Data sharing empowers AI-driven networks based on blockchain technology, taking mobile network operator MNO as an example

Author: Sachin Devmurari

Translation: Huohuo Sauce

Source: Blockchain Base Camp

Editor's note: The original title was "Enable AI-Driven Networks for Data Sharing Based on Blockchain Technology."

Recent advances in artificial intelligence and machine learning algorithms have provided momentum for network automation. Recently, mobile network operators (MNOs) are using artificial intelligence-based modules to automate networks by distributing data authorized in their leased / owned areas.

The emergence of 5G networks has gradually disrupted the traditional network paradigm. Super heterogeneous networks are needed to coordinate and organize various types of network base stations, such as macro base stations, micro base stations, home base stations (Femto), pico base stations (pico), and management of large-scale multiple input Output (MIMO), millimeter wave, or device-to-device (D2D) communication.

However, the problem is that data access is limited for several MNOs. Blockchain-based data sharing can change this situation and enhance the performance of artificial intelligence-driven network systems.

What is an AI-driven network?


Artificial intelligence is not new to us, but earlier versions of artificial intelligence algorithms were limited to certain specific applications, and these applications were limited to the limited computing power of the system.

Then, as artificial intelligence became more applicable, network operators began to explore AI-driven network systems to better organize and distribute networks.

The basic idea is as follows: first, use the clustering method to obtain the optimal partition of the network, and then use the neural network calculation algorithm to obtain the optimal traffic routing. And, with the development of data-driven intelligence, algorithms can now learn by accessing large amounts of data.

Source: AI-powered networks

With the further development of artificial intelligence and computing power, MNOs can now use convolutional neural networks (CNN) and recurrent neural networks (RNN) to create tissue models from large amounts of raw data.

What is blockchain-based data sharing?


With the advent of smart contracts, blockchain-based technologies have become very attractive to many businesses. The basic problem of the early blockchain was verification. Many experts believe that the democratization of data in blockchain-based data sharing is threatening data security.

Smart contracts remove participants' doubts about verification issues and data ownership. The smart contract is first compiled into machine code and uploaded to the blockchain as a transaction. A certain miner packages the transaction, and then other miners verify it by voting on the first block, and then repeat it. After adding data through another client, Perform transaction verification in the second block.

In addition, a third client can read verified data through the blockchain block. As a result, smart contracts are more democratized and data can be verified through a verification system.

Businesses generally prefer licensed smart contracts over public smart contracts because the latter are less secure than licensed smart contracts.

Blockchain-based data sharing in AI-driven networks

  • Blockchain-based data sharing uses AI-driven network smart contracts. The system is divided into three layers.
  • User layer-participants in the system (MNO)
  • System management layer-contains the following components:
  1. MSP (Member Services)
  2. Consensus Nodes
  3. Verifier
  4. Gatekeeper
  5. DataChain
  6. BehaviorChain
  • Data storage layer-cloud-based data storage

Important part of the system layer

Source: System Design

1. MSP (Member Services)

Member services are responsible for issuing membership certificates, authorizing and registering system participants.

It holds a root certificate like the master key and issues a second key (Cu) certificate to registered members. Whenever a new member joins the system, a "Cu" key is provided as a new certificate. The private key is used for identity registration and verification of each member.

In our case, the members are different mobile network operators (MNOs). The identification of each MNO requires a specific certificate provided by the MSP layer.

2. Verifier:

The validator uses the "Cu" certificate issued by MSP for any user calling the API. The application program interface acts as an intermediary between the system and the user. Specific GUIs turn their application ideas into reality.

3. Consensus Nodes:  

The consensus node is responsible for implementing the AI ​​algorithm. Here we integrate the AI ​​algorithm in the blockchain-based data sharing system.

Consensus is used to ensure the consistency of the ledger. Consensus algorithms involve endorsing transactions, where transactions involve compiling raw data into bytecode for the blockchain.

In addition, the order in which transactions are uploaded to the blockchain needs to be determined. During the transaction endorsement process, if both blockchain nodes want to upload transactions to the blockchain, they need to use smart contracts to determine who confirms the transaction.

The method used by Hyperledger Fabric is to sort the transactions in the system. Here, transactions represent patterns and data usage behavior.

4. Gatekeeper:  

The gatekeeper is the bridge between the data layer and the system. Controlling access to the data layer through smart contracts helps maintain the correct data flow and the system's correct access to the original data.

Blockchain (BlockChains): Share its network infrastructure and data access rights to reduce the complexity of expenses and operations.

However, there are also competition and trust issues with multiple MNOs in a real environment, and these issues can be reduced through certificate authorities. In order to authorize higher certificates on shared data, we can use the DataChain and BehaviorChain for the alliance chain similar to the super ledger structure.

The "super ledger" is actually an open source ledger with a modular architecture, which can quickly use components such as consensus nodes and MSPs in the system.

The data chain provides control over data access, and the behavior chain is used to record each piece of data. Therefore, combined, these two blockchains provide data authorization, control of data, and auditing of large amounts of data.

data permission

In any system that allows access to raw data, data permissions are a primary consideration. Data permissions can be divided into four different levels based on their risk factors and other security parameters.

  • Data is only visible to the user (L0)
  • Collective use of data without exposing raw data (L1)
  • Raw data is accessible to definitions and authorized parties (L2)
  • Data is public (L3)

Note: Users can set their own data permission levels for complete permission control

Data structure of the system

In order to speed up the process of data sharing through fast user query and data access, the system specially designed a data structure. Let's first look at how transactions occur in the data chain.

Source: Data Structure

The transactions in the data chain mainly include the following components:

Data owner

Timestamp of transactions that occurred in the block

Data permission levels

Level L2 coded as a hash table

Data hashing-maintain data integrity

Data-to-data link pointer

The transactions in the behavior chain mainly include the following components:

Users requesting access to the data

· Timestamp of when data was accessed.

Data address

Access log summary

How is the system implemented?

Member management

Membership management is accomplished through mutual identification and registration, thereby avoiding malicious activities and ensuring secure data access. This can be done in the following steps: 1) Send a key pair with identity information to the network infrastructure.

2) After the user is authenticated, a digital certificate is issued to the new user for identification.

Data collection  

There are two basic types of data, one related to user privacy and the other not related to user privacy.

The contract uses a validator to verify the identity of the data provider.

The contract then identifies raw data related to user privacy, such as user ID and other data. Once identified, it will be encrypted using an asymmetric key.

The contract sends data to the gatekeeper, which stores the data in the cloud and returns the data address. The contract initiates a data transaction request based on the data address, as shown in the following figure:

This is the pseudocode for the data generation contract.

Data generation contract:

3. Data permission level

As we have already discussed, users can define different data permission levels for others to access data owned by the user. Users can assign data permissions using the following code:

4. Data sharing

If you need to perform data calculations without exposing the original data, you need to form an agency composed of other validators and the government as a potential participant, and apply the algorithm together to avoid malicious access to the data.

However, if you need access to raw data, you need to share your data in the following ways:

Data requests appear with digital certificates and digital signatures.

· The contract verifies the legitimacy of the data request through a validator.

5. Data review

  Each data provider (MNO) receives regular data reports. Any malicious activity or data misuse can be identified through authentication in the system.

The user has full control over the data, so can the data be withdrawn in any malicious activity?

in conclusion

With the advent of 5G networks, an organized and optimal AI-driven network can help MNOs and even enterprises to complete the required data requirements and data intensity.