WO2017054512A1 - Method and device for optimizing communication in heterogeneous cluster - Google Patents

Method and device for optimizing communication in heterogeneous cluster Download PDF

Info

Publication number
WO2017054512A1
WO2017054512A1 PCT/CN2016/085335 CN2016085335W WO2017054512A1 WO 2017054512 A1 WO2017054512 A1 WO 2017054512A1 CN 2016085335 W CN2016085335 W CN 2016085335W WO 2017054512 A1 WO2017054512 A1 WO 2017054512A1
Authority
WO
WIPO (PCT)
Prior art keywords
node
heterogeneous cluster
message
communication
cluster
Prior art date
Application number
PCT/CN2016/085335
Other languages
French (fr)
Chinese (zh)
Inventor
陈前洋
Original Assignee
中兴通讯股份有限公司
Priority date (The priority date 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 date listed.)
Filing date
Publication date
Application filed by 中兴通讯股份有限公司 filed Critical 中兴通讯股份有限公司
Publication of WO2017054512A1 publication Critical patent/WO2017054512A1/en

Links

Images

Classifications

    • HELECTRICITY
    • H04ELECTRIC COMMUNICATION TECHNIQUE
    • H04LTRANSMISSION OF DIGITAL INFORMATION, e.g. TELEGRAPHIC COMMUNICATION
    • H04L65/00Network arrangements, protocols or services for supporting real-time applications in data packet communication
    • H04L65/40Support for services or applications
    • HELECTRICITY
    • H04ELECTRIC COMMUNICATION TECHNIQUE
    • H04LTRANSMISSION OF DIGITAL INFORMATION, e.g. TELEGRAPHIC COMMUNICATION
    • H04L67/00Network arrangements or protocols for supporting network services or applications
    • H04L67/50Network services
    • H04L67/60Scheduling or organising the servicing of application requests, e.g. requests for application data transmissions using the analysis and optimisation of the required network resources

Definitions

  • This paper relates to but not limited to the field of communication technology, and relates to a communication optimization method and device for heterogeneous clusters.
  • the core protocol of the cluster communication component is the Totem protocol stack, which mainly includes SRP (The Totem Secure Remote Password Protocol) and RTP (The Totem redundant ring protocol), and passes the SRP protocol.
  • SRP The Totem Secure Remote Password Protocol
  • RTP The Totem redundant ring protocol
  • a node can obtain a token before sending a message.
  • the number of messages that different nodes can send varies. For example, there are two messages in the cluster. The first node can send these two messages, and the second. A node can only send one message. For the second node, it can also send two messages. To retransmit the message again, if the second node cannot send two messages all the time, the message will be retransmitted until The second node can send 2 messages to complete the communication process. Obviously, retransmission of multiple messages reduces the efficiency of cluster communication.
  • the embodiment of the invention provides a communication optimization method and device for heterogeneous clusters, which solves the problem of retransmitting multiple messages between heterogeneous cluster nodes and reducing the technical efficiency of heterogeneous cluster communication.
  • a communication optimization method for a heterogeneous cluster according to an embodiment of the present invention, where the communication optimization method of the heterogeneous cluster includes:
  • the parameter of the node corresponding to the communication data is adjusted when the number of retransmissions of the communication data reaches a preset number of times; wherein the parameter includes a time when the node receives the token information of the cluster.
  • the step of acquiring communication data of each node in the heterogeneous cluster in the heterogeneous cluster communication state includes:
  • the identification information of the node updates the number of message retransmissions of the node;
  • the communication data of the node includes identification information of the node and a number of message retransmissions corresponding to the node.
  • the step of adjusting parameters of the node corresponding to the communication data when the number of retransmissions of the message having the communication data reaches a preset number of times includes:
  • the method further includes: if the step of acquiring the communication data of each node in the heterogeneous cluster in the heterogeneous cluster communication state, detecting that the new node joins the heterogeneous cluster, acquiring the Describe the identification information of the new node, and record the identification information of the new node into the attribute information of the new node;
  • each node receives the attribute information broadcasted by other nodes, and the received attribute information is consistent with the attribute information corresponding to each node, and enters the heterogeneous cluster communication state.
  • the method further includes: after the step of broadcasting the current attribute information of each node to all the nodes of the heterogeneous cluster, the communication optimization method of the heterogeneous cluster further includes:
  • the current attribute information of the node is broadcast to other nodes in the heterogeneous cluster.
  • the embodiment of the invention provides a communication optimization device for a heterogeneous cluster, and the communication optimization device of the heterogeneous cluster includes:
  • Determining a module set to determine the number of retransmissions of messages in each communication data
  • the adjusting module is configured to adjust a parameter of the node corresponding to the communication data when the number of retransmissions of the communication data reaches a preset number of times; wherein the parameter includes a time when the node receives the token information of the cluster.
  • the obtaining module includes:
  • An update unit configured to: in the process of transmitting a message sequence in each node of the heterogeneous cluster, if it is detected that a node does not receive the preset message in the message sequence, according to the message sequence not received
  • the identification information of the node of the preset message updates the number of message retransmissions of the node;
  • the obtaining unit is configured to obtain communication data of each node, where the communication data of the node includes the identification information of the node and the number of message retransmissions corresponding to the node.
  • the adjusting module includes:
  • a determining unit configured to determine, when the number of retransmissions of the message having the communication data reaches a preset number of times, determining, by the node corresponding to the communication data that the number of retransmissions of the message reaches a preset number of times;
  • the adjusting unit is configured to increase a time when the node receives the token information of the cluster.
  • the communication optimization device of the heterogeneous cluster further includes:
  • a processing module configured to: if it is detected that a new node joins the heterogeneous cluster, acquire identification information of the new node, and record the identification information of the new node into the attribute information of the new node;
  • a broadcast module configured to broadcast, to all the nodes of the heterogeneous cluster, current attribute information of each node, where the attribute information corresponding to each node in the heterogeneous cluster includes identifier information of all the original nodes;
  • the entry module is configured to enter the heterogeneous cluster communication state when each node in the heterogeneous cluster receives the attribute information broadcasted by other nodes and the received attribute information is consistent with the attribute information corresponding to each node.
  • the broadcast module is further configured to: when the node in the heterogeneous cluster does not receive the attribute information broadcast by the other node, broadcast the current node to the other nodes in the heterogeneous cluster. Attribute information.
  • the communication optimization method and device for the heterogeneous cluster determines the number of retransmissions of each communication data in the communication data according to the obtained communication data of each node in the heterogeneous cluster communication state, and has communication
  • the parameters of the node corresponding to the communication data are adjusted to reduce the number of retransmissions of the node, for example, increasing the time at which the node receives the token information of the cluster.
  • the message is retransmitted.
  • the foregoing technical solution reduces the number of message retransmissions in the communication data of the node, and adjusts the parameters of the node when the number of retransmissions of the message reaches a preset number of times, thereby reducing the number of message retransmissions of the node, thereby improving the cluster.
  • Communication efficiency further provides a computer readable storage medium, where the computer readable storage medium stores computer executable instructions, and the computer executable instructions are executed to implement a communication optimization method of a heterogeneous cluster.
  • FIG. 1 is a schematic flowchart of a first embodiment of a communication optimization method for a heterogeneous cluster according to Embodiment 1 of the present invention
  • FIG. 2 is a schematic flowchart of acquiring communication data of each node in the heterogeneous cluster according to Embodiment 2 of the present invention
  • FIG. 3 is a schematic flowchart of adjusting parameters of a node corresponding to the communication data according to Embodiment 3 of the present invention.
  • FIG. 4 is a schematic flowchart of a communication optimization method for a heterogeneous cluster according to Embodiment 4 of the present invention.
  • FIG. 5 is a schematic diagram of functional modules of a communication optimization apparatus for a heterogeneous cluster according to Embodiment 5 of the present invention.
  • FIG. 6 is a schematic diagram of a refinement function module of the acquisition module in FIG. 5;
  • FIG. 7 is a schematic diagram of a refinement function module of the adjustment module in FIG. 5;
  • FIG. 8 is a schematic diagram of a refinement function module of a communication optimization apparatus for a heterogeneous cluster according to Embodiment 5 of the present invention.
  • FIG. 1 is a schematic flowchart diagram of a communication optimization method for a heterogeneous cluster according to Embodiment 1 of the present invention.
  • Step S10 Obtain communication data of each node in the heterogeneous cluster in a heterogeneous cluster communication state
  • the heterogeneous cluster includes multiple nodes.
  • the heterogeneous cluster includes three nodes, which are respectively P1, P2, and P3, and the communication sequence of the cluster is in accordance with P1- The order of P2-P3 is communicated.
  • the communication process in the heterogeneous cluster communication state is as follows: 1) The upper layer application of the heterogeneous cluster delivers a message sequence, such as M1M2M3, after the node P1 obtains the message sequence, buffers the message sequence, and waits for the token to be obtained. After the information token, the token information is delivered to the next node P2, and at this time, the cluster delivers a message sequence, and at this time, the node P1 receives the complete message sequence, and the node P1 communicates normally; 2) After the node P2 receives the token information, the node P2 can deliver the message sequence.
  • a message sequence such as M1M2M3
  • the node P2 if the node P2 misses the message M3 and the token information has arrived at the node P2, the node P2 records the message M3 that needs to be retransmitted and needs to be retransmitted in the storage list corresponding to the token information. P2, and the token information is further passed to the next node P3; 3) after the node P3 obtains the token information, the storage list of the record token information carries the retransmission requirement information, that is, the node P2 needs to retransmit the message M3.
  • the node P3 can receive the complete message sequence at this time, the node P3 can be considered to be normal communication, and then the M3 message is broadcast to the cluster, and the token information is continuously transmitted to the node P1; 4) the node P1 After the token information is obtained, the retransmission request information carried in the storage list of the token information is recorded, that is, the message M3 that the node P2 needs to retransmit, and the token information is transmitted to the node P2.
  • the node P1 Since the node P1 has received the complete message sequence at this time, the node P1 ignores the M3 message broadcast by the node P3; 5) the node P2 acquires the token information, and finds that the token information records itself If the message M3 is not received, then the node P2 acquires the M3 message broadcast by the node P3, so that the received message sequence is complete, and the node P2 clears the message retransmission information in the token information. All nodes in the heterogeneous cluster receive the sequence of messages sent by P1.
  • a new retransmission dynamic statistics module is configured to collect communication data of each node in the heterogeneous cluster in a heterogeneous cluster communication state, and the acquiring the heterogeneous cluster
  • the communication data of each node may be obtained by acquiring communication data of each node in the heterogeneous cluster according to a certain time interval, for example, recording communication data of similar nodes, retransmission requirements, etc., and periodically acquiring each node record. The communication data.
  • Step S20 determining the number of retransmissions of messages in each communication data
  • step S30 when the number of retransmissions of the communication data reaches a preset number of times, the parameters of the node corresponding to the communication data are adjusted; wherein the parameter includes the time when the node receives the token information of the cluster.
  • the embodiment may select each The number of message retransmissions corresponding to each node. For example, if the number of times the message is retransmitted within a period of time is less than the preset number of times, if any node has obvious centralized retransmission characteristics, that is, there is communication data. When the number of retransmissions of the message reaches a preset number of times, the parameters of the node are adjusted to reduce the number of message retransmissions of the node.
  • the window_size parameter of the node may be reduced, the token buffer length may be expanded, or the token interval period parameter may be expanded, and the like.
  • the node The reduction of the window_size parameter can be understood as reducing the information transmission rate of the node. Since the heterogeneous cluster communication process can only optimize the next operation after each message retransmission, the information transmission rate of the node is reduced. The efficiency of the entire communication process can be reduced so that the node can receive messages in the sequence of messages.
  • the communication optimization method for the heterogeneous cluster proposed in this embodiment determines the number of retransmissions of each communication data and the message with the communication data according to the obtained communication data of each node in the heterogeneous cluster communication state.
  • the parameters of the node corresponding to the communication data are adjusted to reduce the number of message retransmissions of the node, such as increasing the time at which the node receives the token information of the cluster, so as to reduce the number of times.
  • the number of message retransmissions of the node rather than during the heterogeneous cluster communication process, if a node does not receive the message, the message is retransmitted, and the scheme retransmits the number of messages in the communication data of the node, and When the number of retransmissions of the message reaches a preset number of times, the parameters of the node are adjusted to reduce the number of retransmissions of the node, thereby improving the communication efficiency of the cluster.
  • the communication optimization method of the heterogeneous cluster according to any of the foregoing embodiments is provided.
  • the step S10 includes:
  • Step S11 in the process of transmitting a message sequence by each node of the heterogeneous cluster, if it is detected that a node does not receive the preset message in the message sequence, according to the pre-received in the message sequence Setting the identifier information of the node of the message to update the number of message retransmissions of the node that does not receive the preset message in the message sequence;
  • Step S12 Acquire communication data of each node, where the communication data of the node includes identification information of the node and a number of message retransmissions corresponding to the node.
  • a communication data table corresponding to each node in the heterogeneous cluster is established, and the communication data table is used to record communication data of each node, and the communication data of each node includes the identification information of the node and the node.
  • the number of corresponding message retransmissions That is, in the process of transmitting a message sequence by each node of the heterogeneous cluster, the number of message retransmissions corresponding to each node is recorded according to the identification information of each node, and if the node is not received, the message sequence is not received.
  • the number of message retransmissions of the node is updated according to the identifier information of the node that does not receive the preset message in the message sequence, and if the node is not detected, the pre-received in the message sequence is not received. Set the message, then protect The number of retransmissions of the message holding the node is unchanged.
  • the communication optimization method of the heterogeneous cluster according to any of the foregoing embodiments is provided.
  • the S30 includes:
  • Step S31 when the number of retransmissions of the message having the communication data reaches a preset number of times, determining that the number of retransmissions of the message reaches a node corresponding to the communication data of the preset number of times;
  • Step S32 adding a time when the node receives the token information of the cluster.
  • the node corresponding to the communication data is determined.
  • the node performs multiple message retransmissions, which may be The communication sequence of the message sequence is faster, or when the node just receives the message sequence, the token information has arrived at the node, and the node can only receive a part of the message sequence, resulting in a message. The message is missed.
  • the embodiment may optionally expand the token interval period parameter, that is, increase the time at which the node receives the token information of the cluster, so as to reduce the number of message retransmissions of the node, that is, the The time for token information transmission in a heterogeneous cluster increases. If the time for transmitting token information in a heterogeneous cluster is 6 seconds, and there are 3 nodes in the heterogeneous cluster, then each node receives the token information. The time is 2 seconds. Now the time that the token information is passed in the heterogeneous cluster is increased to 9 seconds.
  • the time for each node to receive the token information is 3 seconds, which is beneficial to The node receives a complete sequence of messages, thereby reducing the number of message retransmissions for the node.
  • the number of retransmissions of the node may be cleared, and the heterogeneous cluster is continuously counted in a new round of heterogeneous cluster communication. Communication data of each node during communication.
  • the communication optimization method of the heterogeneous cluster proposed in any of the foregoing embodiments, in this embodiment, referring to FIG. 4, the communication optimization method of the heterogeneous cluster also includes:
  • Step S40 in the heterogeneous cluster communication state, acquiring each node in the heterogeneous cluster Before the step of the communication data, if it is detected that the new node joins the heterogeneous cluster, the identification information of the new node is obtained, and the identification information of the new node is recorded in the attribute information of the new node;
  • step S50 the current attribute information of each node is broadcasted to all the nodes of the heterogeneous cluster, where the attribute information corresponding to each node in the heterogeneous cluster includes the identification information of all the original nodes;
  • Step S60 In the heterogeneous cluster, each node receives attribute information broadcasted by other nodes, and when the received attribute information is consistent with the attribute information corresponding to each node, the heterogeneous cluster communication state is entered.
  • the clusters of the original nodes P1, P2, and P3 operate normally. If a new node joins the heterogeneous cluster, such as P4, the identification information of the new node is obtained first, and the new node is obtained. The identification information is recorded in the attribute information of the new node, that is, the local node identification information is recorded in the attribute information of the new node, such as my_proc_set; and the current attribute information of each node is broadcast to all nodes of the heterogeneous cluster. For all nodes of the heterogeneous cluster to synchronize their attribute information to new attribute information containing all node identification information.
  • the attribute information corresponding to each node in the heterogeneous cluster includes the identification information of each node, that is, the attribute information corresponding to the original nodes P1, P2, and P3 is P1P2P3, which is better.
  • An example is as follows: In the heterogeneous cluster, the new node P4 broadcasts a communication message of a heterogeneous cluster dynamic addition and deletion node, such as a JoinMsg message, and the JoinMsg message includes attribute information of the node P4.
  • the nodes P1, P2, and P3 in the heterogeneous cluster receive the JoinMsg message broadcast by the node P4, enter the gather state, and the nodes P1, P2, and P3 update their own attribute information my_proc_set to
  • the attribute information is all synchronized to new attribute information including all node identification information, that is, the original attribute information P1P2P3 of the nodes P1, P2, P3 is updated to new attribute information P1P2P3P4, and the nodes P1, P2, P3 are also clustered.
  • the JoinMsg message is broadcasted.
  • the node P4 also updates its attribute information to the new attribute information P1P2P3P4 when receiving the JoinMsg message broadcast by the nodes P1, P2, and P3.
  • each node when each node performs the broadcast JoinMsg message, as long as the node receives the JoinMsg message broadcast by other nodes, the node will compare the attribute information of the JoinMsg message with its own attribute information, if comparing the my_pro_set and the message in the message.
  • the content of my own my_proc_set is the same, that is, when the received attribute information is consistent with the attribute information corresponding to each node, the JoinMsg message is sent.
  • the sender_id is set to be consistent, that is, the sending address information corresponding to the attribute information of each node is set to be synchronized, displayed in the consensus list, and finally enters the heterogeneous cluster communication state.
  • the method may further include: after the step S50, when the node does not receive the attribute information broadcast by the other node in the heterogeneous cluster, the node is broadcast to other nodes in the heterogeneous cluster. Current attribute information, .
  • retransmission is triggered, that is, message retransmission is performed.
  • message transmission is performed again for other nodes to feed back other nodes.
  • Corresponding attribute information is sent to the node, and each node has a certain time requirement to enter the consensus state.
  • the modification informs the newly added node in JoinMsg to reduce the window_size and expand the token parameter. If it is a new join node entering the consensus state for a long time, then slightly expand the consensus timer duration and so on.
  • FIG. 5 is a schematic diagram of functional modules of a communication optimization apparatus for a heterogeneous cluster according to an embodiment of the present invention.
  • the functional block diagram shown in FIG. 5 is merely an exemplary diagram of an alternative embodiment, and those skilled in the art will surround the heterogeneous cluster communication optimization apparatus shown in FIG. 5.
  • the function module can be easily supplemented by a new function module; the name of each function module is a custom name, which is only used to assist in understanding each program function block of the communication optimization device of the heterogeneous cluster, and is not used to limit the technology of the present application.
  • Solution the core of the technical solution of the present application is the function to be achieved by each functional module of a custom name.
  • Obtaining module 10 configured to acquire each of the heterogeneous clusters in a heterogeneous cluster communication state Communication data of the node;
  • the heterogeneous cluster includes multiple nodes.
  • the heterogeneous cluster includes three nodes, which are respectively P1, P2, and P3, and the communication sequence of the cluster is in accordance with P1.
  • -P2-P3 is in the order of communication, and the communication process in the heterogeneous cluster communication state is as follows: 1) The upper layer application of the heterogeneous cluster delivers a message sequence such as M1M2M3, and the node P1 caches after obtaining the message sequence. The message sequence, after acquiring the token information token, passes the token information to the next node P2, and at this time, the cluster delivers the message sequence, and at this time, the node P1 receives the complete message sequence.
  • the node P1 communicates normally; 2) after the node P2 receives the token information, the node P2 can deliver the message sequence. It should be noted that if the node P2 misses the message M3 and the token information has arrived at the node P2, the node P2 records the message M3 that needs to be retransmitted and needs to be retransmitted in the storage list corresponding to the token information. P2, and the token information is further passed to the next node P3; 3) after the node P3 obtains the token information, the storage list of the record token information carries the retransmission requirement information, that is, the node P2 needs to retransmit the message M3.
  • the node P3 can receive the complete message sequence at this time, the node P3 can be considered to be normal communication, and then the M3 message is broadcast to the cluster, and the token information is continuously transmitted to the After the node P1 obtains the token information, the node P1 records the retransmission requirement information carried in the storage list of the token information, that is, the message M3 that the node P2 needs to retransmit, and the token is The information is delivered to the node P2.
  • the node P1 Since the node P1 has received the complete message sequence at this time, the node P1 ignores the M3 message broadcast by the node P3; 5) the node P2 obtains the token information, and finds the location Recorded from the token information The message M3 has not been received, then the node P2 acquires the M3 message broadcast by the node P3 at this time, so that the received message sequence is complete, and the node P2 clears the message retransmission information in the token information. At this point, all nodes in the heterogeneous cluster have received the message sequence sent by P1.
  • a new retransmission dynamic statistics module is configured to collect communication data of each node in the heterogeneous cluster in a heterogeneous cluster communication state, and the obtaining module 10 obtains the difference.
  • the communication data of each node in the cluster may be configured to obtain communication data of each node in the heterogeneous cluster at a certain time interval, for example, recording communication data of similar nodes, retransmission requirements, etc., and periodically acquiring each The communication data recorded by the nodes.
  • the determining module 20 is configured to determine the number of retransmissions of the message in each communication data
  • the adjusting module 30 is configured to adjust parameters of the node corresponding to the communication data when the number of retransmissions of the communication data reaches a preset number of times; wherein the parameter includes a time when the node receives the token information of the cluster.
  • the determining module 20 directly analyzes data in the background of each node in the heterogeneous cluster to determine communication data corresponding to each node.
  • the number of retransmissions corresponding to each node may be set first. For example, if the number of retransmissions of the message in a period of time is less than the preset number of times, if any node has obvious centralized retransmission characteristics, That is, when the number of retransmissions of the communication data reaches a preset number of times, the adjustment module 30 adjusts the parameters of the node to reduce the number of retransmissions of the node.
  • the adjustment module 30 may reduce the window_size parameter of the node, expand the token cache length, or expand the token interval period parameter, etc., and reduce the window_size parameter of the node may be understood as The information transmission rate of the node is reduced. Because the heterogeneous cluster communication process can only optimize the next operation after each message retransmission, the information transmission rate of the node is reduced, which can reduce the efficiency of the entire communication process. Thereby the node can receive the message in the sequence of messages.
  • the communication optimization device of the heterogeneous cluster proposed in this embodiment determines the number of retransmissions of each communication data and the message with the communication data according to the obtained communication data of each node in the heterogeneous cluster communication state.
  • the parameters of the node corresponding to the communication data are adjusted to reduce the number of message retransmissions of the node, such as increasing the time at which the node receives the token information of the cluster, so as to reduce the number of times.
  • the number of message retransmissions of the node rather than during the heterogeneous cluster communication process, if a node does not receive the message, the message is retransmitted, and the scheme retransmits the number of messages in the communication data of the node, and When the number of retransmissions of the message reaches a preset number of times, the parameters of the node are adjusted to reduce the number of retransmissions of the node, thereby improving the communication efficiency of the cluster.
  • the obtaining module 10 includes:
  • the updating unit 11 is configured to: in the process of transmitting a message sequence in each node of the heterogeneous cluster, if it is detected that a node does not receive the preset message in the message sequence, according to the non-received sequence in the message sequence.
  • the identifier information of the node of the preset message updates the number of message retransmissions of the node that does not receive the preset message in the message sequence;
  • the obtaining unit 13 is configured to acquire communication data of each node, where the communication data of the node includes the identification information of the node and the number of message retransmissions corresponding to the node.
  • a communication data table corresponding to each node in the heterogeneous cluster is established, and the communication data table is used to record communication data of each node, and the communication data of each node includes the identification information of the node and the node. The number of corresponding message retransmissions.
  • the processing unit 12 records the number of message retransmissions corresponding to each node according to the identification information of each node, and if a node is not received, The preset message in the message sequence, the updating unit 11 updates the number of message retransmissions of the node according to the identifier information of the node that does not receive the preset message in the message sequence, if no node is detected yet Receiving the preset message in the message sequence, the number of message retransmissions of the node is kept unchanged.
  • the adjustment module 30 includes:
  • the determining unit 31 is configured to: when the number of retransmissions of the message having the communication data reaches a preset number of times, determine that the number of retransmissions of the message reaches a node corresponding to the communication data of the preset number of times;
  • the adjusting unit 32 is configured to increase the time when the node receives the token information of the cluster.
  • the determining unit 31 determines the node corresponding to the communication data.
  • the node performs multiple times of the message. Retransmission, it may be that the communication sequence of the message sequence is faster, or when the node just receives the message sequence, the token information has arrived at the node, and the node can only receive the message sequence. The part of the message causes the message to be missed. Therefore, the adjusting unit 32 may optionally expand the token interval period parameter, that is, increase the node receiving station.
  • the time of the token information of the cluster is reduced to reduce the number of message retransmissions of the node, that is, the time for transmitting the token information in the heterogeneous cluster is increased, and the time for transmitting the token information in the heterogeneous cluster is 6 seconds.
  • each node receives the token information for 2 seconds, and now the time for transmitting the token information in the heterogeneous cluster is increased to 9 seconds, then The time at which each node receives the token information is 3 seconds, which is advantageous for the node to receive a complete message sequence, thereby reducing the number of message retransmissions of the node.
  • the adjusting unit 32 adjusts parameters of the node corresponding to the communication data
  • the number of retransmissions of the node may be cleared, and statistics are continued in a new round of heterogeneous cluster communication. Communication data of each node in the heterogeneous cluster communication process.
  • the communication optimization apparatus of the heterogeneous cluster includes:
  • the processing module 40 is configured to: if it is detected that the new node joins the heterogeneous cluster, acquire the identification information of the new node, and record the new node identifier information into the attribute information of the new node;
  • the broadcast module 50 is configured to broadcast the current attribute information of each node to all the nodes of the heterogeneous cluster, where the attribute information corresponding to each node in the heterogeneous cluster includes the identification information of all the original nodes;
  • the entering module 60 is configured to enter the heterogeneous cluster communication state when each node in the heterogeneous cluster receives the attribute information broadcasted by other nodes and the received attribute information is consistent with the attribute information corresponding to each node.
  • the clusters of the original nodes P1, P2, and P3 are in normal operation. If a new node, such as P4, joins the heterogeneous cluster, the processing module 40 first obtains the identification information of the new node, and The identification information of the new node is recorded in the attribute information of the new node, that is, the local node identification information is recorded in the attribute information of the new node, such as my_proc_set; meanwhile, the broadcast module 50 is to the heterogeneous cluster. All nodes broadcast the current attribute information of each node, so that all nodes of the heterogeneous cluster synchronize their attribute information into new attribute information including all node identification information.
  • the attribute information corresponding to each node in the heterogeneous cluster includes the identification information of each node, that is, the attribute information corresponding to the original nodes P1, P2, and P3 are P1P2P3, for a better understanding of the embodiment, for example: when the new node P4 joins the heterogeneous cluster, the new node P4 broadcasts a communication message of the heterogeneous cluster dynamic addition and deletion node, such as a JoinMsg message, and the JoinMsg message includes the node P4. Attribute information.
  • the state enters the gather state, and the nodes P1, P2, and P3 update their own.
  • the attribute information my_proc_set is used to synchronize the attribute information to new attribute information including all node identification information, that is, the original attribute information P1P2P3 of the nodes P1, P2, P3 is updated to new attribute information P1P2P3P4, and the nodes P1, P2 are P3 also broadcasts its JoinMsg message to the cluster.
  • the node P4 also updates its attribute information to the new attribute information P1P2P3P4 when receiving the JoinMsg message broadcast by the nodes P1, P2, and P3. It can be understood that When each node broadcasts a JoinMsg message, as long as the node receives the JoinMsg message broadcast by other nodes, it will attribute the attribute information of the JoinMsg message and its own attribute information.
  • the sending address information corresponding to the attribute information of each node sets the synchronization identifier, and is displayed in the consensus list.
  • the broadcast module 50 is further configured to: when the node in the heterogeneous cluster does not receive the attribute information broadcast by the other node, broadcast the current attribute information of the node to other nodes in the heterogeneous cluster. .
  • retransmission is triggered, that is, message retransmission is performed.
  • the broadcast module 50 broadcasts the current attribute information of the node to other nodes, and then transmits the message again for Other nodes feed back the attribute information corresponding to other nodes to the node, and each node has a certain time requirement to enter the consensus state.
  • the modification informs the newly added node in JoinMsg to reduce the window_size and expand the token parameter. If it is a new join node entering the consensus state for a long time, then slightly expand the consensus timer duration and so on.
  • the embodiment of the present invention further provides a computer readable storage medium, where the computer readable storage medium stores computer executable instructions, and the computer executable instructions are executed to implement a communication optimization method of a heterogeneous cluster.
  • each module/unit in the above embodiment may be implemented in the form of hardware, for example, by implementing an integrated circuit to implement its corresponding function, or may be implemented in the form of a software function module, for example, executing a program stored in the memory by a processor. / instruction to achieve its corresponding function.
  • This application is not limited to any specific combination of hardware and software.
  • the above technical solution reduces the number of message retransmissions of the node, thereby improving the communication efficiency of the cluster.

Landscapes

  • Engineering & Computer Science (AREA)
  • Computer Networks & Wireless Communication (AREA)
  • Signal Processing (AREA)
  • Multimedia (AREA)
  • Mobile Radio Communication Systems (AREA)

Abstract

Provided is a method for optimizing communication in a heterogeneous cluster, comprising: when a cluster is in a communication state, acquiring communication data items of respective nodes in the cluster; determining message redelivery counts in respective communication data items; and if the message redelivery count in one of the communication data items reaches a predetermined count, adjusting a parameter of the node corresponding to said communication data item, wherein the parameter comprises a time for the node to receive a token of the cluster.

Description

异构集群的通讯优化方法及装置Communication optimization method and device for heterogeneous cluster 技术领域Technical field
本文涉及但不限于通信技术领域,涉及一种异构集群的通讯优化方法及装置。This paper relates to but not limited to the field of communication technology, and relates to a communication optimization method and device for heterogeneous clusters.
背景技术Background technique
在集群通讯组件的最核心的协议为Totem协议栈,主要包括SRP(The Totem Secure Remote Password protocol,安全远程密码协议)和RRP(The Totem redundant ring Protocol,虚拟路由器冗余协议),并通过SRP协议进行消息传递,即通过一种令牌信息(token)在集群节点中进行传递,节点获取到token后传达上层应用下发的消息序列,而当集群节点的硬件配置如出现较大差异性时,集群成为异构集群。The core protocol of the cluster communication component is the Totem protocol stack, which mainly includes SRP (The Totem Secure Remote Password Protocol) and RTP (The Totem redundant ring protocol), and passes the SRP protocol. Message passing, that is, passing through a token information in a cluster node, the node obtains a token and then conveys a message sequence sent by the upper layer application, and when the hardware configuration of the cluster node is greatly different, The cluster becomes a heterogeneous cluster.
在异构集群环境下,节点获取到token后才能发送消息,而不同节点可以发送的消息数量存在差异性,比如集群中有2个消息,第一个节点可以发送这2个消息,而第二节点只能发送1个消息,为了第二个节点也可以发送2个消息,要再次进行消息重传,若第二个节点一直都无法发送2个消息,就要一直进行消息的重传,直到第二个节点可发送2个消息才完成这个通讯过程,显然,进行多次消息的重传降低了集群通讯效率。In a heterogeneous cluster environment, a node can obtain a token before sending a message. The number of messages that different nodes can send varies. For example, there are two messages in the cluster. The first node can send these two messages, and the second. A node can only send one message. For the second node, it can also send two messages. To retransmit the message again, if the second node cannot send two messages all the time, the message will be retransmitted until The second node can send 2 messages to complete the communication process. Obviously, retransmission of multiple messages reduces the efficiency of cluster communication.
发明内容Summary of the invention
以下是对本文详细描述的主题的概述。本概述并非是为了限制权利要求的保护范围。The following is an overview of the topics detailed in this document. This Summary is not intended to limit the scope of the claims.
本发明实施例提出一种异构集群的通讯优化方法及装置,解决了异构集群节点间进行多次消息的重传,降低了异构集群通讯效率的技术问题。The embodiment of the invention provides a communication optimization method and device for heterogeneous clusters, which solves the problem of retransmitting multiple messages between heterogeneous cluster nodes and reducing the technical efficiency of heterogeneous cluster communication.
本发明实施例提供的一种异构集群的通讯优化方法,所述异构集群的通讯优化方法包括:A communication optimization method for a heterogeneous cluster according to an embodiment of the present invention, where the communication optimization method of the heterogeneous cluster includes:
在异构集群通讯状态下,获取所述异构集群中每个节点的通讯数据;Obtaining communication data of each node in the heterogeneous cluster in a heterogeneous cluster communication state;
确定每个通讯数据中的消息重传次数; Determine the number of message retransmissions in each communication data;
在有通讯数据的消息重传次数达到预设次数时,调整所述通讯数据对应的节点的参数;其中,所述参数包括节点接收所述集群的令牌信息的时间。The parameter of the node corresponding to the communication data is adjusted when the number of retransmissions of the communication data reaches a preset number of times; wherein the parameter includes a time when the node receives the token information of the cluster.
可选地,所述在异构集群通讯状态下,获取所述异构集群中每个节点的通讯数据的步骤包括:Optionally, the step of acquiring communication data of each node in the heterogeneous cluster in the heterogeneous cluster communication state includes:
在所述异构集群的节点传递消息序列的过程中,若检测到有节点未接收到所述消息序列中的预设消息,则根据所述未接收到所述消息序列中的预设信息的节点的标识信息更新所述节点的消息重传次数;In the process of transmitting the message sequence by the node of the heterogeneous cluster, if it is detected that a node does not receive the preset message in the message sequence, according to the preset information that is not received in the message sequence, The identification information of the node updates the number of message retransmissions of the node;
获取各个节点的通讯数据,其中,节点的通讯数据包括所述节点的标识信息以及所述节点对应的消息重传次数。Obtaining communication data of each node, where the communication data of the node includes identification information of the node and a number of message retransmissions corresponding to the node.
可选地,所述在有通讯数据的消息重传次数达到预设次数时,调整所述通讯数据对应的节点的参数的步骤包括:Optionally, the step of adjusting parameters of the node corresponding to the communication data when the number of retransmissions of the message having the communication data reaches a preset number of times includes:
在有通讯数据的消息重传次数达到预设次数时,确定所述消息重传次数达到预设次数的通讯数据对应的节点;When the number of retransmissions of the communication data reaches a preset number of times, determining that the number of retransmissions of the message reaches a node corresponding to the preset number of communication data;
增加所述节点接收所述集群的令牌信息的时间。Increasing the time at which the node receives the token information of the cluster.
可选地,所述方法还包括:所述在异构集群通讯状态下,获取所述异构集群中每个节点的通讯数据的步骤之前若检测到新节点加入所述异构集群,获取所述新节点的标识信息,并将所述新节点的标识信息记录到所述新节点的属性信息中;Optionally, the method further includes: if the step of acquiring the communication data of each node in the heterogeneous cluster in the heterogeneous cluster communication state, detecting that the new node joins the heterogeneous cluster, acquiring the Describe the identification information of the new node, and record the identification information of the new node into the attribute information of the new node;
向所述异构集群的所有节点广播每个节点当前的属性信息,其中,所述异构集群中原先每个节点对应的属性信息都包含原先所有节点的标识信息;Broadcasting the current attribute information of each node to all the nodes of the heterogeneous cluster, where the attribute information corresponding to each node in the heterogeneous cluster includes the identification information of all the original nodes;
在所述异构集群中每个节点接收到其它节点广播的属性信息且接收到的属性信息与每个节点对应的属性信息都一致时,进入异构集群通讯状态。In the heterogeneous cluster, each node receives the attribute information broadcasted by other nodes, and the received attribute information is consistent with the attribute information corresponding to each node, and enters the heterogeneous cluster communication state.
可选地,所述方法还包括:所述向所述异构集群的所有节点广播每个节点当前的属性信息的步骤之后,所述异构集群的通讯优化方法还包括:Optionally, the method further includes: after the step of broadcasting the current attribute information of each node to all the nodes of the heterogeneous cluster, the communication optimization method of the heterogeneous cluster further includes:
在所述异构集群中有节点未接收到其它节点广播的属性信息时,向所述异构集群中其它节点广播所述节点当前的属性信息。When the node does not receive the attribute information broadcast by other nodes in the heterogeneous cluster, the current attribute information of the node is broadcast to other nodes in the heterogeneous cluster.
本发明实施例提供一种异构集群的通讯优化装置,所述异构集群的通讯优化装置包括: The embodiment of the invention provides a communication optimization device for a heterogeneous cluster, and the communication optimization device of the heterogeneous cluster includes:
获取模块,设置为在异构集群通讯状态下,获取所述异构集群中每个节点的通讯数据;Obtaining a module, configured to acquire communication data of each node in the heterogeneous cluster in a heterogeneous cluster communication state;
确定模块,设置为确定每个通讯数据中的消息重传次数;Determining a module, set to determine the number of retransmissions of messages in each communication data;
调整模块,设置为在有通讯数据的消息重传次数达到预设次数时,调整所述通讯数据对应的节点的参数;其中,所述参数包括节点接收所述集群的令牌信息的时间。The adjusting module is configured to adjust a parameter of the node corresponding to the communication data when the number of retransmissions of the communication data reaches a preset number of times; wherein the parameter includes a time when the node receives the token information of the cluster.
可选地,所述获取模块包括:Optionally, the obtaining module includes:
更新单元,设置为在所述异构集群的每个节点传递消息序列过程中,若检测到有节点未接收到所述消息序列中的预设消息,则根据所述未接收到所述消息序列中的预设消息的节点的标识信息更新所述节点的消息重传次数;An update unit, configured to: in the process of transmitting a message sequence in each node of the heterogeneous cluster, if it is detected that a node does not receive the preset message in the message sequence, according to the message sequence not received The identification information of the node of the preset message updates the number of message retransmissions of the node;
获取单元,设置为获取各个节点的通讯数据,其中,节点的通讯数据包括所述节点的标识信息以及所述节点对应的消息重传次数。The obtaining unit is configured to obtain communication data of each node, where the communication data of the node includes the identification information of the node and the number of message retransmissions corresponding to the node.
可选地,所述调整模块包括:Optionally, the adjusting module includes:
确定单元,设置为在有通讯数据的消息重传次数达到预设次数时,确定所述消息重传次数达到预设次数的通讯数据对应的节点;a determining unit, configured to determine, when the number of retransmissions of the message having the communication data reaches a preset number of times, determining, by the node corresponding to the communication data that the number of retransmissions of the message reaches a preset number of times;
调整单元,设置为增加所述节点接收所述集群的令牌信息的时间。The adjusting unit is configured to increase a time when the node receives the token information of the cluster.
可选地,所述异构集群的通讯优化装置还包括:Optionally, the communication optimization device of the heterogeneous cluster further includes:
处理模块,设置为若检测到新节点加入所述异构集群,获取所述新节点的标识信息,并将所述新节点的标识信息记录到所述新节点的属性信息中;a processing module, configured to: if it is detected that a new node joins the heterogeneous cluster, acquire identification information of the new node, and record the identification information of the new node into the attribute information of the new node;
广播模块,用于向所述异构集群的所有节点广播每个节点当前的属性信息,其中,所述异构集群中原先每个节点对应的属性信息都包含原先所有节点的标识信息;a broadcast module, configured to broadcast, to all the nodes of the heterogeneous cluster, current attribute information of each node, where the attribute information corresponding to each node in the heterogeneous cluster includes identifier information of all the original nodes;
进入模块,设置为在所述异构集群中每个节点接收到其它节点广播的属性信息且接收到的属性信息与每个节点对应的属性信息都一致时,进入异构集群通讯状态。The entry module is configured to enter the heterogeneous cluster communication state when each node in the heterogeneous cluster receives the attribute information broadcasted by other nodes and the received attribute information is consistent with the attribute information corresponding to each node.
可选地,所述广播模块,还设置为在所述异构集群中有节点未接收到其它节点广播的属性信息时,向所述异构集群中其它节点广播所述节点当前的 属性信息。Optionally, the broadcast module is further configured to: when the node in the heterogeneous cluster does not receive the attribute information broadcast by the other node, broadcast the current node to the other nodes in the heterogeneous cluster. Attribute information.
本发明实施例提供的异构集群的通讯优化方法及装置,在异构集群通讯状态下,根据获取的每个节点的通讯数据,确定每个通讯数据中的消息重传次数,并在有通讯数据的消息重传次数达到预设次数时,调整所述通讯数据对应的节点的参数,以降低所述节点的消息重传次数,例如增加所述节点接收所述集群的令牌信息的时间,以降低所述节点的消息重传次数,而不是在异构集群通讯过程中,若有节点未接收到消息时,一直进行消息的重传。上述技术方案通过统计节点的通讯数据中的消息重传次数,并在消息重传次数达到预设次数时,调整所述节点的参数,降低了所述节点的消息重传次数,从而提高了集群通讯效率。本发明实施例还提供一种计算机可读存储介质,所述计算机可读存储介质中存储有计算机可执行指令,所述计算机可执行指令被执行时实现异构集群的通讯优化方法。The communication optimization method and device for the heterogeneous cluster provided by the embodiment of the present invention determines the number of retransmissions of each communication data in the communication data according to the obtained communication data of each node in the heterogeneous cluster communication state, and has communication When the number of retransmissions of the data reaches a preset number of times, the parameters of the node corresponding to the communication data are adjusted to reduce the number of retransmissions of the node, for example, increasing the time at which the node receives the token information of the cluster. In order to reduce the number of message retransmissions of the node, instead of during the heterogeneous cluster communication process, if a node does not receive the message, the message is retransmitted. The foregoing technical solution reduces the number of message retransmissions in the communication data of the node, and adjusts the parameters of the node when the number of retransmissions of the message reaches a preset number of times, thereby reducing the number of message retransmissions of the node, thereby improving the cluster. Communication efficiency. The embodiment of the present invention further provides a computer readable storage medium, where the computer readable storage medium stores computer executable instructions, and the computer executable instructions are executed to implement a communication optimization method of a heterogeneous cluster.
在阅读并理解了附图和详细描述后,可以明白其它方面。Other aspects will be apparent upon reading and understanding the drawings and detailed description.
附图说明DRAWINGS
图1为本发明实施例一的异构集群的通讯优化方法第一实施例的流程示意图;1 is a schematic flowchart of a first embodiment of a communication optimization method for a heterogeneous cluster according to Embodiment 1 of the present invention;
图2为本发明实施例二的获取所述异构集群中每个节点的通讯数据的流程示意图;2 is a schematic flowchart of acquiring communication data of each node in the heterogeneous cluster according to Embodiment 2 of the present invention;
图3为本发明实施例三的调整所述通讯数据对应的节点的参数的流程示意图;3 is a schematic flowchart of adjusting parameters of a node corresponding to the communication data according to Embodiment 3 of the present invention;
图4为本发明实施例四的异构集群的通讯优化方法的流程示意图;4 is a schematic flowchart of a communication optimization method for a heterogeneous cluster according to Embodiment 4 of the present invention;
图5为本发明实施例五的异构集群的通讯优化装置的功能模块示意图;5 is a schematic diagram of functional modules of a communication optimization apparatus for a heterogeneous cluster according to Embodiment 5 of the present invention;
图6为图5中获取模块的细化功能模块示意图;6 is a schematic diagram of a refinement function module of the acquisition module in FIG. 5;
图7为图5中调整模块的细化功能模块示意图;7 is a schematic diagram of a refinement function module of the adjustment module in FIG. 5;
图8为本发明实施例五的异构集群的通讯优化装置的细化功能模块示意图。 FIG. 8 is a schematic diagram of a refinement function module of a communication optimization apparatus for a heterogeneous cluster according to Embodiment 5 of the present invention.
本发明目的实现、功能特点及优点将结合实施例,参照附图做进一步说明。The object, features, and advantages of the invention will be further described in conjunction with the embodiments.
具体实施方式detailed description
应当理解,此处所描述的具体实施例仅仅用以解释本申请,并不用于限定本申请。It is understood that the specific embodiments described herein are merely illustrative of the application and are not intended to be limiting.
实施例一Embodiment 1
参照图1,图1为本发明实施例一的异构集群的通讯优化方法的流程示意图。Referring to FIG. 1, FIG. 1 is a schematic flowchart diagram of a communication optimization method for a heterogeneous cluster according to Embodiment 1 of the present invention.
本实施例提出的异构集群的通讯优化方法包括:The communication optimization method for the heterogeneous cluster proposed in this embodiment includes:
步骤S10,在异构集群通讯状态下,获取所述异构集群中每个节点的通讯数据;Step S10: Obtain communication data of each node in the heterogeneous cluster in a heterogeneous cluster communication state;
在本实施例中,所述异构集群中包含多个节点,本实施例可选所述异构集群包括3个节点,分别为P1、P2和P3,且所述集群的通讯顺序按照P1-P2-P3的顺序进行通讯。In this embodiment, the heterogeneous cluster includes multiple nodes. In this embodiment, the heterogeneous cluster includes three nodes, which are respectively P1, P2, and P3, and the communication sequence of the cluster is in accordance with P1- The order of P2-P3 is communicated.
异构集群通讯状态下的通信过程如下:1)所述异构集群的上层应用下发消息序列,如M1M2M3,节点P1在获取到所述消息序列后,缓存所述消息序列,等到获取令牌信息token后,将所述令牌信息传递到下一个节点P2,并且此时所述集群传递消息序列,而此时,所述节点P1接收到完整的消息序列,则节点P1通讯正常;2)节点P2在接收到所述令牌信息后,所述节点P2才能传递所述消息序列。值得注意的是,若所述节点P2漏接了消息M3,而令牌信息已经到达节点P2,那么节点P2在所述令牌信息对应的存储列表中记录需要重传的消息M3以及需重传者P2,并将所述令牌信息继续传递到下一个节点P3;3)在所述节点P3获取到令牌信息后,记录令牌信息的存储列表中携带重传需求信息,即所述节点P2需要重传的消息M3。若此时节点P3若能接收到完整的消息序列,则可认为节点P3通讯正常,并随后向集群广播M3消息,且将所述令牌信息继续传递给所述节点P1;4)节点P1 获取到所述令牌信息后,记录所述令牌信息的存储列表中携带的重传需求信息即所述节点P2需要重传的消息M3,并将所述令牌信息传递给所述节点P2,由于此时所述节点P1已经接收到完整的消息序列,因此所述节点P1忽略节点P3广播的M3消息;5)节点P2获取到所述令牌信息,发现所述令牌信息中记录自己未接收到的消息M3,那么此时所述节点P2获取节点P3广播的M3消息,以使接收到的消息序列完整,并且所述节点P2清理所述令牌信息中的消息重传信息,至此所述异构集群中所有节点都收到了P1发送的消息序列。The communication process in the heterogeneous cluster communication state is as follows: 1) The upper layer application of the heterogeneous cluster delivers a message sequence, such as M1M2M3, after the node P1 obtains the message sequence, buffers the message sequence, and waits for the token to be obtained. After the information token, the token information is delivered to the next node P2, and at this time, the cluster delivers a message sequence, and at this time, the node P1 receives the complete message sequence, and the node P1 communicates normally; 2) After the node P2 receives the token information, the node P2 can deliver the message sequence. It should be noted that if the node P2 misses the message M3 and the token information has arrived at the node P2, the node P2 records the message M3 that needs to be retransmitted and needs to be retransmitted in the storage list corresponding to the token information. P2, and the token information is further passed to the next node P3; 3) after the node P3 obtains the token information, the storage list of the record token information carries the retransmission requirement information, that is, the node P2 needs to retransmit the message M3. If the node P3 can receive the complete message sequence at this time, the node P3 can be considered to be normal communication, and then the M3 message is broadcast to the cluster, and the token information is continuously transmitted to the node P1; 4) the node P1 After the token information is obtained, the retransmission request information carried in the storage list of the token information is recorded, that is, the message M3 that the node P2 needs to retransmit, and the token information is transmitted to the node P2. Since the node P1 has received the complete message sequence at this time, the node P1 ignores the M3 message broadcast by the node P3; 5) the node P2 acquires the token information, and finds that the token information records itself If the message M3 is not received, then the node P2 acquires the M3 message broadcast by the node P3, so that the received message sequence is complete, and the node P2 clears the message retransmission information in the token information. All nodes in the heterogeneous cluster receive the sequence of messages sent by P1.
在本实施例中,可选新增“重发动态统计”模块,用于统计异构集群通讯状态下所述异构集群中每个节点的通讯数据,并且所述获取所述异构集群中每个节点的通讯数据的方式可选按照一定的时间间隔获取所述异构集群中每个节点的通讯数据,例如记录类似节点、重传需求次数等的通讯数据,并定时获取每个节点记录的所述通讯数据。In this embodiment, a new retransmission dynamic statistics module is configured to collect communication data of each node in the heterogeneous cluster in a heterogeneous cluster communication state, and the acquiring the heterogeneous cluster The communication data of each node may be obtained by acquiring communication data of each node in the heterogeneous cluster according to a certain time interval, for example, recording communication data of similar nodes, retransmission requirements, etc., and periodically acquiring each node record. The communication data.
步骤S20,确定每个通讯数据中的消息重传次数;Step S20, determining the number of retransmissions of messages in each communication data;
步骤S30,在有通讯数据的消息重传次数达到预设次数时,调整所述通讯数据对应的节点的参数;其中,所述参数包括节点接收所述集群的令牌信息的时间。In step S30, when the number of retransmissions of the communication data reaches a preset number of times, the parameters of the node corresponding to the communication data are adjusted; wherein the parameter includes the time when the node receives the token information of the cluster.
在本实施例中,在获取到每个节点的通讯数据后,在异构集群中的每个节点的后台直接分析数据,以确定每个节点对应的通讯数据,本实施例可选先对每个节点对应的消息重传次数进行设置,如设置节点在一段时间内消息重传的次数应小于预设的次数,如果任一个节点存在明显的集中重传特征,也就是说在有通讯数据的消息重传次数达到预设次数时,则将该节点的参数进行调整,以降低所述节点的消息重传次数。In this embodiment, after the communication data of each node is obtained, the data is directly analyzed in the background of each node in the heterogeneous cluster to determine the communication data corresponding to each node, and the embodiment may select each The number of message retransmissions corresponding to each node is set. For example, if the number of times the message is retransmitted within a period of time is less than the preset number of times, if any node has obvious centralized retransmission characteristics, that is, there is communication data. When the number of retransmissions of the message reaches a preset number of times, the parameters of the node are adjusted to reduce the number of message retransmissions of the node.
可以理解的是,当有通讯数据的消息重传次数达到预设次数,说明所述通讯数据对应的节点在当前环境下节点消息序列能力较低,有可能是消息序列的通讯速率较快,或者是在节点刚接收到所述消息序列的消息时,令牌信息已经到达该节点了,此时所述节点只能将接收消息序列中的一部分消息,导致消息漏接。因此,本实施例可选将所述节点的window_size参数减少,将token缓存长度扩大,或者是将token间隔周期参数扩大等等。将所述节点 的window_size参数减少可理解为将所述节点的信息发送速率降低,由于异构集群通讯过程中,每一次消息重传后只能优化接下来的操作,因此将所述节点的信息发送速率降低,可以降低整个通讯过程的效率,从而使得所述节点可以接收到消息序列中的消息。It can be understood that when the number of retransmissions of the communication data reaches a preset number of times, it indicates that the node corresponding to the communication data has a lower capability of the node message sequence in the current environment, and may be that the communication rate of the message sequence is faster, or When the node just receives the message sequence, the token information has arrived at the node. At this time, the node can only receive a part of the message sequence, causing the message to be missed. Therefore, in this embodiment, the window_size parameter of the node may be reduced, the token buffer length may be expanded, or the token interval period parameter may be expanded, and the like. The node The reduction of the window_size parameter can be understood as reducing the information transmission rate of the node. Since the heterogeneous cluster communication process can only optimize the next operation after each message retransmission, the information transmission rate of the node is reduced. The efficiency of the entire communication process can be reduced so that the node can receive messages in the sequence of messages.
本实施例提出的异构集群的通讯优化方法,在异构集群通讯状态下,根据获取的每个节点的通讯数据,确定每个通讯数据中的消息重传次数,并在有通讯数据的消息重传次数达到预设次数时,调整所述通讯数据对应的节点的参数,以降低所述节点的消息重传次数,如增加所述节点接收所述集群的令牌信息的时间,以降低所述节点的消息重传次数,而不是在异构集群通讯过程中,若有节点未接收到消息时,一直进行消息的重传,本方案通过统计节点的通讯数据中的消息重传次数,并在消息重传次数达到预设次数时,调整所述节点的参数以降低所述节点的消息重传次数,从而提高了集群通讯效率。The communication optimization method for the heterogeneous cluster proposed in this embodiment determines the number of retransmissions of each communication data and the message with the communication data according to the obtained communication data of each node in the heterogeneous cluster communication state. When the number of retransmissions reaches a preset number of times, the parameters of the node corresponding to the communication data are adjusted to reduce the number of message retransmissions of the node, such as increasing the time at which the node receives the token information of the cluster, so as to reduce the number of times. The number of message retransmissions of the node, rather than during the heterogeneous cluster communication process, if a node does not receive the message, the message is retransmitted, and the scheme retransmits the number of messages in the communication data of the node, and When the number of retransmissions of the message reaches a preset number of times, the parameters of the node are adjusted to reduce the number of retransmissions of the node, thereby improving the communication efficiency of the cluster.
实施例二Embodiment 2
可选地,为了提高异构集群的通讯的灵活性,基于前述任一实施例提出的异构集群的通讯优化方法,在本实施例中,参照图2,所述步骤S10包括:Optionally, in order to improve the communication flexibility of the heterogeneous cluster, the communication optimization method of the heterogeneous cluster according to any of the foregoing embodiments is provided. In this embodiment, referring to FIG. 2, the step S10 includes:
步骤S11,在所述异构集群的每个节点传递消息序列的过程中,若检测到有节点未接收到所述消息序列中的预设消息,则根据未接收到所述消息序列中的预设消息的节点的标识信息更新所述未接收所述消息序列中的预设消息的节点的消息重传次数;Step S11, in the process of transmitting a message sequence by each node of the heterogeneous cluster, if it is detected that a node does not receive the preset message in the message sequence, according to the pre-received in the message sequence Setting the identifier information of the node of the message to update the number of message retransmissions of the node that does not receive the preset message in the message sequence;
步骤S12,获取每个节点的通讯数据,其中,节点的通讯数据包括所述节点的标识信息以及所述节点对应的消息重传次数。Step S12: Acquire communication data of each node, where the communication data of the node includes identification information of the node and a number of message retransmissions corresponding to the node.
在本实施例中,建立异构集群中每个节点对应的通讯数据表,所述通讯数据表用于记录每个节点的通讯数据,每个节点的通讯数据包括该节点的标识信息以及该节点对应的消息重传次数。也就是说,在所述异构集群的每个节点传递消息序列过程中,根据每个节点的标识信息记录每个节点对应的消息重传次数,若检测到有节点未接收到所述消息序列中的预设消息,根据所述未接收到所述消息序列中的预设消息的节点的标识信息更新所述节点的消息重传次数,若没有检测到节点未接收所述消息序列中的预设消息,那么保 持节点的消息重传次数不变。In this embodiment, a communication data table corresponding to each node in the heterogeneous cluster is established, and the communication data table is used to record communication data of each node, and the communication data of each node includes the identification information of the node and the node. The number of corresponding message retransmissions. That is, in the process of transmitting a message sequence by each node of the heterogeneous cluster, the number of message retransmissions corresponding to each node is recorded according to the identification information of each node, and if the node is not received, the message sequence is not received. And the preset message in the message, the number of message retransmissions of the node is updated according to the identifier information of the node that does not receive the preset message in the message sequence, and if the node is not detected, the pre-received in the message sequence is not received. Set the message, then protect The number of retransmissions of the message holding the node is unchanged.
实施例三Embodiment 3
可选地,为了提高异构集群的通讯的灵活性,基于前述任一实施例提出的异构集群的通讯优化方法,在本实施例中,参照图3,所述S30包括:Optionally, in order to improve the flexibility of the communication of the heterogeneous cluster, the communication optimization method of the heterogeneous cluster according to any of the foregoing embodiments is provided. In this embodiment, referring to FIG. 3, the S30 includes:
步骤S31,在有通讯数据的消息重传次数达到预设次数时,确定所述消息重传次数达到预设次数的通讯数据对应的节点;Step S31, when the number of retransmissions of the message having the communication data reaches a preset number of times, determining that the number of retransmissions of the message reaches a node corresponding to the communication data of the preset number of times;
步骤S32,增加所述节点接收所述集群的令牌信息的时间。Step S32, adding a time when the node receives the token information of the cluster.
在本实施例中,在有通讯数据的消息重传次数达到预设次数时,确定所述通讯数据对应的节点,根据前述任一实施例可知,所述节点进行多次消息重传,有可能是消息序列的通讯速率较快,或者是在节点刚接收到所述消息序列的消息时,令牌信息已经到达该节点了,此时所述节点只能将接收消息序列中的一部分消息,导致消息漏接,因此,本实施例可选将token间隔周期参数扩大,也就是说增加所述节点接收所述集群的令牌信息的时间,以降低所述节点的消息重传次数,即将所述异构集群中令牌信息传递的时间增加,原先在异构集群中传递令牌信息的时间若是6秒,且所述异构集群中有3个节点,那么每个节点接收到令牌信息的时间即为2秒,现在将令牌信息在所述异构集群中传递的时间增加到9秒,则每个节点接收到令牌信息的时间即为3秒,那么有利于所述节点接收完整的消息序列,从而降低了所述节点的消息重传次数。可选地,在调整完所述通讯数据对应的节点的参数后,可将所述节点的消息重传次数清零,并在新一轮异构集群通讯过程中,继续统计所述异构集群通讯过程中各个节点的通讯数据。In this embodiment, when the number of retransmissions of the communication data reaches a preset number of times, the node corresponding to the communication data is determined. According to any of the foregoing embodiments, the node performs multiple message retransmissions, which may be The communication sequence of the message sequence is faster, or when the node just receives the message sequence, the token information has arrived at the node, and the node can only receive a part of the message sequence, resulting in a message. The message is missed. Therefore, the embodiment may optionally expand the token interval period parameter, that is, increase the time at which the node receives the token information of the cluster, so as to reduce the number of message retransmissions of the node, that is, the The time for token information transmission in a heterogeneous cluster increases. If the time for transmitting token information in a heterogeneous cluster is 6 seconds, and there are 3 nodes in the heterogeneous cluster, then each node receives the token information. The time is 2 seconds. Now the time that the token information is passed in the heterogeneous cluster is increased to 9 seconds. The time for each node to receive the token information is 3 seconds, which is beneficial to The node receives a complete sequence of messages, thereby reducing the number of message retransmissions for the node. Optionally, after the parameters of the node corresponding to the communication data are adjusted, the number of retransmissions of the node may be cleared, and the heterogeneous cluster is continuously counted in a new round of heterogeneous cluster communication. Communication data of each node during communication.
实施例四Embodiment 4
可选地,为了提高异构集群的通讯的灵活性,基于前述任一实施例提出的异构集群的通讯优化方法,在本实施例中,参照图4,所述异构集群的通讯优化方法还包括:Optionally, in order to improve the flexibility of communication of the heterogeneous cluster, the communication optimization method of the heterogeneous cluster proposed in any of the foregoing embodiments, in this embodiment, referring to FIG. 4, the communication optimization method of the heterogeneous cluster Also includes:
步骤S40,所述在异构集群通讯状态下,获取所述异构集群中每个节点 的通讯数据的步骤之前,若检测到新节点加入所述异构集群,获取所述新节点的标识信息,并将所述新节点的标识信息记录到所述新节点的属性信息中;Step S40, in the heterogeneous cluster communication state, acquiring each node in the heterogeneous cluster Before the step of the communication data, if it is detected that the new node joins the heterogeneous cluster, the identification information of the new node is obtained, and the identification information of the new node is recorded in the attribute information of the new node;
步骤S50,向所述异构集群的所有节点广播每个节点当前的属性信息,其中,所述异构集群中原先每个节点对应的属性信息都包含原先所有节点的标识信息;In step S50, the current attribute information of each node is broadcasted to all the nodes of the heterogeneous cluster, where the attribute information corresponding to each node in the heterogeneous cluster includes the identification information of all the original nodes;
步骤S60,在所述异构集群中每个节点接收到其它节点广播的属性信息,且接收到的属性信息与每个节点对应的属性信息都一致时,进入异构集群通讯状态。Step S60: In the heterogeneous cluster, each node receives attribute information broadcasted by other nodes, and when the received attribute information is consistent with the attribute information corresponding to each node, the heterogeneous cluster communication state is entered.
在本实施例中,原先节点P1、P2、P3组成的集群正常运行,若此时新节点如P4加入所述异构集群中,则先获取新节点的标识信息,并将所述新节点的标识信息记录到所述新节点的属性信息中,即在所述新节点的属性信息如my_proc_set中记录本节点标识信息;同时,向所述异构集群的所有节点广播每个节点当前的属性信息,以供所述异构集群的所有节点将其属性信息都同步为包含所有节点标识信息的新属性信息。In this embodiment, the clusters of the original nodes P1, P2, and P3 operate normally. If a new node joins the heterogeneous cluster, such as P4, the identification information of the new node is obtained first, and the new node is obtained. The identification information is recorded in the attribute information of the new node, that is, the local node identification information is recorded in the attribute information of the new node, such as my_proc_set; and the current attribute information of each node is broadcast to all nodes of the heterogeneous cluster. For all nodes of the heterogeneous cluster to synchronize their attribute information to new attribute information containing all node identification information.
值得注意的是,所述异构集群中原先每个节点对应的属性信息都包含原先每个节点的标识信息,即所述原先节点P1、P2、P3对应的属性信息都为P1P2P3,为更好理解本实施例,举例如下:在新节点P4加入所述异构集群中,新节点P4广播异构集群动态增删节点的通讯消息如JoinMsg消息,所述JoinMsg消息包括节点P4的属性信息,此时,所述异构集群中原先的3个节点P1、P2、P3接收到所述节点P4广播的JoinMsg消息时,进入gather状态,并且所述节点P1、P2、P3更新自己的属性信息my_proc_set,以将属性信息都同步为包含所有节点标识信息的新属性信息,即将所述节点P1、P2、P3原先的属性信息P1P2P3更新为新属性信息P1P2P3P4,并且所述节点P1、P2、P3也会向集群广播其JoinMsg消息,此时所述节点P4也会在接收到节点P1、P2、P3广播的JoinMsg消息时,将其属性信息更新为新属性信息P1P2P3P4。可以理解的是,在每个节点进行广播JoinMsg消息时,只要节点接收到其它节点广播的JoinMsg消息,就会将JoinMsg消息的属性信息与自身的属性信息进行比对,若比较消息中的my_pro_set和自己的my_proc_set内容相同,即接收到的属性信息与每个节点对应的属性信息一致时,则把JoinMsg消息 中的sender_id设置成一致(consensus),也就是将每个节点的属性信息对应的发送地址信息设置同步标识,并在consensus列表中显示,最后进入异构集群通讯状态。It is to be noted that the attribute information corresponding to each node in the heterogeneous cluster includes the identification information of each node, that is, the attribute information corresponding to the original nodes P1, P2, and P3 is P1P2P3, which is better. An example is as follows: In the heterogeneous cluster, the new node P4 broadcasts a communication message of a heterogeneous cluster dynamic addition and deletion node, such as a JoinMsg message, and the JoinMsg message includes attribute information of the node P4. When the original three nodes P1, P2, and P3 in the heterogeneous cluster receive the JoinMsg message broadcast by the node P4, enter the gather state, and the nodes P1, P2, and P3 update their own attribute information my_proc_set to The attribute information is all synchronized to new attribute information including all node identification information, that is, the original attribute information P1P2P3 of the nodes P1, P2, P3 is updated to new attribute information P1P2P3P4, and the nodes P1, P2, P3 are also clustered. The JoinMsg message is broadcasted. At this time, the node P4 also updates its attribute information to the new attribute information P1P2P3P4 when receiving the JoinMsg message broadcast by the nodes P1, P2, and P3. It can be understood that, when each node performs the broadcast JoinMsg message, as long as the node receives the JoinMsg message broadcast by other nodes, the node will compare the attribute information of the JoinMsg message with its own attribute information, if comparing the my_pro_set and the message in the message. The content of my own my_proc_set is the same, that is, when the received attribute information is consistent with the attribute information corresponding to each node, the JoinMsg message is sent. The sender_id is set to be consistent, that is, the sending address information corresponding to the attribute information of each node is set to be synchronized, displayed in the consensus list, and finally enters the heterogeneous cluster communication state.
可选地,所述方法还可以包括:所述步骤S50之后,在所述异构集群中有节点未接收到其它节点广播的属性信息时,向所述异构集群中其它节点广播所述节点当前的属性信息,。Optionally, the method may further include: after the step S50, when the node does not receive the attribute information broadcast by the other node in the heterogeneous cluster, the node is broadcast to other nodes in the heterogeneous cluster. Current attribute information, .
在异构集群环境中,不是所有的节点的consensus列表都能够迅速更新成一致的,有可能出现的情况是,节点P2没有接收到节点P3广播的JoinMsg消息,因此节点P2的consensus[p3]=false,没有接收P3的JoinMsg消息,此时就会引发重传,也就是进行消息重传,通过向其它节点广播所述节点当前的属性信息,再一次进行消息传递,以供其它节点反馈其它节点对应的属性信息至所述节点中,每个节点进入consensus状态都有一定时间要求。如果是已经在集群中的节点进入consensus状态时间长、接收JoinMsg时间间隔长,则修改在JoinMsg中通知新加入节点减少window_size、扩大token参数。如果是新加入节点进入consensus状态时间长、则稍微扩大consensus定时器时长等等。In a heterogeneous cluster environment, not all nodes' consensus lists can be quickly updated to be consistent. It is possible that node P2 does not receive the JoinMsg message broadcast by node P3, so node P2's consensus[p3]= False, there is no JoinMsg message for receiving P3. At this time, retransmission is triggered, that is, message retransmission is performed. By broadcasting the current attribute information of the node to other nodes, message transmission is performed again for other nodes to feed back other nodes. Corresponding attribute information is sent to the node, and each node has a certain time requirement to enter the consensus state. If the node that has been in the cluster enters the consensus state for a long time and the receiving JoinMsg interval is long, the modification informs the newly added node in JoinMsg to reduce the window_size and expand the token parameter. If it is a new join node entering the consensus state for a long time, then slightly expand the consensus timer duration and so on.
实施例五Embodiment 5
参照图5,图5为本发明实施例的异构集群的通讯优化装置的功能模块示意图。Referring to FIG. 5, FIG. 5 is a schematic diagram of functional modules of a communication optimization apparatus for a heterogeneous cluster according to an embodiment of the present invention.
需要强调的是,对本领域的技术人员来说,图5所示功能模块图仅仅是一个可选实施例的示例图,本领域的技术人员围绕图5所示的异构集群的通讯优化装置的功能模块,可轻易进行新的功能模块的补充;每功能模块的名称是自定义名称,仅用于辅助理解该异构集群的通讯优化装置的每个程序功能块,不用于限定本申请的技术方案,本申请的技术方案的核心是,每个自定义名称的功能模块所要达成的功能。It should be emphasized that, for those skilled in the art, the functional block diagram shown in FIG. 5 is merely an exemplary diagram of an alternative embodiment, and those skilled in the art will surround the heterogeneous cluster communication optimization apparatus shown in FIG. 5. The function module can be easily supplemented by a new function module; the name of each function module is a custom name, which is only used to assist in understanding each program function block of the communication optimization device of the heterogeneous cluster, and is not used to limit the technology of the present application. Solution, the core of the technical solution of the present application is the function to be achieved by each functional module of a custom name.
本实施例提出的一种异构集群的通讯优化装置包括:A communication optimization device for a heterogeneous cluster proposed in this embodiment includes:
获取模块10,设置为在异构集群通讯状态下,获取所述异构集群中每个 节点的通讯数据;Obtaining module 10, configured to acquire each of the heterogeneous clusters in a heterogeneous cluster communication state Communication data of the node;
在本实施例中,所述异构集群中的包含多个节点,本实施例可选所述异构集群包括3个节点,分别为P1、P2和P3,且所述集群的通讯顺序按照P1-P2-P3的顺序进行通讯,而异构集群通讯状态下的通信过程如下:1)所述异构集群的上层应用下发消息序列如M1M2M3,节点P1在获取到所述消息序列后,缓存所述消息序列,等到获取令牌信息token后,将所述令牌信息传递到下一个节点P2,并且此时所述集群传递消息序列,而此时,所述节点P1接收到完整的消息序列,则节点P1通讯正常;2)在节点P2接收到所述令牌信息后,所述节点P2才能传递所述消息序列。值得注意的是,若所述节点P2漏接了消息M3,而令牌信息已经到达节点P2,那么节点P2在所述令牌信息对应的存储列表中记录需要重传的消息M3以及需重传者P2,并将所述令牌信息继续传递到下一个节点P3;3)在所述节点P3获取到令牌信息后,记录令牌信息的存储列表中携带重传需求信息,即所述节点P2需要重传的消息M3,若此时节点P3若能接收到完整的消息序列,则可认为节点P3通讯正常,并随后向集群广播M3消息,且将所述令牌信息继续传递给所述节点P1;4)节点P1获取到所述令牌信息后,记录所述令牌信息的存储列表中携带的重传需求信息即所述节点P2需要重传的消息M3,并将所述令牌信息传递给所述节点P2,由于此时所述节点P1已经接收到完整的消息序列,因此所述节点P1忽略节点P3广播的M3消息;5)节点P2获取到所述令牌信息,发现所述令牌信息中记录自己未接收到的消息M3,那么此时所述节点P2获取节点P3广播的M3消息,以使接收到的消息序列完整,并且所述节点P2清理所述令牌信息中的消息重传信息,至此所述异构集群中所有节点都收到了P1发送的消息序列。In this embodiment, the heterogeneous cluster includes multiple nodes. In this embodiment, the heterogeneous cluster includes three nodes, which are respectively P1, P2, and P3, and the communication sequence of the cluster is in accordance with P1. -P2-P3 is in the order of communication, and the communication process in the heterogeneous cluster communication state is as follows: 1) The upper layer application of the heterogeneous cluster delivers a message sequence such as M1M2M3, and the node P1 caches after obtaining the message sequence. The message sequence, after acquiring the token information token, passes the token information to the next node P2, and at this time, the cluster delivers the message sequence, and at this time, the node P1 receives the complete message sequence. Then, the node P1 communicates normally; 2) after the node P2 receives the token information, the node P2 can deliver the message sequence. It should be noted that if the node P2 misses the message M3 and the token information has arrived at the node P2, the node P2 records the message M3 that needs to be retransmitted and needs to be retransmitted in the storage list corresponding to the token information. P2, and the token information is further passed to the next node P3; 3) after the node P3 obtains the token information, the storage list of the record token information carries the retransmission requirement information, that is, the node P2 needs to retransmit the message M3. If the node P3 can receive the complete message sequence at this time, the node P3 can be considered to be normal communication, and then the M3 message is broadcast to the cluster, and the token information is continuously transmitted to the After the node P1 obtains the token information, the node P1 records the retransmission requirement information carried in the storage list of the token information, that is, the message M3 that the node P2 needs to retransmit, and the token is The information is delivered to the node P2. Since the node P1 has received the complete message sequence at this time, the node P1 ignores the M3 message broadcast by the node P3; 5) the node P2 obtains the token information, and finds the location Recorded from the token information The message M3 has not been received, then the node P2 acquires the M3 message broadcast by the node P3 at this time, so that the received message sequence is complete, and the node P2 clears the message retransmission information in the token information. At this point, all nodes in the heterogeneous cluster have received the message sequence sent by P1.
在本实施例中,可选新增“重发动态统计”模块,用于统计异构集群通讯状态下所述异构集群中每个节点的通讯数据,并且所述获取模块10获取所述异构集群中每个节点的通讯数据的方式可选按照一定的时间间隔获取所述异构集群中每个节点的通讯数据,例如记录类似节点、重传需求次数等的通讯数据,并定时获取每个节点记录的所述通讯数据。In this embodiment, a new retransmission dynamic statistics module is configured to collect communication data of each node in the heterogeneous cluster in a heterogeneous cluster communication state, and the obtaining module 10 obtains the difference. The communication data of each node in the cluster may be configured to obtain communication data of each node in the heterogeneous cluster at a certain time interval, for example, recording communication data of similar nodes, retransmission requirements, etc., and periodically acquiring each The communication data recorded by the nodes.
确定模块20,设置为确定每个通讯数据中的消息重传次数; The determining module 20 is configured to determine the number of retransmissions of the message in each communication data;
调整模块30,设置为在有通讯数据的消息重传次数达到预设次数时,调整所述通讯数据对应的节点的参数;其中,所述参数包括节点接收所述集群的令牌信息的时间。The adjusting module 30 is configured to adjust parameters of the node corresponding to the communication data when the number of retransmissions of the communication data reaches a preset number of times; wherein the parameter includes a time when the node receives the token information of the cluster.
在本实施例中,在所述获取模块10获取到每个节点的通讯数据后,所述确定模块20在异构集群中的每个节点后台直接分析数据,以确定每个节点对应的通讯数据,本实施例可选先对每个节点对应的消息重传次数进行设置,如节点在一段时间内消息重传的次数应小于预设的次数,如果任一个节点存在明显的集中重传特征,也就是说在有通讯数据的消息重传次数达到预设次数时,则所述调整模块30将该节点的参数进行调整,以降低所述节点的消息重传次数,可以理解的是,当有通讯数据的消息重传次数达到预设次数,说明所述通讯数据对应的节点在当前环境下节点消息序列能力较低,有可能是消息序列的通讯速率较快,或者是在节点刚接收到所述消息序列的消息时,令牌信息已经到达该节点了,此时所述节点只能将接收消息序列中的一部分消息,导致消息漏接。因此,本实施例可选所述调整模块30将所述节点的window_size参数减少,将token缓存长度扩大,或者是将token间隔周期参数扩大等等,将所述节点的window_size参数减少可理解为将所述节点的信息发送速率降低,由于异构集群通讯过程中,每一次消息重传后只能优化接下来的操作,因此将所述节点的信息发送速率降低,可以降低整个通讯过程的效率,从而使得所述节点可以接收到消息序列中的消息。In this embodiment, after the obtaining module 10 obtains the communication data of each node, the determining module 20 directly analyzes data in the background of each node in the heterogeneous cluster to determine communication data corresponding to each node. In this embodiment, the number of retransmissions corresponding to each node may be set first. For example, if the number of retransmissions of the message in a period of time is less than the preset number of times, if any node has obvious centralized retransmission characteristics, That is, when the number of retransmissions of the communication data reaches a preset number of times, the adjustment module 30 adjusts the parameters of the node to reduce the number of retransmissions of the node. It is understood that when there is The number of retransmissions of the communication data reaches a preset number, indicating that the node corresponding to the communication data has a lower capability of the node message sequence in the current environment, and may be that the communication rate of the message sequence is faster, or the node just received the node. When the message sequence message is described, the token information has arrived at the node, and the node can only receive a part of the message sequence, resulting in cancellation. Bobble. Therefore, in this embodiment, the adjustment module 30 may reduce the window_size parameter of the node, expand the token cache length, or expand the token interval period parameter, etc., and reduce the window_size parameter of the node may be understood as The information transmission rate of the node is reduced. Because the heterogeneous cluster communication process can only optimize the next operation after each message retransmission, the information transmission rate of the node is reduced, which can reduce the efficiency of the entire communication process. Thereby the node can receive the message in the sequence of messages.
本实施例提出的异构集群的通讯优化装置,在异构集群通讯状态下,根据获取的每个节点的通讯数据,确定每个通讯数据中的消息重传次数,并在有通讯数据的消息重传次数达到预设次数时,调整所述通讯数据对应的节点的参数,以降低所述节点的消息重传次数,如增加所述节点接收所述集群的令牌信息的时间,以降低所述节点的消息重传次数,而不是在异构集群通讯过程中,若有节点未接收到消息时,一直进行消息的重传,本方案通过统计节点的通讯数据中的消息重传次数,并在消息重传次数达到预设次数时,调整所述节点的参数以降低所述节点的消息重传次数,从而提高了集群通讯效率。 The communication optimization device of the heterogeneous cluster proposed in this embodiment determines the number of retransmissions of each communication data and the message with the communication data according to the obtained communication data of each node in the heterogeneous cluster communication state. When the number of retransmissions reaches a preset number of times, the parameters of the node corresponding to the communication data are adjusted to reduce the number of message retransmissions of the node, such as increasing the time at which the node receives the token information of the cluster, so as to reduce the number of times. The number of message retransmissions of the node, rather than during the heterogeneous cluster communication process, if a node does not receive the message, the message is retransmitted, and the scheme retransmits the number of messages in the communication data of the node, and When the number of retransmissions of the message reaches a preset number of times, the parameters of the node are adjusted to reduce the number of retransmissions of the node, thereby improving the communication efficiency of the cluster.
可选地,为了提高异构集群的通讯的灵活性,在本实施例中,参照图6,所述获取模块10包括:Optionally, in order to improve the flexibility of the communication of the heterogeneous cluster, in this embodiment, referring to FIG. 6, the obtaining module 10 includes:
更新单元11,设置为在所述异构集群的每个节点传递消息序列过程中,若检测到有节点未接收到所述消息序列中的预设消息,根据未接收到所述消息序列中的预设消息的节点的标识信息更新所述未接收所述消息序列中的预设消息的节点的消息重传次数;The updating unit 11 is configured to: in the process of transmitting a message sequence in each node of the heterogeneous cluster, if it is detected that a node does not receive the preset message in the message sequence, according to the non-received sequence in the message sequence The identifier information of the node of the preset message updates the number of message retransmissions of the node that does not receive the preset message in the message sequence;
获取单元13,设置为获取每个节点的通讯数据,其中,节点的通讯数据包括所述节点的标识信息以及所述节点对应的消息重传次数。The obtaining unit 13 is configured to acquire communication data of each node, where the communication data of the node includes the identification information of the node and the number of message retransmissions corresponding to the node.
在本实施例中,建立异构集群中每个节点对应的通讯数据表,所述通讯数据表用于记录每个节点的通讯数据,每个节点的通讯数据包括该节点的标识信息以及该节点对应的消息重传次数。也就是说,在所述异构集群的每个节点传递消息序列过程中,所述处理单元12根据每个节点的标识信息记录每个节点对应的消息重传次数,若检测到有节点未接收所述消息序列中的预设消息,所述更新单元11根据所述未接收所述消息序列中的预设消息的节点的标识信息更新所述节点的消息重传次数,若没有检测到节点未接收所述消息序列中的预设消息,那么保持节点的消息重传次数不变。In this embodiment, a communication data table corresponding to each node in the heterogeneous cluster is established, and the communication data table is used to record communication data of each node, and the communication data of each node includes the identification information of the node and the node. The number of corresponding message retransmissions. That is, in the process of transmitting a message sequence by each node of the heterogeneous cluster, the processing unit 12 records the number of message retransmissions corresponding to each node according to the identification information of each node, and if a node is not received, The preset message in the message sequence, the updating unit 11 updates the number of message retransmissions of the node according to the identifier information of the node that does not receive the preset message in the message sequence, if no node is detected yet Receiving the preset message in the message sequence, the number of message retransmissions of the node is kept unchanged.
可选地,为了提高异构集群的通讯的灵活性,在本实施例中,参照图7,所述调整模块30包括:Optionally, in order to improve the flexibility of the communication of the heterogeneous cluster, in this embodiment, referring to FIG. 7, the adjustment module 30 includes:
确定单元31,设置为在有通讯数据的消息重传次数达到预设次数时,确定所述消息重传次数达到预设次数的通讯数据对应的节点;The determining unit 31 is configured to: when the number of retransmissions of the message having the communication data reaches a preset number of times, determine that the number of retransmissions of the message reaches a node corresponding to the communication data of the preset number of times;
调整单元32,设置为增加所述节点接收所述集群的令牌信息的时间。The adjusting unit 32 is configured to increase the time when the node receives the token information of the cluster.
在本实施例中,在有通讯数据的消息重传次数达到预设次数时,所述确定单元31确定所述通讯数据对应的节点,根据前述任一实施例可知,所述节点进行多次消息重传,有可能是消息序列的通讯速率较快,或者是在节点刚接收到所述消息序列的消息时,令牌信息已经到达该节点了,此时所述节点只能将接收消息序列中的一部分消息,导致消息漏接,因此,本实施例可选所述调整单元32将token间隔周期参数扩大,也就是说增加所述节点接收所 述集群的令牌信息的时间,以降低所述节点的消息重传次数,即将所述异构集群中令牌信息传递的时间增加,原先在异构集群中传递令牌信息的时间若是6秒,且所述异构集群中有3个节点,那么每个节点接收到令牌信息的时间即为2秒,现在将令牌信息在所述异构集群中传递的时间增加到9秒,则每个节点接收到令牌信息的时间即为3秒,那么有利于所述节点接收完整的消息序列,从而降低了所述节点的消息重传次数。可选地,在所述调整单元32调整完所述通讯数据对应的节点的参数后,可将所述节点的消息重传次数清零,并在新一轮异构集群通讯过程中,继续统计所述异构集群通讯过程中各个节点的通讯数据。In this embodiment, when the number of retransmissions of the communication data reaches a preset number of times, the determining unit 31 determines the node corresponding to the communication data. According to any of the foregoing embodiments, the node performs multiple times of the message. Retransmission, it may be that the communication sequence of the message sequence is faster, or when the node just receives the message sequence, the token information has arrived at the node, and the node can only receive the message sequence. The part of the message causes the message to be missed. Therefore, the adjusting unit 32 may optionally expand the token interval period parameter, that is, increase the node receiving station. The time of the token information of the cluster is reduced to reduce the number of message retransmissions of the node, that is, the time for transmitting the token information in the heterogeneous cluster is increased, and the time for transmitting the token information in the heterogeneous cluster is 6 seconds. And the three nodes in the heterogeneous cluster, then each node receives the token information for 2 seconds, and now the time for transmitting the token information in the heterogeneous cluster is increased to 9 seconds, then The time at which each node receives the token information is 3 seconds, which is advantageous for the node to receive a complete message sequence, thereby reducing the number of message retransmissions of the node. Optionally, after the adjusting unit 32 adjusts parameters of the node corresponding to the communication data, the number of retransmissions of the node may be cleared, and statistics are continued in a new round of heterogeneous cluster communication. Communication data of each node in the heterogeneous cluster communication process.
可选地,为了提高异构集群的通讯的灵活性,在本实施例中,参照图8,所述异构集群的通讯优化装置包括:Optionally, in order to improve the flexibility of the communication of the heterogeneous cluster, in this embodiment, referring to FIG. 8, the communication optimization apparatus of the heterogeneous cluster includes:
处理模块40,设置为若检测到新节点加入所述异构集群,获取所述新节点的标识信息,并将所述新节点标识信息记录到所述新节点的属性信息中;The processing module 40 is configured to: if it is detected that the new node joins the heterogeneous cluster, acquire the identification information of the new node, and record the new node identifier information into the attribute information of the new node;
广播模块50,设置为向所述异构集群的所有节点广播每个节点当前的属性信息,其中,所述异构集群中原先每个节点对应的属性信息都包含原先所有节点的标识信息;The broadcast module 50 is configured to broadcast the current attribute information of each node to all the nodes of the heterogeneous cluster, where the attribute information corresponding to each node in the heterogeneous cluster includes the identification information of all the original nodes;
进入模块60,设置为在所述异构集群中每个节点接收到其它节点广播的属性信息且接收到的属性信息与每个节点对应的属性信息都一致时,进入异构集群通讯状态。The entering module 60 is configured to enter the heterogeneous cluster communication state when each node in the heterogeneous cluster receives the attribute information broadcasted by other nodes and the received attribute information is consistent with the attribute information corresponding to each node.
在本实施例中,原先节点P1、P2、P3组成的集群正常运行,若此时新节点如P4加入所述异构集群中,则所述处理模块40先获取新节点的标识信息,并将所述新节点的标识信息记录到所述新节点的属性信息中,即在所述新节点的属性信息如my_proc_set中记录本节点标识信息;同时,所述广播模块50向所述异构集群的所有节点广播每个节点当前的属性信息,以供所述异构集群的所有节点将其属性信息都同步为包含所有节点标识信息的新属性信息。In this embodiment, the clusters of the original nodes P1, P2, and P3 are in normal operation. If a new node, such as P4, joins the heterogeneous cluster, the processing module 40 first obtains the identification information of the new node, and The identification information of the new node is recorded in the attribute information of the new node, that is, the local node identification information is recorded in the attribute information of the new node, such as my_proc_set; meanwhile, the broadcast module 50 is to the heterogeneous cluster. All nodes broadcast the current attribute information of each node, so that all nodes of the heterogeneous cluster synchronize their attribute information into new attribute information including all node identification information.
值得注意的是,所述异构集群中原先每个节点对应的属性信息都包含原先每个节点的标识信息,即所述原先节点P1、P2、P3对应的属性信息都为 P1P2P3,为更好理解本实施例,举例如下:在新节点P4加入所述异构集群中,新节点P4广播异构集群动态增删节点的通讯消息如JoinMsg消息,所述JoinMsg消息包括节点P4的属性信息,此时,所述异构集群中原先的3个节点P1、P2、P3接收到所述节点P4广播的JoinMsg消息时,进入gather状态,并且所述节点P1、P2、P3更新自己的属性信息my_proc_set,以将属性信息都同步为包含所有节点标识信息的新属性信息,即将所述节点P1、P2、P3原先的属性信息P1P2P3更新为新属性信息P1P2P3P4,并且所述节点P1、P2、P3也会向集群广播其JoinMsg消息,此时所述节点P4也会在接收到节点P1、P2、P3广播的JoinMsg消息时,将其属性信息更新为新属性信息P1P2P3P4,可以理解的是,在各个节点进行广播JoinMsg消息时,只要节点接收到其它节点广播的JoinMsg消息,就会将JoinMsg消息的属性信息与自身的属性信息进行比对,若比较消息中的my_pro_set和自己的my_proc_set内容相同,即接收到的属性信息与各个节点对应的属性信息一致时,则把JoinMsg消息中的sender_id设置成一致(consensus),也就是将各个节点的属性信息对应的发送地址信息设置同步标识,并在consensus列表中显示,最后所述进入模块60进入异构集群通讯状态。It is to be noted that the attribute information corresponding to each node in the heterogeneous cluster includes the identification information of each node, that is, the attribute information corresponding to the original nodes P1, P2, and P3 are P1P2P3, for a better understanding of the embodiment, for example: when the new node P4 joins the heterogeneous cluster, the new node P4 broadcasts a communication message of the heterogeneous cluster dynamic addition and deletion node, such as a JoinMsg message, and the JoinMsg message includes the node P4. Attribute information. At this time, when the original three nodes P1, P2, and P3 in the heterogeneous cluster receive the JoinMsg message broadcast by the node P4, the state enters the gather state, and the nodes P1, P2, and P3 update their own. The attribute information my_proc_set is used to synchronize the attribute information to new attribute information including all node identification information, that is, the original attribute information P1P2P3 of the nodes P1, P2, P3 is updated to new attribute information P1P2P3P4, and the nodes P1, P2 are P3 also broadcasts its JoinMsg message to the cluster. At this time, the node P4 also updates its attribute information to the new attribute information P1P2P3P4 when receiving the JoinMsg message broadcast by the nodes P1, P2, and P3. It can be understood that When each node broadcasts a JoinMsg message, as long as the node receives the JoinMsg message broadcast by other nodes, it will attribute the attribute information of the JoinMsg message and its own attribute information. For comparison, if the content of my_pro_set in the comparison message is the same as the content of its own my_proc_set, that is, the received attribute information is consistent with the attribute information corresponding to each node, the sender_id in the JoinMsg message is set to be consistent, that is, The sending address information corresponding to the attribute information of each node sets the synchronization identifier, and is displayed in the consensus list. Finally, the entering module 60 enters the heterogeneous cluster communication state.
可选地,所述广播模块50,还设置为在所述异构集群中有节点未接收到其它节点广播的属性信息时,向所述异构集群中其它节点广播所述节点当前的属性信息。Optionally, the broadcast module 50 is further configured to: when the node in the heterogeneous cluster does not receive the attribute information broadcast by the other node, broadcast the current attribute information of the node to other nodes in the heterogeneous cluster. .
在异构集群环境中,不是所有的节点的consensus列表都能够迅速更新成一致的,有可能出现的情况是,节点P2没有接收到节点P3广播的JoinMsg消息,因此节点P2的consensus[p3]=false,没有接收P3的JoinMsg消息,此时就会引发重传,也就是进行消息重传,所述广播模块50通过向其它节点广播所述节点当前的属性信息,再一次进行消息传递,以供其它节点反馈其它节点对应的属性信息至所述节点中,每个节点进入consensus状态都有一定时间要求。如果是已经在集群中的节点进入consensus状态时间长、接收JoinMsg时间间隔长,则修改在JoinMsg中通知新加入节点减少window_size、扩大token参数。如果是新加入节点进入consensus状态时间长、则稍微扩大consensus定时器时长等等。 In a heterogeneous cluster environment, not all nodes' consensus lists can be quickly updated to be consistent. It is possible that node P2 does not receive the JoinMsg message broadcast by node P3, so node P2's consensus[p3]= False, there is no JoinMsg message for receiving P3. At this time, retransmission is triggered, that is, message retransmission is performed. The broadcast module 50 broadcasts the current attribute information of the node to other nodes, and then transmits the message again for Other nodes feed back the attribute information corresponding to other nodes to the node, and each node has a certain time requirement to enter the consensus state. If the node that has been in the cluster enters the consensus state for a long time and the receiving JoinMsg interval is long, the modification informs the newly added node in JoinMsg to reduce the window_size and expand the token parameter. If it is a new join node entering the consensus state for a long time, then slightly expand the consensus timer duration and so on.
需要说明的是,在本文中,术语“包括”、“包含”或者其任何其它变体意在涵盖非排他性的包含,从而使得包括一系列要素的过程、方法、物品或者装置不仅包括那些要素,而且还包括没有明确列出的其它要素,或者是还包括为这种过程、方法、物品或者装置所固有的要素。在没有更多限制的情况下,由语句“包括一个……”限定的要素,并不排除在包括该要素的过程、方法、物品或者装置中还存在另外的相同要素。It is to be understood that the term "comprises", "comprising", or any other variants thereof, is intended to encompass a non-exclusive inclusion, such that a process, method, article, or device comprising a series of elements includes those elements. It also includes other elements not explicitly listed, or elements that are inherent to such a process, method, article, or device. An element that is defined by the phrase "comprising a ..." does not exclude the presence of additional equivalent elements in the process, method, item, or device that comprises the element.
上述本发明实施例序号仅仅为了描述,不代表实施例的优劣。The serial numbers of the embodiments of the present invention are merely for the description, and do not represent the advantages and disadvantages of the embodiments.
本发明实施例还提供一种计算机可读存储介质,所述计算机可读存储介质中存储有计算机可执行指令,所述计算机可执行指令被执行时实现异构集群的通讯优化方法。The embodiment of the present invention further provides a computer readable storage medium, where the computer readable storage medium stores computer executable instructions, and the computer executable instructions are executed to implement a communication optimization method of a heterogeneous cluster.
本领域普通技术人员可以理解上述方法中的全部或部分步骤可通过程序来指令相关硬件(例如处理器)完成,所述程序可以存储于计算机可读存储介质中,如只读存储器、磁盘或光盘等。可选地,上述实施例的全部或部分步骤也可以使用一个或多个集成电路来实现。相应地,上述实施例中的各模块/单元可以采用硬件的形式实现,例如通过集成电路来实现其相应功能,也可以采用软件功能模块的形式实现,例如通过处理器执行存储于存储器中的程序/指令来实现其相应功能。本申请不限制于任何特定形式的硬件和软件的结合。本领域的普通技术人员应当理解,可以对本申请的技术方案进行修改或者等同替换,而不脱离本申请技术方案的精神和范围,均应涵盖在本申请的权利要求范围当中。One of ordinary skill in the art will appreciate that all or a portion of the above steps may be performed by a program to instruct related hardware, such as a processor, which may be stored in a computer readable storage medium, such as a read only memory, disk or optical disk. Wait. Alternatively, all or part of the steps of the above embodiments may also be implemented using one or more integrated circuits. Correspondingly, each module/unit in the above embodiment may be implemented in the form of hardware, for example, by implementing an integrated circuit to implement its corresponding function, or may be implemented in the form of a software function module, for example, executing a program stored in the memory by a processor. / instruction to achieve its corresponding function. This application is not limited to any specific combination of hardware and software. A person skilled in the art should understand that the technical solutions of the present application can be modified or equivalent, without departing from the spirit and scope of the technical solutions of the present application, and should be included in the scope of the claims of the present application.
工业实用性Industrial applicability
上述技术方案降低了所述节点的消息重传次数,从而提高了集群通讯效率。 The above technical solution reduces the number of message retransmissions of the node, thereby improving the communication efficiency of the cluster.

Claims (10)

  1. 一种异构集群的通讯优化方法,所述异构集群的通讯优化方法包括:A communication optimization method for a heterogeneous cluster, wherein the communication optimization method of the heterogeneous cluster includes:
    在异构集群通讯状态下,获取所述异构集群中每个节点的通讯数据;Obtaining communication data of each node in the heterogeneous cluster in a heterogeneous cluster communication state;
    确定每个通讯数据中的消息重传次数;Determine the number of message retransmissions in each communication data;
    在有通讯数据的消息重传次数达到预设次数时,调整所述通讯数据对应的节点的参数;其中,所述参数包括节点接收所述集群的令牌信息的时间。The parameter of the node corresponding to the communication data is adjusted when the number of retransmissions of the communication data reaches a preset number of times; wherein the parameter includes a time when the node receives the token information of the cluster.
  2. 如权利要求1所述的异构集群的通讯优化方法,其中,所述在异构集群通讯状态下,获取所述异构集群中每个节点的通讯数据的步骤包括:The communication optimization method of the heterogeneous cluster according to claim 1, wherein the step of acquiring communication data of each node in the heterogeneous cluster in the heterogeneous cluster communication state comprises:
    在所述异构集群的节点传递消息序列的过程中,若检测到有节点未接收到所述消息序列中的预设消息,则根据所述未接收到所述消息序列中的预设信息的节点的标识信息更新所述节点的消息重传次数;In the process of transmitting the message sequence by the node of the heterogeneous cluster, if it is detected that a node does not receive the preset message in the message sequence, according to the preset information that is not received in the message sequence, The identification information of the node updates the number of message retransmissions of the node;
    获取各个节点的通讯数据,其中,节点的通讯数据包括所述节点的标识信息以及所述节点对应的消息重传次数。Obtaining communication data of each node, where the communication data of the node includes identification information of the node and a number of message retransmissions corresponding to the node.
  3. 如权利要求1所述的异构集群的通讯优化方法,其中,所述在有通讯数据的消息重传次数达到预设次数时,调整所述通讯数据对应的节点的参数的步骤包括:The communication optimization method of the heterogeneous cluster according to claim 1, wherein the step of adjusting the parameters of the node corresponding to the communication data when the number of retransmissions of the communication data reaches a preset number of times comprises:
    在有通讯数据的消息重传次数达到预设次数时,确定所述消息重传次数达到预设次数的通讯数据对应的节点;When the number of retransmissions of the communication data reaches a preset number of times, determining that the number of retransmissions of the message reaches a node corresponding to the preset number of communication data;
    增加所述节点接收所述集群的令牌信息的时间。Increasing the time at which the node receives the token information of the cluster.
  4. 如权利要求1-3任一项所述的异构集群的通讯优化方法,所述方法还包括:所述在异构集群通讯状态下,获取所述异构集群中每个节点的通讯数据的步骤之前若检测到新节点加入所述异构集群,获取所述新节点的标识信息,并将所述新节点的标识信息记录到所述新节点的属性信息中;The method for optimizing communication of a heterogeneous cluster according to any one of claims 1 to 3, further comprising: obtaining, in a heterogeneous cluster communication state, communication data of each node in the heterogeneous cluster If the new node is detected to join the heterogeneous cluster, the identification information of the new node is obtained, and the identification information of the new node is recorded in the attribute information of the new node.
    向所述异构集群的所有节点广播每个节点当前的属性信息,其中,所述异构集群中原先每个节点对应的属性信息都包含原先所有节点的标识信息;Broadcasting the current attribute information of each node to all the nodes of the heterogeneous cluster, where the attribute information corresponding to each node in the heterogeneous cluster includes the identification information of all the original nodes;
    在所述异构集群中每个节点接收到其它节点广播的属性信息且接收到的属性信息与每个节点对应的属性信息都一致时,进入异构集群通讯状态。 In the heterogeneous cluster, each node receives the attribute information broadcasted by other nodes, and the received attribute information is consistent with the attribute information corresponding to each node, and enters the heterogeneous cluster communication state.
  5. 如权利要求4所述的异构集群的通讯优化方法,所述方法还包括:所述向所述异构集群的所有节点广播每个节点当前的属性信息的步骤之后,所述异构集群的通讯优化方法还包括:The method for optimizing communication of a heterogeneous cluster according to claim 4, further comprising: after the step of broadcasting current attribute information of each node to all nodes of the heterogeneous cluster, the heterogeneous cluster Communication optimization methods also include:
    在所述异构集群中有节点未接收到其它节点广播的属性信息时,向所述异构集群中其它节点广播所述节点当前的属性信息。When the node does not receive the attribute information broadcast by other nodes in the heterogeneous cluster, the current attribute information of the node is broadcast to other nodes in the heterogeneous cluster.
  6. 一种异构集群的通讯优化装置,所述异构集群的通讯优化装置包括:A communication optimization device for a heterogeneous cluster, the communication optimization device of the heterogeneous cluster includes:
    获取模块,设置为在异构集群通讯状态下,获取所述异构集群中每个节点的通讯数据;Obtaining a module, configured to acquire communication data of each node in the heterogeneous cluster in a heterogeneous cluster communication state;
    确定模块,设置为确定每个通讯数据中的消息重传次数;Determining a module, set to determine the number of retransmissions of messages in each communication data;
    调整模块,设置为在有通讯数据的消息重传次数达到预设次数时,调整所述通讯数据对应的节点的参数;其中,所述参数包括节点接收所述集群的令牌信息的时间。The adjusting module is configured to adjust a parameter of the node corresponding to the communication data when the number of retransmissions of the communication data reaches a preset number of times; wherein the parameter includes a time when the node receives the token information of the cluster.
  7. 如权利要求6所述的异构集群的通讯优化装置,其中,所述获取模块包括:The communication optimization device of the heterogeneous cluster of claim 6, wherein the obtaining module comprises:
    更新单元,设置为在所述异构集群的每个节点传递消息序列过程中,若检测到有节点未接收到所述消息序列中的预设消息,则根据所述未接收到所述消息序列中的预设消息的节点的标识信息更新所述节点的消息重传次数;An update unit, configured to: in the process of transmitting a message sequence in each node of the heterogeneous cluster, if it is detected that a node does not receive the preset message in the message sequence, according to the message sequence not received The identification information of the node of the preset message updates the number of message retransmissions of the node;
    获取单元,设置为获取各个节点的通讯数据,其中,节点的通讯数据包括所述节点的标识信息以及所述节点对应的消息重传次数。The obtaining unit is configured to obtain communication data of each node, where the communication data of the node includes the identification information of the node and the number of message retransmissions corresponding to the node.
  8. 如权利要求6所述的异构集群的通讯优化装置,其中,所述调整模块包括:The communication optimization device of the heterogeneous cluster of claim 6, wherein the adjustment module comprises:
    确定单元,设置为在有通讯数据的消息重传次数达到预设次数时,确定所述消息重传次数达到预设次数的通讯数据对应的节点;a determining unit, configured to determine, when the number of retransmissions of the message having the communication data reaches a preset number of times, determining, by the node corresponding to the communication data that the number of retransmissions of the message reaches a preset number of times;
    调整单元,设置为增加所述节点接收所述集群的令牌信息的时间。The adjusting unit is configured to increase a time when the node receives the token information of the cluster.
  9. 如权利要求6-8任一项所述的异构集群的通讯优化装置,所述异构集群的通讯优化装置还包括:The communication optimization device of the heterogeneous cluster according to any one of claims 6-8, wherein the communication optimization device of the heterogeneous cluster further comprises:
    处理模块,设置为若检测到新节点加入所述异构集群,获取所述新节点 的标识信息,并将所述新节点的标识信息记录到所述新节点的属性信息中;a processing module, configured to acquire the new node if it detects that a new node joins the heterogeneous cluster Identification information, and recording the identification information of the new node into the attribute information of the new node;
    广播模块,用于向所述异构集群的所有节点广播每个节点当前的属性信息,其中,所述异构集群中原先每个节点对应的属性信息都包含原先所有节点的标识信息;a broadcast module, configured to broadcast, to all the nodes of the heterogeneous cluster, current attribute information of each node, where the attribute information corresponding to each node in the heterogeneous cluster includes identifier information of all the original nodes;
    进入模块,设置为在所述异构集群中每个节点接收到其它节点广播的属性信息且接收到的属性信息与每个节点对应的属性信息都一致时,进入异构集群通讯状态。The entry module is configured to enter the heterogeneous cluster communication state when each node in the heterogeneous cluster receives the attribute information broadcasted by other nodes and the received attribute information is consistent with the attribute information corresponding to each node.
  10. 如权利要求9所述的异构集群的通讯优化装置,所述广播模块,还设置为在所述异构集群中有节点未接收到其它节点广播的属性信息时,向所述异构集群中其它节点广播所述节点当前的属性信息。 The communication optimization device of the heterogeneous cluster according to claim 9, wherein the broadcast module is further configured to: when a node in the heterogeneous cluster does not receive attribute information broadcast by other nodes, to the heterogeneous cluster Other nodes broadcast the current attribute information of the node.
PCT/CN2016/085335 2015-09-30 2016-06-08 Method and device for optimizing communication in heterogeneous cluster WO2017054512A1 (en)

Applications Claiming Priority (2)

Application Number Priority Date Filing Date Title
CN201510638932.3 2015-09-30
CN201510638932.3A CN106559458A (en) 2015-09-30 2015-09-30 The communication optimization method and device of isomeric group

Publications (1)

Publication Number Publication Date
WO2017054512A1 true WO2017054512A1 (en) 2017-04-06

Family

ID=58417337

Family Applications (1)

Application Number Title Priority Date Filing Date
PCT/CN2016/085335 WO2017054512A1 (en) 2015-09-30 2016-06-08 Method and device for optimizing communication in heterogeneous cluster

Country Status (2)

Country Link
CN (1) CN106559458A (en)
WO (1) WO2017054512A1 (en)

Cited By (1)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN109428688A (en) * 2017-09-04 2019-03-05 中国移动通信有限公司研究院 A kind of HARQ number of retransmissions Dynamic Configuration, device and medium

Citations (4)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN101594308A (en) * 2009-07-03 2009-12-02 华为技术有限公司 Message transmitting method and system
CN102647323A (en) * 2012-03-28 2012-08-22 华为技术有限公司 Flow control method and device as well as clustering system
WO2014031989A1 (en) * 2012-08-23 2014-02-27 Interdigital Patent Holdings, Inc. Operating with multiple schedulers in a wireless system
CN104469966A (en) * 2014-12-08 2015-03-25 哈尔滨工业大学深圳研究生院 Random access method and system of TD-LTE virtual private network based on dynamic priorities

Patent Citations (4)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN101594308A (en) * 2009-07-03 2009-12-02 华为技术有限公司 Message transmitting method and system
CN102647323A (en) * 2012-03-28 2012-08-22 华为技术有限公司 Flow control method and device as well as clustering system
WO2014031989A1 (en) * 2012-08-23 2014-02-27 Interdigital Patent Holdings, Inc. Operating with multiple schedulers in a wireless system
CN104469966A (en) * 2014-12-08 2015-03-25 哈尔滨工业大学深圳研究生院 Random access method and system of TD-LTE virtual private network based on dynamic priorities

Cited By (2)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN109428688A (en) * 2017-09-04 2019-03-05 中国移动通信有限公司研究院 A kind of HARQ number of retransmissions Dynamic Configuration, device and medium
CN109428688B (en) * 2017-09-04 2021-11-09 中国移动通信有限公司研究院 HARQ retransmission time dynamic configuration method, device and medium

Also Published As

Publication number Publication date
CN106559458A (en) 2017-04-05

Similar Documents

Publication Publication Date Title
EP2515501B1 (en) Media clock negotiation
WO2012124128A1 (en) Information processing device, sensor system, program and recording medium
CN103544303B (en) A kind of method of data synchronization, system and equipment
JP2002124935A (en) Method and system for controlling retransmission for multicast distribution service, retransmission controller, radio base station and radio terminal equipment
JP2002141903A (en) Node configuration information management method and wireless network system
CN101507183B (en) Decoupled connections
CN103096183B (en) A kind of highly efficient stream media transmission method
CN110881135A (en) Method, device, equipment and medium for optimizing microphone-connected transmission protocol
WO2020051862A1 (en) Broadcast storm prevention method and apparatus
WO2015096802A1 (en) Method, device, and server for message transmission
WO2017054512A1 (en) Method and device for optimizing communication in heterogeneous cluster
US9781691B2 (en) Wireless communication device, non-transitory computer readable medium, and wireless communication system
US9161287B2 (en) Technique for efficient message delivery in Ad Hoc, mesh, wireless computer networks
US20110137971A1 (en) Data collection method and system
WO2018219305A1 (en) Message retransmission method and device
Maxemchuk Reliable multicast with delay guarantees
WO2015188481A1 (en) Unread message processing method, device and system
US9307509B1 (en) Transmitting radio signals using simultaneous retransmission
WO2016165289A1 (en) Inter-board data processing method and device
US10756844B2 (en) Devices and method for wirelessly broadcasting media packets
US20180199090A1 (en) Information processing system and server apparatus
CN111262706B (en) Data transmission method, server and storage device
JP2015080197A (en) Communication control device and communication control method
WO2018036005A1 (en) Data packet transmission control method, base station, terminal, and system
WO2016090804A1 (en) Information acquisition method, device and router

Legal Events

Date Code Title Description
121 Ep: the epo has been informed by wipo that ep was designated in this application

Ref document number: 16850137

Country of ref document: EP

Kind code of ref document: A1

NENP Non-entry into the national phase

Ref country code: DE

122 Ep: pct application non-entry in european phase

Ref document number: 16850137

Country of ref document: EP

Kind code of ref document: A1