Оффлайн
- Регистрация
- 1 Мар 2015
- Сообщения
- 14,205
- Баллы
- 155
Brief History of DBMS
Database Management System (DBMS) is a fundamental software that manages and organizes data in a unified, reliable, and efficient manner, ensuring the security and integrity of the data while providing efficient data querying capabilities.
Database refers to a collection of data organized, stored, and managed according to its data structure.
Note:
The term Database is often used interchangeably with Database Management System (DBMS), with a conceptual distinction made only when both appear simultaneously. Our documentation may feature such interchangeable usage.
Hierarchical Database and Network Database were the two main types of the first-generation database management system.In a hierarchical database, data is organized in a tree structure (similar to a file system), while in a network database, data is organized in a many-to-many network structure.
In 1970, E.F. Codd first proposed the Relational Model, providing a unified and concise data model for database management systems. A database that conforms to the relational model is called a Relational Database, where data is stored in relational tables composed of rows (tuples) and columns (attributes). The corresponding DBMS is known as Relational Database Management System (RDBMS), which is still the most mainstream type of database management system today.
Brief History of YashanDB
YashanDB, the Yashan database management system, is a new type of database management system independently designed and developed by Shenzhen Institute of Computing Sciences (SICS). Based on classical database theory, it incorporates original theories of bounded computation, approximate computation, parallel scalability, and cross-model integration computation, meeting the demands for high performance, high concurrency, and high security in critical industries such as finance, government, and energy.
The development of YashanDB can be divided into the following stages:
2013~2018 Original Theory System Proof and Foundation
Pioneered the theory of big data computational complexity, laying the foundation for the complexity of big data queries.
Developed original theory systems, establishing a theoretical foundation, with over 60 A-class papers.
Topics include bounded computation, incremental computation, approximate computation, parallel computation, big data quality, cross-model integration, correlation analysis, Logic+a1, and big data computational complexity theory.
Achieved the grand slam for best paper and the ten year inspection award at top database conferences.
2019 Official Operation of SICS
Shenzhen Institute of Computing Sciences officially commenced operations, initiating the engineering of bounded computation theory.
A group of young scientists from globally renowned universities and experienced engineers from well-known companies gathered.
Achieved preliminary verification of theoretical feasibility, while the product was showcased at CNCC, and DeepTech conducted an interview.
2020~2021 Prototype Verification, Product and Engineering Implementation
Completed feasibility verification of distributed analysis prototype development .
Completed the first phase of original theory, core technology, and self-developed systems for YashanDB, that achieved core system replacement capability.
2022 Authoritative Testing Certification, First Year of Marketization
Certification of compatibility with mainstream domestic processors, servers, OS, and middleware.
Won the "Top Ten Hardcore Technologies" award at the Fifth Digital China Construction Summit.
A 100% self-managed kernel code rate, authority tested by institutions under the Ministry of Industry and Information Technology (MIIT).
Trusted databases for centralized transaction/distributed analysis, conducted evaluations by China Academy of Information and Communications Technology (CAICT).
Over 30 customers across various industries including finance, government, energy, transportation, and state-owned enterprises.
Completed feasibility verification for prototype development for advanced cluster.
2023~Present Market Replication, High-end Substitution
Penetration and scale replication in industries such as finance, operators, government, and energy.
Release of YashanDB for Cluster and Commercialization with distributed deployment of YashanDB.
Official release of v22.2, with comprehensive enhancements in functionality, performance, and stability.
Strategic cooperation agreement signed with China Academy of Industrial Internet (CAII).
CMMI Maturity Level 3 certification passed by Shenzhen Yashan Technology.
Level 3 of information security Level Protection.
Release of GIS solution, jointly with Zhongdi Digital.
First appearance at the Trusted Database Development Conference, list of the "China Database Industry Map" and contributing editor of the "Database Development Research Report".
One of the first co-construction units for the telecommunications industry working group under the Database Application Innovation Laboratory (DBL).
Deployment Forms of YashanDB
Standalone (Primary-Standby) Deployment (abbreviated as Standalone Deployment)
YashanDB with Standalone Deployment combines traditional relational database theory with innovative underlying engine technology, suitable for centralized transactional business scenarios, supporting primary/standby forms.
Distributed Cluster Deployment (abbreviated as Distributed Deployment)
YashanDB with Distributed Deployment is a native distributed database inheriting standalone capabilities, suitable for distributed analysis business scenarios, supporting primary/standby forms.
YAC Deployment
YashanDB with YAC is a multi-active cluster based on shared storage, providing compute/storage scaling and financial-grade high availability capabilities, suitable for high-end core transactional business scenarios.
Core Features of YashanDB
Row-based/Column-based Storage
Supports Heap Table(Row-based), TAC Table(Colum-based) and LSC Table(Colum-based).
Provides list/range/hash/interval partition types and various composite partitioning capabilities.
Supports vectorization calculation.
Transaction Management
Supports ACID for completed transaction, fine-grained lock management, and data consistency for statements reading/writing. Supports transaction isolation levels including read committed and serialized, undo self-management, and MVCC.
High Performance Query
Provides an optimizer model based on costs and rules.
Supports distributed execution method based on MPP. Features data sorting, sparse indexing, pre-reading/caching, data compression on storage technologies , and partition pruning, parallel querying, and condition pushdown on SQL engines, resulting in high performance queries.
Database Replication
Supports both synchronous and asynchronous replication modes. Provides various replication strategies such as maximize protection and maximize performance.
Backup and Recovery
Provides physical backup and logical backup capabilities. Supports full backup and incremental backup. Supports Point-In-Time Recovery.
High Availability
Provides one-primary/multiple-standby capability, as well as cascade-standby. Supports switchover and failover based on leader election.
Flashback
Provides flashback query, flashback modification, and flashback drop based on recycle bin capabilities.
General SQL Capabilities
Supports common SQL syntax, complying with ANSI SQL standards. Offers a rich library of functions and various data types. Supports high-performance plan operators.
PL
Supports stored procedures, user defined functions, anonymous blocks, advanced packages, jobs, triggers, and more.
Cohesive Memory
Supports core technology for cohesive memory, used for coordinated read-write access of data blocks across instances in YAC and various concurrent controls not related to data.
File System Capabilities
Provides file system services directly managing raw devices. Offers parallel file read-write capabilities for multi-node clusters in YAC.
Spatial Data Management
Supports the ST_GEOMETRY data type, used for storing and accessing geometric objects complying with the SFA SQL standard established by the Open Geospatial Consortium (OGC).
Online Scaling
Supports online scaling in distributed deployment forms, catering to customer business development needs.
Applicable Scenarios for YashanDB
Online Transaction Processing
YashanDB's centralized transaction system focuses on the development of next-generation infrastructure hardware and software, driven by core application scenarios in critical industries, providing a high-performance and reliable database foundation to meet the business needs of high-concurrency online transaction processing.
Online Transactions with Extremely High Performance Requirements
These scenarios have strict requirements for the timeliness and accuracy of transaction processing, requiring high-reliability protection for data. YashanDB features refined transaction management capabilities, leveraging a robust storage foundation combined with a high-speed SQL engine to achieve excellent transaction processing performance while ensuring strong consistency of data.
7*24 Uninterrupted Service
YashanDB's high availability capability, through a multi-replica log synchronization mechanism, can synchronize and persist data in real-time across multiple data centers, implementing online primary-standby switching through automatic leader election based on the Raft protocol, ensuring a stable and continuous operational state of the system without user perception.
Centralized Enterprise Management
YashanDB's HTAP hybrid load model is based on "the same data, the same engine", supporting both online real-time transactions and real-time analyses simultaneously, providing massive data's real-time online analysis capabilities while ensuring high-concurrency online throughput.
Seamless Switching in High Availability Scenarios
Industries like finance, telecommunications, and electricity have high demands for availability, often using high-availability architectures like Oracle RAC that rely on shared storage. YAC possesses similar high-availability architecture capabilities; when a single point of failure occurs in the cluster, another node will take over with transparent client switching, achieving RPO of 0 and RTO in just over ten seconds, ensuring core business continuity without interruption.
Massive Data Analysis
YashanDB's distributed analysis system is based on bounded computation theory for instant analysis, focusing on solving bottleneck issues related to big data computation efficiency, real-time data analysis, mass data storage costs, and data silos (3V: Volume, Velocity, Variety), providing customers with flexible traditional data warehouse acceleration and one-stop data warehouse solutions.
Massive Steady-State Data Analysis
YashanDB's LSC table organizes data in a columnar structure, supporting hot/cold data (also known as "mutable/stable Data") separation and high-compression object storage. By utilizing data sorting, sparse indexing, filtering, and other technologies, it achieves high-performance queries for massive data while supporting mutable data (or called hot data) area writes to enhance transaction performance, and supports silent conversion and fusion querying between hot data areas and cold data areas, primarily focusing on interactive analysis scenarios for massive steady-state data.
Ad-Hoc Interactive Self-Service Analysis Scenarios
YashanDB conducts interactive exploratory analysis on business, and provides customers with second-level response experiences for query analysis through our fully self-developed technologies including cost-effective column-based storage engine, vectorization execution engine, and efficient distributed algorithms.
Real-Time Personalized Recommendation Scenarios
YashanDB provides multi-dimensional analysis based on massive historical user behavior log data, supporting real-time entry of behavior logs into the database, completing second-level calculations of historical tags and real-time tags, and supporting user selection through various tag combinations to increase marketing success rates.
Database Management System (DBMS) is a fundamental software that manages and organizes data in a unified, reliable, and efficient manner, ensuring the security and integrity of the data while providing efficient data querying capabilities.
Database refers to a collection of data organized, stored, and managed according to its data structure.
Note:
The term Database is often used interchangeably with Database Management System (DBMS), with a conceptual distinction made only when both appear simultaneously. Our documentation may feature such interchangeable usage.
Hierarchical Database and Network Database were the two main types of the first-generation database management system.In a hierarchical database, data is organized in a tree structure (similar to a file system), while in a network database, data is organized in a many-to-many network structure.
In 1970, E.F. Codd first proposed the Relational Model, providing a unified and concise data model for database management systems. A database that conforms to the relational model is called a Relational Database, where data is stored in relational tables composed of rows (tuples) and columns (attributes). The corresponding DBMS is known as Relational Database Management System (RDBMS), which is still the most mainstream type of database management system today.
Brief History of YashanDB
YashanDB, the Yashan database management system, is a new type of database management system independently designed and developed by Shenzhen Institute of Computing Sciences (SICS). Based on classical database theory, it incorporates original theories of bounded computation, approximate computation, parallel scalability, and cross-model integration computation, meeting the demands for high performance, high concurrency, and high security in critical industries such as finance, government, and energy.
The development of YashanDB can be divided into the following stages:
2013~2018 Original Theory System Proof and Foundation
Pioneered the theory of big data computational complexity, laying the foundation for the complexity of big data queries.
Developed original theory systems, establishing a theoretical foundation, with over 60 A-class papers.
Topics include bounded computation, incremental computation, approximate computation, parallel computation, big data quality, cross-model integration, correlation analysis, Logic+a1, and big data computational complexity theory.
Achieved the grand slam for best paper and the ten year inspection award at top database conferences.
2019 Official Operation of SICS
Shenzhen Institute of Computing Sciences officially commenced operations, initiating the engineering of bounded computation theory.
A group of young scientists from globally renowned universities and experienced engineers from well-known companies gathered.
Achieved preliminary verification of theoretical feasibility, while the product was showcased at CNCC, and DeepTech conducted an interview.
2020~2021 Prototype Verification, Product and Engineering Implementation
Completed feasibility verification of distributed analysis prototype development .
Completed the first phase of original theory, core technology, and self-developed systems for YashanDB, that achieved core system replacement capability.
2022 Authoritative Testing Certification, First Year of Marketization
Certification of compatibility with mainstream domestic processors, servers, OS, and middleware.
Won the "Top Ten Hardcore Technologies" award at the Fifth Digital China Construction Summit.
A 100% self-managed kernel code rate, authority tested by institutions under the Ministry of Industry and Information Technology (MIIT).
Trusted databases for centralized transaction/distributed analysis, conducted evaluations by China Academy of Information and Communications Technology (CAICT).
Over 30 customers across various industries including finance, government, energy, transportation, and state-owned enterprises.
Completed feasibility verification for prototype development for advanced cluster.
2023~Present Market Replication, High-end Substitution
Penetration and scale replication in industries such as finance, operators, government, and energy.
Release of YashanDB for Cluster and Commercialization with distributed deployment of YashanDB.
Official release of v22.2, with comprehensive enhancements in functionality, performance, and stability.
Strategic cooperation agreement signed with China Academy of Industrial Internet (CAII).
CMMI Maturity Level 3 certification passed by Shenzhen Yashan Technology.
Level 3 of information security Level Protection.
Release of GIS solution, jointly with Zhongdi Digital.
First appearance at the Trusted Database Development Conference, list of the "China Database Industry Map" and contributing editor of the "Database Development Research Report".
One of the first co-construction units for the telecommunications industry working group under the Database Application Innovation Laboratory (DBL).
Deployment Forms of YashanDB
Standalone (Primary-Standby) Deployment (abbreviated as Standalone Deployment)
YashanDB with Standalone Deployment combines traditional relational database theory with innovative underlying engine technology, suitable for centralized transactional business scenarios, supporting primary/standby forms.
Distributed Cluster Deployment (abbreviated as Distributed Deployment)
YashanDB with Distributed Deployment is a native distributed database inheriting standalone capabilities, suitable for distributed analysis business scenarios, supporting primary/standby forms.
YAC Deployment
YashanDB with YAC is a multi-active cluster based on shared storage, providing compute/storage scaling and financial-grade high availability capabilities, suitable for high-end core transactional business scenarios.
Core Features of YashanDB
Row-based/Column-based Storage
Supports Heap Table(Row-based), TAC Table(Colum-based) and LSC Table(Colum-based).
Provides list/range/hash/interval partition types and various composite partitioning capabilities.
Supports vectorization calculation.
Transaction Management
Supports ACID for completed transaction, fine-grained lock management, and data consistency for statements reading/writing. Supports transaction isolation levels including read committed and serialized, undo self-management, and MVCC.
High Performance Query
Provides an optimizer model based on costs and rules.
Supports distributed execution method based on MPP. Features data sorting, sparse indexing, pre-reading/caching, data compression on storage technologies , and partition pruning, parallel querying, and condition pushdown on SQL engines, resulting in high performance queries.
Database Replication
Supports both synchronous and asynchronous replication modes. Provides various replication strategies such as maximize protection and maximize performance.
Backup and Recovery
Provides physical backup and logical backup capabilities. Supports full backup and incremental backup. Supports Point-In-Time Recovery.
High Availability
Provides one-primary/multiple-standby capability, as well as cascade-standby. Supports switchover and failover based on leader election.
Flashback
Provides flashback query, flashback modification, and flashback drop based on recycle bin capabilities.
General SQL Capabilities
Supports common SQL syntax, complying with ANSI SQL standards. Offers a rich library of functions and various data types. Supports high-performance plan operators.
PL
Supports stored procedures, user defined functions, anonymous blocks, advanced packages, jobs, triggers, and more.
Cohesive Memory
Supports core technology for cohesive memory, used for coordinated read-write access of data blocks across instances in YAC and various concurrent controls not related to data.
File System Capabilities
Provides file system services directly managing raw devices. Offers parallel file read-write capabilities for multi-node clusters in YAC.
Spatial Data Management
Supports the ST_GEOMETRY data type, used for storing and accessing geometric objects complying with the SFA SQL standard established by the Open Geospatial Consortium (OGC).
Online Scaling
Supports online scaling in distributed deployment forms, catering to customer business development needs.
Applicable Scenarios for YashanDB
Online Transaction Processing
YashanDB's centralized transaction system focuses on the development of next-generation infrastructure hardware and software, driven by core application scenarios in critical industries, providing a high-performance and reliable database foundation to meet the business needs of high-concurrency online transaction processing.
Online Transactions with Extremely High Performance Requirements
These scenarios have strict requirements for the timeliness and accuracy of transaction processing, requiring high-reliability protection for data. YashanDB features refined transaction management capabilities, leveraging a robust storage foundation combined with a high-speed SQL engine to achieve excellent transaction processing performance while ensuring strong consistency of data.
7*24 Uninterrupted Service
YashanDB's high availability capability, through a multi-replica log synchronization mechanism, can synchronize and persist data in real-time across multiple data centers, implementing online primary-standby switching through automatic leader election based on the Raft protocol, ensuring a stable and continuous operational state of the system without user perception.
Centralized Enterprise Management
YashanDB's HTAP hybrid load model is based on "the same data, the same engine", supporting both online real-time transactions and real-time analyses simultaneously, providing massive data's real-time online analysis capabilities while ensuring high-concurrency online throughput.
Seamless Switching in High Availability Scenarios
Industries like finance, telecommunications, and electricity have high demands for availability, often using high-availability architectures like Oracle RAC that rely on shared storage. YAC possesses similar high-availability architecture capabilities; when a single point of failure occurs in the cluster, another node will take over with transparent client switching, achieving RPO of 0 and RTO in just over ten seconds, ensuring core business continuity without interruption.
Massive Data Analysis
YashanDB's distributed analysis system is based on bounded computation theory for instant analysis, focusing on solving bottleneck issues related to big data computation efficiency, real-time data analysis, mass data storage costs, and data silos (3V: Volume, Velocity, Variety), providing customers with flexible traditional data warehouse acceleration and one-stop data warehouse solutions.
Massive Steady-State Data Analysis
YashanDB's LSC table organizes data in a columnar structure, supporting hot/cold data (also known as "mutable/stable Data") separation and high-compression object storage. By utilizing data sorting, sparse indexing, filtering, and other technologies, it achieves high-performance queries for massive data while supporting mutable data (or called hot data) area writes to enhance transaction performance, and supports silent conversion and fusion querying between hot data areas and cold data areas, primarily focusing on interactive analysis scenarios for massive steady-state data.
Ad-Hoc Interactive Self-Service Analysis Scenarios
YashanDB conducts interactive exploratory analysis on business, and provides customers with second-level response experiences for query analysis through our fully self-developed technologies including cost-effective column-based storage engine, vectorization execution engine, and efficient distributed algorithms.
Real-Time Personalized Recommendation Scenarios
YashanDB provides multi-dimensional analysis based on massive historical user behavior log data, supporting real-time entry of behavior logs into the database, completing second-level calculations of historical tags and real-time tags, and supporting user selection through various tag combinations to increase marketing success rates.