CN111130879A - PBFT algorithm-based cluster exception recovery method - Google Patents
PBFT algorithm-based cluster exception recovery method Download PDFInfo
- Publication number
- CN111130879A CN111130879A CN201911349305.2A CN201911349305A CN111130879A CN 111130879 A CN111130879 A CN 111130879A CN 201911349305 A CN201911349305 A CN 201911349305A CN 111130879 A CN111130879 A CN 111130879A
- Authority
- CN
- China
- Prior art keywords
- view
- node
- nodes
- recovery
- cluster
- 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.)
- Granted
Links
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
- H04L41/06—Management of faults, events, alarms or notifications
- H04L41/0654—Management of faults, events, alarms or notifications using network fault recovery
- H04L41/0668—Management of faults, events, alarms or notifications using network fault recovery by dynamic selection of recovery network elements, e.g. replacement by the most appropriate element after failure
-
- 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
- H04L41/06—Management of faults, events, alarms or notifications
- H04L41/0654—Management of faults, events, alarms or notifications using network fault recovery
-
- 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
Abstract
The invention discloses a PBFT algorithm-based cluster anomaly recovery method. In a cluster network composed of 3f +1 nodes, the PBFT algorithm can tolerate the occurrence of an abnormality of only a maximum of f nodes at the same time, but in an actual deployment network, the entire cluster often enters an abnormal state due to network delay, network disconnection, node restart, or the like. And the abnormal node tries to initiate a view change request, before the cluster network is recovered, the view change request is finally overtime, at the moment, the abnormal node enters a state to be recovered to initiate a recovery request, until the cluster network is recovered, more than 2f +1 nodes to be recovered can select a consistent view value, a new main node is determined, and the abnormal recovery of the cluster can be completed by waiting for a new view starting message of the main node. The method solves the problem that the PBFT algorithm cannot be automatically recovered after cluster failure, and greatly improves the usability of the PBFT algorithm.
Description
Technical Field
The invention relates to a PBFT algorithm, in particular to a cluster exception recovery method based on the PBFT algorithm.
Background
PBFT is an abbreviation for Practical Byzantine Fault Tolerance, namely: a practical byzantine fault-tolerant algorithm. The algorithm is proposed by Miguel Castro and Barbara Liskov in 1999, solves the problems of feasible theory and low actual efficiency on the basis of the original Byzantine fault-tolerant algorithm, really realizes the Byzantine attack problem in a distributed asynchronous consensus system, and greatly promotes the application of the Byzantine fault-tolerant algorithm in the actual system. To be able to tolerate the presence of f byzantine nodes, PBFT requires at least 3f +1 nodes for the entire network.
The PBFT is actually a state machine replica algorithm that can solve the byzantine problem, i.e., each node is a state machine whose system guarantees service state consistency, and the state service consistency of the state machines in the whole network enables the state machines to implement consistent operation.
The PBFT algorithm has two major protocols:
(1) and the three-stage consensus protocol ensures that the block packed by the main node completes the consistency confirmation of the packed result through the three-stage consensus and ensures the system activity of coping with Byzantine attack.
(2) The view change protocol ensures that the system can select a new host node for consensus through the protocol under the condition that the host node is malicious, and the original consensus state is not destroyed.
Meanwhile, in order to ensure that the confirmed consistency state of each node is persistent and the garbage capacity in the memory is kept within a certain limit, the PBFT also designs a set of check point mechanism.
However, the current PBFT algorithm has the following problems when applied to practical scenes: when cluster nodes need to be down and restarted or a cluster network is seriously paralyzed and partitioned, the PBFT algorithm cannot timely complete the self-recovery of the cluster, so that a method for recovering cluster abnormality needs to be designed.
Disclosure of Invention
Aiming at the defects of the prior art, the invention provides a PBFT algorithm-based cluster abnormity recovery method, which can rapidly recover a cluster under the condition of cluster restart or cluster network paralysis, and the specific technical scheme is as follows:
a PBFT algorithm-based cluster exception recovery method, the cluster is composed of 3f +1 nodes, wherein the PBFT algorithm can tolerate Byzantine errors of f nodes at most; the method comprises the following steps:
(1) the cluster enters an abnormal state due to restarting or network failure and the like, and all nodes of the cluster become abnormal nodes;
(2) each abnormal node sends a view change request to all nodes in the whole network, then waits for the view change requests of other nodes within a specified time and counts, and if 2f +1 view change requests containing the same view value are received within the specified time, the recovery is completed by a view change recovery method in a PBFT algorithm; if 2f +1 view change requests containing the same view value cannot be received within the specified time, the view change is overtime; the view change request includes the ID of the node, the target view value of the node, and the like.
(3) The abnormal node with the overtime view change request immediately enters a state to be recovered to become a node to be recovered;
(4) the node to be recovered sends a recovery request to all nodes in the whole network, wherein the recovery request comprises the ID of the node and the target view value of the node;
(5) the node to be recovered waits for the recovery requests of other nodes within the specified time and counts, and if 2f +1 recovery requests containing the same view value are received within the specified time, the step (6) is carried out to confirm the new view; if more than f +1 contained view values are received within the specified time and are larger than the target view value of the node, selecting the minimum view value as the target view value of the node to be restored and re-entering the step (4); if 2f +1 recovery requests containing the same view value can not be received within the specified time, returning to the step (4), and retransmitting the recovery request once, wherein the target view value is unchanged;
(6) and constructing a new view by using the target view value, entering the new view by all nodes to be restored which receive 2f +1 restoration requests containing the same view value in a specified time, determining a new main node according to the view value of the new view, and taking the rest nodes as slave nodes. The main node in the new view completes the abnormal recovery through a view change recovery method in a PBFT algorithm and constructs a new view starting message for whole network broadcasting according to the received 2f +1 recovery requests; the slave nodes in the new view wait for the new view starting message of the master node within a specified time;
(7) if the slave node in the new view receives a new view starting message of the master node within the specified time, the step (8) is carried out to complete the view change; if the new view starting message of the master node cannot be received within the specified time, adding 1 to the target view value of the master node, and re-entering the step (4);
(8) the slave node in the new view completes abnormal recovery by a view change recovery method in a PBFT algorithm, and updates the view value of the slave node to be a target view value; and recovering the whole cluster to work normally until all the abnormal nodes recover.
Further, the view change request in the step (2) and the restoration request in the step (4) further include digital signature information of the node.
Further, the counting rule in the step (2) and the step (5) is that only one ticket can be counted by one node ID, and only one ticket can be counted by the node if recovery requests from the same node are received for multiple times.
Further, the new view start message in step (6) may include the aggregated signatures of 2f +1 nodes.
Further, in the step (6), it is determined that a new master node determines a next master node in a polling manner, and when polling is performed, the view value is incremented by 1 each time, and the master node IDs change in turn.
The invention has the following beneficial effects: the method of the invention is based on PBFT algorithm, a cluster abnormal recovery flow is newly added, in the actual PBFT deployment network, the cluster restarting or cluster network paralysis often occurs, under the condition, a normal node can carry out abnormal state and firstly tries the view change method of PBFT, the view change often times out and can not be completed because the network paralysis can be recovered within a certain time, the method of the invention enables the node to enter the abnormal recovery flow at the moment, and continuously sends abnormal recovery information until more than 2f +1 nodes send recovery information containing the same view value together when the network is recovered, at the moment, the abnormal recovery of the cluster can be completed by selecting a new node. The PBFT algorithm-based cluster anomaly recovery method provided by the invention improves the problem that the recovery cannot be realized after the cluster is paralyzed in the classical PBFT algorithm, greatly improves the recovery efficiency after the cluster is anomalous, and improves the availability of the PBFT algorithm.
Drawings
Fig. 1 is a schematic flow chart of cluster exception recovery.
Detailed Description
The present invention will be described in detail below with reference to the accompanying drawings and preferred embodiments, and the objects and effects of the present invention will become more apparent, and the present invention will be further described in detail below with reference to the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are merely illustrative of the invention and are not intended to limit the invention.
The invention provides a PBFT algorithm-based cluster anomaly recovery method. In a cluster network composed of 3f +1 nodes, the PBFT algorithm can tolerate the occurrence of an abnormality of only a maximum of f nodes at the same time, but in an actual deployment network, the entire cluster often enters an abnormal state due to network delay, network disconnection, node restart, or the like. The nodes entering the exception will first try to initiate a view change request, however, before the cluster network recovers, the view change request will eventually time out, resulting in the cluster failing to recover autonomously. And the abnormal node enters a state to be recovered to initiate a recovery request until the cluster network is recovered, more than 2f +1 nodes to be recovered can select a consistent view value, a new main node is determined, and the abnormal recovery of the cluster can be completed by waiting for a new view starting message of the main node. The method solves the problem that the PBFT algorithm cannot be automatically recovered after cluster failure, and greatly improves the usability of the PBFT algorithm.
The specific implementation flow of the method is as follows:
(1) the cluster enters an abnormal state due to restarting or network failure and the like, and all nodes of the cluster become abnormal nodes;
(2) each abnormal node sends a view change request to all nodes in the whole network, then waits for the view change requests of other nodes within a specified time and counts, and if 2f +1 view change requests containing the same view value are received within the specified time, the recovery is completed by a view change recovery method in a PBFT algorithm; before the cluster network is recovered, 2f +1 view change requests containing the same view value cannot be received within a specified time, so that the view change is overtime; the view change request includes the ID of the node, the target view value of the node, and the like.
(3) The abnormal node with the overtime view change request immediately enters a state to be recovered to become a node to be recovered;
(4) the node to be recovered sends a recovery request to all nodes in the whole network, wherein the recovery request comprises an ID of the node, a target view value of the node and the like;
(5) the node to be recovered waits for the recovery requests of other nodes within the specified time and counts, and if 2f +1 recovery requests containing the same view value are received within the specified time, the step (6) is carried out to confirm the new view; if more than f +1 contained view values are received within the specified time and are larger than the target view value of the node, selecting the minimum view value as the target view value of the node to be restored and re-entering the step (4); if 2f +1 recovery requests containing the same view value cannot be received within a specified time due to cluster restarting or cluster network paralysis and the like, returning to the step (4), and resending the recovery request once, wherein the target view value is unchanged;
(6) and constructing a new view by using the target view value, entering the new view by all nodes to be restored which receive 2f +1 restoration requests containing the same view value in a specified time, determining a new main node according to the view value of the new view, and taking the rest nodes as slave nodes. The flow of recovering the anomaly of the master node and the slave node in the new view is shown in fig. 1, wherein the master node in the new view completes the anomaly recovery by a view change recovery method in a PBFT algorithm and constructs a new view start message for broadcasting over the whole network according to the received 2f +1 recovery requests; the slave nodes in the new view wait for the new view starting message of the master node within a specified time;
(7) if the slave node in the new view receives a new view starting message of the master node within the specified time, the step (8) is carried out to complete the view change; if the new view starting message of the master node cannot be received within the specified time, adding 1 to the target view value of the master node, and re-entering the step (4);
(8) the slave node in the new view completes abnormal recovery by a view change recovery method in a PBFT algorithm, and updates the view value of the slave node to be a target view value; and recovering the whole cluster to work normally until all the abnormal nodes recover.
Preferably, the view change request in the step (2) and the restoration request in the step (4) each further include digital signature information of a node.
The counting rule in the step (2) and the step (5) is that only one ticket can be counted by one node ID, and only one ticket can be counted by the node if recovery requests from the same node are received for multiple times. Avoiding repeated counting of tickets.
The new view start message in step (6) will contain the aggregated signature of 2f +1 nodes. Aggregated signatures can greatly reduce the size of the message body that needs to be transmitted, compared to simply adding the signatures of 2f +1 nodes together.
In the step (6), determining a new master node and determining the next master node according to a polling mode, wherein the ID of the master node is view mod N, the view is a view value, and N is the number of nodes; in polling, the view value is incremented by 1 each time, and the master node IDs change in turn.
It will be understood by those skilled in the art that the foregoing is only a preferred embodiment of the present invention, and is not intended to limit the invention, and although the invention has been described in detail with reference to the foregoing examples, it will be apparent to those skilled in the art that various changes in the form and details of the embodiments may be made and equivalents may be substituted for elements thereof. All modifications, equivalents and the like which come within the spirit and principle of the invention are intended to be included within the scope of the invention.
Claims (5)
1. A PBFT algorithm-based cluster exception recovery method, the cluster is composed of 3f +1 nodes, wherein the PBFT algorithm can tolerate Byzantine errors of f nodes at most; the method is characterized by comprising the following steps:
(1) the cluster enters an abnormal state due to restarting or network failure and the like, and all nodes of the cluster become abnormal nodes;
(2) each abnormal node sends a view change request to all nodes in the whole network, then waits for the view change requests of other nodes within a specified time and counts, and if 2f +1 view change requests containing the same view value are received within the specified time, the recovery is completed by a view change recovery method in a PBFT algorithm; if 2f +1 view change requests containing the same view value cannot be received within the specified time, the view change is overtime; the view change request includes the ID of the node, the target view value of the node, and the like.
(3) The abnormal node with the overtime view change request immediately enters a state to be recovered to become a node to be recovered;
(4) the node to be recovered sends a recovery request to all nodes in the whole network, wherein the recovery request comprises the ID of the node and the target view value of the node;
(5) the node to be recovered waits for the recovery requests of other nodes within the specified time and counts, and if 2f +1 recovery requests containing the same view value are received within the specified time, the step (6) is carried out to confirm the new view; if more than f +1 contained view values are received within the specified time and are larger than the target view value of the node, selecting the minimum view value as the target view value of the node to be restored and re-entering the step (4); if 2f +1 recovery requests containing the same view value can not be received within the specified time, returning to the step (4), and retransmitting the recovery request once, wherein the target view value is unchanged;
(6) and constructing a new view by using the target view value, entering the new view by all nodes to be restored which receive 2f +1 restoration requests containing the same view value in a specified time, determining a new main node according to the view value of the new view, and taking the rest nodes as slave nodes. The main node in the new view completes the abnormal recovery through a view change recovery method in a PBFT algorithm and constructs a new view starting message for whole network broadcasting according to the received 2f +1 recovery requests; the slave nodes in the new view wait for the new view starting message of the master node within a specified time;
(7) if the slave node in the new view receives a new view starting message of the master node within the specified time, the step (8) is carried out to complete the view change; if the new view starting message of the master node cannot be received within the specified time, adding 1 to the target view value of the master node, and re-entering the step (4);
(8) the slave node in the new view completes abnormal recovery by a view change recovery method in a PBFT algorithm, and updates the view value of the slave node to be a target view value; and recovering the whole cluster to work normally until all the abnormal nodes recover.
2. The PBFT algorithm-based cluster exception recovery method of claim 1, wherein the view change request in step (2) and the recovery request in step (4) further comprise digital signature information of nodes.
3. The PBFT algorithm-based cluster anomaly recovery method according to claim 1, wherein the counting rule in the step (2) and the step (5) is that only one ticket can be counted for one node ID, and only one ticket can be counted for the node if recovery requests from the same node are received for multiple times.
4. The PBFT algorithm-based cluster exception recovery method as claimed in claim 1, wherein the new view start message in step S6 includes an aggregate signature of 2f +1 nodes.
5. The PBFT algorithm-based cluster exception recovery method as claimed in claim 1, wherein in step (6), a new master node is determined to be the next master node in a polling manner, and when polling is performed, the view value is incremented by 1 each time, and the master node IDs are sequentially changed in turn.
Priority Applications (2)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201911349305.2A CN111130879B (en) | 2019-12-24 | 2019-12-24 | PBFT algorithm-based cluster exception recovery method |
| CN202211201652.2A CN115632933A (en) | 2019-12-24 | 2019-12-24 | PBFT algorithm-based cluster exception recovery method |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201911349305.2A CN111130879B (en) | 2019-12-24 | 2019-12-24 | PBFT algorithm-based cluster exception recovery method |
Related Child Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202211201652.2A Division CN115632933A (en) | 2019-12-24 | 2019-12-24 | PBFT algorithm-based cluster exception recovery method |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN111130879A true CN111130879A (en) | 2020-05-08 |
| CN111130879B CN111130879B (en) | 2022-11-18 |
Family
ID=70500071
Family Applications (2)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201911349305.2A Active CN111130879B (en) | 2019-12-24 | 2019-12-24 | PBFT algorithm-based cluster exception recovery method |
| CN202211201652.2A Pending CN115632933A (en) | 2019-12-24 | 2019-12-24 | PBFT algorithm-based cluster exception recovery method |
Family Applications After (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202211201652.2A Pending CN115632933A (en) | 2019-12-24 | 2019-12-24 | PBFT algorithm-based cluster exception recovery method |
Country Status (1)
| Country | Link |
|---|---|
| CN (2) | CN111130879B (en) |
Cited By (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| WO2021068919A1 (en) * | 2019-10-10 | 2021-04-15 | 杭州趣链科技有限公司 | Pbft algorithm-based improved method for active recovery of single node from anomaly |
| CN112929354A (en) * | 2021-01-28 | 2021-06-08 | 恒宝股份有限公司 | Practical Byzantine fault-tolerant anti-attack deadlock method and device |
| EP3933640A1 (en) * | 2020-07-03 | 2022-01-05 | Alipay (Hangzhou) Information Technology Co., Ltd. | Consensus node changing method and related apparatus based on honey badger byzantine fault tolerance consensus |
| US11251967B2 (en) * | 2020-07-03 | 2022-02-15 | Alipay (Hangzhou) Information Technology Co., Ltd. | Consensus methods and systems in consortium blockchain |
| CN116915796A (en) * | 2023-09-14 | 2023-10-20 | 杭州趣链科技有限公司 | Autonomous recovery method and device after cluster view bifurcation and electronic equipment |
Families Citing this family (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN116991623B (en) * | 2023-08-30 | 2024-01-02 | 杭州趣链科技有限公司 | Block chain node exception recovery method and device, electronic equipment and storage medium |
Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN108182635A (en) * | 2017-12-18 | 2018-06-19 | 深圳前海微众银行股份有限公司 | Block chain common recognition method, system and computer readable storage medium |
| WO2018161901A1 (en) * | 2017-03-10 | 2018-09-13 | 阿里巴巴集团控股有限公司 | Consensus method and device |
| WO2019170169A2 (en) * | 2019-06-05 | 2019-09-12 | Alibaba Group Holding Limited | Consensus system and method |
| CN110460484A (en) * | 2019-10-10 | 2019-11-15 | 杭州趣链科技有限公司 | A kind of single node exception active restoration methods based on PBFT algorithm improvement |
| CN110493198A (en) * | 2019-07-26 | 2019-11-22 | 北京工业大学 | A method of it is attacked based on Sybil in PBFT algorithm defence block chain is improved |
-
2019
- 2019-12-24 CN CN201911349305.2A patent/CN111130879B/en active Active
- 2019-12-24 CN CN202211201652.2A patent/CN115632933A/en active Pending
Patent Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| WO2018161901A1 (en) * | 2017-03-10 | 2018-09-13 | 阿里巴巴集团控股有限公司 | Consensus method and device |
| CN108182635A (en) * | 2017-12-18 | 2018-06-19 | 深圳前海微众银行股份有限公司 | Block chain common recognition method, system and computer readable storage medium |
| WO2019170169A2 (en) * | 2019-06-05 | 2019-09-12 | Alibaba Group Holding Limited | Consensus system and method |
| CN110493198A (en) * | 2019-07-26 | 2019-11-22 | 北京工业大学 | A method of it is attacked based on Sybil in PBFT algorithm defence block chain is improved |
| CN110460484A (en) * | 2019-10-10 | 2019-11-15 | 杭州趣链科技有限公司 | A kind of single node exception active restoration methods based on PBFT algorithm improvement |
Non-Patent Citations (1)
| Title |
|---|
| 雷长剑等: "志愿云环境下的拜占庭容错研究", 《计算机工程》 * |
Cited By (9)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| WO2021068919A1 (en) * | 2019-10-10 | 2021-04-15 | 杭州趣链科技有限公司 | Pbft algorithm-based improved method for active recovery of single node from anomaly |
| US11841778B2 (en) | 2019-10-10 | 2023-12-12 | Hangzhou Qulian Technology Co., Ltd. | Method and system for active failure recovery of single node improved based on PBFT algorithm, computer device and storage medium |
| EP3933640A1 (en) * | 2020-07-03 | 2022-01-05 | Alipay (Hangzhou) Information Technology Co., Ltd. | Consensus node changing method and related apparatus based on honey badger byzantine fault tolerance consensus |
| US11251967B2 (en) * | 2020-07-03 | 2022-02-15 | Alipay (Hangzhou) Information Technology Co., Ltd. | Consensus methods and systems in consortium blockchain |
| US11258851B2 (en) | 2020-07-03 | 2022-02-22 | Alipay (Hangzhou) Information Technology Co., Ltd. | Consensus node changing method and related apparatus based on honey badger byzantine fault tolerance consensus mechanism |
| CN112929354A (en) * | 2021-01-28 | 2021-06-08 | 恒宝股份有限公司 | Practical Byzantine fault-tolerant anti-attack deadlock method and device |
| CN112929354B (en) * | 2021-01-28 | 2022-04-08 | 恒宝股份有限公司 | Practical Byzantine fault-tolerant anti-attack deadlock method and device |
| CN116915796A (en) * | 2023-09-14 | 2023-10-20 | 杭州趣链科技有限公司 | Autonomous recovery method and device after cluster view bifurcation and electronic equipment |
| CN116915796B (en) * | 2023-09-14 | 2023-12-12 | 杭州趣链科技有限公司 | Autonomous recovery method and device after cluster view bifurcation and electronic equipment |
Also Published As
| Publication number | Publication date |
|---|---|
| CN111130879B (en) | 2022-11-18 |
| CN115632933A (en) | 2023-01-20 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN111130879B (en) | PBFT algorithm-based cluster exception recovery method | |
| US20180150501A1 (en) | Database system, server device, computer program product, and information processing method | |
| CN110460484B (en) | Single-node abnormal active recovery method improved based on PBFT algorithm | |
| CN102394914A (en) | Cluster brain-split processing method and device | |
| CN110032478B (en) | Method, device and system for real-time synchronization of data of main and standby centers and storage medium | |
| CN103580915A (en) | Method and device for determining main control node of trunking system | |
| CN111427728A (en) | State management method, main/standby switching method and electronic equipment | |
| CN109783277B (en) | Disaster recovery end, production end and data recovery method between disaster recovery end and production end | |
| CN110633168A (en) | Data backup method and system for distributed storage system | |
| CN111752488B (en) | Management method and device of storage cluster, management node and storage medium | |
| CN105095008A (en) | Distributed task fault redundancy method suitable for cluster system | |
| CN111083049B (en) | User table item recovery method and device, electronic equipment and storage medium | |
| CN113055196B (en) | Message processing method, network device and computer readable storage medium | |
| CN110635941A (en) | Database node cluster fault migration method and device | |
| EP2432193A2 (en) | Method of data replication in a distributed data storage system and corresponding device | |
| JP2003337717A (en) | Fault recovery synchronizing system of online transaction process | |
| CN111338857A (en) | Byzantine fault-tolerant consensus protocol | |
| CN111555858A (en) | Practical Byzantine fault-tolerant consensus method based on block chain type storage | |
| CN109194521B (en) | Flow forwarding method and equipment | |
| CN116232893A (en) | Consensus method and device of distributed system, electronic equipment and storage medium | |
| CN112068978B (en) | Method and device for prolonging timing period of VIEW-CHANGE secondary start timer | |
| CN115378815A (en) | Data recovery method, device, network equipment and storage medium | |
| CN115314361A (en) | Server cluster management method and related components thereof | |
| CN112491633B (en) | Fault recovery method, system and related components of multi-node cluster | |
| CN113064768B (en) | Method and device for switching fragment nodes in block chain 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 | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |