Research And Implementation Of Memory Management Subsystem In Distributed Cache

With the rapid development of computer science and technology,the distributed cache is widely used in order to improve the performance of the system and to meet the increasing demands of the users.The distributed cache can speed up the processing speed of data by putting the data in a high speed memory.In the design of the distributed cache architecture,the memory management subsystem is the foundation and the core component,the design and implementation of memory management subsystem have a great impact on the performance of distributed cache.This paper firstly introduces the background and significance of the subject,and points out the important meaning of the research on the memory management subsystem in the specific project.Then this paper researches the current status at home and abroad,analyzes memory management mechanism of the mainstream distributed cache system in detail,then introduces the related technologies involved in the memory management subsystem.After in-depth understanding of memory management mechanism of the mainstream distributed cache system and related technologies,Combined with the specific project requirements,this paper describes the development requirements of the memory management subsystem from two aspects of function and performance.Then,the overall structure of the memory management subsystem is studied and designed,and the internal function modules and the mutual relations are determined.The specific content includes data storage structure and two level hash table structure designed and implemented based on data characteristics of project,reference counting function used to saving memory space,efficient memory management strategy based on the idea of pre allocating and memory pool in order to ensure the fast allocation of memory blocks and the stability of the system.A variety of data operation interface and its internal logic are also designed and implemented.At last this paper tests the system from two aspects of the function and performance.In the test of function,the function of each module is verified mainly through the data query,insert,and delete operations.In the test of performance,the performance of the system is mainly investigated under different data length and different reading and writing ratio and compared with the mainstream distributed cache such as redis and memcached.The results show that system is stable and has a high performance for reading and writing.