An optimization strategy of distributed metadata based on cache

With the development of science and technology, the scale of data is getting larger and larger, and the demand for efficient storage and access of data is getting higher and higher. Distributed storage system is one of the important technologies, which is widely used in scientific computing and data-intensive application. However, due to the large amount of data access, metadata access in distributed storage systems can easily become a performance bottleneck. Therefore, reasonable metadata management is of great significance for improving the I/O performance of the entire system. Metadata caching technology can effectively improve throughput and reduce latency. At the same time, how to balance metadata load among metadata servers is one of the key technologies of metadata management. Aiming at the problem of metadata optimization in distributed storage system, this research proposes an effective solution from the perspective of cache design and load balancing, and has achieved good results through experimental verification.