CN108737175B

CN108737175B - Node management method and implementation system thereof

Info

Publication number: CN108737175B
Application number: CN201810484136.2A
Authority: CN
Inventors: 胡凝; 丛宏雷; 李喜平; 王大双
Original assignee: Shanghai Distributed Technologies Co ltd
Current assignee: Shanghai Distributed Technologies Co ltd
Priority date: 2018-05-19
Filing date: 2018-05-19
Publication date: 2021-04-23
Anticipated expiration: 2038-05-19
Also published as: CN108737175A

Abstract

The invention provides a node management method and an implementation system. The consensus nodes monitor the current block height, establish connection to form a connection network, and handshake the consensus nodes in the connection network. And if no response data packet is received after the preset time, judging that the node response is overtime, and recording the handshake time to form a recording queue. And if the overtime times exceed the preset times, judging the connection blockage of the node, and sending an abnormal request data packet to the consensus node. And when the number of the connected and blocked consensus nodes exceeds the preset blocking number, the consensus nodes quit the consensus network. The block chain network can work normally when the number of nodes is too large, the network of partial network nodes is not good, and the low computation force nodes accumulate the whole network and slowly go out of blocks.

Description

Node management method and implementation system thereof

Technical Field

The invention relates to the technical field of block chains (Blockchain), in particular to a node management method and an implementation system thereof.

Background

A blockchain is generally understood to be a distributed book, which is essentially a database of assets that can be shared at multiple sites, at different geographic locations, or across a network of multiple organizations. Participants in a network may obtain a copy of a unique, real ledger. A blockchain may also be understood to be a decentralized, distributed state machine. The state machine is a machine for recording states, and by means of distributed consistency algorithms, the states can be guaranteed to be switched point to point under the condition that no center exists, and the states can be uniformly and consistently switched into a new state, and the switching is based on transactions approved by all people.

The consensus network in the block chain usually adopts a connection-oriented mode, and a few adopt a connectionless mode. In most block chain networks, a consensus network needs each node to establish connection to form a consensus channel, consensus is achieved by broadcasting a consensus data packet, particularly in algorithms such as BFT, a node packs transactions into alternative blocks and broadcasts the alternative blocks to the consensus network through the consensus data packet, and all consensus nodes need to forward the data packet, so that huge network overhead is formed.

The block chain network packs the transactions into blocks through each node, namely a plurality of transactions are packed into alternative blocks and added into the block chain through distributed consistent consensus, and then the transactions in the blocks are executed, namely the account is dropped. The blockchain network comprises a large number of consensus nodes, and accounting nodes, consensus candidate nodes, light nodes and the like can be arranged according to requirements. Records, distributed consistent consensus and execution of transactions in the block chain network all require computing power, and computing power of a single node is based on hardware performance and comprises network throughput, storage throughput, computing power and data algorithms. Each node forms a block chain network, embodies the computing power of the whole network, and is reflected in the service throughput (TPS, transactionapsecond, transaction number/second) and the block output speed.

The factors affecting the performance of the blockchain network are mainly two:

1. when the number of the consensus nodes is enough and exceeds the required number, the process of consensus occupies a large amount of network resources, and further causes network jitter (also called Packet jitter), that is, the block chain network is unstable, reduces the efficiency, and even cannot achieve the consensus to generate blocks.

2. In the account drop process after the consensus is completed, due to the diversity of the transactions, the operation of a large number of transactions and intelligent contracts, a large number of computing resources are occupied, node performance bottlenecks may exist, blocks cannot be generated, and the whole network is slowed down.

At present, the block chain nodes cannot describe the network condition through the self condition, and when the nodes are too many to form broadcast storms or some nodes have network jitter to influence the operation of the whole block chain network, no corresponding automatic solution is provided. The block chain network is difficult to judge which nodes are problem nodes and which nodes are normal nodes, and even if the abnormal consensus nodes of the network perceive that the performance of the nodes is low and the network condition is not good, the abnormal consensus nodes of the network are not willing to quit the network, so that the block chain network can not work normally continuously.

Therefore, it is necessary to develop a new node management method to solve the above technical problems.

Disclosure of Invention

The technical problem to be solved by the invention is as follows: when the number of the consensus nodes is excessive or the network states and performances of some consensus nodes are not good, network jitter occurs, so that the block chain network cannot work normally.

In order to solve the above technical problem, the present invention provides a node management method, which includes the following steps.

And (3) node connection step: the method comprises the steps that consensus nodes in a consensus network monitor the height of a current block, at least three consensus nodes are connected to form a connection network, each consensus node in the connection network shakes hands with the consensus nodes except the consensus nodes, the shaking hands are used for sending detection data packets to the nodes which are connected, and the nodes which receive the detection data packets send response data packets to the nodes which send the detection data packets;

and a timeout judging step: setting preset time, when one consensus node does not receive a response data packet of another consensus node after the preset time is exceeded, judging that the response of the node is overtime, recording the handshake time of each handshake by the consensus node to form a recording queue, wherein the handshake time is the interval time between the sending of a detection data packet and the receiving of a response data packet;

and (3) node screening: setting a preset number of times and a preset blocking number, recording the response overtime times of the consensus nodes except the consensus nodes in the connection network by the nodes sending the detection data packets, judging the connection blocking of the nodes when the response overtime times of one node exceeds the preset number, sending an abnormal request data packet to the consensus nodes, wherein the handshake times do not exceed the preset number at the same block height;

a decision making step: the common recognition nodes receiving the abnormal request data packet count and handshake time of nodes except the abnormal request data packet to determine the number of the common recognition nodes which are blocked in connection, when the number of the common recognition nodes which are blocked in connection is smaller than the preset blocking number, the common recognition nodes which are blocked in connection send smooth data packets to the nodes which send the abnormal request, and when the number of the common recognition nodes which are blocked in connection exceeds the preset blocking number, the common recognition nodes exit the common recognition network.

The method has the advantages that the network which gives a problem is judged by intelligently judging which of the common identification nodes gives a problem according to the rule, the network which gives a problem is judged to temporarily exit the node, the block chain network is ensured to work normally when the number of the nodes is too large, the network of partial network nodes is not good, and the low-computation-force nodes accumulate the whole network to give blocks slowly, meanwhile, the number of detection data packets for detecting the network condition is limited, and the monitoring mechanism is prevented from broadcasting too much to influence the network.

Preferably, in the node connection step, a network monitoring channel is established, and the detection data packet, the response data packet, the abnormal request data packet, and the clear data packet are sent through the network monitoring channel. The beneficial effects are that: and a network monitoring channel is additionally arranged, so that the method is suitable for different kinds of consensus networks.

Preferably, in the node connection step and the node screening step, a consensus algorithm traversing all consensus nodes is adopted, and the detection data packet, the response data packet, the abnormal request data packet, and the unobstructed data packet are sent through a consensus broadcast channel. The beneficial effects are that: in the consensus network of the BFT similar consensus protocol, a consensus channel needs to be established, all nodes need to forward the consensus data packet in the consensus process, and the network overhead is saved by using the original channel.

Preferably, in the node connecting step and the node screening step, the sensing packet and the response packet are part of a common identification packet.

Preferably, in the timeout determining step, the consensus node records handshake time of each handshake with each node, and forms a record queue by sorting the records according to handshake time from high to low, and the record that has not received the response packet is arranged at the forefront of the record queue.

Preferably, in the timeout determining step, the consensus node records the handshake time of each handshake with each node, and the records with each node are respectively sorted from high to low according to the handshake time to form a record queue.

Preferably, in the timeout determining step and the node screening step, a predetermined number of nodes is set, the consensus node checks the response-overtime consensus nodes recorded in the queue, and sequentially sends the abnormal request packets to the consensus nodes recording that the number of response-overtime consensus nodes in the queue exceeds the predetermined number, and the number of the sent abnormal request packets does not exceed the predetermined number of nodes. The beneficial effects are that: prevent the block chain network from being unable to work due to the fact that a large number of nodes exit.

Preferably, the consensus node broadcasts the record queue to a connection network while sending an abnormal request packet, receives the consensus node of the record queue, and compares the record queue with its own record queue to determine the consensus node with the connection block. The beneficial effects are that: the link conditions of the self node and other nodes are compared, and the nodes which fail are easier to judge.

Preferably, a predetermined block time is set, the consensus nodes in the consensus network simultaneously listen to the block speed, and when the block time exceeds the predetermined block time, the consensus nodes with the response timeout times exceeding the predetermined times in the recording queue sequentially send the abnormal request data packets. The beneficial effects are that: the low-computation-effort nodes are tired of the whole network and cannot go out of the block, and still trigger an exit mechanism to stabilize the network.

Preferably, in the decision step, after receiving at least two abnormal request data packets, the consensus node determines the number of the connection-blocked consensus nodes according to the record queue statistics and handshake time of nodes except for the consensus node itself, sends a clear data packet to the node sending the abnormal request when the connection-blocked consensus nodes are smaller than the predetermined blocking number, and exits the consensus network when the connection-blocked consensus nodes exceed the predetermined blocking number. The beneficial effects are that: more than two nodes inform the nodes of carrying out connection judgment only when the network is overtime, and the cost of connection judgment is reduced.

Preferably, the decision step further comprises a self-checking step: setting preset self-checking time, sending an abnormal request data packet to the node common identification points except the node common identification points by the common identification nodes beyond the preset self-checking time without receiving a smooth data packet, counting the handshake time of the nodes except the node common identification points to determine the number of the common identification nodes which are connected and blocked, and exiting the common identification network when the number of the common identification nodes which are connected and blocked exceeds the preset blocking number. The beneficial effects are that: the common node is not connected with other nodes smoothly, possibly the common node is in a problem of the other side, and possibly the common node is in a problem of the common node itself, if the connection with a plurality of nodes is blocked, the common node is possibly in a problem of the common node itself, and self-checking is provided to better maintain the network.

Preferably, in the self-checking step, at least two abnormal request data packets sent by the common node do not receive a clear data packet after exceeding the predetermined self-checking time, and count handshake time with nodes other than the common node to determine the number of the common node with the blocked connection, and exit the common network when the number of the common node with the blocked connection exceeds the predetermined blocking number. The beneficial effects are that: and more than two abnormal requests are sent to perform connection judgment, so that the overhead of connection judgment is reduced.

Preferably, the number of the non-accessible blocks is set, the consensus nodes in the consensus network simultaneously sense and record the block speed, when the number of times of the block-out time exceeds the preset block-out time exceeds the number of the non-accessible blocks, the node manager is informed and an abnormal alarm is broadcast.

Preferably, the decision step further includes a node self-control step, where the consensus nodes in the consensus network simultaneously monitor the total number of the consensus nodes in the consensus network, and when the total number of the remaining consensus nodes is less than the total number of the consensus fault tolerance, the node manager is notified and an abnormal alarm is broadcast.

Preferably, the node in the blockchain network delegates the interest obtaining interest value to become a consensus node in the consensus network, and the consensus node obtains the benefit by performing consensus and the interest value.

Preferably, in the decision step, the consensus node that exits the consensus network randomly delegates its own rights to the consensus nodes other than itself in the connection network to become a delegating node, the consensus node that accepts the delegation has a superimposed rights value, the superimposed rights value is a superposition of its own rights value and the rights value that accepts the delegation, and the consensus node that accepts the delegation obtains a profit by performing the consensus and the superimposed rights value.

Preferably, the joint consensus node calculates the profit by using a profit function according to the right and benefit value every time a round of joint consensus is performed, wherein the profit function is a spindle-shaped function.

Preferably, the node accepting the delegation sends a detection data packet to the delegation node, the delegation node receiving the detection data packet sends a response data packet to the node accepting the delegation, when the node accepting the delegation does not receive the response data packet of the delegation node after the preset time, the right entrusted by the delegation node is randomly delegated to a consensus node except the consensus node in the connection network, when the node accepting the delegation receives the response data packet of the delegation node in the preset time, the right entrusted by the delegation node and the income are returned to the delegation node, and the income of the delegation node and the node accepting the delegation node is distributed in equal proportion according to the respective right value. The beneficial effects are that: the problem nodes are encouraged to exit through revenue distribution.

Preferably, a deduction coefficient is set, the deduction of the entrusted node is multiplied by the deduction coefficient to obtain deduction, the deduction is taken out from the income of the entrusted node, and the deduction is distributed to the nodes in the connection network. The beneficial effects are that: the quitted nodes do not participate in consensus after all, and the income coefficients of the quitted nodes are reduced to balance the network.

Preferably, the waiting time is set, after the waiting time, the consensus node exiting the network sends the detection data packet to the consensus node in the connection network, the consensus node receiving the detection data packet sends a response data packet to the consensus node from which the detection data packet originates, and when the response data packet of the consensus node in the connection network is received in the preset time, the join application is sent to the connection network. The beneficial effects are that: and after the network is recovered, returning to the consensus network.

Preferably, each consensus node in the connecting network handshakes with the consensus node exiting the network, agrees to join the connecting network when the handshake time is less than the predetermined time, and joins the connecting network when the consensus node in the connecting network exceeds 1/2.

The invention also provides a system for realizing the node management method, which comprises a monitoring unit, a connecting unit, a queue unit, a decision unit and a parameter setting unit;

the monitoring unit is used for monitoring the block height information in the block chain network and sending the block height information to the connecting unit;

the connection unit is used for receiving the block height information of the monitoring unit, establishing connection with the common identification nodes except the monitoring unit to form a connection network, periodically performing handshaking and recording handshaking time to form a recording queue, storing the recording queue to the queue unit, judging that the response of one common identification node is overtime when the response of the node exceeds a preset time and does not receive a response data packet of the other common identification node, judging that the connection of the node is blocked when the response of the node exceeds the preset time, and sending an abnormal request data packet to the common identification node; receiving a response data packet of a node which is connected with a block and sending the response data packet to the decision unit, receiving an abnormal request data packet sent by a common node except the node, receiving a smooth instruction of the decision unit and sending a smooth data packet to the node which sends the abnormal request;

the queue unit is used for storing a record queue;

the decision unit is used for reading the recording queue in the queue unit, making a decision according to the recording queue, receiving an abnormal request data packet, counting handshake time of nodes except for the abnormal request data packet to determine the number of the common identification nodes with the connection blockage, sending a smooth instruction to the connection unit when the common identification nodes with the connection blockage are smaller than the preset blockage number, and quitting the abnormal request data packet from the common identification network when the number of the common identification nodes with the connection blockage exceeds the preset blockage number;

the parameter setting unit is used for setting the preset time, the preset blocking quantity and the preset times.

Preferably, the system also comprises a node admission unit, a workload statistic unit and a profit distribution unit;

the parameter setting unit is used for setting waiting time;

after a self node enters a connection network, a node admission unit receives an addition application of nodes outside the connection network and performs handshake, when the handshake time is less than the preset time, an agreement instruction is sent to the connection unit, the connection unit receives the agreement instruction and broadcasts the agreement instruction to the connection network, the connection unit receives the agreement instruction of a common node except the self and sends the agreement instruction to the node admission unit, the node admission unit counts the number of the agreement instruction, when the common node in the connection network exceeds 1/2, the node admission unit receives the network access application, receives the common node interest authorization of the connection network, adds the node to the common network and distributes an interest value, the node admission unit sends an acceptance instruction, and the connection unit receives the acceptance instruction and broadcasts the acceptance instruction to a block chain network;

after the self node exits the consensus network, the node admission unit sends a joining application instruction to the connection unit after the waiting time;

the connection unit receives the joining application instruction, sends a detection data packet to the consensus node in the connection network, and sends a joining application to the connection network when receiving a response data packet of the consensus node in the connection network in the preset time; the connection unit sends the receiving instruction to the node access unit after receiving the receiving instruction, and the node access unit receives the rights and interests entrusts of the node and distributes rights and interests values;

the workload counting unit counts the workload of the self-consensus node and sends the workload to the connecting unit, and the connecting unit broadcasts the workload; the connection unit receives the workloads broadcasted by the consensus nodes except the connection unit and sends the workloads to the workload statistical unit, and the workload statistical unit counts the workloads of all the consensus nodes;

and the profit distribution unit calculates and distributes the profits of all the consensus nodes according to the workload and the equity value.

Preferably, the node admission unit further comprises a voting unit, a self right and interest delegating unit and a rights and interests delegating unit of other people;

the voting unit receives an adding application and handshake of nodes outside a connection network, sends an agreement instruction when the handshake time is less than the preset time, receives an agreement instruction of a consensus node except the voting unit, receives a network access application when the consensus node in the connection network exceeds 1/2 agreement, and sends an acceptance instruction;

the self-interest entrusting unit receives the interest entrusting of the self-node, allocates an interest value, and randomly entrusts the own interest to the consensus nodes except the self-node in the connection network after the self-node exits the consensus network;

the other-person rights and interests entrusting unit accepts the rights and interests entrustment of the nodes exiting the network when accepting the rights and interests entrustment of the nodes exiting the network, allocates rights and interests values, sends a detection data packet to the entrusting node, randomly entrusts the rights and interests entrusted by the entrusting node to the consensus nodes except the self-acquirement nodes in the connection network when the preset time is exceeded and the response data packet of the entrusting node is not received, returns the rights and interests entrusted by the entrusting node and the profits of the self-authorization node to the entrusting node when the preset time is exceeded, and allocates the profits of the entrusting node and the self-authorization node in equal proportion according to the respective rights and interests values.

Preferably, the interception unit further comprises a height interception module, a speed interception module and a total interception module;

the height interception module is used for intercepting the height of a current block in a block chain network;

the speed interception module intercepts the block speed, and when the block-out time exceeds a preset block-out time, the speed interception module sequentially sends an abnormal request data packet to the consensus nodes with the response overtime times exceeding the preset times in the recording queue;

the total number interception module intercepts the total number of the consensus nodes in the consensus network, and informs a node manager and broadcasts an abnormal alarm when the total number of the remaining consensus nodes is less than the total number of the consensus fault tolerance.

Preferably, the decision unit further comprises a judgment module and a self-checking module;

the judging module receives the abnormal request data packet, counts the handshake time of the nodes except the abnormal request data packet to determine the number of the common identification nodes which are blocked, sends a smooth data packet to the node which sends the abnormal request when the number of the common identification nodes which are blocked is less than the preset blocking number, and exits the common identification network when the number of the common identification nodes which are blocked exceeds the preset blocking number;

the self-checking module sends an abnormal request data packet which does not receive a smooth data packet when the preset self-checking time is exceeded, the handshake time of the nodes except the self-checking module is counted to determine the number of the common identification nodes which are blocked by connection, and when the number of the common identification nodes which are blocked by connection exceeds the preset blocking number, the self-checking module exits the common identification network.

The node management method and the implementation system thereof provided by the invention provide a hierarchical solution for each step and each unit, so as to ensure that the block chain network can work normally when the number of nodes is excessive, the network of partial network nodes is not good, and the low-calculation-force nodes delay the whole network and get out of blocks slowly.

Drawings

Fig. 1 is a flowchart illustrating a node management method according to an embodiment of the present invention.

Fig. 2 is a block diagram of an implementation system of the node management method of the present invention.

Detailed Description

In order to make the objects, technical solutions and advantages of the embodiments of the present invention clearer, the technical solutions in the embodiments of the present invention will be clearly and completely described below with reference to the drawings in the embodiments of the present invention, and it is obvious that the described embodiments are some, but not all, embodiments of the present invention. All other embodiments, which can be derived by a person skilled in the art from the embodiments given herein without making any creative effort, shall fall within the protection scope of the present invention.

Referring to fig. 1, to solve the problems in the prior art, the present invention provides an embodiment of a node management method, which includes the following steps.

Node connection step S1: the method comprises the steps that consensus nodes in a consensus network monitor the height of a current block, at least three consensus nodes are connected to form a connection network, each consensus node in the connection network shakes hands with the consensus nodes except the consensus nodes, the shaking hands are used for sending detection data packets to the nodes which are connected, and the nodes which receive the detection data packets send response data packets to the nodes which send the detection data packets;

timeout determining step S2: setting preset time, when one consensus node does not receive a response data packet of another consensus node after the preset time is exceeded, judging that the response of the node is overtime, recording the handshake time of each handshake by the consensus node to form a recording queue, wherein the handshake time is the interval time between the sending of a detection data packet and the receiving of a response data packet;

node screening step S3: setting a preset number of times and a preset blocking number, recording the response overtime times of the consensus nodes except the consensus nodes in the connection network by the nodes sending the detection data packets, judging the connection blocking of the nodes when the response overtime times of one node exceeds the preset number, sending an abnormal request data packet to the consensus nodes, wherein the handshake times do not exceed the preset number at the same block height;

decision step S4: the common recognition nodes receiving the abnormal request data packet count and handshake time of nodes except the abnormal request data packet to determine the number of the common recognition nodes which are blocked in connection, when the number of the common recognition nodes which are blocked in connection is smaller than the preset blocking number, the common recognition nodes which are blocked in connection send smooth data packets to the nodes which send the abnormal request, and when the number of the common recognition nodes which are blocked in connection exceeds the preset blocking number, the common recognition nodes exit the common recognition network.

In the node connection step S1, generally, all the consensus nodes form a consensus network, and all the consensus nodes in the consensus network are connected to form a connection network, at which point the consensus network and the connection network have the same range. Specifically, in a consensus network, different areas are provided, the network states have large differences, the consensus nodes and adjacent consensus nodes form a connection network, and one consensus network forms N connection networks, that is, one consensus network is divided into N connection networks.

The connection and handshake may select an existing connection-oriented internet protocol.

Corresponding to fig. 1, the present invention further provides a system for implementing a node management method, please refer to fig. 2, including a listening unit 201, a connection unit 202, a queue unit 203, a decision unit 204, and a parameter setting unit 205;

the monitoring unit 201 is configured to monitor the block height information in the block chain network and send the block height information to the connection unit;

the connection unit 202 is configured to receive the block height information of the monitoring unit 201, establish a connection with a common node other than the monitoring unit to form a connection network, periodically perform a handshake and record handshake time to form a record queue, store the record queue in the queue unit 203, determine that a response of one common node is overtime when the common node does not receive a response packet of another common node after exceeding a predetermined time, determine that the connection of the node is blocked when the number of times that the response of the one node is overtime exceeds a predetermined number of times, and send an abnormal request packet to the common node; receiving a response data packet of a node which is connected with a block and sending the response data packet to the decision unit, receiving an abnormal request data packet sent by a common node except the node, receiving a smooth instruction of the decision unit 204 and sending a smooth data packet to the node which sends the abnormal request;

the queue unit 203 is used for storing a record queue;

the decision unit 204 is configured to read the record queue in the queue unit 203, make a decision according to the record queue, receive an abnormal request packet, count handshake time of nodes other than the record queue to determine the number of common identification nodes blocked in connection, send a clear instruction to the connection unit 202 when the number of common identification nodes blocked in connection is less than a predetermined blocking number, and exit the common identification network when the number of common identification nodes blocked in connection exceeds the predetermined blocking number;

the parameter setting unit 205 is configured to set the predetermined time, the predetermined blocking number, and the predetermined number of times.

The node management method provided by the invention is mainly used for node management of the block chain network, so that the implementation system of the node management method provided by the invention is mainly implemented by each node of the block chain network based on various methods. The interception unit is for example implemented by each node of the block chain network based on an interception method. The connection unit is realized by a method for connecting each node in a block chain network, for example. The decision unit is for example implemented by a node in the blockchain network based on a decision method. The block parameter unit is implemented, for example, by each node of the blockchain network based on a method for setting parameters differently.

The consensus node refers to a node which has the authority to perform consensus in the block chain network. The block height in the blockchain network refers to the number of blocks between a block and the first block in the blockchain network. For example, if the block height of the last block that has been identified is r-1, then the block height of the last block that has been identified in the round is r. The proposed nodes constructing the candidate blocks of different block heights may be the same or different.

Preferably, in the node connection step, a network monitoring channel is established, and the detection data packet, the response data packet, the abnormal request data packet, and the clear data packet are sent through the network monitoring channel.

The method is characterized in that a connection-oriented service mode is additionally adopted outside block link point consensus, a connection network adopts connection-oriented service, namely, a communication line is established in advance when two communication parties communicate, and the procedures comprise three procedures of connection establishment, connection use and connection release. The TCP protocol is a connection service oriented protocol.

Preferably, in the node connection step and the node screening step, a consensus algorithm traversing all consensus nodes is adopted, and the detection data packet, the response data packet, the abnormal request data packet, and the unobstructed data packet are sent through a consensus broadcast channel.

In the byzantine consensus algorithms such as BFT, PBFT, VBFT, etc., because each node of the consensus protocol needs to broadcast and forward each other, similar to the method adopted by the connection network, handshaking can be performed by using the data packets forwarded by the consensus.

In some embodiments, the handshake time setting is the same as the timeout time of the ping command, the overtime data packet is recorded as the timeout, and no specific handshake time is recorded.

Preferably, in the timeout determining step and the node screening step, a predetermined number of nodes is set, the consensus node checks the response-overtime consensus nodes recorded in the queue, and sequentially sends the abnormal request packets to the consensus nodes recording that the number of response-overtime consensus nodes in the queue exceeds the predetermined number, and the number of the sent abnormal request packets does not exceed the predetermined number of nodes.

Preferably, the consensus node broadcasts the record queue to a connection network while sending an abnormal request packet, receives the consensus node of the record queue, and compares the record queue with its own record queue to determine the consensus node with the connection block.

Preferably, a predetermined block time is set, the consensus nodes in the consensus network simultaneously listen to the block speed, and when the block time exceeds the predetermined block time, the consensus nodes with the response timeout times exceeding the predetermined times in the recording queue sequentially send the abnormal request data packets.

Preferably, in the decision step, after receiving at least two abnormal request data packets, the consensus node determines the number of the connection-blocked consensus nodes according to the record queue statistics and handshake time of nodes except for the consensus node itself, sends a clear data packet to the node sending the abnormal request when the connection-blocked consensus nodes are smaller than the predetermined blocking number, and exits the consensus network when the connection-blocked consensus nodes exceed the predetermined blocking number.

Preferably, the method further comprises the following self-checking steps: setting preset self-checking time, sending abnormal request data packets to the node common identification points except the node common identification points by the common identification nodes beyond the preset self-checking time without receiving unobstructed data packets, counting handshake time of handshake with the node common identification points except the node common identification points to determine the quantity of the common identification nodes with blocked connection, sending the unobstructed data packets to the node sending the abnormal request when the common identification nodes with blocked connection are smaller than the preset blocked quantity, and quitting the common identification nodes from the common identification network when the common identification nodes with blocked connection exceed the preset blocked quantity.

Preferably, in the self-checking step, at least two abnormal request data packets sent by the consensus node do not receive a clear data packet beyond the preset self-checking time, and count handshake time of nodes except the consensus node to determine the number of the connection blocked consensus nodes, when the connection blocked consensus node is smaller than the preset blocking number, send the clear data packet to the node sending the abnormal request, and when the connection blocked consensus node exceeds the preset blocking number, exit the consensus network

In the POW and POS consensus algorithm, theoretically, the consensus fault-tolerant total number, or the consensus fault-tolerant total number is 1, and in the byzantine consensus algorithm such as BFT, PBFT, VBFT, etc., the consensus fault-tolerant total number is 2/3 of the number of the consensus network nodes, that is, 2/3 where the remaining nodes are smaller than the number of the consensus network nodes makes the blockchain network inoperable, notifies the node manager and broadcasts an alarm.

The spindle-shaped function is a function that in a positive number part, the Y value corresponding to an X value is the highest, and the Y values larger than the X value and smaller than the X value are sequentially reduced.

In some embodiments, the fusiform function is a small degree of freedom chi-squared function or a normal distribution function.

Preferably, the node accepting the delegation sends a detection data packet to the delegation node, the delegation node receiving the detection data packet sends a response data packet to the node accepting the delegation, when the node accepting the delegation does not receive the response data packet of the delegation node after the preset time, the right entrusted by the delegation node is randomly delegated to a consensus node except the consensus node in the connection network, when the node accepting the delegation receives the response data packet of the delegation node in the preset time, the right entrusted by the delegation node and the income are returned to the delegation node, and the income of the delegation node and the node accepting the delegation node is distributed in equal proportion according to the respective right value.

Preferably, a deduction coefficient is set, the deduction of the entrusted node is multiplied by the deduction coefficient to obtain deduction, the deduction is taken out from the income of the entrusted node, and the deduction is distributed to the nodes in the connection network.

Preferably, the waiting time is set, after the waiting time, the consensus node exiting the network sends the detection data packet to the consensus node in the connection network, the consensus node receiving the detection data packet sends a response data packet to the consensus node from which the detection data packet originates, and when the response data packet of the consensus node in the connection network is received in the preset time, the consensus node sends a join application to the connection network

the parameter setting unit is used for setting waiting time;

the self-checking module sends an abnormal request data packet which does not receive a smooth data packet after the preset self-checking time, counts the handshake time of the nodes except the self to determine the number of the common identification nodes which are blocked, sends the smooth data packet to the node which sends the abnormal request when the common identification nodes which are blocked are less than the preset blocking number, and exits the common identification network when the common identification nodes which are blocked are more than the preset blocking number.

Unless otherwise defined, technical or scientific terms used in the claims and the specification should have the ordinary meaning as understood by those of ordinary skill in the art to which the invention belongs.

Although the present invention has been described with reference to the preferred embodiments, it is not limited thereto. Various changes and modifications may be effected therein by one skilled in the art without departing from the spirit and scope of the invention as defined in the appended claims. It will be evident to those skilled in the art that the invention is not limited to the details of the foregoing illustrative embodiments, and that the present invention may be embodied in other specific forms without departing from the spirit or essential attributes thereof. The present embodiments are therefore to be considered in all respects as illustrative and not restrictive, the scope of the invention being indicated by the appended claims rather than by the foregoing description, and all changes which come within the meaning and range of equivalency of the claims are therefore intended to be embraced therein.

Furthermore, it should be understood that although the present description refers to embodiments, not every embodiment may contain only a single embodiment, and such description is for clarity only, and those skilled in the art should integrate the description, and the embodiments may be combined as appropriate to form other embodiments understood by those skilled in the art.

Claims

1. A node management method is characterized by comprising the following steps:

a decision making step: and the common recognition node receiving the abnormal request data packet determines the number of the common recognition nodes with connection blockage according to the record queue statistics and the handshake time of the node handshakes except the common recognition node, when the common recognition nodes with connection blockage are smaller than the preset blockage number, the common recognition node sends a smooth data packet to the node sending the abnormal request, and when the common recognition nodes with connection blockage exceed the preset blockage number, the common recognition node exits the common recognition network.

2. The node management method according to claim 1, wherein in the node connection step, a network monitoring channel is established, and the probe packet, the response packet, the abnormal request packet, and the clear packet are transmitted through the network monitoring channel.

3. The node management method according to claim 1, wherein in the node connection step and the node selection step, a consensus algorithm traversing all consensus nodes is used, and the sensing packet, the response packet, the abnormal request packet, and the clear packet are transmitted through a consensus broadcast channel.

4. The node management method according to claim 3, wherein in the node connection step and the node selection step, the probe packet and the response packet are part of a common identification packet.

5. The node management method according to claim 1, wherein in the timeout determining step, the consensus node records the handshake time of each handshake with each node, and forms a record queue by sorting the records from high to low in the handshake time, and the record that has not received the response packet is arranged at the forefront of the record queue.

6. The node management method according to claim 5, wherein in the timeout determining step, the consensus node records a handshake time of each handshake with each node, and forms a record queue by sorting records with each node from high to low according to the handshake time.

7. The node management method according to claim 5, wherein in the timeout determining step and the node screening step, a predetermined number of nodes is set, the consensus node checks the consensus nodes with response timeout recorded in the queue, and sequentially transmits the abnormal request packets to the consensus nodes with response timeout exceeding the predetermined number in the recording queue, and the number of transmitted abnormal request packets does not exceed the predetermined number of nodes.

8. The node management method according to any one of claims 1 to 7, wherein the consensus node broadcasts the record queue to the connection network while sending an abnormal request packet, receives the consensus node of the record queue, and compares the record queue with its own record queue to determine the consensus node with the connection congestion.

9. The node management method according to claim 5, wherein a predetermined block time is set, the consensus nodes in the consensus network simultaneously listen for the block speed, and when the block time exceeds the predetermined block time, the consensus nodes whose number of response times out exceeds a predetermined number in the record queue sequentially transmit the abnormal request packet.

10. The node management method according to claim 8, wherein in the decision step, after receiving at least two abnormal request data packets, the consensus node counts handshake times of nodes except for the consensus node itself according to the record queue to determine the number of the connection-blocked consensus nodes, when the connection-blocked consensus node is smaller than a predetermined blocking number, sends a clear data packet to the node which sends the abnormal request, and when the connection-blocked consensus node exceeds the predetermined blocking number, exits the consensus network.

11. The node management method according to claim 8, wherein the deciding step further comprises a self-checking step of: setting preset self-checking time, sending abnormal request data packets to the node common-recognition points except the node common-recognition points by the common-recognition points except the node common-recognition points, counting the handshake time of the node common-recognition points except the node common-recognition points, determining the number of the common-recognition points which are blocked by connection according to the record queue, and quitting the common-recognition points when the number of the common-recognition points which are blocked by connection exceeds the preset blocking number.

12. The node management method according to claim 11, wherein in the self-checking step, at least two abnormal request packets sent by the common node do not receive a clear packet beyond the predetermined self-checking time, and count handshake times of handshake with nodes other than the common node to determine the number of the common node with the blocked connection, and when the number of the common node with the blocked connection exceeds the predetermined blocking number, the common node exits the common network.

13. The node management method according to claim 9, wherein the number of the non-block-out is set, the consensus node in the consensus network simultaneously senses the block speed and records the block speed, and when the number of times of the block-out exceeding the predetermined block-out time exceeds the number of the non-block-out, the node manager is notified and an abnormal alarm is broadcast.

14. The node management method according to claim 8, wherein the deciding step further comprises a node self-control step, wherein the consensus nodes in the consensus network simultaneously monitor the total number of the consensus nodes in the consensus network, and when the total number of the remaining consensus nodes is less than the total number of the consensus faults, the node manager is notified and an abnormal alarm is broadcast.

15. The node management method according to claim 1, wherein the node in the blockchain network entrusts the equity obtaining equity value to a consensus node in the consensus network, and the consensus node obtains a profit by performing consensus and equity value.

16. The node management method according to claim 15, wherein in the deciding step, the consensus node which exits the consensus network randomly grants its own right to the consensus nodes other than itself in the connection network to become a delegating node, the consensus node which accepts the delegation has a superimposed right value, the superimposed right value is a superposition of the own right value and the authorization right value, and the consensus node which accepts the delegation obtains a profit by performing the consensus and the superimposed right value.

17. The node management method according to claim 16, wherein the consensus node calculates a profit by a profit function based on the equity value every time the consensus node performs a round of consensus, the profit function being a spindle-shaped function.

18. The node management method according to claim 16 or 17, wherein the node that accepts the delegation transmits a probe packet to the delegating node, the delegating node that receives the probe packet transmits a response packet to the node that accepts the delegation, and when the node that accepts the delegation does not receive the response packet from the delegating node for a predetermined time, randomly delegates the right granted by the delegating node to a consensus node other than itself in the connection network, and when the node that accepts the delegation receives the response packet from the delegating node for the predetermined time, returns the right and the profit delegated by the delegating node to the delegating node, and the profits of the delegating node and the node that accepts the delegation are allocated in equal proportion to the respective right values.

19. The node management method according to claim 18, wherein a deduction coefficient is set, the deduction coefficient is multiplied by the profit of the principal node to obtain a deduction profit, the deduction profit is taken out from the profit of the principal node, and the deduction profit is distributed to the nodes in the connected network.

20. The node management method according to claim 1, wherein a waiting time is set, the consensus node which exits the network sends a detection packet to the consensus node in the connection network after the waiting time, the consensus node which receives the detection packet sends a response packet to the consensus node from which the detection packet originated, and when the response packet is received from the consensus node in the connection network at the predetermined time, sends a join application to the connection network.

21. The node management method of claim 20, wherein each of the consensus nodes in the connecting network handshake with the consensus node of the exiting network, and the node of the exiting network joins the connecting network when the handshake time is less than the predetermined time and the consensus node in the connecting network joins the connecting network when the consensus node in the connecting network exceeds 1/2.

22. A system for realizing a node management method is characterized by comprising a monitoring unit, a connection unit, a queue unit, a decision unit and a parameter setting unit;

the queue unit is used for storing a record queue;

23. The system for implementing a node management method according to claim 22, further comprising a node admission unit, a workload statistics unit, and a profit distribution unit;

the parameter setting unit is used for setting waiting time;

24. The system for implementing the node management method as claimed in claim 23, wherein the node admission unit further includes a voting unit, a self right delegation unit, and a others right delegation unit;

25. The system for implementing the node management method according to claim 22, wherein the interception unit further comprises an altitude interception module, a speed interception module, and a total number interception module;

26. The system for implementing a node management method according to claim 22, wherein the decision unit further comprises a judgment module and a self-check module;