Re ad /w ri te S pl it ti ng Read/write splitting can be used to cope with business access with high stress. Sharding‐ Sphere provides flexible read/write splitting capabilities and can achieve read other, and mul‐ tiple components can be used together by overlaying. It includes data sharding, read/write splitting, data encryption and shadow database and so on. The user‐defined feature can be fully customized faced the bottleneck with increasing TPS. For the application with massive concurrence read but less write in the same time, we can divide the database into a primary database and a replica database. The

READWRITE_SPLITTING dataSourceGroups: (+): # 读写分离逻辑数据源名称，默认使用 Groovy 的行表达式 SPI 实现来 解析 write_data_source_name: # 写库数据源名称，默认使用 Groovy 的行表达式 SPI 实现来解析 read_data_source_names: # 读库数据源名称，多个从数据源用逗号分隔，默认使用 使用读写分离数据源 配置示例 rules: - !READWRITE_SPLITTING dataSourceGroups: readwrite_ds: writeDataSourceName: write_ds readDataSourceNames: - read_ds_0 - read_ds_1 transactionalReadQueryStrategy: PRIMARY loadBalancerName: dataSourceConfig = new ReadwriteSplittingDataSourceRuleConfiguration( "demo_read_query_ds", "demo_write_ds", Arrays.asList("demo_read_ds_ 0", "demo_read_ds_1"), "demo_weight_lb"); Properties algorithmProps

Availability Type . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 39 Dynamic Read/Write Splitting . . . . . . . . . . . . . . . . . . . . . . . . . . . . 39 3.4.7 Limitations . . . . . single database, which can achieve data security across underlying data sources. Read/write Split‐ ting Read/write splitting can be used to cope with business access with high stress. Based on its understanding topological awareness of the underlying database, ShardingSphere provides flexible and secure read/write splitting capabilities and can achieve load balancing for read access. High Avail‐ ability High

Re ad /w ri te S pl it ti ng Read/write splitting can be used to cope with business access with high stress. Sharding‐ Sphere provides flexible read/write splitting capabilities and can achieve read other, and mul‐ tiple components can be used together by overlaying. It includes data sharding, read/write splitting, data encryption and shadow database and so on. The user‐defined feature can be fully customized faced the bottleneck with increasing TPS. For the application with massive concurrence read but less write in the same time, we can divide the database into a primary database and a replica database. The

Availability Type . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 38 Dynamic Read/Write Splitting . . . . . . . . . . . . . . . . . . . . . . . . . . . . 38 3.4.7 Limitations . . . . . underlying data sources. Read/write Split‐ ting Read/write splitting can be used to cope with business access with high stress. Sharding‐ Sphere provides flexible read/write splitting capabilities and other, and mul‐ tiple components can be used together by overlaying. It includes data sharding, read/write splitting, database high availability, data encryption and shadow database and so on. The user‐defined

READWRITE_SPLITTING dataSources: (+): # 读写分离逻辑数据源名称，默认使用 Groovy 的行表达式 SPI 实现来 解析 write_data_source_name: # 写库数据源名称，默认使用 Groovy 的行表达式 SPI 实现来解析 read_data_source_names: # 读库数据源名称，多个从数据源用逗号分隔，默认使用 使用读写分离数据源 配置示例 rules: - !READWRITE_SPLITTING dataSources: readwrite_ds: writeDataSourceName: write_ds readDataSourceNames: - read_ds_0 - read_ds_1 transactionalReadQueryStrategy: PRIMARY loadBalancerName: dataSourceConfig = new ReadwriteSplittingDataSourceRuleConfiguration( "demo_read_query_ds", "demo_write_ds", Arrays.asList("demo_read_ds_ 0", "demo_read_ds_1"), "demo_weight_lb"); Properties algorithmProps

参数解释 读写分离 rules: - !READWRITE_SPLITTING dataSources: (+): # 读写分离逻辑数据源名称 write_data_source_name: # 写库数据源名称 read_data_source_names: # 读库数据源名称，多个从数据源用逗号分隔 transactionalReadQueryStrategy 使用读写分离数据源 配置示例 rules: - !READWRITE_SPLITTING dataSources: readwrite_ds: writeDataSourceName: write_ds readDataSourceNames: - read_ds_0 - read_ds_1 transactionalReadQueryStrategy: PRIMARY loadBalancerName: dataSourceConfig = new ReadwriteSplittingDataSourceRuleConfiguration( "demo_read_query_ds", "demo_write_ds", Arrays.asList("demo_read_ds_ 0", "demo_read_ds_1"), "demo_weight_lb"); Properties algorithmProps

READWRITE_SPLITTING dataSources: (+): # 读写分离逻辑数据源名称 static-strategy: # 读写分离类型 write-data-source-name: # 写库数据源名称 read-data-source-names: # 读库数据源名称，多个从数据源用逗号分隔 loadBalancerName: # 负载均衡算法名称 (+): # 读写分离逻辑数据源名称 dynamic-strategy: # 读写分离类型 auto-aware-data-source-name: # 数据库发现逻辑数据源名称 write-data-source-query-enabled: # 从库全部下线，主库是否承担读流量 loadBalancerName: # 负载均衡算法名称 # 负载均衡算法配置 loadBalancers: rules: - !READWRITE_SPLITTING dataSources: readwrite_ds: staticStrategy: writeDataSourceName: write_ds readDataSourceNames: - read_ds_0 - read_ds_1 loadBalancerName: random loadBalancers: random:

the item to be sorted which has its own order, merger ordering only has the time complexity of O(mn(log m)), and the number of shard m is generally small enough to be considered as O(n), with a very low transactions after sharding; • Support RC isolation level; • Rollback transaction according to undo log; • Support recovery committing transaction automatically after the service is down. Unsupported faced the bottleneck with increasing TPS. For the application with massive concurrence read but less write in the same time, we can divide the database into a primary database and a replica database. The

databases in the same log . . . . . . . . . . . . . . . . . . . . . . 215 To distinguish databases and users in the same log . . . . . . . . . . . . . . . . . 216 To split into different log files . . . . the item to be sorted which has its own order, merger ordering only has the time complexity of O(mn(log m)), and the number of shard m is generally small enough to be considered as O(n), with a very low transactions after sharding; • Support RC isolation level; • Rollback transaction according to undo log; • Support recovery committing transaction automatically after the service is down. Unsupported