CN116055279A - Weighted voting consensus method based on network time delay - Google Patents
Weighted voting consensus method based on network time delay Download PDFInfo
- Publication number
- CN116055279A CN116055279A CN202211630098.XA CN202211630098A CN116055279A CN 116055279 A CN116055279 A CN 116055279A CN 202211630098 A CN202211630098 A CN 202211630098A CN 116055279 A CN116055279 A CN 116055279A
- Authority
- CN
- China
- Prior art keywords
- node
- master node
- network
- time
- delay
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Pending
Links
- 238000000034 method Methods 0.000 title claims abstract description 24
- 238000004891 communication Methods 0.000 claims abstract description 18
- 230000004044 response Effects 0.000 claims description 41
- 230000000977 initiatory effect Effects 0.000 claims description 14
- 230000007246 mechanism Effects 0.000 abstract description 3
- 230000001360 synchronised effect Effects 0.000 abstract description 3
- 230000006872 improvement Effects 0.000 description 10
- 230000003111 delayed effect Effects 0.000 description 2
- 238000005192 partition Methods 0.000 description 2
- 230000008569 process Effects 0.000 description 2
- 238000010586 diagram Methods 0.000 description 1
- 238000011156 evaluation Methods 0.000 description 1
- 230000003993 interaction Effects 0.000 description 1
- 230000007704 transition Effects 0.000 description 1
Images
Classifications
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04L—TRANSMISSION OF DIGITAL INFORMATION, e.g. TELEGRAPHIC COMMUNICATION
- H04L41/00—Arrangements for maintenance, administration or management of data switching networks, e.g. of packet switching networks
- H04L41/30—Decision processes by autonomous network management units using voting and bidding
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04L—TRANSMISSION OF DIGITAL INFORMATION, e.g. TELEGRAPHIC COMMUNICATION
- H04L43/00—Arrangements for monitoring or testing data switching networks
- H04L43/08—Monitoring or testing based on specific metrics, e.g. QoS, energy consumption or environmental parameters
- H04L43/0852—Delays
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04L—TRANSMISSION OF DIGITAL INFORMATION, e.g. TELEGRAPHIC COMMUNICATION
- H04L47/00—Traffic control in data switching networks
- H04L47/10—Flow control; Congestion control
- H04L47/12—Avoiding congestion; Recovering from congestion
- H04L47/122—Avoiding congestion; Recovering from congestion by diverting traffic away from congested entities
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04L—TRANSMISSION OF DIGITAL INFORMATION, e.g. TELEGRAPHIC COMMUNICATION
- H04L67/00—Network arrangements or protocols for supporting network services or applications
- H04L67/01—Protocols
- H04L67/10—Protocols in which an application is distributed across nodes in the network
- H04L67/1095—Replication or mirroring of data, e.g. scheduling or transport for data synchronisation between network nodes
Landscapes
- Engineering & Computer Science (AREA)
- Computer Networks & Wireless Communication (AREA)
- Signal Processing (AREA)
- Environmental & Geological Engineering (AREA)
- Small-Scale Networks (AREA)
Abstract
The invention belongs to the technical field of distributed cloud computing, and particularly relates to a weighted voting consensus method based on network time delay, which is used for electing a master node for coordinating the whole communication of a cluster so as to solve the problem of consistency writing of a multi-copy state machine. The conventional voting-based distributed consensus algorithm cannot consider the key factor of network delay, which affects the overall communication performance of the cluster, and if a node with poor network obtains a plurality of votes to be a master node, the consensus efficiency of the round period of the cluster can be greatly reduced. The invention discloses a dynamic weighting voting mechanism based on network time delay, wherein each node is switched among a plurality of states of a slave node, a candidate node and a master node according to a consensus protocol. The method can elect a master node for coordinating the whole communication of the cluster, ensures that the network of the master node is unobstructed and the whole communication time delay between the master node and other nodes is smaller in each round, enables the writing operation of the whole cluster to be faster and synchronous, and improves the hysteresis of the reading operation.
Description
Technical Field
The invention belongs to the technical field of distributed cloud computing, and particularly relates to a weighted voting consensus method based on network time delay, which is used for solving the consistency writing problem of a multi-copy state machine.
According to the invention, according to a network delay dynamic weighting voting mechanism, a node with the lowest overall communication delay in the cluster network topology is selected in real time as a master node so as to coordinate the final consistency writing of the multi-copy state machine and achieve data consensus.
Background
The consensus algorithm is to solve the problem of duplicate state machine consistency in a distributed system. For a distributed system, it is desirable to achieve the following characteristics:
bringing the data closer to the user geographically, thereby reducing access delay;
when part of the components fail, the system can still work continuously, so that the usability is improved;
and extends to multiple machines to provide data access services simultaneously, thereby improving read throughput.
Therefore, the design of the consensus algorithm needs to consider the usage scenario described above. For each duplicate state machine, starting with the same state and executing the same instructions in the same order, then eventually an agreed upon result can be reached. However, in this process, various errors may exist, so in order to make the replica state machine achieve final consistency in the partially synchronous network environment, the industry proposes a distributed consensus algorithm based on voting, so that each node can autonomously coordinate to complete the distributed read-write operation on the premise of adhering to the protocol of the consensus algorithm. However, the algorithm fails to consider the key factor of network delay, which affects the overall communication performance of the cluster, if a node with poor network obtains a plurality of votes to be a master node, the consensus efficiency of the round period of the cluster is greatly slowed down, network partitions are easy to appear even, and further, writing operations of local area clients are delayed and cannot be submitted, and the read information is seriously delayed.
The invention discloses a consensus method based on network delay weighted voting, which introduces a node network delay dynamic evaluation mechanism, ensures that the main node network responsible for coordination of each round is unobstructed and has smaller overall communication delay with other nodes of a cluster, so that the writing operation of the whole cluster is faster and synchronous, and the hysteresis of reading operation is improved.
Disclosure of Invention
In order to solve at least one of the technical problems, the invention provides a consensus method based on network delay weighted voting. The aim of the invention is achieved by the following technical scheme:
the invention provides a weighted voting consensus method based on network time delay, which comprises a master node, candidate nodes, slave nodes and clients, and specifically comprises the following steps:
s1, each node is in a slave node state initially, each node maintains a heartbeat timer with random time amount, and when the slave node timer is overtime, the node is switched to a candidate node state;
s2, the candidate node increases the current random number, initiates a succession type election, firstly throws a vote for the candidate node, broadcasts a voting request to other nodes through network equipment, and records the time of initiating the request;
s3, if other nodes receiving the voting request do not vote in a new period, and the latest operation code serial number of the candidate node is not lower than the latest operation code serial number of the candidate node, the candidate node responds to the instruction of voting;
s4, after the candidate node receives the response, recording the node to which the response belongs and receiving response time, and when the candidate receives more than half of tickets, the candidate node becomes a master node and calculates to obtain initial weighted network time delay;
s5, the master node periodically initiates broadcasting, informs all other nodes of the cluster, and declares the identity of the master node and the initial weighted network delay of the master node;
s6, when all the write operations of the client are sent to the slave node, the write operations are redirected to the master node, and the master node uniformly schedules the write operations. The master node attaches a serial number to the operation code to be sent, broadcasts a pre-writing request to other nodes, and records the time of initiating the request;
s7, receiving a pre-writing request from a node, setting the operation code into a pre-writing state, responding to the master node, and judging whether the network delay of the current master node is too high at the moment so as to determine whether to initiate bullet type election;
s8, after receiving the response, the master node records the node to which the response belongs and receives the response time, when the master node receives more than half of the responses, calculates new weighted network time delay, submits the corresponding operation code, then the response client informs that the operation code is submitted, and broadcasts an operation code submitting request to the slave node and the latest weighted network time delay of the master node;
s9, the slave node submits the operation code after receiving the submitting request.
As a further improvement, the heartbeat timer is a software program with a countdown function, and is regarded as overtime when the countdown is finished, and the countdown is reset to an initial value each time the slave node receives information from other candidate nodes or the master node.
As a further improvement, the period of time is a period of time from the beginning of one round of election to the beginning of the next round of election, and the period of time is a strictly increasing integer and uniquely corresponds to each round of period of time.
As a further improvement, the serial number is a strictly increasing integer, and uniquely corresponds to each operation code.
As a further improvement, in the step S4, the method for calculating the weighted network delay is as follows: the total node number is represented by n, the weighted network delay is represented by E (t), the initiating request time is represented by t0, the node to which the response belongs is represented by i, the received i node response time is represented by ti, E (t) =Σ ((ti-t 0)/2)/k, i epsilon [1, k ]; k=n/2+1 (n is an even number) or (n+1)/2 (n is an odd number). The initial weighted network time delay of the master node is E (t) 0, and in the mth round of communication, the latest weighted network time delay of the master node is E (t) m.
As a further improvement, in the step S7, the step of initiating the bullet and the election from the node includes the steps of:
s71, the current master node network delay is larger than the slave node heartbeat timing, and the follow-up communication round has higher overtime risk, so that new round of bullet and new round of election is initiated in advance, the current arbitrary number is increased, a vote is thrown for the user, voting requests are broadcast to other nodes through network equipment, and the time for initiating the requests is recorded;
s72, if other nodes receiving the voting request do not vote in a new period, and the latest operation code serial number of the node for starting the bullet is not lower than the latest operation code serial number of the node for starting the bullet, the node for starting the bullet responds to represent voting;
s73, after receiving the response, the bullet starting node records the node to which the response belongs and receives the response time, and when more than half of the candidate receives the ticket, the initial weighted network time delay can be calculated;
s74, the bullet starting node switches to the state of the main node successfully if the initial weighted network time delay of the bullet starting node is smaller than the latest weighted network time delay of the current main node, and broadcasts to inform other nodes, if the initial weighted network time delay of the bullet starting node is larger than the latest weighted network time delay of the current main node, other operations are not generated, and the current main node period is still followed.
The invention changes the direct voting of the election process in the consensus algorithm into the weighted voting based on the network time delay, and in the period of the rotation stage, not only the traditional succession mode main node rotation is supported, but also the elastic type main node rotation based on the network time delay is introduced. The network delay factor is fully considered in the election of the master node, so that the communication efficiency between the master node and other slave nodes in each period of wilting can be greatly improved, the total delay consumed by each round of communication in the cluster is reduced, and the effectiveness of cluster communication is improved. By dynamically updating the network delay parameters of the master node, the slave node can monitor the network state of the master node at any time, when the network state of the master node tends to be congested, a new master node with excellent network condition can be selected in time through the election of the elastic record in a short time, the next election is not needed to be carried out until the communication of the incumbent master node is overtime like the old consensus algorithm, and the stability of trunking communication is ensured.
Drawings
The invention will be further described with reference to the accompanying drawings, in which embodiments do not constitute any limitation of the invention, and other drawings can be obtained by one of ordinary skill in the art without inventive effort from the following drawings. FIG. 1 is a node state transition diagram of the present invention; FIG. 2 is a master election flow chart of the present invention; FIG. 3 is a flow chart of the distributed data read/write method of the present invention.
Detailed Description
In order to better understand the technical solutions of the present invention, the following description will be made in detail with reference to the accompanying drawings and specific embodiments, and it should be noted that, without conflict, the embodiments of the present application and features in the embodiments may be combined with each other.
Referring to fig. 1, the state machines of all copies in the cluster according to the present invention respectively appear in the form of slave nodes, candidate nodes and master nodes, and can mutually switch when different conditions are satisfied, and the specific steps are as follows:
s1, starting a state, wherein the node automatically enters a slave node state when being started.
S2, after the slave node is started, a timer with overtime election is started, and when the timer expires, the slave node is switched to the candidate node state to initiate election.
S3, after entering the candidate node state, election starts, but if a new master node is not selected before the next election timeout arrives, the new election still keeps being restarted in the candidate node state.
And S4, when the nodes in the candidate state receive more than half of node votes through the succession election or receive more than half of node votes through the bullet election and the weighted network delay is smaller, the switching state is changed into a new main node.
S5, if the node in the candidate state receives the message from the master node or the message with the higher arbitrary number, the node in the candidate state indicates that the master node is already in the slave node state, and the node is switched back to the slave node state.
S6, if the message from the higher tenn is received in the master node state, switching to the slave node state. Most of this occurs in the presence of network partitions.
The invention provides a weighted voting consensus method based on network time delay, which is shown in combination with fig. 2, and comprises a master node, candidate nodes and slave nodes, wherein the method comprises the following specific steps:
s1, each node is in a slave node state initially, each node maintains a heartbeat timer with random time amount, and when the slave node timer is overtime, the node is switched to a candidate node state;
s2, the candidate node increases the current random number, initiates a succession type election, firstly throws a vote for the candidate node, broadcasts a voting request to other nodes through network equipment, and records the time of initiating the request;
s3, if other nodes receiving the voting request do not vote in a new period, and the latest operation code serial number of the candidate node is not lower than the latest operation code serial number of the candidate node, the candidate node responds to the instruction of voting;
s4, after the candidate node receives the response, recording the node to which the response belongs and receiving response time, and when the candidate receives more than half of tickets, the candidate node becomes a master node and calculates to obtain initial weighted network time delay;
s5, the master node periodically initiates broadcasting, informs all other nodes of the cluster, and declares the identity of the master node and the initial weighted network delay of the master node;
s6, when all the write operations of the client are sent to the slave node, the write operations are redirected to the master node, and the master node uniformly schedules the write operations. The master node attaches a serial number to the operation code to be sent, broadcasts a pre-writing request to other nodes, and records the time of initiating the request;
s7, receiving a pre-writing request from a node, setting the operation code into a pre-writing state, responding to the master node, and judging whether the network delay of the current master node is too high at the moment so as to determine whether to initiate bullet type election;
s8, after receiving the response, the master node records the node to which the response belongs and receives the response time, when the master node receives more than half of the responses, calculates new weighted network time delay, submits the corresponding operation code, then the response client informs that the operation code is submitted, and broadcasts an operation code submitting request to the slave node and the latest weighted network time delay of the master node;
s9, the slave node submits the operation code after receiving the submitting request.
As a further improvement, the heartbeat timer is a software program with a countdown function, and is regarded as overtime when the countdown is finished, and the countdown is reset to an initial value each time the slave node receives information from other candidate nodes or the master node.
As a further improvement, the period of time is a period of time from the beginning of one round of election to the beginning of the next round of election, and the period of time is a strictly increasing integer and uniquely corresponds to each round of period of time.
As a further improvement, the serial number is a strictly increasing integer, and uniquely corresponds to each operation code.
As a further improvement, in the step S4, the method for calculating the weighted network delay is as follows: the total node number is represented by n, the weighted network delay is represented by E (t), the initiating request time is represented by t0, the node to which the response belongs is represented by i, the received i node response time is represented by ti, E (t) =Σ ((ti-t 0)/2)/k, i epsilon (1, k); k=n/2+1 (n is an even number) or (n+1)/2 (n is an odd number). The initial weighted network time delay of the master node is E (t) 0, and in the mth round of communication, the latest weighted network time delay of the master node is E (t) m.
As a further improvement, in the step S7, the step of initiating the bullet and the election from the node includes the steps of:
s71, the current master node network delay is larger than the slave node heartbeat timing, and the follow-up communication round has higher overtime risk, so that new round of bullet and new round of election is initiated in advance, the current arbitrary number is increased, a vote is thrown for the user, voting requests are broadcast to other nodes through network equipment, and the time for initiating the requests is recorded;
s72, if other nodes receiving the voting request do not vote in a new period, and the latest operation code serial number of the node for starting the bullet is not lower than the latest operation code serial number of the node for starting the bullet, the node for starting the bullet responds to represent voting;
s73, after receiving the response, the bullet starting node records the node to which the response belongs and receives the response time, and when more than half of the candidate receives the ticket, the initial weighted network time delay can be calculated;
s74, the bullet starting node switches to the state of the main node successfully if the initial weighted network time delay of the bullet starting node is smaller than the latest weighted network time delay of the current main node, and broadcasts to inform other nodes, if the initial weighted network time delay of the bullet starting node is larger than the latest weighted network time delay of the current main node, other operations are not generated, and the current main node period is still followed.
With reference to fig. 3, the weighted voting consensus method based on network delay provided by the invention can ensure the final consistency of data through the read-write interaction with the client, and specifically comprises the following steps:
s1, a client accesses a near node in a cluster and initiates a write operation request;
s2, when the slave node receives a write request from the client, redirecting the write request to the master node;
s3, the master node initiates a write-ready request, and broadcasts an operation code to other nodes of the cluster;
s4, when the slave node receives a write-ready request from the master node, the operation code is put into a write-ready state and responds to the master node;
s5, when the master node receives more than half of responses for completing preparation for writing, the operation code can be executed, and the response client side informs that the operation code is submitted;
s6, the master node initiates an operation code executing request to other nodes of the cluster;
and S7, executing the operation code when the slave node receives the operation code executing request from the slave node.
Claims (7)
1. The weighted voting consensus method based on the network time delay is characterized by comprising the following specific steps of:
s1, each node is in a slave node state initially, each node maintains a heartbeat timer with random time amount, and when the slave node timer is overtime, the node is switched to a candidate node state;
s2, the candidate node increases the current random number, initiates a succession type election, firstly throws a vote for the candidate node, broadcasts a voting request to other nodes through network equipment, and records the time of initiating the request;
s3, if other nodes receiving the voting request do not vote in a new period, and the latest operation code serial number of the candidate node is not lower than the latest operation code serial number of the candidate node, the candidate node responds to the instruction of voting;
s4, after the candidate node receives the response, recording the node to which the response belongs and receiving response time, and when the candidate receives more than half of tickets, the candidate node becomes a master node and calculates to obtain initial weighted network time delay;
s5, the master node periodically initiates broadcasting, informs all other nodes of the cluster, and declares the identity of the master node and the initial weighted network delay of the master node;
s6, when all the write operations of the client are sent to the slave node, the write operations are redirected to the master node, and the master node uniformly schedules the write operations. The master node attaches a serial number to the operation code to be sent, broadcasts a pre-writing request to other nodes, and records the time of initiating the request;
s7, receiving a pre-writing request from a node, setting the operation code into a pre-writing state, responding to the master node, and judging whether the network delay of the current master node is too high at the moment so as to determine whether to initiate bullet type election;
s8, after receiving the response, the master node records the node to which the response belongs and receives the response time, when the master node receives more than half of the responses, calculates new weighted network time delay, submits the corresponding operation code, then the response client informs that the operation code is submitted, and broadcasts an operation code submitting request to the slave node and the latest weighted network time delay of the master node;
s9, the slave node submits the operation code after receiving the submitting request.
2. The method of claim 1, wherein the heartbeat timer is a software program with a countdown function, and the countdown is reset to an initial value each time a slave node receives information from another candidate node or a master node when the countdown is deemed to be a timeout.
3. The method of claim 1, wherein the period of time is a period of time from a start of a round of election to a start of a next round of election, and the period of time is a strictly increasing integer, and the period of time corresponds to each round of period of time.
4. The network latency based weighted voting consensus method of claim 1, wherein the relay election is initiated when a slave node heartbeat timer expires and information from a master node has not been received.
5. The network latency based weighted voting consensus method of claim 1 wherein the sequence number is a strictly increasing integer uniquely corresponding to each opcode.
6. The method of weighting voting consensus based on network delay according to claim 1, wherein in the step S4, the method of calculating the weighted network delay is: the total node number is represented by n, the weighted network delay is represented by E (t), the initiating request time is represented by t0, the node to which the response belongs is represented by i, the received i node response time is represented by ti, E (t) =Σ ((ti-t 0)/2)/k, i epsilon [1, k ]; k=n/2+1 (n is even) or (n+1)/2 (n is odd), the initial weighted network delay of the master node is E (t) 0, and in the mth round of communication, the latest weighted network delay of the master node is E (t) m.
7. The method of weighted voting consensus based on network delay according to claim 1, wherein in step S7, the bullet from the node is launched, comprising the steps of:
s71, the current master node network delay is larger than the slave node heartbeat timing, and the follow-up communication round has higher overtime risk, so that new round of bullet and new round of election is initiated in advance, the current arbitrary number is increased, a vote is thrown for the user, voting requests are broadcast to other nodes through network equipment, and the time for initiating the requests is recorded;
s72, if other nodes receiving the voting request do not vote in a new period, and the latest operation code serial number of the node for starting the bullet is not lower than the latest operation code serial number of the node for starting the bullet, the node for starting the bullet responds to represent voting;
s73, after receiving the response, the bullet starting node records the node to which the response belongs and receives the response time, and when more than half of the candidate receives the ticket, the initial weighted network time delay can be calculated;
s74, the bullet starting node switches to the state of the main node successfully if the initial weighted network time delay of the bullet starting node is smaller than the latest weighted network time delay of the current main node, and broadcasts to inform other nodes, if the initial weighted network time delay of the bullet starting node is larger than the latest weighted network time delay of the current main node, other operations are not generated, and the current main node period is still followed.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN202211630098.XA CN116055279A (en) | 2022-12-19 | 2022-12-19 | Weighted voting consensus method based on network time delay |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN202211630098.XA CN116055279A (en) | 2022-12-19 | 2022-12-19 | Weighted voting consensus method based on network time delay |
Publications (1)
Publication Number | Publication Date |
---|---|
CN116055279A true CN116055279A (en) | 2023-05-02 |
Family
ID=86120821
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN202211630098.XA Pending CN116055279A (en) | 2022-12-19 | 2022-12-19 | Weighted voting consensus method based on network time delay |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN116055279A (en) |
Citations (4)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
US20200028750A1 (en) * | 2018-07-17 | 2020-01-23 | Software Ag | System and/or method for maintaining highly-available, consistent, partition-tolerant clusters using client voters |
CN113194146A (en) * | 2021-05-06 | 2021-07-30 | 杭州复杂美科技有限公司 | Leader node determination method, computer device, and storage medium |
US11100743B1 (en) * | 2017-12-30 | 2021-08-24 | S&S Crypto Technologies | Blockchain-based election system |
CN115102967A (en) * | 2022-06-15 | 2022-09-23 | 西安工业大学 | Consensus method with high consensus efficiency and distributed system |
-
2022
- 2022-12-19 CN CN202211630098.XA patent/CN116055279A/en active Pending
Patent Citations (4)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
US11100743B1 (en) * | 2017-12-30 | 2021-08-24 | S&S Crypto Technologies | Blockchain-based election system |
US20200028750A1 (en) * | 2018-07-17 | 2020-01-23 | Software Ag | System and/or method for maintaining highly-available, consistent, partition-tolerant clusters using client voters |
CN113194146A (en) * | 2021-05-06 | 2021-07-30 | 杭州复杂美科技有限公司 | Leader node determination method, computer device, and storage medium |
CN115102967A (en) * | 2022-06-15 | 2022-09-23 | 西安工业大学 | Consensus method with high consensus efficiency and distributed system |
Non-Patent Citations (4)
Title |
---|
S. SINGH: "Electing leaders based upon performance: the delay model", 《11TH INTERNATIONAL CONFERENCE ON DISTRIBUTED COMPUTING SYSTEMS》, 6 August 2002 (2002-08-06) * |
刘克猛: "面向区块链的实用拜占庭容错算法优化研究", 《中国优秀硕士学位论文全文数据库 信息科技辑》, 15 January 2022 (2022-01-15) * |
刘易: "基于地理信息的分簇式路由协议", 《湖南科技大学学报(自然科学版)》, 25 December 2012 (2012-12-25) * |
王海勇;郭凯璇;潘启青;: "基于投票机制的拜占庭容错共识算法", 计算机应用, no. 06, 29 January 2019 (2019-01-29) * |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
CN110677485B (en) | Dynamic layered Byzantine fault-tolerant consensus method based on credit | |
US5261085A (en) | Fault-tolerant system and method for implementing a distributed state machine | |
US7711825B2 (en) | Simplified Paxos | |
US11102084B2 (en) | Fault rectification method, device, and system | |
US11271814B2 (en) | Online capacity-expanding and online capacity-reducing methods and apparatuses for distributed consensus system | |
US8005888B2 (en) | Conflict fast consensus | |
KR20170134576A (en) | Method and apparatus for selecting time-based nodes | |
CN110505319A (en) | A kind of RS485 is from device address auto-allocation method and system | |
CA2393852C (en) | Data sharing method, terminal, and medium on which program is recorded | |
CN108829720B (en) | Data processing method and device | |
CN110855737B (en) | Consistency level controllable self-adaptive data synchronization method and system | |
CN112799867B (en) | Election method based on distributed storage system and distributed storage system | |
CN114461438A (en) | Distributed database disaster recovery system and method of asymmetric center mode | |
CN110635941A (en) | Database node cluster fault migration method and device | |
CN110417882B (en) | Method and device for determining main node and storage medium | |
CN114726867B (en) | Hot standby multi-main method based on Lift | |
CN113486042B (en) | Data processing method, device, computer readable medium and electronic equipment | |
CN109921897B (en) | Triggering method and device for workload certification calculation, calculating equipment and storage medium | |
CN113440863B (en) | Processing method and device for multiplayer online game, electronic equipment and storage medium | |
CN113220421B (en) | Management method, management server and management system of server cluster | |
CN116055279A (en) | Weighted voting consensus method based on network time delay | |
CN113204424A (en) | Method and device for optimizing Raft cluster and storage medium | |
US10848549B1 (en) | Leaderless, parallel, and topology-aware protocol for achieving consensus | |
CN116232893A (en) | Consensus method and device of distributed system, electronic equipment and storage medium | |
US7260611B2 (en) | Multi-leader distributed system |
Legal Events
Date | Code | Title | Description |
---|---|---|---|
PB01 | Publication | ||
PB01 | Publication | ||
SE01 | Entry into force of request for substantive examination | ||
SE01 | Entry into force of request for substantive examination |