Research on information classification and storage in cloud computing data center based on group collaboration intelligent clustering

Abstract

In order to overcome the problems of long time and low accuracy of traditional methods, a cloud computing data center information classification and storage method based on group collaborative intelligent clustering was proposed. The cloud computing data center information is collected in real time through the information acquisition terminal, and the collected information is transmitted. The optimization function of information classification storage location was constructed by using the group collaborative intelligent clustering algorithm, and the optimal solutions of all storage locations were evolved to obtain the elite set. According to the information attribute characteristics, different information was allocated to different elite sets to realize the classified storage of information in the cloud computing data center. The experimental results show that the longest time of information classification storage is only 0.6 s, the highest information loss rate is 10.0%, and the highest accuracy rate is more than 80%.

Keywords

Group collaborative intelligent clustering cloud computing information classification storage information acquisition terminal optical network

1. Introduction

In the context of cloud computing data gradually becoming diversified and complex, information presents multi-attribute characteristics, and it is no longer possible to make an effective judgment on the data type only by analyzing a single data. Therefore, the research on big data analysis technology oriented to diversified cloud computing data arises at the historic moment [17, 20]. However, with the gradual expansion of the amount of information in the network, a large amount of information needs to be transmitted across layers, and in this process, it will inevitably be disturbed by network attacks, leading to certain risks in information processing [5]. Due to the continuous increase of information in the cloud computing environment, all kinds of information show the characteristics of highly dynamic changes. How to ensure the real-time transmission, classification and storage of information is of great significance [3, 7, 18].

Reference [12] proposed a storage method of network library information based on knowledge unit mining. Firstly, a spatial model was established to classify different types of network library information for different types. Secondly, multi-path information allocation is realized according to the classification results, so that different types of information are allocated to different channels. The load characteristics of the channel are decomposed to realize the load balance of information transmission. Finally, the information storage model is established to effectively store different types of information. The experimental results show that this method has the advantage of less channel anomalies in the information classification and storage process, but it has the problem of high information loss rate, which affects the mining of library resources. Reference [13] proposed a robot trajectory data classification storage based on pattern recognition method, gives the overall design framework, according to the framework system hardware mainly includes data acquisition module, data processing module and data storage module, in the heart of the data acquisition module circuit design XC2VP30 chip as the core chip, and the data is stored by FLASH chip. On the basis of hardware design, pattern recognition method is used to construct data sample set, the data in the set is classified by membership degree calculation, and the data classification results are transmitted to the data storage module through FPGA, so as to complete the classified storage and processing of data. The experimental results show that this method has high throughput and can adapt to the data task of massive data, but it has the problem of low accuracy of data classification. Reference [14] proposed a classification and storage method for offshore multi-source transmission data under cloud computing. Data denoising processing was realized by wavelet transform method. Data similarity and support were calculated based on the denoising results and association rules. Determine the location of data storage, store the classified data, and complete the overall design of the method. The experimental results show that this method has the advantage of less conformity consumption, but it has the problem of long time of data classification and storage.

Because the above methods ignore the clustering processing of cloud computing data center information, resulting in low accuracy of information classification and storage. Therefore, in order to solve the problems existing in these methods, this paper introduces group collaborative intelligent clustering and designs a new cloud computing data center information classification and storage method. The overall design scheme of this method is as follows:

(1) The cloud computing data center information is collected in real time through the information acquisition terminal, and the collected information is transmitted through the access layer, backbone network and aggregation layer using optical network as the carrier.

(2) The optimization function of information classification storage location is constructed by using the group collaborative intelligent clustering algorithm, and the optimal solutions of all storage locations are evolved to obtain the elite set. Different information is allocated to different elite sets according to their attributes and characteristics to realize the classified storage and processing of information in the cloud computing data center.

(3) Time consuming, information loss rate and accuracy of information classification and storage in cloud computing data centers of different methods are compared through simulation experiments.

2. Design of cloud computing data center information classification storage method

2.1. Cloud computing data center information collection

In order to greatly improve the efficiency of information classification and storage in cloud computing data centers, multiple types of information need to be collected to provide an information basis for subsequent information processing. This paper mainly collects information in real time by designing an information acquisition terminal. The chip of the information acquisition terminal chooses STM32F103 as the main chip, and transmits data to the acquisition center with the help of GPRS software. This step adopts the form of timed acquisition, with the specific time being transmitted once every 15 s [6, 10]. After every 20 continuous times, the information is transmitted to the information monitoring center once, that is, the database will undergo an update every 300 s, so as to ensure the real-time information. The circuit module of information acquisition terminal is shown in Fig. 1.

Fig. 1.

Information acquisition terminal circuit module diagram.

According to Fig. 1, when information acquisition terminal is not received within 15 s information, indicates the terminal offline, at this point, you should consider whether information transmission channel noise interference, in order to realize the information acquisition terminal to reconnect, should restart the terminal, to avoid information transmission link blocking problem [4]. When the terminal is restarted, the information transmission instruction will be received again. At this time, the normal operation can be continued. If the normal information transmission state is still not restored after reconnection, the terminal needs to be shut down to detect whether the chip has problems. Therefore, the information acquisition system has many characteristics of high efficiency and high precision, which can solve the problem of low efficiency of current information acquisition methods, and can be further promoted in practice.

2.2. Cloud computing data center information transmission

At present, although the transmission network has generally achieved 4G network coverage, the gradual increase of data has brought great pressure to the network information transmission. However, the current information transmission hardware equipment is limited by its service life, and network congestion and transmission failure will occur in the actual transmission process [9]. To this end, it is necessary to ensure the continuity of information transmission.

After collecting cloud computing data center information, the information is transmitted and processed. This paper mainly uses optical network as the carrier for information transmission. Optical transmission network is a network composed of multiple nodes and links, and the transmission nodes in the network have the functions of sending and receiving information [15]. The transmission network is mainly composed of three parts, namely, the access layer, the backbone network and the aggregation layer. As the basic layer of the optical network, the backbone network is mainly responsible for the task of data transmission. In general, it realizes transmission function through intensive optical wave multiplexing equipment. The aggregation layer realizes information aggregation through GEPTN devices, which can also be called the access layer. The information transmission process is shown in Fig. 2.

Fig. 2.

Flow chart of information transmission.

In order to solve the problem of large information transmission capacity, the concept of expansion capacity is put forward. Capacity expansion of an optical transmission network is the process of adding new stations to the network and cutting off the optical fiber of adjacent stations to add new stations. In the actual operation, the optical network expansion process is a complex software and hardware process [2, 16]. In addition to the hardware need to break the fiber, fiber, connection and other operations, but also in the software configuration of network management data.

With the development of cloud computing, there will be more and more cloud computing data centers, leading to the rapid growth of information. In order to meet the needs of network operation and improve the capacity of optical transmission network, in addition to the realization of information transmission, it is also necessary to carry out effective scheduling and management of a large amount of information.

2.3. Information classification and storage based on group collaborative intelligent clustering

When the traditional method is used for information classification and storage, the problem of local optimal solution often appears. If the local optimal solution is obtained, the search for the next optimal solution will be stopped immediately, thus no in-depth calculation will be carried out, resulting in poor convergence of the algorithm [19]. Aiming at this problem, based on the results of information collection and information transmission, this paper adopts the group collaborative intelligent clustering algorithm to carry out clustering calculation and obtain the optimal solution. This algorithm can ensure the information interaction between subpopulations and can carry out information retrieval in each subpopulation to avoid the occurrence of local optimal solution. In addition, this algorithm has the characteristics of simple operation process, high calculation efficiency and high precision, and can realize the effective classified storage of information [8, 11].

Group synergy intelligent clustering method is a kind of cooperative coevolution to groups of PSO algorithm, this paper mainly uses group synergy intelligent clustering algorithm to construct information classified storage location optimization function, and evolution of all storage location optimal solution for elite collection, according to the attributes of information distribution of different information to different elite in the collection, realize cloud computing data center information classification storage and processing.

The optimization function of information classification storage location is built by using the group collaborative intelligent clustering algorithm. The specific description of the function is as follows: $\begin{matrix} (1) & P (i) = \sum_{i = 1}^{m} P_{i} + rand (0, 1) \times (X_{max} - X_{min}) - X_{ij} \end{matrix}$

In the above formula, $X_{ij}$ represents the distance between the information classification storage node i and j, $X_{max}^{j}$ and $X_{min}^{j}$ respectively represent the maximum and minimum capacity of the information storage location where the node is located, $P_{i}$ represents the solution of the objective function, and $rand (0, 1)$ represents a random number.

The population with a particle number of N was divided into M subpopulations, and the general PSO algorithm was used for local search of each subpopulation. During the retrieval process, the efficiency and position inside the subpopulation were continuously adjusted to obtain the optimal solution of the information classification storage location optimization function. When under the condition of evolution to the R-th generation, the first child population will get a local optimal solution, expressed in $P_{g 1}$ , on the basis of this will be passed on to the next is the solution population, and by using $P_{g 1}$ pairs of the child within populations adapt to function with the lowest particle is updated, the child population will exist at this time period for the evolution of the R, get the local optimal solution of the child population of $P_{g 2}$ , and the solution is passed to the next subpopulation, and so on, finally the last subpopulation will pass the local optimal solution $P_{gi}$ to the first subpopulation. At this time, it is necessary to judge whether the solution meets the accuracy requirements. If it does, evolution will stop; if it does not, evolution will continue. In this process, the interval between each subpopulation is R generations, and information transmission and exchange can be realized between adjacent populations, and this cycle can be carried out until the termination conditions are met and the optimal solution of information classification storage location in the cloud computing data center is obtained.

The specific realization process of group collaborative intelligent clustering is shown in Fig. 3.

Fig. 3.

Clustering method of group collaborative intelligence.

When the population evolution to R, under the condition of fixed cycle before a child population will a child back feedback local optimal solution, after a child population evolution with Shared information, to ensure that each population of particles in the position of the optimal solution, thus increasing the convergence efficiency of the algorithm, so the iteration, get information classification storage location of the final solution.

The specific implementation process of cloud computing data center information classification and storage based on group collaborative intelligent clustering is as follows:

Analyze the attribute characteristics of cloud computing data center information samples [1], and the specific calculation formula is as follows: $\begin{matrix} (2) & W = \frac{λ}{2 a} \sum_{i = 1}^{M} ‖ y^{i} ‖ \end{matrix}$

Where, λ represents the attribute feature vector of the information sample in the cloud computing data center, and a represents the similarity coefficient of the information sample, M represents the number of subpopulations; $y^{i}$ represents the i-th cloud computing data center.

(2) Calculate the individual adaptation value of each subpopulation. The specific formula is as follows: $\begin{matrix} (3) & K_{i} = \sum_{p = 1}^{n} \sum_{m = 1}^{n} (t_{p} - o_{m}) \end{matrix}$

In the formula, $t_{p}$ represents the master subpopulation and $o_{m}$ represents the auxiliary subpopulation. The optimal solution of each subpopulation is updated according to the calculation results.

(3) Evolve the optimal solution of all information classification storage locations, and then get the elite set. According to the information attribute characteristics, different cloud computing data center information is allocated to different elite sets, so as to realize classified storage processing and get the final result of information classification storage.

3. Experimental verification

3.1. Experimental scheme design

In order to verify the effectiveness of the information classification storage method for cloud computing data center based on group collaborative intelligent clustering, a simulation experiment was designed. The specific experimental scheme was designed as follows:

(1) Experimental environment: Simulation experiment was conducted in Intel Xeon Gold6254@3.10GHz(X2)CPU, 768Gbram, 2× Teslav100GPU, operating system was WindowsServer2019, and programming language was Matlab.

(2) Experimental data: In the process of the experiment, in order to facilitate data statistics, the experimental data were first determined. The experimental data came from the ClickHouse cloud database, which is an open source column database, and the database kernel is fully compatible with the open source community version. The data in the database has the characteristics of security, cluster dynamic scalability, etc., and is a reliable platform for cloud data analysis. This data set contains a total of 3 attributes, 5 types and classes, a total of 8700 data points, which contains certain noise, accounting for up to 3.2%. 5 different types of data are randomly selected from this database for processing. On this basis, the length of the simulation experimental data is set to ensure that all the simulation experimental data can be input to the simulation platform, and the optimal simulation parameters are taken as the initial simulation parameters.

The above data are carried in a simulation platform, and the framework of the simulation platform is shown in Fig. 4.

Fig. 4.

Simulation platform framework.

During the experiment, because of interference will be affected by various factors, can make the results produce certain error, in order to avoid the error data acquisition and data calculation error to the negative effects of the experiment result, will be repeated calculation in the process of experiment, and according to the calculation results are normalized for many times, in order to reduce the error for the effects of the experimental results.

(3) Experimental method: The method of this paper was compared with the reference [12] method and reference [13] method to carry out comparative experimental verification.

(4) Experimental indicators: Time consuming of information classification and storage, information loss rate and accuracy of information classification and storage in cloud computing data center are taken as experimental indicators. Among them, the shorter the time of classified storage of information in cloud computing data center, the higher the efficiency of classified storage; The lower the information loss rate, the better the information classification and storage effect; The higher the accuracy of cloud computing data center information classification and storage, the higher the storage accuracy.

3.2. Analysis of experimental results

3.2.1. Comparison of information classification and storage time of different methods

In order to verify the advantages of the method presented in this paper, first of all, the time consuming of information classification storage was taken as an experimental index to compare the classification storage effects of different methods. The time consuming of information classification storage can reflect the actual working efficiency of different methods. The shorter the time, the higher the efficiency of the method, that is, the better the application effect. Comparison results of information classification storage time of different methods are shown in Fig. 5.

Fig. 5.

Comparison of information classification storage time of different methods.

It can be seen from the analysis of Fig. 5 that, in the case of the same number of iterations, the information classification storage time of the reference [12] method and reference [13] method has a great change and a significant increase in the time consumption, while the information classification storage time of the method in this paper has a gentle change. When the number of iterations is 6, the time consumption of reference [12] method is 2.2 s, the time consumption of reference [13] method is 1.7 s and the time consumption of the method in this paper is 0.4 s. The longest time of reference [12] method was 3.7 s, the longest time of reference [13] method was 4.1 s, and the time of the method in this paper was 0.6 s. Through comparison, it can be seen that the method presented in this paper is significantly better than traditional methods in terms of information classification and storage time, indicating that this method has higher work efficiency and can realize cloud computing data center information classification and storage in a shorter time. The reason is that the method of cloud computing data center information through data collection terminal for real-time acquisition, with optical network as the carrier, through the access layer, backbone and convergence layer on gathering the information transmission, to realize the cloud computing data center information classification storage, so this method has lower information classification storage time consuming.

3.2.2. Comparison of information loss rates between different methods

Secondly, the information loss rate in the process of information classification storage is taken as the experimental index to compare the classification storage effect of different methods. The information loss rate can reflect the effectiveness of different methods. The lower the information loss rate, the more comprehensive the information obtained by the method, and the stronger the effectiveness of the results. Comparison results of information loss rates of different methods are shown in Fig. 6.

Fig. 6.

Comparison of information loss rates of different methods.

It can be seen from the analysis of Fig. 6 that in the process of information classification and storage, the information loss rate is basically not affected by the amount of data, and the information loss rate of different methods is basically maintained within a certain range without an obvious trend of change. Among them, the maximum information loss rate of reference [12] method is 13.4%, the maximum information loss rate of reference [13] method is 14.1%, and the maximum information loss rate of the method of this paper is 10.0%. Through comparison, it can be seen that the information loss rate of the method in this paper is lower, indicating that the information processing results obtained by this method are more comprehensive. The reason is that the method collects the information of the cloud computing data center in real time through the information acquisition terminal, and transmits the collected information. The optimization function of information classification storage location is built by using the group collaborative intelligent clustering algorithm to realize the classified storage of information in the cloud computing data center. Therefore, the information loss rate of this method in the process of information classification storage is low.

3.2.3. Comparison of information classification and storage accuracy of different methods

Finally, the accuracy of information classification and storage is taken as the experimental index to compare the effects of different methods. The accuracy of information classification and storage can reflect the effectiveness of different methods. The higher the accuracy of information classification and storage, the more reliable the information processed by the method is. The comparison results of the accuracy of information classification and storage by different methods are shown in Fig. 7.

Fig. 7.

Comparison of information classification and storage accuracy of different methods.

The analysis of Fig. 7 shows that the accuracy of information classification and storage of the method presented in this paper basically shows a trend of continuous growth with a large range of change, while the change trend of the accuracy of classification and storage of reference [12] method and reference [13] method has no obvious change rule, and there is a certain gap with the method of this paper. Through comparison, it can be seen that the highest accuracy of the method in this paper reaches over 80%, while the accuracy of the traditional method is always lower than 45%, which clearly shows the superiority of the method in this paper. Because the method uses group synergy intelligent clustering algorithm to construct information classified storage location optimization function, and evolution of all storage location optimal solution for elite collection, according to the information attributes will assign different information to different elite in the collection, information classification storage cloud computing data center, so the method has higher storage information classification accuracy.

4. Conclusion

In order to solve the problems of high information loss rate, low accuracy of data classification and long time of data classification and storage existing in traditional methods, this paper designs a cloud computing data center information classification and storage method with the help of group collaborative intelligent clustering algorithm. The information collection terminal is used to collect the information of the cloud computing data center, and the optical network is used to transmit the information. Then, the group collaborative intelligent clustering algorithm is used to carry out the clustering calculation on the information to obtain the optimal solution, so as to ensure the information interaction between sub-populations and avoid the occurrence of local optimal solution. The optimal solution of clustering is evolved by calculating the individual fitness value of the population, and the optimal individual is further obtained. Based on this, the structure of radial basis neural network is built, and the information classification storage and processing of cloud computing data center is realized by using this structure. The experimental results show that the longest time of information classification storage is only 0.6 s, the highest information loss rate is 10.0%, and the highest accuracy rate is more than 80%, indicating that the method can effectively improve the accuracy of information classification storage, the information loss rate is low, and the operation time is short.

Although the method in this paper has improved the traditional method and achieved significant research results, in view of the current development of cloud computing, it will continue to generate more types of information, and how to update different types of information is the key research content to be faced in the future.

Footnotes

Acknowledgements

This work was supported by Heilongjiang Provincial Philosophy and Social Science Research Plan Project under grand no 18TQE569; Research team and platform support plan of Heilongjiang Bayi Agricultural University under grand no ZDTDJH201801.

References

Arunkumar,

Karthikeyan and

Karthik, An iterative processing scheme for enhancing the map reduce using map information storage in wireless environment, Wireless Personal Communications111(3) (2020), 1575–1587. doi:10.1007/s11277-019-06941-7.

Gao,

Liu and

Guo, A data mining method using deep learning for anomaly detection in cloud computing environment, Mathematical Problems in Engineering2020(1) (2020), 1–11. doi:10.1155/2020/6677301.

Gomathi and

N.P.

Karlekar, Ontology and hybrid optimization based SVNN for privacy preserved medical data classification in cloud, International Journal on Artificial Intelligence Tools28(3) (2019), 1950009–1950018. doi:10.1142/S021821301950009X.

Gupta,

B.P.

Kaur and

Jangra, An efficient method for fault tolerance in cloud environment using encryption and classification, Soft Computing23(24) (2019), 13591–13602. doi:10.1007/s00500-019-03896-6.

Hou,

Liu and

Hu, Research on secure transmission and storage of energy IoT information based on blockchain, Peer-to-Peer Networking and Applications13(4) (2020), 1225–1235.

Jagdish,

Viloria and

Vargas, Modeling software architecture design on data storage security in cloud computing environments, Journal of Intelligent and Fuzzy Systems39(6) (2020), 1–8.

Jyothi,

Bhargavi and

Mani, Analysis of various security methods of virtual machines in cloud computing, Journal of Computational and Theoretical Nanoscience17(12) (2020), 5368–5377. doi:10.1166/jctn.2020.9430.

Kim,

H.B.

Jin and

Baym, Metastable hybridization-based DNA information storage to allow rapid and permanent erasure, Nature Communications11(1) (2020), 5008–5015. doi:10.1038/s41467-020-18842-6.

Kumar,

Anand and

N.K.

Trivedi, Load balancing in cloud computing – a classification, Solid State Technology63(6) (2020), 3467–3476.

10.

Li,

Han and

Tang, Logisticschain: A blockchain-based secure storage scheme for logistics data, Mobile Information Systems22(3) (2021), 1–15.

11.

Li and

Guo, Research on ship data big data parallel scheduling algorithm based on cloud computing, Journal of Coastal Research94(1) (2019), 535–542. doi:10.2112/SI94-106.1.

12.

D.L.

Liu,

J.G.

Wang and

Zhang, Research on network library information storage model based on knowledge unit mining, Journal of the China Society for Scientific and Technical Information39(02) (2020), 51–57.

13.

Liu,

M.Z.

Wang and

Liu, Design of robot moving trajectory data classification and storage system based on pattern recognition, Modern Electronics Technique43(04) (2020), 145–148.

14.

Q.J.

Luo,

Z.F.

Zhang and

W.D.

Xie, Research on classification and storage method of multi-source data transmission at sea under cloud computing, Ship Science and Technology42(04) (2020), 179–181.

15.

Mijuskovic,

Chiumento and

Bemthuis, Resource management techniques for cloud/fog and edge computing: An evaluation framework and classification, Sensors21(5) (2021), 1832–1840. doi:10.3390/s21051832.

16.

Ogiela, Transformative computing in advanced data analysis processes in the cloud, Information Processing & Management57(5) (2020), 102260–102271. doi:10.1016/j.ipm.2020.102260.

17.

Xu, Classification and storage method of marine multi-source transmission data under cloud computing, Journal of Coastal Research115(1) (2020), 84–96. doi:10.2112/JCR-SI115-025.1.

18.

Yao, Research on campus network cloud storage open platform based on cloud computing and big data technology, Journal of Intelligent and Fuzzy Systems38(1) (2019), 1–9.

19.

Zhao and

Zhao, Research on the management model of marine economic information resources in the big data era based on cloud computing, Journal of Coastal Research94(1) (2019), 316–325. doi:10.2112/SI94-065.1.

20.

Zheng,

Duan and

Wang, Research on distance transform and neural network Lidar information sampling classification-based semantic segmentation of 2D indoor room maps, Sensors21(4) (2021), 1365–1372. doi:10.3390/s21041365.