Batch processing request-based Microsoft SQL server database cluster

Abstract

The invention relates to a batch processing request-based Microsoft SQL server database cluster. The cluster consists of two parts, wherein a gateway and backups thereof are arranged at the front end of the cluster, and two or more of database server nodes are arranged at the rear end of the cluster; each node is provided with a structured query language (SQL) Server example and a cluster proxy and is not provided with a shared disk structure; and each node has an independent and consistent data set, and the cluster is only provided with a host. The gateway is used for managing the integral cluster; and for each client link, the gateway opens a database link corresponding to each node simultaneously to receive, analyze and optimize all client requests, wherein the client requests are divided into two classes, namely the read-only query requests and the data updating requests; for the read-only query requests, the gateway sends the requests to any one node to execute the requests according to a set load balancing algorithm so as to realize the real database access load balancing and improve the performance of a system; and the updating requests are sent to the host of the cluster to be executed, and are subjected to synchronous data by the host to other nodes to keep the data set of each node real-time and consistent, eliminate a single-point fault source and realize system fault tolerance. The integral cluster can improve the performance, availability and reliability of the system obviously and solve various problematic issues of the performance, single-point faults, zero data loss and the like in the field of databases, and is a comprehensive database cluster scheme.

Description

technical field [0001] The invention relates to the field of database clusters, in particular to a method for a Microsoft SQL Server database cluster based on client batch processing requests and having three functions of fault tolerance, load balancing and automatic optimization. Background technique [0002] The database system is the core of the information system. In the field of database clusters, there are currently several solutions on the market: asynchronous replication, failover clusters and Oracle's application-level clusters (RAC), but these solutions cannot improve data reliability and system performance. Simultaneous progress has been made in the three aspects of system availability and system availability. [0003] Asynchronous data replication includes transactional replication, mirroring, snapshots, and third-party data replication technologies. These technologies shorten the time interval between data updates, but they have certain defects in reliability, ...

AI Technical Summary

Problems solved by technology

[0002] The database system is the core of the information system. In the field of database clusters, there are several solutions on the market: asynchronous replication, failover clusters and Oracle application-level clusters (RAC), but these solutions cannot achieve simultaneous progress in three aspects of data reliability, system performance and system availability.

[0003] Asynchronous data replication includes transactional replication, mirroring, snapshots, and third-party data replication technologies. These technologies shorten the time interval between data updates, but there are certain limitations in reliability Defect, does not have the load balancing feature, does not help the overall performance of the system, but will bring a certain performance overhead to the system

[0004] The Microsoft SQL Server database comes with a failover cluster, which is based on the Windows system failover cluster, and its function is the same as that of many third-party databases on the market. Similar to the fault tolerance scheme, it only provides system availability, which does not help to improve system performance and reliability, and also causes waste of hardware resources

[0005] Although Oracle's application-level cluster (RAC) improves system performance, it also has problems such as complex configuration and maintenance, single point of failure, and I / O bottlenecks

Images