KunDB
Transwarp Distributed Transactional Database
SQL Compatibility
/
Strong Consistency
/
High Performance
/
High Availability
Distributed Transactional Database Highly Compatible with MySQL and Oracle
KunDB provides complete capability and reliable migration solutions
Eight Reasons to Choose KunDB as the Transactional Database
SQL Compliance & ACID
Supports standard SQL and compatible with MySQL,Oracle dialects and ecology, and Oracle PL/SQL. Provides distributed transaction support and rigorously ensures consistency in distributed architecture. Be completely transparent to back-end distributed architecture using in business operations and follow centralized database user habits, to reduce migration costs and learning costs.
Performance Scaling
Distributed architecture ensures excellent horizontal scalability and high concurrency, for the easy handling of highly concurrent heavy traffic, avoiding the bottleneck caused by database performance in the system.
Highly Available and Distributed Deployment
Two deployment modes: the highly available mode and distributed mode. The highly available deployment mode provides highly available storage for multiple storage node components and is fully compatible with MySQL while the distributed mode at the bottom layer provides distributed storage with horizontal partitioning, with high compatibility with MySQL and Oracle and support for distributed query. Highly available mode and distributed mode are easily switched after deployment.
Private Deployment and Cloud Support
The database combines distributed architecture and container technology to achieve the service and flexible arrangement of database components, taking into account the private deployment and cloud deployment scenarios, with rapid expansion and contraction, elastic scheduling, and easy adaptation of heterogeneous hardware environment and full use of hardware resources.
Excellent Performance in High-concurrency Transaction Scenarios
In high concurrent transaction scenarios, small scale KunDB cluster can easily achieve tens of thousands of TPS. Take TPCC benchmark test data, the transaction processing capacity of as few as 3 KunDB nodes can replace Oracle.
Ad-hoc Query and Distributed Computing
Built-in Transwarp self-developed distributed computing engine can read the latest data for ad-hoc query and complex analysis without the ETL process.
Security and Reliability
Transwarp self-developed domestic database supports domestic CPUs, servers and other independent and controllable hardware platforms and ensures the security of data information. It provides multi-dimensional security of the database in terms of user authentication, discretionary access control (DAC) and mandatory access control (MAC), link and data encryption, operation and SQL audit, etc., to strictly guarantee the security and integrity of data and business.
High Availability and Disaster Tolerance
With built-in fault recovery and off-site disaster recovery mechanism, the system is highly available through multi-copy + failover to guarantee 7x24 continuous service.
New-generation Distributed Database Core Capabilities
Core Components
SQL Access
SQL Access
Connecting to networks, managing SQL parse and assigning routes
Optimizer
Optimizer
To optimize SQL compilation results, include logical optimization and physical optimization, support of various optimization models (such as RBO and CBO)
Datastore Cluster
Datastore Cluster
The actual data-storing databases of each partition supports MVCC and transactions, ensures multi-copy consistency by master-slave synchronization, and supports fast master/standby fail over
SQL Compiler
SQL Compiler
Provides SQL syntax parsing, compilation and optimization to be compatible with MySQL and Oracle syntax
Efficient Compute Engine
Efficient Compute Engine
Concurrently reads the partition table data in each storage layer for ad-hoc query and distributed computing
Global Transaction Manager
Global Transaction Manager
Distributed transaction management unit, providing global transaction management capabilities and Snap-shots, implementing distributed transactions based on 2PC and global transaction management, ensuring ACID attributes of transaction, supporting RR, Serializable and other transaction isolation levels
PL/SQL Compiler
PL/SQL Compiler
Provides PL/SQL syntax parsing, compilation and optimization, and compatibility with Oracle PL/SQL.
Scheduling Executor
Scheduling Executor
Routes operations that do not cross Shard to the corresponding storage database cluster; handles complex analysis and transaction operations across data partitions
Metadata Management
Metadata Management
Provides metadata persistence and query service within the database, built on distributed consistent storage
SQL access
Connecting to networks, managing SQL parse and assigning routes
SQL compiler
Provides SQL syntax parsing, compilation and optimization to be compatible with MySQL and Oracle syntax
PL/SQL compiler
Provides PL/SQL syntax parsing, compilation and optimization, and compatibility with Oracle PL/SQL.
Optimizer
To optimize SQL compilation results, include logical optimization and physical optimization, support of various optimization models (such as RBO and CBO)
Scheduling Executor
Routes operations that do not cross Shard to the corresponding storage database cluster; handles complex analysis and transaction operations across data partitions
Efficient Compute Engine
Concurrently reads the partition table data in each storage layer for ad-hoc query and distributed computing
Datastore Cluster
The actual data-storing databases of each partition supports MVCC and transactions, ensures multi-copy consistency by master-slave synchronization, and supports fast master/standby fail over
Global Transaction Manager
Distributed transaction management unit, provids global transaction management capabilities and Snap-shots, implements distributed transactions based on 2PC and global transaction management, ensures ACID attributes of transaction, supports RR, Serializable and other transaction isolation levels
Metadata Management
Provides metadata persistence and query service within the database, built on distributed consistent storage
Core Functions
SQL Engine
Horizontal Slicing
User Management
Disaster Backup and Recovery
Data and Operations Management Tools
Heterogeneous Platforms Hybrid Deployment
Authorization
Distributed Transaction
Online Scaling
Data Encryption
Multi-tenant Support
Ad-hoc Query and Distributed Computing
SQL Engine
Horizontal Slicing
User Management
Disaster Backup and Recovery
Data and Operations Management Tools
Heterogeneous Platforms Hybrid Deployment
Authorization
Distributed Transaction
Online Scaling
Data Encryption
Multi-tenant Support
Ad-hoc Query and Distributed Computing
Outstanding Performance in Financial Scenario Performance Testing
SQL Compatibility, High Concurrency, Strong Consistency, High Availability, Hybrid Workload
Technical Scenarios
Highly Concurrent Read and Write
KunDB, as a transaction database, has done multi-layer optimization for data reading and writing; the distributed cluster easily supports tens of thousands of TPS. With a stable architecture and data redundancy, the high availability of the cluster is guaranteed. It is highly suitable for transactional scenarios with high concurrent reading and writing demand for data and ACID guarantee, such as the core business systems of finance, telecom, energy, etc.; it is also suitable for scenarios with high concurrency and large-capacity storage demand for database, such as IoT, log storage and other systems.
Hybrid Transactional and Analytical Processing
KunDB has built-in analysis enhancement engine and supports flexible distributed computing, reduces the overhead of data synchronization compared to HTAP composed of heterogeneous databases. Combined with Transwarp self-developed analytic engine technology and unified metadata management service, it can also handle complex analytic SQL as a transaction library with high performance. It is very suitable for scenarios with analytical processing, such as fund clearing, supervision and audit, etc.
Service-oriented SOA
The adoption of micro-services and other architectures to reconstruct the centralized architecture causes problem in distributed transaction consistency. To meet the problem, KunDB can be used as a unified underlying database. The business is still built with micro-services, what's different is that all micro-services are connected to KunDB in a unified method. Developers only need to consider the business implementation and merge the transaction logic of each micro-service in the same transaction, and KunDB will ensure consistency of the distributed transactions.
Industry Scenarios
Core Transaction System in Finance
With an aim to be a financial-grade distributed database with strong consistency guarantee, KunDB provides high concurrency and high stability performance guarantee for core transaction system in finance, and also meets the demand of partial batch query computing. Through cross-engine room and off-site disaster recovery solutions, KunDB can achieve RPO=0 and meet the requirements of up to hundreds of systems in financial core trading system for database.
Telecommunication Billing System
Billing systems require extremely high stability and transient response capability of database. KunDB provides high concurrent read and write capability through horizontal data splitting and transaction concurrency control technology with distributed architecture, and ensures reliable operation of database through rich operation and maintenance tools such as KunDBA.
E-government System
The development of Internet-based e-government system requires high concurrent database with large-capacity storage. KunDB's horizontal scaling capability and multi-tenancy capability provide easy to expand and manage infrastructure services for the incubation and development of government Internet and many innovative services.
Medical Information System
KunDB cooperates with mainstream medical solution providers and tries to replace core systems such as HIS and HRP with domestic alternatives. Through distributed data storage and Oracle PL/SQL support, KunDB provides support for adaption of medical information systems to domestic database and for regional medical center construction.
Success Stories
Government office system replacing Oracle
A provincial government system uses KunDB to replace the Oracle appliance, solving the problem of insufficient performance caused by 30+ system accessing databases, handling new Internet services and complex analysis requirements.
• 32-node X86 server replaced the 6-node Oracle appliance, saving 50% of hardware and software cost.
• Concurrent transaction processing and analysis processing capabilities are improved by 5~10 times
• 32-node X86 server replaced the 6-node Oracle appliance, saving 50% of hardware and software cost.
• Concurrent transaction processing and analysis processing capabilities are improved by 5~10 times
Saving
50%
Hardware and software costs
Enhance
5-10 times faster
Concurrent transaction processing and analysis processing capabilities
Transwarp, Shaping the Future Data World