Design And Implementation Of Efficient Cache System For Cloud Service Of Internet Of Things

Compared with the traditional Internet,the Internet of Things has more access devices and higher requirements for processing delays,which poses a challenge to the computing power of IoT cloud services.The high-cost way of increasing the processing power by adding physical machines is not the best choice.Using cache technology,in the form of software,it has more realistic value and significance to improve the performance of the system under the existing hardware conditions.In this context,based on the existing IoT cloud service system,designing and implementing a cache system with high performance is the focus of this thesis.The introduction part of this thesis mainly introduces the research background of the thesis and the research status of the in-memory database and cache technology.The second part summarizes the common problems of caching,and then summarizes the related technologies used in the implementation of the caching system,including the Redis and the aspect-oriented programming technology.The third part mainly describes the design of an efficient caching system for IoT cloud services.Firstly,the usage scenarios and data characteristics of the IoT cloud service are summarized,and the performance bottleneck of the existing system is deeply studied,the necessity of using the cache system to solve the current performance problem is pointed out.Then according to the layered architecture of the existing IoT cloud service system,three levels of cache granularity schemes are proposed.Through analysis and comparison,the cache granularity oriented to the database interface level is finally selected.On the basis of the cache granularity design,considering the data characteristics in the IoT scenario and the characteristics of various storage types of Redis,the cache data is divided into two types,namely,hot and cold,and the storage structure is designed separately.Finally,according to the design of the storage structure,different dirty cache clearing strategies are designed for the hot and cold cache data,and a method for clearing the serialized cache data is proposed.The fourth part implements the cache system based on the system design.In order to reduce the intrusion into existing systems,the caching system uses the idea of aspect-oriented programming(AOP)and proposes an implementation of the AOP framework in the JavaScript language.Based on the implemented AOP framework,the cache system is modularized;mainly includes a cache configuration module,a cache statistics module,a cache update module,and a Redis adaptation module.After completing implement caching system,a lot of stress tests are performed on the performance of the cache system.The test results show that the IoT cloud service system effectively reduces the interface processing delay after joining the cache system,and has obvious improvement in concurrent performance.