A High-Throughput Blockchain System and Performance Optimization Model for 6G Networks

Abstract

The present invention provides a high-throughput blockchain system for 6G networks. The blockchain is used to achieve decentralized network device secure access authentication and secure sharing management of data; digital twins are integrated in the edge network to analyze, predict, make decisions, and schedule local network conditions and mobile base station information; through federated learning, scheduling and trusted sharing of large-scale network infrastructures are realized. In addition, the present invention establishes a throughput optimization model for the high-throughput blockchain system for 6G networks, derives the time spent on computing and communication in the blockchain network during consensus, analyzes the factors affecting the system throughput, and proposes an optimization method for throughput. Finally, the present invention also establishes a security performance analysis model for the high-throughput blockchain system for 6G networks, describes the measures taken by the high-throughput blockchain system for 6G networks against Byzantine attacks, and analyzes the security performance of the system when it is attacked by malicious nodes.

Description

Technical Field

[0001] The present invention relates to a high-throughput blockchain and performance optimization model for 6G networks, which can be used in the field of blockchain technology. Background Art

[0002] The development of mobile communication networks is increasingly driven by business application requirements. With the rise of applications such as high-definition video, AR / VR, etc., the proportion of video traffic in the total network traffic will exceed 80%. The huge user demand has brought great challenges to the existing mobile communication network architecture, and the network is required to have efficient, secure, and scalable content distribution capabilities. The 5G network has realized the transformation from communication-oriented to service-oriented, and 6G will comprehensively support the digitization of the entire world based on 5G, and combined with the development of technologies such as artificial intelligence, realize the intelligent ubiquitous access, comprehensively empower all things, and promote society to move towards the "digital twin" world of the combination of virtual and reality, and realize the beautiful vision of "digital twin, intelligent ubiquity". These scenarios will require new network capabilities such as security inborn, intelligence inborn, and digital twin, which also put higher requirements on data privacy.

[0003] The future 6G network needs to be interconnected with traditional networks. Among them, a large number of diverse nodes (such as macro base stations, small base stations, terminals, etc.) have different communication characteristics, caching capabilities, computing capabilities, and load conditions. Therefore, it is necessary to coordinate different nodes to achieve distributed network resource complementarity and on-demand networking to meet the complex business requirements of future holographic communication, intelligent industry, etc. However, these distributed network resources may belong to different enterprises, operators, individuals, or third parties, etc. Therefore, a new set of trust and security management solutions is needed to ensure the secure and reliable sharing of resources, the secure circulation of data, and privacy protection.

[0004] In the future 6G era, there are complex application scenarios such as holographic communication and intelligent industry, which need to cope with massive terminal access, huge data transmission, and differentiated service requirements such as ultra-low latency, ultra-high bandwidth, and ultra-high reliability, and will present the characteristics of full spectrum, full application, full coverage, and strong security. Summary of the Invention

[0005] The purpose of the present invention is to solve the above problems existing in the prior art, and propose a high-throughput blockchain and performance optimization model for 6G networks.

[0006] The purpose of the present invention will be achieved through the following technical solutions: A high-throughput blockchain system for 6G networks, the 6G network is divided into three layers, including a core cloud layer, an edge cloud layer, and a device layer.

[0007] The core cloud layer is mainly composed of cloud computing centers provided by each operator, responsible for the overall scheduling, management and supervision of the system, leading the MECs acting as master nodes in each shard chain to complete federated learning, training the global network device scheduling model, authenticating and registering new nodes when they appear in the network, and classifying the new nodes into each shard according to the registration information;

[0008] The edge cloud layer consists of high-performance mobile edge computing nodes (MECs) and various base stations. The MECs are responsible for connecting the core cloud layer and the device layer in the communication network. The main tasks are to manage the shard chain consensus within the shard and relay federated learning messages, and can also participate in the main chain consensus to obtain the global model;

[0009] The device layer consists of sensors, gateway devices, intelligent terminals, fixed base stations and mobile base stations, which is the source of data for decision-making and aggregating models in the network, responsible for collecting data, issuing decisions and forming local models. There are several shards in the device layer, and each shard contains several clusters that cover all the devices within the shard. Each cluster has a gateway or intelligent terminal as the cluster head, responsible for receiving data sent by sensors and issuing corresponding instructions to mobile base stations.

[0010] Preferably, the blockchain network adopts a sharded blockchain design, including a main chain and several shard chains,

[0011] The shard chain is jointly maintained by the mobile edge computing nodes (MECs) in the edge cloud layer of the same shard and the gateway and intelligent terminal nodes in the device layer. The stored content includes the digital twin models maintained in each sub-cluster in the device layer and a large number of unstructured data summaries composed of original data and instructions issued to mobile base stations;

[0012] The main chain is jointly maintained by the cloud computing centers in the core cloud layer and the MECs in the edge cloud layer, responsible for managing all shards and nodes in the blockchain network and storing the federated learning model;

[0013] The shard chain combines digital twins to establish local models through the data sent by sensors to gateways and intelligent terminals to complete the scheduling of network infrastructure. The main chain combines federated learning to aggregate the local models in each shard chain into a global model, realizing the scheduling and trusted sharing of large-scale network infrastructure.

[0014] Preferably, the network infrastructure scheduling includes the following steps:

[0015] S1: When a new network device enters the network, it first needs to submit a registration application to the nearest cloud computing center. After completing the registration, its tasks and shards in the network will be determined according to its configuration information. The network device includes sensors, intelligent terminals, gateway devices and servers;

[0016] S2: The cluster heads of each cluster continuously receive raw data from the devices within the cluster. When a new round of work cycle starts, the gateway at the device layer transmits the real data of the physical space collected in the previous work cycle to the digital twin virtual space.

[0017] S3: The virtual space analyzes and predicts based on the received data file to update the digital twin state.

[0018] S4: When there is network congestion or a sign of a network service request with a higher rate in the network, exceeding the service capacity of the fixed macro base station, scheduling instructions are sent to the mobile base station according to the updated model to provide temporary high-speed network services for users in the specified area and update the digital twin state.

[0019] S5: If scheduling of network infrastructure is required in the current work cycle, after the scheduling is completed, the device layer uploads the updated digital twin state and the data file summary used in the current work cycle to the sharding chain, and consensus on data and models is completed within the shard.

[0020] S6: At the start of the work cycle, the cloud computing center in the core cloud completes the election of the work cycle leader, and the leader of the current work cycle assigns federated learning tasks to each shard in the network.

[0021] S7: Each shard updates its local model according to the digital twin state updated and consensus in the sharding chain.

[0022] S8: The elected cloud computing center collects sufficient local models.

[0023] S9: The cloud computing center aggregates the local models and publishes the aggregated global model and the local models of each shard as transactions to the main chain network for consensus.

[0024] Preferably, the data received from the sensors includes service categories and network conditions.

[0025] The present invention also discloses a performance optimization model for a high-throughput blockchain system for 6G networks, derives the time spent on computing and communication in the blockchain network during consensus, evaluates the working efficiency of the network by analyzing the consensus time and transaction throughput of the blockchain network, analyzes that the main factor affecting the system throughput is the number of shards, and solves the optimal number of shards and the optimal system throughput under different network scales.

[0026] The present invention also discloses the measures taken by the high-throughput blockchain system for 6G networks against Byzantine attacks, and analyzes the security performance of the system when attacked by malicious nodes through the attack model and the defense model.

[0027] Compared with the prior art by adopting the above technical solutions, the present invention has the following technical effects:

[0028] (1) In this system, the blockchain is used to implement decentralized network device security access authentication and secure sharing management of data, provide customized security services for sensitive data, and adopts the design of sharded blockchain to improve the throughput of the system as much as possible; digital twins are integrated in the edge network to connect the digital space and the physical space in real time, analyze, predict and make decisions on the local network conditions and mobile base station information, and achieve efficient intelligent scheduling self-organization of the mobile network.

[0029] Through federated learning, while protecting personal data privacy, an overall model of network intelligent scheduling is established, taking into account the scheduling and optimization of both the overall and local aspects. The integration of edge digital twins and global federated learning realizes the scheduling and trusted sharing of large-scale network infrastructure.

[0030] (2) An optimization model for the throughput of a high-throughput blockchain system for 6G networks is established. The time spent on computing and communication in the consensus of the blockchain network is deduced, the working efficiency of the network is evaluated by analyzing the consensus time and transaction throughput of the blockchain network, the factors affecting the system throughput are analyzed, and an optimization method for the throughput is proposed.

[0031] (3) A security performance analysis model for a high-throughput blockchain system for 6G networks is established, and the measures taken by the high-throughput blockchain system for 6G networks against Byzantine attacks are described. Through the attack model and the defense model, the security performance of the system when it is attacked by malicious nodes is analyzed. BRIEF DESCRIPTION OF THE DRAWINGS

[0032] Figure 1 It is a schematic diagram of the architecture of a high-throughput blockchain system for 6G networks provided by the present invention.

[0033] Figure 2 It is a schematic diagram of the sharded blockchain structure provided by the present invention.

[0034] Figure 3 It is a schematic diagram of the consensus process of the sharded blockchain network in the present invention.

[0035] Figure 4 It is the consensus process of the PBFT algorithm analyzed in the present invention. DETAILED DESCRIPTION OF THE EMBODIMENTS

[0036] The objectives, advantages and features of the present invention will be illustrated and explained by the non-limiting description of the following preferred embodiments. These embodiments are only typical examples of applying the technical solutions of the present invention, and all technical solutions formed by equivalent replacement or equivalent transformation fall within the scope of protection required by the present invention.

[0037] The present invention discloses a high-throughput blockchain and performance optimization model for 6G networks. The present invention assumes that future 6G networks require native support and distributed deployment for AI platforms, and it is necessary to introduce digital twin technology to generate digital twin network entities of 6G networks, build an intelligent endogenous data foundation for 6G networks, achieve decentralized real-time closed-loop control of various network management and applications, and at the same time form a new distributed intelligent paradigm. Through end-edge-cloud collaboration, a globally optimal model is obtained. On the basis of realizing endogenous distributed intelligence, the security access of network devices and the secure storage of privacy data are ensured through blockchain.

[0038] The high-throughput blockchain system designed by the present invention for 6G networks is as Figure 1 shown. This system is divided into a 6G network and a blockchain network. The 6G network is mainly divided into three layers from top to bottom: the core cloud layer, the edge cloud layer, and the device layer. The core cloud layer is mainly composed of the cloud computing center 1 provided by each operator. The edge cloud layer is composed of high-performance mobile edge computing nodes MEC2 and various base stations 3 and 4. The device layer includes a large number of gateways 5, intelligent terminals, and sensors.

[0039] The blockchain network is used to realize the infrastructure sharing of the decentralized network. It is composed of a main chain and several shard chains 6. The shard chains are responsible for storing the idle resource information of the base stations, including idle computing resources, storage resources, and device resources, as well as the access information of gateways and intelligent terminals, including the accessed base stations, service types, and traffic usage conditions, etc. The digital twin models formed by these information are directly stored on the shard blockchain as formatted data, while the original data and unstructured data such as network infrastructure scheduling commands are stored locally, and only the digest is uploaded to the chain; the main chain collects the local models of each shard and integrates them into a global model and uploads it to the chain.

[0040] The blockchain network is composed of some important nodes in the 6G network. Among them, the cloud computing center (core cloud node) in the core cloud layer and the MEC nodes in the edge cloud layer jointly maintain the main chain. The MEC in the edge cloud layer and the gateways and intelligent terminals (device layer nodes) in the device layer are divided into different blockchain shards according to device information and geographical locations, and jointly maintain their respective shard chains. It should be noted that the MEC node is both a replica node 9 of the main chain and the main node 10 of the shard chain.

[0041] Among them, the 6G network is as follows: Core cloud layer: The core cloud layer is mainly composed of cloud computing centers, which are responsible for the overall scheduling, management, and supervision of the system, and lead the MECs serving as the main nodes in each shard chain to complete federated learning and train the global network device scheduling model. Multiple operators can each provide cloud computing centers to achieve cross-network intelligent scheduling and sharing of network devices, save costs, and improve the utilization efficiency of devices and resources.

[0042] The functions undertaken by the core cloud layer mainly include:

[0043] (1) Node registration: When new network devices (including smart terminals, gateway devices, and servers) enter the network, they first submit a registration application to the nearest cloud computing center. After evaluating the device security, the cloud computing center issues an access transaction in the main chain. The access transaction contains the device's registration ID and configuration information, and the configuration information will be used to determine the device's tasks and shards in the network. When the access transaction completes consensus and is recorded on the main chain, the cloud computing center sends a registration success message to the registered device, including the device's registration ID and sharding status.

[0044] (2) Federated learning task allocation: At the beginning of each round of federated learning, the cloud computing center selects a subset of devices from the device layer gateways and smart terminal nodes to participate in this round of training. The cloud computing center assigns federated learning tasks to the gateway nodes in the device subset and sends detailed configuration information, such as data structure, shared state, and model parameters. The purpose of federated learning is to collect local digital twin models within each shard for training a global network device scheduling model to facilitate cross-shard scheduling of network devices. For example, federated learning can be used to obtain the traffic conditions and resource usage of each shard to better optimize the blockchain sharding method or better predict future network usage to schedule network devices in advance and improve the response speed of services.

[0045] The gateway nodes and smart terminals at the device layer will perform local calculations according to the received configuration information and their digital twins, and send the model updates to the cloud computing center for aggregation. After the cloud computing center receives sufficient updated models, it will publish the aggregation results and the collected models to the main chain for storage.

[0046] Edge cloud layer: The edge cloud layer consists of high-performance mobile edge computing nodes MEC and various base stations. MEC is responsible for connecting the core cloud layer and the device layer in the communication network. Its main tasks are to manage the shard chain consensus within the shard and relay federated learning messages, and it can also participate in the main chain consensus to obtain the global model.

[0047] MEC can use the wireless access network to provide cloud computing functions nearby, creating a service environment with high performance, low latency, and high bandwidth, accelerating the rapid download of various contents, services, and applications in the network, and enabling consumers to enjoy an uninterrupted high-quality network experience. It is the core for the system to achieve intelligent scheduling of network devices while considering both the global and local aspects.

[0048] Device layer: The device layer consists of sensors, gateway devices, smart terminals, fixed base stations, and mobile base stations. The device layer is the source of data for decision-making and aggregating models in the network, responsible for collecting data, issuing decisions, and forming models.

[0049] There are several shards in the device layer. Each shard contains several clusters 7, which cover all the devices within the shard. Each cluster has a gateway or intelligent terminal as the cluster head, responsible for receiving the data sent by the sensors (including service category, network status, etc.) and issuing corresponding instructions to the mobile base station. Usually, all the devices in the shard are covered by the signal of a fixed macro base station. However, there may be peak periods of mobile communication network usage at certain locations within the shard area (such as the venue of a festival event, a tourist attraction during holidays), and the macro base station cannot well meet the network needs of users. The mobile base station can arrive at the scene to provide high-quality network services.

[0050] The device layer integrates digital twins. Each cluster head manages a digital twin within its sub-cluster and updates the status of the digital twin according to the real-time data provided by the user terminal, conducts analysis, prediction, judgment and decision-making, and finally completes the scheduling of the mobile base station. The status update of the digital twin and the summary of the real-time data will be uploaded to the shard chain for consensus and stored on the chain. Each device within the shard has a preference list to store the identity of its preferred cluster head, and the preference is determined by physical distance, bit error rate, etc.

[0051] In order to improve the scalability and transaction throughput of the blockchain, and enable the blockchain to meet the requirements of ubiquitous access and high-speed services in the 6G mobile network on the basis of realizing functions such as distributed data consistent storage, difficult to tamper with, and preventing transaction repudiation, and to achieve secure access authentication of devices and secure sharing of device scheduling data in the 6G network, the blockchain of the present invention adopts a sharding design, and its structure is as Figure 2 shown. The entire blockchain system consists of a main chain and several shard chains.

[0052] The blockchain network is designed as follows:

[0053] Shard chain: The shard chain is jointly maintained by the mobile edge computing nodes MEC in the edge cloud of the same shard and the cluster heads in the device layer. The main content stored includes the digital twin models maintained in each sub-cluster in the device layer and a large number of unstructured data summaries composed of the original data and the instructions issued to the mobile base station.

[0054] As a structured data, the digital twin model can be directly stored in the blockchain database. For large-capacity unstructured data, if directly uploaded to the blockchain, on the one hand, it will double the communication pressure and reduce the efficiency of the blockchain network operation. On the other hand, it is not necessary to upload a large amount of unprocessed raw data for storage. Therefore, only the hash value of the raw data and the identity information of the cluster head are used as the data summary for the blockchain operation, while the raw data is stored in the local cluster head. If needed, the corresponding raw data can be queried according to the identity information of the cluster head. By comparing whether the hash value of the data file is consistent with the hash value stored on the blockchain, it can be determined whether the original data file has been modified, which improves the information writing speed and query speed to a certain extent and ensures the data accessibility and traceability.

[0055] Main chain: The main chain is jointly maintained by the cloud computing center in the core cloud layer and the MEC in the edge cloud layer, responsible for managing all shards and nodes in the blockchain network. When a new network device applies to enter the blockchain network, its identity is verified by the cloud computing center, and the registration information of the new node will be stored in the main chain. Since the MEC in the edge cloud participates in the consensus of both the main chain and the shard chain at the same time, all authorized nodes in the entire system can obtain the registration information of the new node.

[0056] In addition, the main chain is also responsible for storing the federated learning model. After the aggregation of the federated learning model is completed, the cloud computing center publishes the aggregated global model and the local models of each shard as transactions to the main chain network for consensus, realizing intelligent scheduling that takes into account the overall situation and balances the whole and the part.

[0057] In the high-throughput blockchain for 6G network designed by the present invention, the shard chain and the main chain work simultaneously, integrating digital twin and federated learning, and realizing efficient intelligent scheduling and deployment of network infrastructure that takes into account both the global and local aspects.

[0058] The specific working process is as follows:

[0059] Step S1: When a new network device (including sensors, intelligent terminals, gateway devices, servers, etc.) enters the network, it first needs to submit a registration application to the nearest cloud computing center. After registration, its tasks and shards in the network will be determined according to its configuration information.

[0060] Step S2: The cluster head of each sub-cluster continuously receives raw data from the devices within the sub-cluster. When a new working cycle starts, the gateway at the device layer transmits the real data of the physical space collected in the previous working cycle to the digital twin virtual space.

[0061] Step S3: The virtual space analyzes and predicts based on the received data file to update the digital twin state.

[0062] Step S4: When network congestion or signs of requests for higher-speed network services appear in the network, exceeding the service capacity of the fixed macro base station, a scheduling instruction is sent to the mobile base station according to the updated model to provide temporary high-speed network services for users in the specified area, and the digital twin status is updated;

[0063] Step S5: If network infrastructure scheduling is required in the current working cycle, after the scheduling is completed, the device layer uploads the updated digital twin status and the data file summary used in the current working cycle to the sharding chain, and consensus on data and models is completed within the shard;

[0064] Step S6: At the beginning of the working cycle, the cloud computing center in the core cloud completes the election of the person in charge of the working cycle, and the person in charge of the current working cycle assigns federated learning tasks to each shard in the network;

[0065] Step S7: Each shard updates the local model according to the digital twin status updated through consensus in the sharding chain;

[0066] Step S8: The elected cloud computing center collects sufficient local models;

[0067] Step S9: The cloud computing center aggregates the local models, and publishes the aggregated global model and the local models of each shard as transactions to the main chain network for consensus.

[0068] The following provides an optimization model for the performance of a high-throughput blockchain system for 6G networks. In the high-throughput blockchain for 6G networks described in the technical solution, the blockchain network consists of a main chain and several sharding chains. Since the nodes in the main chain are composed of cloud computing centers and MECs provided by major operators and other network service providers, this technical solution assumes that these node positions are relatively fixed and use wired transmission, while the nodes in the sharding chain are mainly gateway devices and mobile terminals of users, so this technical solution assumes that these nodes use wireless transmission, and the nodes in the same sharding blockchain form a single-hop network. The main chain and all sharding chains independently use the Practical Byzantine Fault Tolerance consensus algorithm (PBFT) for consensus.

[0069] The consensus process is as Figure 3 shown. After the sharding chains complete consensus respectively, the main chain completes the global consensus, and the average time required to complete one round of consensus is as shown in Equation (1).

[0070]

[0071] Among them, is the total average consensus time for all shards to complete sharding chain consensus, T mainIt is the average consensus time consumed by the main chain to integrate the global model after completing the sharded chain consensus. Since each shard in the system starts the sharded chain consensus synchronously, can be expressed as

[0072]

[0073] where S is the number of sharded chains in the system, is the average time required for the i-th shard to complete the sharded chain consensus.

[0074] There are many factors affecting the blockchain consensus efficiency. Under the condition of unchanged hardware, the most important factors are the working efficiency of the consensus algorithm and the number of nodes in the blockchain. The PBFT Practical Byzantine Fault Tolerance consensus algorithm adopted by this technical solution has excellent performance when the number of nodes is not too large. Therefore, how to shard to minimize the total average consensus time of the sharded chain and the average consensus time T main of the main chain has become the key to improving the consensus efficiency of this blockchain.

[0075] For the i-th shard, the average time required to complete the consensus The consensus process and the formula corresponding to the average time spent of the PBFT consensus algorithm are as Figure 4 shown. Among them, Pre-prepare, Prepare, and Commit are the three stages of the PBFT algorithm consensus, and each stage consists of two parts: calculation and communication. Assume that there are n nodes in a shard, n = 3f + 1, and f is the number of tolerable Byzantine nodes.

[0076] Request: At the beginning of the consensus stage, the replica node i sends a Request message to the primary node p, and the format is where TXN represents the transaction information submitted by the node i to the primary node p, i represents the ID of the replica node, represents the digital signature performed by the private key of i, represents the message authentication code MAC that can be verified by the nodes i and p. Since other workloads and queuing rules are not the focus of this invention, this technical solution assumes that the queuing time can be ignored. Therefore, the average calculation time for the primary node to verify the Request message is

[0077]

[0078] where, b i (t) refers to the number of transactions sent by the replica node i to the primary node p, α refers to the average calculation time-consuming for generating or verifying a MAC, b tot refers to the number of transactions in a block. The reason for the sum with b i(t) is not exactly the same because some abnormal transactions are discarded when verifying the MAC, and θ refers to the average computing time for verifying a digital signature.

[0079] In a sharded blockchain, each node transmits data through a wireless network. Assume the probability of a message transmission failure is p e , then the probability of correct transmission is p s = 1 - p e .

[0080] Therefore, the average communication time from the replica node to the primary node is

[0081]

[0082] where M R is the size of the Request message, W i，p is the maximum transmission bandwidth between the replica node and the primary node, T0 is the average time for two nodes to establish a connection, and T i，p is the average delay caused by link quality and the distance between nodes.

[0083] From formulas (3) and (4), the average total time for this stage can be obtained as

[0084]

[0085] Pre-prepare: The primary node p sorts all normal transactions and generates an unverified block, and then broadcasts a Pre-prepare message P is the ID of the primary node, refers to the MAC generated by p that is suitable for verification by all nodes. So the average computing time of the primary node is

[0086]

[0087] where δ is the average computing time for generating a digital signature. After verifying the block, the replica node also needs to verify each transaction in the block. So the average computing time for the replica node to verify the Pre-prepare message is

[0088]

[0089] Because it is a broadcast, the average time consumed for a replica node to receive the message sent by the primary node can be expressed as

[0090]

[0091] where M P-p is the size of the Pre-prepare message, W p，ris the maximum transmission bandwidth between the primary node and the replica nodes, T p，r is the average latency caused by link quality and the distance between nodes. So the total average time consumption for this stage is:

[0092]

[0093] Prepare: After verification, the replica nodes send a Prepare message to all nodes After each node receives 2f messages, it proceeds to the next step. Here, D(m) is the depth of the block, s is the ID of the target node, and r is the ID of the source replica node. So the average computing time for a node to generate a prepare message is

[0094]

[0095] And the average computing time for verifying after receiving the Prepare messages from other nodes is

[0096]

[0097] The average time consumed by a node to complete communication in this stage is

[0098]

[0099] where M P is the size of the Prepare message, W r，s is the maximum transmission bandwidth between the source replica node and the target node, T r，s is the average latency caused by link quality and the distance between nodes. Since the source replica node has to generate n - 1 MACs, the total average time consumption for this stage is:

[0100]

[0101] Commit: After all nodes receive 2f Prepare messages, they broadcast a Commit message. After collecting 2f Commit messages (including their own), they proceed to the next step. Each node has to generate n - 1 MACs and verify 2f MACs. So, the average computing time for each node to generate a Commit message is

[0102]

[0103] And the average computing time for verifying the received Commit messages is

[0104]

[0105] The average time consumed by a node to complete communication in this stage is

[0106]

[0107] Therefore, the total average time consumption in this stage is

[0108]

[0109] At this point, the consensus process ends. Therefore, for the i-th shard, the average time required to complete consensus is

[0110]

[0111] Therefore, the number of transactions processed per unit time TPS in a single shard chain can be expressed as:

[0112]

[0113] After completing the shard chain consensus, the primary nodes of each shard will package the information to be uploaded as a transaction and send it to the primary node in the main chain, and participate in the consensus of the main chain as a replica node of the main chain. The main chain also uses the PBFT consensus algorithm. The calculation method of the average time required to generate and verify the consensus messages in each stage is similar to that in the shard chain. It should be noted that the main chain nodes use optical fiber to transmit data, and only the data packet size and transmission bandwidth will affect the average communication time. The average consensus time of each stage in the main chain can be expressed as:

[0114]

[0115] Among them, is the number of all transactions in a block, S is the number of shard chains, N main is the total number of nodes in the main chain, N main = S + M, M is the number of core cloud nodes in the main chain, and F is the number of tolerable Byzantine nodes in the main chain. Therefore, the total average time consumption of the main chain consensus is:

[0116]

[0117] According to Equation (1), the average time required for the high-throughput blockchain described in the present invention for a 6G network to complete one round of work is:

[0118]

[0119]

[0120] And the number of transactions processed per unit time TPS of the system can be expressed as:

[0121]

[0122] The performance optimization model of this technical solution assumes that the total number of nodes in the area remains unchanged at N, with the maximum TPS as the goal.

[0123] To achieve the maximum TPS, it is necessary to minimize the consensus time per round while increasing the number of transactions processed in each consensus round. In a round of sharded chain consensus, each cluster head submits a transaction to its primary node. Therefore, the number of transactions in shard i is determined by the number of cluster heads in shard i, and the sum of the number of transactions in all sharded chains is determined by the number of shards S in the system; in a round of main chain consensus, the primary nodes of each sharded chain submit a transaction to the main chain primary node, so the total amount is determined by the number of shards S in the system. The number of shards S also determines the average size of each shard, and thus determines the consensus time for one round.

[0124] In summary, the number of transactions processed per unit time TPS in this system is determined by the number of shards S. Therefore, the mathematical description of this problem is

[0125]

[0126] s.t. 1 < S < (N - M) / 4

[0127]

[0128] T main < τ max

[0129] S ∈ Z

[0130] M is the number of core cloud nodes in the main chain. Therefore, N - M is the total number of sharded chain nodes composed of gateway nodes and intelligent terminal nodes in the system. To ensure the security of the PBFT consensus algorithm, the number of nodes within a shard must not be less than 4. Therefore, the number of shards S is less than (N - M) / 4. At the same time, to ensure the user's network experience, the consensus time of each blockchain (including the main chain and sub-chains) must not exceed τ max , τ max is the maximum tolerable consensus time, that is, in the case of normal communication or no attack, the consensus time will not exceed τ with a probability of P (for example, P = 99.9%). max .

[0131] In the case of uniform node distribution, 200 points are randomly scattered within a square area with a side length of 1 kilometer as edge cloud nodes and device layer nodes in the blockchain network. Their specific responsibilities will be allocated separately after sharding is completed. One cloud computing center is placed at the center of the square area as the core cloud node. Through numerical analysis, it can be found that when the number of nodes in the system is kept constant, the TPS changes with the number of shards in the system. When the number of shards is 16, that is, the average number of nodes in each shard is 11.5, the TPS of the system reaches the maximum value. The hollow circles in the figure represent the average TPS values of the system under different numbers of shards obtained by Monte Carlo simulation. When the number of shards is less than 16, the TPS of the system increases rapidly with the increase in the number of shards, and then decreases gently with the increase in the number of shards.

[0132] To better analyze the influence of network scale on the optimal sharding and the division of responsibilities of nodes in the blockchain network, the optimal TPS, the optimal sharding situation, and the optimal node division under different network scales are summarized in Table 1.

[0133] Table 1 Parameters under different network scales

[0134] Total number of nodes 100 200 300 400 500 600 700 Optimal TPS 11470 12308 12689 12925 13088 13207 13298 Optimal number of shards 11 16 19 22 25 27 29 Number of nodes in the edge cloud layer 11 16 19 22 25 27 29 Number of nodes in the device layer 89 184 281 378 475 573 671 Average number of nodes per shard 8.0909 11.5 14.789 17.181 19 21.222 23.137

[0135] Total number of nodes 800 900 1000 2000 3000 4000 5000 Optimal TPS 13374 13438 13493 13780 13910 14041 14110 Optimal number of shards 31 33 35 49 61 70 78 Number of nodes in the edge cloud layer 31 33 35 49 61 70 78 Number of nodes in the device layer 769 867 965 1951 2939 3930 4922 Average number of nodes per shard 24.806 26.272 27.571 39.816 48.180 56.142 63.102

[0136] It can be seen that the optimal TPS gradually increases with the increase in network scale. This is because the high-throughput blockchain for 6G designed in this technical solution adopts the design of sharded blockchain. With the increase in network scale, the optimal number of shards also increases, and the number of transactions processed in parallel by the system also increases accordingly. Therefore, even though due to the increase in network scale, the scales of the sharded chain and the main chain and the single working cycle all increase correspondingly, the overall TPS of the system is still gradually increasing. When the total number of nodes is 1000, the peak value of the system throughput increases by 17.6% compared with when the number of nodes is 100. However, the upward trend gradually decreases because the communication complexity of the PBFT consensus algorithm is O(n 2 )), and the time consumed in communication increases sharply with the increase in the number of nodes. When the number of nodes in the system reaches a certain scale, the communication tasks in the PBFT consensus algorithm are greatly aggravated, and the time consumed in communication limits the further growth of TPS.

[0137] In addition, to reflect the superiority of the high-throughput blockchain for 6G and the throughput performance optimization model proposed in this technical solution, a comparison of throughput performance will be made among the blockchain with optimal sharding, fixed sharding, and traditional non-sharding. When the number of nodes is large, the throughput performance of traditional non-sharding is very poor, and as the number of nodes continues to increase, the performance continues to decline; in contrast, the blockchain with sharding design has a great improvement in throughput performance.

[0138] When the number of nodes is 100, the system throughput of the system with 20 fixed shards reaches 33 times that of the traditional non-sharded system. If the optimal sharding design of the present invention is adopted, the system throughput can be further increased by 53.1% compared with the fixed sharding, that is, 50 times that of the non-sharded system. The blockchain with fixed sharding can reach the same level as the optimal sharding when the network scale reaches a certain value, but under other network scales, the throughput performance drops rapidly. Therefore, when sharding design is adopted and an appropriate number of shards is set, the throughput performance of the blockchain can be effectively improved.

[0139] The following provides an analysis model for the security performance of a high-throughput blockchain system for 6G networks.

[0140] Suppose there are N blockchain nodes in total, and the attacker attacks a nodes at the same time. The attack speed is t a , that is, the time for the attacker to break through the firewall of the node itself and then tamper with the node information. The verification is initiated during the process of transaction request processing. If a new verification is initiated during the period from the attacker launching the attack to the successful tampering, the abnormal behavior of the node will be detected by the system and the attack fails. Suppose the time intervals for the initiation of requests and the initiation of verifications are both t d obeys an exponential distribution with parameter λ.

[0141] For the next attack time t a the probability that there is at least one request is

[0142]

[0143] During any attack time t a the probability that the attacker successfully attacks a blockchain node is

[0144]

[0145] Then during this period, X×r blockchain nodes are invaded. Suppose the number of nodes restored in each attack cycle is Y. When the total number of successfully attacked nodes does not exceed 1 / 3, the system runs normally. Therefore, in the attack cycle i, M(i) nodes are invaded

[0146]

[0147] The probability that the attacker successfully attacks a node is

[0148]

[0149] For the attacker, the cost is the cost C of invading the node a1 and the penalty C for the attack being detected a2; The benefit is that the intrusion node attack is successful, B a ;

[0150] For the defender, the cost is the cost C generated by the attacker's successful attack on the node d ; The benefit is that the node operates normally, B d1 And the benefit B of finding the problem node led by the master node d2 .

[0151] Therefore, the utility function of the attacker can be obtained as

[0152] u a = B a × PT - C a1 - C a2 × (1 - PT) #(29)

[0153] The utility function of the defender is

[0154]

[0155] In blockchain, the Byzantine Generals (BG) problem is how to reach a consensus among untrusted nodes. In the BG problem, a group of general armies in Byzantine regions surround the city. Some generals like to attack, while others prefer to retreat. However, if only some of the generals attack the city, the attack will fail. Therefore, they must reach a consensus agreement on attacking or retreating. However, there are likely to be traitors and spies of the enemy in the army, who will influence the generals' decisions or tamper with information when transmitting messages. The Byzantine problem lies in how the remaining loyal generals can reach a consistent agreement in a distributed environment without being affected by traitors when it is known that there are members rebelling.

[0156] As a blockchain network, blockchain is also distributed and faces similar problems. In blockchain, there is no central node to ensure that the ledgers on distributed nodes are the same. Therefore, protocols are needed to ensure the consistency of ledgers in different nodes. At the same time, the distributed network may also face hacker attacks, trying to tamper with transaction information and prevent transaction consensus. Therefore, the blockchain network of this technical solution adopts the Practical Byzantine Fault Tolerance (PBFT) consensus algorithm, which is a consistency algorithm based on message passing. The algorithm needs to go through five stages to reach consistency, and these stages may be repeated due to failures.

[0157] Assume that the total number of nodes is 3f + 1, where f is the number of faulty or Byzantine nodes that can be tolerated. It can be seen that in the PBFT algorithm, the more nodes there are, the stronger the security. In the blockchain network designed in this technical solution, in order to improve the throughput of the system, a sharding design is adopted. However, if the number of nodes in a shard chain is too small, it may lead to the failure of consensus within the shard when there are faulty or Byzantine nodes. Therefore, the present invention limits the minimum number of nodes in a shard, so that even if there are faulty or Byzantine nodes in the shard, the consensus can be completed normally.

[0158] The work of the blockchain nodes in this technical solution is functionally divided into two stages. One is the normal operation stage, which verifies the entities in the system, that is, the prevention stage. The other is the stage of dealing with problems when an attack is detected. In the prevention stage, the security mechanism of the blockchain network mainly includes two aspects: preventing illegal nodes from accessing and preventing attackers from tampering with the information in the shared database. The main way is to conduct identity authentication by exchanging information data among blockchain nodes. Each node has its own verification information and stores the verification information of other nodes. Through encrypted data transmission and information verification, consensus is obtained to achieve the effect of preventing tampering and forgery. At the same time, the consistent data in the blockchain network is distributedly stored, making it difficult to tamper with the information of a single node.

[0159] In the processing stage, the node detects abnormal traffic, judges whether there is an attack behavior, and mitigates it through smart contracts to achieve automatic response to attacks. The smart contract will regularly check the node status, traverse each state machine, transaction, and trigger condition included in each contract item by item; push the transactions that meet the conditions to the queue to be verified and wait for consensus; the transactions that do not meet the trigger conditions will continue to be stored in the transaction pool.

[0160] This technical solution also enhances the system security by limiting the consensus time and adding a recovery mechanism. Limiting the consensus time means that in order to ensure the security and anti-tampering of the consensus process, it is necessary to ensure that the consensus time of each block is within a certain range.

[0161] This technical solution sets a maximum time delay τ max . If the consensus time is less than τ max , then the block is considered valid. If the consensus time is greater than τ max , then the messages in the consensus process of the block may be intercepted and tampered with or there may be a communication failure, and the consensus is considered to have failed. Ignore the block and skip the consensus of this round.

[0162] The recovery mechanism mainly takes measures after some nodes are found to have problems by normal nodes. After each round of consensus, if the master node receives reports from more than one-third of the replica nodes that a certain node has exceeded the response time, the master node broadcasts a message to other replica nodes to request contact with that node. If the node is working properly and receives the message, it replies with a short <(OK, v, n) encryption> to confirm the identity of the node; if it fails to contact, it replies with a node failure message to the master node, suspecting that the node is a faulty or Byzantine error node.

[0163] If the master node receives point failure messages sent by more than one-third of the replica nodes, to avoid greater harm, the master node sends an isolation command to other nodes, requiring the devices under that node to be assigned to other nodes and rejecting transactions initiated by that node. If the node is considered normal, the master node broadcasts a confirmation message, which contains the identity authentication information of the replica nodes that sent the node failure messages. If a replica node sends a node failure message to the master node but does not receive an isolation command for the faulty node or a confirmation message containing the identity authentication information of this node, it will send the node failure message again after waiting for a period of time. When f + 1 nodes among all nodes agree to this request, the instruction is notified to the node using the identity proof, so that the problem node is effectively isolated and becomes a "bare commander", and then maintenance personnel will conduct subsequent inspections and repairs on the problem node.

[0164] In terms of anti-destruction, this system can respond quickly and stop losses in a timely manner when under attack. For each shard, as long as the number of faulty nodes does not exceed one-third, it can work normally. Even if the number of faulty nodes exceeds one-third of the number of nodes in the shard, only that shard will stop working, limiting the risk and loss to a relatively small range. Each shard records various activities of the system and records timestamps. When an attack occurs, it can achieve self-checking, self-verifying, and self-recovering; after the attack is mitigated, it will conduct activity analysis and trace the root cause of the nodes that generated the attack activities.

[0165] In terms of information confidentiality and integrity, this system uses blockchain technology. Each node manages its own private key and distributes the public key. Each block stores an encrypted fragment of node information, and privacy protection can be achieved without any third-party access and control of data. The blockchain is composed of multiple distributed databases that can be read, added, and not deleted, connected in a chain structure to maintain a record list of blocks. Each block contains a timestamp and a link to the previous block, so the blockchain provides a security guarantee that once there is a record, the data cannot be modified.

[0166] In the future 6G era, there are complex application scenarios such as holographic communication and intelligent industry, which need to cope with massive terminal access, huge data transmission, as well as differentiated service requirements such as ultra-low latency, ultra-high bandwidth, and ultra-high reliability, and will present the characteristics of full spectrum, full application, full coverage, and strong security. In response to the above requirements, this technical solution designs a high-throughput blockchain system for 6G networks, integrates the blockchain into the 6G infrastructure endogenously, combines digital twin and federated learning, realizes the secure sharing and intelligent scheduling of network infrastructure on the basis of ensuring the privacy, security, and credibility of data, achieves the deep integration of the blockchain and wireless communication, breaks the trust barrier between "humans-machines-things-networks", improves the efficiency and security of wireless networks, and establishes a more comprehensive system performance optimization model to evaluate and optimize system performance.

[0167] The blockchain is used to achieve decentralized network device secure access authentication and secure sharing management of data; digital twin is integrated in the edge network to analyze, predict, make decisions, and schedule local network conditions and mobile base station information; through federated learning, the scheduling and trusted sharing of large-scale network infrastructure are realized. In addition, this technical solution establishes a throughput optimization model for the high-throughput blockchain system for 6G networks, deduces the time spent by the blockchain network in computing and communication during consensus, analyzes the factors affecting system throughput, and proposes an optimization method for throughput. Finally, this technical solution also establishes a security performance analysis model for the high-throughput blockchain system for 6G networks, describes the measures taken by the high-throughput blockchain system for 6G networks against Byzantine attacks, and analyzes the security performance of the system when it is attacked by malicious nodes.

[0168] The blockchain is used to achieve decentralized network device secure access authentication and secure sharing management of data; digital twin is integrated in the edge network to analyze, predict, make decisions, and schedule local network conditions and mobile base station information; through federated learning, the scheduling and trusted sharing of large-scale network infrastructure are realized. In addition, this invention establishes a throughput optimization model for the high-throughput blockchain system for 6G networks, deduces the time spent by the blockchain network in computing and communication during consensus, analyzes the factors affecting system throughput, and proposes an optimization method for throughput. Finally, this invention also establishes a security performance analysis model for the high-throughput blockchain system for 6G networks, describes the measures taken by the high-throughput blockchain system for 6G networks against Byzantine attacks, and analyzes the security performance of the system when it is attacked by malicious nodes.

[0169] It is apparent to those skilled in the art that the present invention is not limited to the details of the above-described exemplary embodiments, and that the present invention can be implemented in other specific forms without departing from the spirit and essential characteristics of the present invention. Therefore, in all respects, the embodiments should be regarded as exemplary and non-restrictive. The scope of the present invention is defined by the appended claims rather than the above description. Accordingly, all changes that fall within the meaning and scope of the equivalent elements of the claims are intended to be embraced within the present invention, and any reference signs in the claims should not be construed as limiting the claims involved.

[0170] In addition, it should be understood that although this specification is described in terms of embodiments, not every embodiment contains only one independent technical solution. This narrative manner of the specification is merely for clarity. Those skilled in the art should regard the specification as a whole, and the technical solutions in the various embodiments can also be appropriately combined to form other embodiments that can be understood by those skilled in the art. There are still various embodiments of the present invention, and all technical solutions formed by equivalent transformation or equivalent substitution fall within the protection scope of the present invention.

Claims

1. A high-throughput blockchain system for 6G networks, characterized in that: The 6G network is divided into three layers, including the core cloud layer, the edge cloud layer, and the device layer. The core cloud layer is mainly composed of cloud computing centers provided by each operator, responsible for the overall scheduling, management, and supervision of the system, and leading the MECs serving as the primary nodes in each shard chain to complete federated learning, training the global network device scheduling model. When a new node appears in the network, it authenticates and registers the new node, and distributes the new node into each shard according to the registration information. The edge cloud layer consists of high-performance mobile edge computing nodes (MECs) and various base stations. The MEC is responsible for connecting the core cloud layer and the device layer in the communication network. Its main task is to manage the shard chain consensus within the shard and relay federated learning messages, and participate in the main chain consensus to obtain the global model. The device layer consists of sensors, gateway devices, intelligent terminals, fixed base stations, and mobile base stations. It is the source of data for decision-making and aggregating models in the network, responsible for collecting data, issuing decisions, and forming local models. There are several shards in the device layer, and each shard contains several clusters that cover all the devices within the shard. Each cluster has a gateway or intelligent terminal as the cluster head, responsible for receiving data sent by sensors and issuing corresponding instructions to the mobile base station. The blockchain network adopts a sharded blockchain design, including a main chain and several shard chains. The shard chain is jointly maintained by the MECs in the edge cloud layer and the gateway and intelligent terminal nodes in the device layer within the same shard. The stored content includes the digital twin models maintained in each sub-cluster in the device layer and a large number of unstructured data summaries composed of raw data and instructions issued to the mobile base station. The main chain is jointly maintained by the cloud computing centers in the core cloud layer and the MECs in the edge cloud layer, responsible for managing all shards and nodes in the blockchain network and storing the federated learning model. The shard chain combines digital twins to establish a local model based on the data sent by sensors to the gateway and intelligent terminal, completing the scheduling of network infrastructure. The main chain combines federated learning to aggregate the local models in each shard chain into a global model, realizing the scheduling and trusted sharing of large-scale network infrastructure.

2. The high-throughput blockchain system for 6G networks according to claim 1, characterized in that: The network infrastructure scheduling includes the following steps: S1: When a new network device enters the network, it first submits a registration application to the nearest cloud computing center. After completing the registration, its tasks and shards in the network will be determined according to its configuration information. The network device includes sensors, intelligent terminals, gateway devices, and servers. S2: The cluster head of each sub-cluster continuously receives raw data from the devices within the sub-cluster. When a new working cycle starts, the gateway in the device layer transmits the real data of the physical space collected in the previous working cycle to the digital twin virtual space. S3: The virtual space analyzes and predicts based on the received data file, updating the digital twin state. S4: When there is network congestion or a sign of a network service request with a higher rate in the network, exceeding the service capacity of the fixed macro base station, a scheduling instruction is sent to the mobile base station according to the updated model to provide temporary high-speed network services for users in the specified area and update the digital twin status; S5: If network infrastructure scheduling is required in the current working cycle, after the scheduling is completed, the device layer uploads the digital twin status update and the data file summary used in the current working cycle to the sharding chain, and data and model consensus are completed within the shard; S6: At the beginning of the working cycle, the cloud computing center in the core cloud will complete the election of the person in charge of the working cycle, and the person in charge of the current working cycle will allocate federated learning tasks to each shard in the network; S7: Each shard updates the local model according to the digital twin status update consensus in the sharding chain; S8: The elected cloud computing center collects the local models; S9: The cloud computing center aggregates the local models and publishes the aggregated global model and the local models of each shard as transactions to the main chain network for consensus.

3. The high-throughput blockchain system for 6G networks according to claim 1, characterized in that: Responsible for receiving data sent by sensors, including service categories and network conditions.

Images