Research on Lossless Expansion Algorithm for Cloud Native Storage
Distributed storage is a common and important product in multi-cloud data centers.In the cloud native era,the application iteration cycle is shortened,and the storage capacity and adjustment speed are also higher requirements.Cloud-native storage occupies the mainstream market due to its high scalability,robustness,and high performance of its architecture.However,cluster expansion and shrinkage lead to long-term data migration in the storage pool,leading to high performance loss,unstable cluster performance and other problems affecting business experience.This paper proposes an optimization algorithm by adding a logical label mechanism to dynamically map storage paths to physical media.In this way,incremental data can be dumped onto the expanded storage medium,avoiding mass data migration,and reducing the performance loss and cluster instability probability of the cloud native storage in the expansion or reduction scenario.
万方数据
维普
