In terms of research on cloud security technology, it mainly includes: •Cloud computing security: It mainly analyzes the cloud itself and the application services involved, focusing on its corresponding security issues. This mainly involves how to effectively implement security isolation, ensure the security of Internet user data, and how to effectively protect it. Malicious network attacks, improve the system security of cloud computing platforms, as well as user access authentication and corresponding information transmission auditing, security and other aspects. •Ensuring the security of cloud infrastructure: It mainly involves how to utilize the corresponding resources of corresponding Internet security infrastructure equipment to effectively optimize cloud services, thereby ensuring that expected security protection requirements are met. •Cloud security technology services: Focus on how to ensure the security service requirements of Internet end users, and can effectively implement client computer virus prevention and other related services. Judging from the development of cloud security architecture, if the security level of cloud computing service providers is not high, service users will need to have stronger security capabilities and assume more management responsibilities.

Cloud security technology should be considered from the perspectives of openness, security, and architecture.

• Massive data: The data volume of big data is huge, jumping from TB level to PB level (1PB=1024TB), EB level (IE B=1024PB), and even reaching ZB level (IZ B=I024EB). • Diverse data types: Big data has many data types, generally divided into structured data and unstructured data. Compared with the text-based structured data that has been easily stored in the past, there are more and more unstructured data, including web logs, audio, video, pictures, geographical location information, etc. These multiple types of data have limited data processing capabilities. Higher demands were made. • Low data value density: The level of data value density is inversely proportional to the size of the total data address. Take video as an example. For an hour-long video, under continuous and uninterrupted monitoring, the useful data may only be one or two seconds. How to "purify" the value of data more quickly through powerful machine algorithms has become an urgent problem to be solved in the current context of big data. • Fast data processing speed: In order to quickly mine data value from huge amounts of data, it is generally required to process different types of data quickly. This is the most significant feature of big data that distinguishes traditional data mining.

Data collection technology mainly obtains data information from websites through distributed crawling, distributed high-speed and high-reliability data collection, and high-speed whole-network data mapping technology. In addition to what is contained within the network, collection of network streams can be processed using bandwidth management techniques such as DPI or DFI. Data integration technology is based on data collection and entity recognition to achieve high-quality integration of data into information. Data integration technology includes multi-source and multi-modal information integration models, heterogeneous data intelligent conversion models, intelligent pattern extraction and pattern matching algorithms for heterogeneous data integration, automatic fault-tolerant mapping and conversion models and algorithms, and correctness verification methods for integrated information. Usability assessment methods for integrated information, etc. Data cleaning technology generally removes unreasonable and erroneous data based on correctness conditions and data constraint rules, repairs important information, and ensures data integrity. Including data correctness semantic model, association model and data constraint rules, data error model and error recognition learning framework, automatic detection and repair algorithms for different error types, evaluation model and evaluation method of error detection and repair results, etc.

Distributed computing emerged with the development of distributed systems. Its core is to decompose tasks into many small parts and assign them to multiple computers for processing. Through the mechanism of parallel work, it can save overall computing time and improve computing efficiency. the goal of. Currently, the mainstream distributed computing systems include Hadoop, Spark and Storm. Hadoop is commonly used for offline complex big data processing. Spark is often used for offline fast big data processing, while Storm is often used for online real-time big data processing. Big data analysis and mining technology mainly refers to improving existing data mining and machine learning technology; developing new data mining technologies such as data network mining, specific group mining, and graph mining; innovating big data such as data connection and similarity connection of basic objects Fusion technology; breakthroughs in field-oriented big data mining technologies such as user interest analysis, network behavior analysis, and emotional semantic analysis.

Big data management technology mainly focuses on big data storage, big data collaboration, security and privacy. Big data storage technology mainly has three aspects. 1) A new database cluster using MPP architecture realizes big data storage through multiple big data processing technologies such as column storage and coarse-grained indexes, and an efficient distributed computing model; 2) Relevant big data technologies are derived around Hadoop to deal with data and scenarios that are difficult to process with traditional relational databases, and support big data storage and analysis by extending and encapsulating Hadoop; 3) Based on integrated servers, storage devices, operating systems, and database management systems, a big data all-in-one machine with good stability and scalability is realized. Collaborative management technology of multiple data centers is another important direction of big data research. Through the distributed workflow engine, workflow scheduling and load balancing are realized, and storage and computing resources of multiple data centers are integrated to provide support for building a big data service platform.

Big data application and service technology mainly includes analysis application technology and visualization technology. Big data analysis applications are mainly business-oriented analysis applications. Based on distributed Haichi data analysis and mining, big data analysis application technology is driven by business needs, carries out special data analysis for different types of business needs, and provides users with highly available and easy-to-use data analysis services. Visualization helps people explore and understand complex data through interactive visual representations. Big data visualization technology mainly focuses on text visualization technology, network (graph) visualization technology, spatiotemporal data visualization technology, multi-dimensional data visualization and interactive visualization, etc. In terms of technology, it mainly focuses on In Situ Interactive Analysis, data representation, uncertainty quantification and domain-oriented visualization tool libraries.

• Multi-centralization: The verification, accounting, storage, maintenance and transmission of data on the chain all rely on the distributed system structure. Pure mathematical methods are used to replace centralized organizations to build trust relationships between multiple distributed nodes, thereby establishing Trusted distributed systems. • Multi-party maintenance: The incentive mechanism ensures that all nodes in the distributed system can participate in the verification process of data blocks, and select specific nodes through the consensus mechanism to add newly generated blocks to the blockchain. • Time series data: Blockchain uses a chain structure with timestamp information to store data information, and adds time dimension attributes to the data information, thereby achieving traceability of data information. • Smart contracts: Blockchain technology can provide users with flexible script codes to support the creation of new smart contracts. • Cannot be tampered with: In the blockchain system, because the subsequent blocks between adjacent blocks can verify the preceding blocks, if the data information of a certain block is tampered with, the block and all its contents need to be modified recursively. However, the cost of each hash recalculation is huge and must be completed within a limited time, so the non-tamperability of the data on the chain can be guaranteed. • Open consensus: In the blockchain network, each physical device can serve as a node in the network, and any node can join freely and have a complete copy of the database. • Safe and trustworthy: Data security can be achieved by encrypting data on the chain based on asymmetric encryption technology. Each node in the distributed system uses the computing power generated by the blockchain consensus algorithm to resist external attacks and ensure data on the chain. It cannot be tampered with or forged, so it has high confidentiality, credibility and security.