thresh‐ old, DBA’s operation pressure will also increase. The time cost of data backup and data recovery will be more uncontrollable with increasing amount of data. Generally, it is a relatively reasonable Though sharding has solved problems such as performance, availability and single‐node backup and recovery, its distributed architecture has also introduced some new problems as acquiring profits. One problem Alibaba Group and Ant Finance. The goal of AT transaction is to provide incremental transaction ACID semantics under the micro‐service architec‐ ture, so that developers can use distributed transactions as

splitting can be used to cope with business access with high stress. Based on its understanding of SQL semantics and the topological awareness of the underlying database, ShardingSphere provides flexible and thresh‐ old, DBA’s operation pressure will also increase. The time cost of data backup and data recovery will be more uncontrollable with increasing amount of data. Generally, it is a relatively reasonable Challenges Although data sharding solves problems regarding performance, availability, and backup recovery of single points, the distributed architecture has introduced new problems while gaining benefits

splitting capabilities and can achieve read access load balancing based on the understanding of SQL semantics and the ability to perceive the un‐ derlying database topology. High Avail‐ ability High availability thresh‐ old, DBA’s operation pressure will also increase. The time cost of data backup and data recovery will be more uncontrollable with increasing amount of data. Generally, it is a relatively reasonable Challenges Although data sharding solves problems regarding performance, availability, and backup recovery of single points, the distributed architecture has introduced new problems while gaining benefits

splitting capabilities and can achieve read access load balancing based on the understanding of SQL semantics and the ability to perceive the underlying database topology. Da ta M ig ra ti on Data migration thresh‐ old, DBA’s operation pressure will also increase. The time cost of data backup and data recovery will be more uncontrollable with increasing amount of data. Generally, it is a relatively reasonable Challenges Although data sharding solves problems regarding performance, availability, and backup recovery of single points, the distributed architecture has introduced new problems while gaining benefits

splitting capabilities and can achieve read access load balancing based on the understanding of SQL semantics and the ability to perceive the underlying database topology. Da ta M ig ra ti on Data migration thresh‐ old, DBA’s operation pressure will also increase. The time cost of data backup and data recovery will be more uncontrollable with increasing amount of data. Generally, it is a relatively reasonable Challenges Although data sharding solves problems regarding performance, availability, and backup recovery of single points, the distributed architecture has introduced new problems while gaining benefits

+…+ ?? ?1 + ?2+… + ??+…+ ??−1 (? < n − 1 < n) SAVEPOINT用于持久化各个子任务的执行状态 补偿行为：Forward Recovery/Backward Recovery/混合 执行路径 ?1 + ?2+… + ??+…+ ??(? < n) ?1 + ?2+… + ??+ ?? + ??−1 …+ ?1(? < n) ?1 + ?2+…

thresh‐ old, DBA’s operation pressure will also increase. The time cost of data backup and data recovery will be more uncontrollable with increasing amount of data. Generally, it is a relatively reasonable Though sharding has solved problems such as performance, availability and single‐node backup and recovery, its distributed architecture has also introduced some new problems as acquiring profits. One problem sharding; • Support RC isolation level; • Rollback transaction according to undo log; • Support recovery committing transaction automatically after the service is down. Unsupported • Do not support other

thresh‐ old, DBA’s operation pressure will also increase. The time cost of data backup and data recovery will 4.2. Sharding 20 Apache ShardingSphere document, v5.0.0 be more uncontrollable with increasing Though sharding has solved problems such as performance, availability and single‐node backup and recovery, its distributed architecture has also introduced some new problems as acquiring profits. One problem sharding; • Support RC isolation level; • Rollback transaction according to undo log; • Support recovery committing transaction automatically after the service is down. Unsupported • Do not support other

thresh‐ old, DBA’s operation pressure will also increase. The time cost of data backup and data recovery will be more uncontrollable with increasing amount of data. Generally, it is a relatively reasonable Though sharding has solved problems such as performance, availability and single‐node backup and recovery, its distributed architecture has also introduced some new problems as acquiring profits. One problem sharding; • Support RC isolation level; • Rollback transaction according to undo log; • Support recovery committing transaction automatically after the service is down. Unsupported • Do not support other

ShardingSphere document, v5.1.2 ├ ├ ├──status ├ ├ ├ ├──UUID ├ ├ ├ ├──.... ├ ├ ├──xa_recovery_id ├ ├ ├ ├──recovery_id ├ ├ ├ ├ ├──UUID ├ ├ ├ ├──.... ├ ├ ├──worker_id ├ ├ ├ ├──UUID ├ ├ ├ ├──.... ├