Research load balancing technology of distributed database based on consistent hash

Consistent hash algorithm is applied to build a distributed database load balancing model, but this model should deal with thousands of user requests, handle 10 billion data information, along with low-latency response scenarios, all of this is a grim challenge. When load balancing control mechanism is built by the consistency hash, some nodes in the database cluster are overloaded, but some are idle. These load imbalances can seriously do great damage to the overall performance of distributed database system. This paper proposes a detailed description of variance mathematics model about dynamic load balancing, the core is to track system load, evaluation, classification and storage of each node in distributed cluster. This algorithm controls mutual feedback between node load states, idle data node allots the item of overload node, overall suppression of single point overload. After by experimental simulation, compared with the auxiliary loop hash model, this algorithm improves load balancing efficiency by 30% and settles disputes about distributed database load imbalance based on consistent hashing.