The following is the transcript of Wei Kewei, CTO of Inspur KaiwuDB, who was invited to give a speech at the Trusted Database Development Conference held in Beijing on July 4th. Welcome to like, bookmark, and follow!
Create a multi-mode engine, the road to AIoT database exploration
0 1 What is "complex"?
In the era of Industry 4.0, the Internet of Things industry drives the market of data elements to grow continuously and become a new long-term growth point. A truly competitive Internet of Things enterprise must not only be able to cope with the high-speed writing, storage, and analysis of multi-modal mass data brought about by the development of the Internet of Things, but more importantly, provide more value-added services based on the huge amount of data . This makes the enterprise's demand for databases not only stay on a time-series database with excellent performance, but also requires the database to have different engine capabilities including transaction analysis engines to support online transactions and data analysis.
Throughout the domestic database market, multi-database application scenarios are becoming more and more common, but this model has brought many usage problems: for developers, it is necessary to have best practices and solid technologies for developing different models; for DBAs, it is necessary to clearly Clarify the operation and maintenance rules of different databases, and the cost of development and operation is high; in addition, data transmission and conversion between different databases will involve multiple copy management, data is prone to errors during conversion and transmission, and it is also very difficult in terms of performance and consistency Guaranteed.
Multi-database challenges in AIoT scenarios
Facing the complex challenges of data storage and governance brought about by multi-modal data in the Internet of Things era, how to integrate these "weapons"—including the integration of hardware and software, the fusion of computing models, and the fusion of storage models, so as to accelerate data Valuation and improving user data consumption capabilities are major issues that need to be solved urgently in today's databases.
02 Such "simple"!
In view of the above situation, Wei Kewei proposed: KaiwuDB truly realizes "simplification" through "multi-mode architecture".
As the industry's first "AIoT-oriented distributed multi-mode database" product, KaiwuDB is based on unified SQL syntax, unified database commands, unified development and maintenance tools, unified security certification, etc. to achieve true multi-mode integration, high-performance timing engine , Flexible deployment of machine learning engines. Relying on core technologies such as "in-place computing" and "native AI" to support time-series data, memory data, and relational data in the same database for unified storage, processing, and AI intelligent analysis, "one database for multiple purposes" stores complex data The management process is greatly simplified .
KaiwuDB multi-mode architecture
Under the multi-mode architecture, developers only need to deal with one database, and operation and maintenance personnel only need to operate and maintain one script to manage the database. Different data models can be converted between the same set of databases to ensure data consistency and conversion performance. In this way, going from multi-database to multi-mode will greatly reduce the cost of user development and operation and maintenance, and truly realize "cost reduction and efficiency improvement" .
Simplifying applications with innovative database technology and continuously creating value for users' digital transformation and cost reduction and efficiency increase are the direction of the joint efforts of databases at this stage. Defining the database from the perspective of business rather than data model is an important starting point for KaiwuDB's distributed multi-mode database architecture design.
The "simplification" multi-mode architecture provides a new development direction for databases, especially for AIoT scenarios . In the future, KaiwuDB will continue to deepen the multi-mode architecture, strive to promote the data value mining of user enterprises through more technical practices, and empower enterprises to upgrade digitally and intelligently.