NoSQL databases: Consistency of data

NoSQL database: Consistent read consistency for data

• Strong consistency
  When visiting any node in the cluster at any time, the data results are consistent;

• User consistency
  For the same user, access to the data obtained during the cluster is consistent;
  Resolve user consistency: Use sticky sessions to bind sessions to specific nodes for processing;
  This will reduce the performance of the load balancer;

• Finally consistent
  The data synchronization between nodes in the cluster is inconsistent with the temporary data, but the data synchronization is finished, and finally has the consistency;

Update consistency

• Pessimistic Way

  使用写锁大幅减少系统响应能力可能导致死锁

• Optimistic Way

  先让冲突发生。再检測顺序自己主动合并的处理方式极具“领域特定”问题

Easing "consistency constraints"

• Cap theorem
  Consistency (consistency), availability (availability), and partition tolerance (Partition tolerance), 3 properties can only meet 2 at the same time;

Explanation of partition tolerance: clusters are still available if the cluster is divided into multiple communications failures

• CA system

    单server  集群中出现”分区“，就不可用

• Pa/pc

    集群出现”分区“时，须要在”一致性“ 和“可用性”间权衡  通常会牺牲部分一致性（eg：使用终于一致性），保证可用性

Relaxing "persistence" constraints

more stringent durability. means a lot of other performance losses;

• Sacrificing "persistence" for better performance
• Replication "Persistence" failure

    主节点故障。未同步到从节点的数据丢失  主节点恢复，故障期间更新的数据冲突  解决方式：针对单个请求指定其所需的持久性

Attached Mind Map

References

The essence of NoSQL

Posted by: Big CC | 02jul,2014
Blog: blog.me115.com [Subscribe]
Weibo: Sina Weibo

NoSQL databases: Consistency of data