Functional Features
Business Transparent Access
In the traditional master-replica node model, each primary UDB node and secondary node has a separate connection address. Users need to configure each address in their business program to direct write requests to the primary UDB node and read requests to the secondary nodes.
The UDB database proxy middleware integrates multiple master-replica nodes into a single cluster. By providing an additional proxy middleware address, read and write operations can be performed on the primary and secondary nodes. This approach makes the Read-Write statement forwarding logic transparent to the business, enabling easy implementation of Read-Write separation without the need to modify the business system.
Flexible Dynamic Scaling
Compared to the single-node deployment of open-source Read-Write separation middleware in the industry, the UDB Read-Write Separation Proxy offers dynamic scaling capabilities. It allows for adjusting the number of nodes or upgrading/downgrading machine configurations to meet business requirements. Additionally, it employs distributed load balancing products like ULB for load balancing and disaster recovery, enhancing the overall robustness and stability of the system.
Multiple Read Weight Distribution Strategies
In a console that enables one-click activation or deactivation of Read-Write separation functionality. It also provides four read weight distribution strategies, allowing customers to flexibly choose and configure based on their specific business needs. These strategies are designed to optimize database operations and cater to different usage scenarios.
- Primary and Secondary Balance: Distributes read requests evenly between the primary and secondary databases.
- Primary Focused: Directs all read requests to the primary database only, with no requests to the secondary databases.
- Secondary Only Balance: Distributes read requests evenly across all secondary databases, excluding the primary database.
- Custom: Allows customers to define their own distribution ratios for read requests.
SQL Throttling
This feature enables the activation of SQL throttling after employing a database proxy. It allows flexible configuration based on business needs and helps control the number of time-consuming SQL requests, thus ensuring database performance isn't degraded by excessive time-consuming SQL queries.
Automatic Fault Tolerance
The database proxy performs health checks on all nodes in the master-replica replication cluster. If the primary node is found unavailable, write requests and system commands to that node are denied. Should a secondary node become unavailable or experience data delays beyond a configurable threshold, it's removed from the distribution target list until it recovers or the delay falls below the threshold.
A single MySQL cluster can create multiple proxies
To ensure smooth operation across different business units without resource monopolization by any single unit, multiple proxies can be created within the same MySQL cluster for isolation and management. This setup enhances system performance and scalability, ensuring fair resource utilization. The use of multiple proxies also improves system scalability, allowing the addition of new proxies to manage increased loads without expanding the entire cluster. This flexibility and scalability effectively meet varying business needs, maintaining system stability and reliability.
