cloud security challenges - operators of the era of leaking data privacy
Security has been the main problems restricting the development of cloud computing, many enterprises choose to non-critical application deployment to the public cloud, critical applications deployed in the local private cloud - for public cloud security concerns are a major technical reasons, encourage the growth of hybrid cloud enterprise for public cloud security is the most fundamental concerns about cloud computing operators to data confidentiality issues - enterprise their applications and data on the public cloud, public cloud operators leak data, enterprise is unable to prevent,
For operators confidentiality concerns challenges has become a time of cloud computing technology, cloud computing, data of owners and users are separated, more and more enterprises and government in the migration to the cloud environment, the system and data hosted on a third party cloud platform,
Is considered to solve this problem: homomorphic encryption is the best, the most promising technology,
homomorphic encryption, encryption process
Homomorphic encryption (Homomorphic encryption, HE) is based on the theory of mathematical problems in computational complexity of the cryptography technology, its purpose is to solve the outsourcing computing (outsourced computation), data privacy and security problems in the intuitive definition of Homomorphic encryption is a form of data encryption, it allows people to the encrypted data to a specific algebraic operation to get the output of the encryption, the encryption is the result of the output of the decryption is in line with the operation results of expressly specified and a Homomorphic encryption scheme consists of four processes:
(1) the homomorphic encryption system: system administrators to generate the parameters required by each process and the key, establish a homomorphic encryption system,
(2) clear encryption process: data owner using the public key cipher to encrypt plaintext data, in order to protect the data privacy,
(3) the ciphertext calculation process: data processing using the public key to homomorphism ciphertext is computed cipher text output,
(4) the cipher decryption process: data owner by using the private key to decrypt the cipher text output to obtain the calculation results,
homomorphic applications - cloud, chain blocks, AI...
Homomorphic Encryption is still a very new and frontier areas, the theory for the first time in 1978, is considered one of the holy grail in the field of cryptography, until now still out of reach, like a legendary computer scientist important development is the nearest Craig Gentry in 2009 first put forward in the thesis the first Fully Homomorphic Encryption in the field of digital (Fully Homomorphic Encryption, FHE),
Homomorphic encryption solves the data calculated with the data belongs to the main body of privacy problem, therefore, the technology can be used in all computing services and inconsistent data belongs to the subject in the scene, such as: cloud computing outsourcing services, training and reasoning of artificial intelligence, the chain of blocks of decentralization, iot data acquisition and processing, political votes and elections, etc., the following focuses on homomorphic encryption application in depth study and chain block and its effect,
Deep learning as an important direction of research and application in the field of artificial intelligence, in recent years in the algorithm, data and calculate three engine driven to get rapid development, as one of the important driving force of the AI data, privacy and security has become a key challenge, including training data security, data privacy and reasoning model application is confidential, many companies are selling tag data, but these data are AI algorithms company buy to do what use almost can't control, later this year in June, Microsoft removed the world's largest face recognition database MS Celeb, a direct reason is unable to control the use of the data,
AI is in its infancy now, the future as the AI industrial division continuously detailed, protect data security and privacy in the process of AI applications will also be a common industry problems, apparently homomorphic encryption technology is very suitable for solving this problem, homomorphic encryption technology can be executed in the ciphertext state deep learning of the neural network model training, reasoning and confidential data application, the AI will be a hot field in homomorphic encryption,
Block chain is as an important breakthrough of the core technology independent innovation in our country, is the financial, supply chain, public security, the Internet of things, and other fields are playing a more and more important role in the financial sector block chain + pay (international clearing and settlement), block chain + insurance claims, block chain + exchange, block chain + digital assets such as applications that value in the process of the interaction between people's trust to the trust of the people and technology, however, block chain trust in the application process is not yet complete, especially the lack of protection of users' privacy, for example, block chain + payment application, verify the effectiveness of the transaction will cause the transaction data at the same time, the trading account (or anonymous account) content such as leakage, therefore, block in validation and application of the chain of the contradiction between data privacy protection, and the application of homomorphic encryption method can provide privacy protection block chain, in the guarantee validity verified at the same time realize the block chain data privacy protection,
wave for the research and application of homomorphic encryption
As the most potential future data privacy solution, in the process of application of homomorphic encryption and theoretical study, the wave in the efforts for the development and mature contribute own strength,
Application research, the wave of current open source implementation of the homomorphic encryption tracking and research, including HElib of IBM and Microsoft's SEAL homomorphism computing, at present, the application of homomorphic encryption two key bottleneck problems: one is the high complexity, computational efficiency is low, 2 it is to support the homomorphism of less computing operation, it is because of homomorphic encryption technology on the computing complexity and homomorphism security features, has attracted more and more power into the exploration of its application in the study, in addition, in the depth of the privacy of a wave of learning areas focus on tracking the Intel nGraph HE - Transformer: use the Microsoft SEAL as the back-end infrastructure technology, realize the front-end different deep learning framework for homomorphic encryption calculation of seamless invocation, to protect the data privacy in artificial intelligence applications, however, in the process of research, found that the depth of the privacy protection technology distance learning real commercial is still a long way to go, the most critical in reasoning is the stage of calculation time delay for a few minutes, therefore, the wave is dedicated to the study at present stage using heterogeneous speed up platform improve the efficiency of HE - the calculation of the Transformer, in order to achieve acceptable calculation and response latency,
Wave of theoretical research, focusing on development of the international forefront of cryptography and information security, tracking the top-level meetings (such as the Crypto, EuroCrypto, AsiaCrypto and CF, etc.) the most advanced research progress of homomorphic encryption, at the same time, the wave is also closely related the domestic each big university research team (such as password state key laboratory of science and technology, Chinese Academy of Sciences by letter, zhejiang wanli college, Beijing university of posts and telecommunications, university of electronic science and technology, etc.), strive for more cooperation opportunities for homomorphic encryption in domestic development, implementation and their contributions to the standardization, in addition, the wave is also in the practice of their own scientific research mission, and strive to on the basis of the existing methods, open up new research ideas, break through the old application bottlenecks, such as: in depth study of the application of image recognition and classification data characteristics, design and implementation can not only meet the requirements of the safety and accuracy, and can guarantee the performance of the new methods and new technologies,

Now, is in the midst of the IT technology and the application of phase change, cloud, big data, such as artificial intelligence technology to the development of human society from the informationization, and upgrade to intelligent data as a key production in the modern era of information and future intelligent information has become a personal, group and important economic core strategic assets and capital of the country, at present, the protection of data privacy and security have become the whole society need and important problem to be solved, thus the bottleneck of homomorphic encryption once breakthroughs, will have broad prospects!