making using AI, etc. As a distributed, multi-mode AIoT database product that supports cloud-edge-device collaboration, KaiwuDB has a patented "in-place computing" technology and is based on MMIO (memory mapping) technology to achieve "whatever is stored is used" and supports millions of data in seconds. Into the database, tens of millions of records can be queried with millisecond-level response. Data exploration of 2 billion records can be completed in 1 second, and 15-level drill-down of 5 million recorded data can be completed in 10 seconds, effectively ensuring high-speed warehousing and rapid query of massive time series data in IoT scenarios. KaiwuDB multi-mode architecture supports the unified storage, processing and AI intelligent analysis of time series data, memory data and relational data in the same database. It realizes automatic selection and automatic conversion of models based on unified operators and communication protocols, which can realize "one-time" five". At the same time, KaiwuDB also has features such as SQL support, data compression, intelligent pre-computing, subscription publishing, cluster deployment, etc., as well as high availability, low cost, and easy operation and maintenance.

Come from practice and go to the scene. Wei Kewei also shared the discrete manufacturing IIoT solution jointly created by KaiwuDB and a heavy industry company in Shandong, which effectively helped the group's factory build unified equipment data collection service capabilities, data compression and hierarchical storage capabilities, real-time analysis of equipment operating status, and start-up. Stop control, cycle parameter adjustment capabilities, etc. Improve the efficiency of data-driven operation management by establishing a factory-level data sharing mechanism.

Practice goes a long way, and things are accomplished. As a new force in the domestic database industry, KaiwuDB explores the needs of the AIoT field, deeply explores the innovative data scenarios constantly derived from the Internet of Things industry, and strives to become a database that better understands the Internet of Things; but it not only focuses on the Internet of Things, but also focuses on digital energy, Digital government, finance and other industries have also been successfully implemented. Today, as the open source ecosystem at home and abroad is developing rapidly, KaiwuDB will also strive to strengthen its own technological innovation, explore more possibilities with partners in the ecosystem, promote the sustainable development of the database industry, and jointly create new value for data.

The above is the detailed content of Inspur KaiwuDB: 'One step faster” – creating a database that better understands the Internet of Things. For more information, please follow other related articles on the PHP Chinese website!