Abstract: Homomorphic encryption, as a key technology to realize data privacy computing, has wide application requirements and some feasible application schemes in cloud computing, block chain, privacy computing and other fields.

This article is shared from huawei cloud community “homomorphic encryption in the application of federated computing”, author: life is long, knowledge is long.

1, what is homomorphic encryption

Homomorphic encryption is a form of encryption that allows one to perform a specific form of algebraic operation on the ciphertext and still get the same result as if the same operation were performed on the plain text. In other words, the technology allows people to process encrypted data and get the right results without decrypting the data throughout the process. The implementation effect of homomorphic encryption is shown in figure 1.

CraigGentry, who was the first to construct Fully Homomorphic Encryption [Gen09], gave the following definition:

A way to delegate processing of your data, without giving awayaccess to it.

It can be seen from the above figure that homomorphic encryption differs from general encryption schemes in that it focuses on data processing security. Homomorphic encryption provides a function for processing encrypted data.

A little abstract? Let’s take a real life example. A user named Alice bought a chunk of gold that she wanted workers to shape into a necklace. But workers might steal gold in the process. After all, even a gram of gold is worth a lot of money. So could there be away that workers could delegate processing of your data without giving away access to it? Alice: Yes, there is.

  • Alice locks the gold in an airtight box, which is fitted with a glove.

  • A worker can wear this glove and process the gold inside the box. But the box was locked, so not only could the worker not get the gold, but any gold that fell during the process.

  • After processing. Alice takes back the box, opens the lock, and gets the gold.

The box looks something like this:

The corresponding relation here is:

  • Box: Encryption algorithm

  • Lock on the box: user key

  • Put the gold in the box and lock it: encrypt the data using a homomorphic encryption scheme

  • Processing: the application of homomorphism, in the condition of the data is not available to directly process the encryption results

  • Unlock: decrypt the result and get the result directly after processing

Source: www.zhihu.com/question/27…

2. Application of homomorphic encryption in federated computing

With the development of blockchain, privacy computing and other emerging fields and their higher requirements for privacy protection, homomorphic encryption has been applied to more diverse fields. Huawei Trusted Intelligent Computing TICS introduces homomorphic encryption in federated data analysis to ensure computing security.

Before the introduction of homomorphic encryption, TICS combines traditional symmetric and asymmetric encryption algorithms to ensure the security of data transmission. When the ciphertext data is sent to the aggregator node for multi-party computation, it needs to be decrypted first. After the computation is complete, the ciphertext data is encrypted and transmitted. This ensures the security of data transmission, but cannot ensure the security of computation.

The calculation process before the introduction of homomorphic encryption:

Calculation process after homomorphic encryption:

Use homomorphic encryption in TICS

On the Alliance management page, the administrator enables High-level Privacy Protection. After this function is enabled, if the SQL statement has operations supported by the Paillier homomorphic encryption algorithm, the operations will be calculated by the homomorphic ciphertext, and other calculation processes remain unchanged.

Create a job and execute the corresponding SQL job.

Execute the job, you can see the TICS system DAG diagram, showing the entire process of homomorphic encryption. The output is consistent with the non-homomorphic scenario.

Click to follow, the first time to learn about Huawei cloud fresh technology ~