| With the ongoing growth of State Grid Corporation of China’s business operations,and the emergence of new power grid equipment and products,the amount of data generated by the corporation has significantly increased.Additionally,the continued development and adoption of emerging technologies like the Internet and the Internet of Things has also contributed to the rise in data volume.This requires that the State Grid Corporation of China must use these data effectively,and through big data analysis technology,conduct in-depth mining of the data covering the whole range of power production,supply and trading,so as to obtain the value behind the data to support accurate decision-making in various aspects such as dispatching,energy planning,market trading,etc.,and also monitor the operation status of power grid equipment in time to improve the efficiency of power grid operation.However,in the face of so much data,how to store and read data efficiently is also a problem that the State Grid Corporation of China has to face.In response to this problem,this paper aims to store the frequently accessed hot data in the cache through cold and hot data identification,so as to speed up the data access speed,and improve the response speed of the system through efficient cache replacement strategy.Firstly,the traditional cold and hot data recognition algorithm and cache replacement algorithm are introduced in detail,and their limitations are evaluated critically in this paper.Then the principle of Newton’s cooling law is introduced,and the feasibility of applying it to data temperature calculation is analyzed.Based on this,a temperature model employing Newton’s cooling law is constructed to quantify the heat of data in the computer.Secondly,by closely examining the correspondence between the sequence created by the user’s data access requests and the discretized time series,the study employs the Markov chain model to forecast forthcoming data blocks with potential access.Additionally,the research adopts a weighted Markov chain model to enhance the precision of the prediction model.The experimental outcomes demonstrate that the proposed weighted Markov chain model outperforms existing models in terms of predicting accuracy.Finally,the temperature model and the prediction model are applied to the cache replacement algorithm,and the cache replacement algorithm T-PGDSF based on the data cold and hot degree is proposed.Furthermore,comprehensive testing experiments are conducted to evaluate the efficacy of the algorithm using diversified data sets including simulated and actual collected data sets,and a comparative analysis is performed between T-PGDSF and traditional cache replacement algorithms with respect to multiple performance indicators.After analyzing the experimental results,the cache replacement algorithm based on data cooling and heating presented in this paper has good performance under various workloads,but it also has the disadvantage of large time consumption.This study holds great guiding significance for future research endeavors conducted by the State Grid Corporation of China concerning the storage and reading of hot and cold data. |
