A Balance Regulation Algorithm for Reliable Data in Hierarchical Private Cloud Architecture

Abstract

At present, enterprises need data for data analysis. They mostly use a kind of point-to-point data transmission form, which does not have regulation mechanism in the process. However, it has the problem of low data reliability, including two parts as follows: 1) Faulty sensor affects the collected data amount of the terminal server; 2) offensive data invade the data in transmission. In view of this, we propose the hierarchical private cloud architecture, including three aspects as follows. Firstly, we use distributed computing and virtualization capabilities of cloud computing to realize the hierarchical transmission of data. Secondly, through this mechanism of hierarchical transmission and classification algorithm of machine learning, we realize hierarchical filtering of offensive data. Finally, by combining hierarchical transmission mechanism with threshold value, classification algorithm, and limit tolerance mechanism, we regulate the data amount to monitor fault sensor in real time. Experiments are conducted to assess the proposed architecture’s performance. The results show that each layer acts as a protective screen to counterattack the offensive data, which shows good robustness, real-time, and adaptive ability. Moreover, to compare with OM mode, the identification efficiency of fault sensor of TM mode is improved by 2 times. Also, TM mode improves 33.33% identification acuity, which is suitable for the enterprises that are mainly based on streaming computing. In summary, the hierarchical private cloud architecture achieves the filtering of offensive data and the real-time identification of faulty sensor, which guarantees the security, accuracy, and integrity of the data transmission process.

Keywords

• Balance regulation
• cloud computing
• machine learning
• threshold value