Replica Selection And Consistency Maintenance Strategy In Data Grid Environment

Data grid technology can manage independent,heterogeneous,and dispersed data uniformly,and solve the problem of storage and management of a large number of data resources.In the data grid environment,when a user sends a data access request,it is a problem of copy selection that needs to select the best copy from the data copies according to certain conditions.At the same time,when the data source is updated,how to ensure that the data content of multiple copies of the data source is consistent is a copy consistency maintenance problem.Reasonable copy selection and consistency maintenance strategies can reduce data access latency and bandwidth consumption to a certain extent,and achieve load balancing among storage nodes effectively.In the data grid environment,the paper designs the replica selection strategy and replica consistency maintenance strategy,and implements them under the simulation platform Optorsim.The main work of the thesis is as follows:1.A copy selection strategy is designed based on the improved ant colony algorithm.In view of the high access cost of the current copy selection strategy,which is likely to cause network congestion,the paper analyzes the main influencing factors that affect the performance of copy selection,and combines the ant colony foraging process with the copy selection process in the grid environment.The pheromone is initialized,the optimal probability copy selection is optimized to random probability copy selection,and the copy pheromone is dynamically updated according to the copy access situation,and the best copy is finally selected.2.A copy consistency maintenance strategy is designed based on timestamp.In view of the long access delay and serious waste of network bandwidth in the current copy consistency maintenance strategy,time stamp technology is introduced to determine whether a copy is the latest based on whether the time stamp is the same as the data source,and at the same time,update conflict detection is set,and the copy The update process is divided into updates outside the grid area and updates within the grid area.The primary copy outside the grid area is updated in real time according to the update of the data source.In the grid area,the access to the copy is high and the load of the copy node is small.Update,the remaining copies are updated when the user visits.3.The two improved algorithms proposed in the paper are implemented in the OptorSim simulation platform.Based on the improved ant colony algorithm,the copy selection strategy is based on the number of copies and access methods.18 sites are simulated under three simulation scenarios.The experimental results show that The access time of algorithm copy is better than other algorithms,which reduces data access delay and bandwidth consumption to a certain extent,and realizes load balancing among storage nodes.The time-stamp-based copy consistency maintenance strategy simulates 40 sites.The experimental results show that compared with other algorithms,the number of copies of this algorithm achieves better results,and the total operation time is better than other algorithms,which shortens the response time to a certain extent.Reduce bandwidth consumption.