WO2013170592A1 - 一种令牌周转控制方法、装置及系统 - Google Patents

一种令牌周转控制方法、装置及系统 Download PDF

Info

Publication number
WO2013170592A1
WO2013170592A1 PCT/CN2012/084234 CN2012084234W WO2013170592A1 WO 2013170592 A1 WO2013170592 A1 WO 2013170592A1 CN 2012084234 W CN2012084234 W CN 2012084234W WO 2013170592 A1 WO2013170592 A1 WO 2013170592A1
Authority
WO
WIPO (PCT)
Prior art keywords
node
token
turnover
record information
message
Prior art date
Application number
PCT/CN2012/084234
Other languages
English (en)
French (fr)
Inventor
张明
Original Assignee
华为技术有限公司
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 华为技术有限公司 filed Critical 华为技术有限公司
Publication of WO2013170592A1 publication Critical patent/WO2013170592A1/zh

Links

Classifications

    • HELECTRICITY
    • H04ELECTRIC COMMUNICATION TECHNIQUE
    • H04LTRANSMISSION OF DIGITAL INFORMATION, e.g. TELEGRAPHIC COMMUNICATION
    • H04L12/00Data switching networks
    • H04L12/28Data switching networks characterised by path configuration, e.g. LAN [Local Area Networks] or WAN [Wide Area Networks]
    • H04L12/42Loop networks
    • H04L12/427Loop networks with decentralised control
    • H04L12/433Loop networks with decentralised control with asynchronous transmission, e.g. token ring, register insertion

Definitions

  • the present invention relates to the field of token turnover control, and in particular, to a token turnover control method, apparatus and system. Background technique
  • the totem single-ring ordered protocol provides a mechanism for sending messages to the globally ordered upper-layer services in the domain. This means that the message requires reliability and order guarantee.
  • On top of the sending domain is a logical token passing ring.
  • the token passed in the logical token passing ring is used to control access to the ring, and only The node holding the token can send a message.
  • Each node within the logical token delivery ring can send multiple messages each time it gets a token. When all nodes have no messages to send, the token continues to loop inside the ring.
  • the node needs to wait for the token to send the message each time. This causes the node to send a message with a certain delay.
  • Embodiments of the present invention provide a token turnover control method, apparatus, and system to reduce message delay and improve message throughput.
  • an embodiment of the present invention provides a token turnover control method, including:
  • an embodiment of the present invention further provides a token turnover method, including:
  • the message to be sent is sent; based on the number of messages sent by the node in the token rotation process of the current round, the number of messages sent in the current round is recorded to In the token turnover record information corresponding to the node identifier in the token;
  • the updated token is delivered to the corresponding next node according to the node turnover sequence indicated by the node turnover record information in the token.
  • the embodiment of the present invention further provides a token turnover control apparatus, including: an ear module, configured to acquire each node in the current logical token delivery ring in the past M round token rotation process History information of the message transmission status;
  • an update module configured to update, according to the history information, a node rotation sequence indicated in the node turnover record information included in the token, where the node rotation record information is used to indicate that the token is in a logical token delivery ring Node rotation order within;
  • a sending module configured to deliver the updated token to the corresponding next node according to a node rotation sequence indicated by the updated node turnover record information.
  • an embodiment of the present invention further discloses a node that sends a message based on token turnover, including:
  • a communication module configured to receive a to-be-sent message after receiving the token passed in the logical token delivery ring
  • a recording module configured to record, according to the number of messages sent by the node in the token rotation process of the current round, the number of messages sent in the current round to the node turnover record information corresponding to the node identifier in the token ;
  • the communication module is further configured to deliver the updated token to the corresponding next node according to the node rotation sequence indicated by the node turnover record information in the token.
  • the embodiment of the present invention further provides a cluster system using a token ring mechanism, including multiple nodes, where the multiple nodes include: a master node and a slave node, where
  • the master node is configured to acquire history record information of a message sent by each node in the current logical token delivery ring during a past M round of token turnover; and according to the historical record information, the node included in the token
  • the node rotation sequence indicated in the turnover record information is updated, and the node rotation record information is used to indicate a node rotation order of the token in the logical token delivery ring;
  • the node turnover sequence indicated by the subsequent node turnover record information transfers the updated token to the corresponding next node;
  • the slave node after receiving the token passed in the logical token delivery ring, sends a message to be sent, based on the number of messages sent by the node in the token turnover process of the current round, and the message sent by the current round
  • the quantity is recorded in the node turnover record information corresponding to the node identifier in the token, and the updated token is delivered to the corresponding one according to the node rotation order indicated by the node turnover record information in the token.
  • the order of the tokens in the nodes can be updated according to the historical transmission status of each node, which can enable the nodes with message sending requirements to send messages faster, reduce the backlog of messages on these nodes, reduce message delay, and improve node performance. And availability, improved message throughput.
  • FIG. 1 is a schematic flow chart of a first embodiment of a token turnover control method of the present invention
  • FIG. 2 is a schematic flow chart of an embodiment of a token turnover method according to an embodiment of the present invention
  • FIG. 3 is a schematic flow chart of a second embodiment of a token turnover control method of the present invention.
  • FIG. 4 is a schematic diagram showing the sequence of transmission of a token in a logical token delivery ring in a node according to an embodiment of the present invention
  • FIG. 5 is a schematic structural view of a first embodiment of a token turnover control device of the present invention
  • FIG. 6 is a schematic structural view of a second embodiment of the token turnover control device of the present invention
  • FIG. 7 is a token-based turnover of the present invention.
  • FIG. 8 is a schematic structural diagram of a cluster system using a token ring mechanism according to the present invention.
  • FIG. 9 is a schematic structural diagram of a node included in a cluster system using the token ring mechanism of the present invention shown in FIG. 8.
  • the technical solutions in the embodiments of the present invention are clearly and completely described in the following with reference to the accompanying drawings in the embodiments of the present invention. It is obvious that the described embodiments are only a part of the embodiments of the present invention, but not all of the embodiments. All other embodiments obtained by those skilled in the art based on the embodiments of the present invention without creative efforts are within the scope of the present invention.
  • a logical token delivery ring includes multiple nodes, including a responsible token creation token originating node.
  • the token initiating node also maintains in the token a node turnover record information indicating the node rotation order of the token within the token delivery ring.
  • the token is passed between the nodes in the logical token delivery ring, and each node on the ring sends a message when receiving the token.
  • the method of this embodiment specifically includes:
  • S101 Obtain historical record information of message sending status of each node in the current logical token delivery ring during past M round token turnover.
  • the manner in which the S101 obtains the record information of the historical transmission status of each node in the logical token delivery ring may be:
  • the token originating node may save the node turnover record information in the token, and the node turnover record information may be a node turnover table.
  • the form indicates that, in the process of issuing the token for the turnaround, each node records the number of messages sent in the current round to the token corresponding to the node identifier after acquiring the token and sending the message.
  • the node identifier may be a name of a corresponding node, or other identifier that can uniquely represent a node.
  • the token initiating node After each round of token turnover, when the token returns to the token originating node, the token initiating node periodically updates and records the historical transmission status of each node according to the node turnover record information obtained after the current round of looping. After the M round of token turnover ends, the statistical result recorded for each node is the history information of the message transmission status of each node.
  • the value of the M can be adjusted and changed according to the specific conditions of the nodes in the logical token delivery ring, so as to better reduce the message transmission delay of each node.
  • the method for obtaining the record information of the historical transmission status of each node in the S101 may be: setting the save node turnover record information directly in the token, after obtaining the token in each round, according to the token
  • the number of message transmissions of each node recorded in the node turnover record information is updated.
  • the number of accumulated messages corresponding to each node in the history record information locally maintained by the token initiation node is initiated; after the M round of token turnover ends,
  • the number of accumulated message transmissions corresponding to each node recorded in the history information is used as a history of message transmission status of each node in the logical token delivery ring.
  • record information may be expressed in the form of a history record table, and may also be represented by a script file or metadata.
  • S102 Update, according to the history information, a node rotation sequence indicated in the node turnover record information included in the token.
  • the node turnover record information is used to indicate a node rotation order of the token within the logical token delivery ring.
  • determining that the logical token delivery ring is in the historical point rotation sequence skipping these rarely sending messages or not sending the message during the subsequent N rounds of token turnover
  • a node that is, does not pass a token to a node that transmits or does not send a message, exploits the right of these nodes to send a message, facilitates sending of data by other nodes that frequently send messages, and reduces the message transmission delay on the node that normally sends the message.
  • S103 Pass the updated token to the corresponding next node according to the node rotation order indicated by the updated node turnover record information. That is, according to the new node turnover sequence, some nodes that skip the sending or not sending the message are skipped.
  • the present invention has the following advantages:
  • the order of rotation of the tokens in the nodes can be updated according to the historical transmission status of each node, so that nodes with larger message transmission requirements can send messages faster, reduce the backlog of messages on these nodes, reduce message delay, and improve Node performance and availability, improved message throughput.
  • a token turnover method performed by each node in a logical token delivery ring includes:
  • the current node After receiving the token passed in the logical token delivery ring, the current node sends a message to be sent.
  • the node can send one or more messages.
  • S202 Record, according to the number of messages sent by the node in the token rotation process of the current round, the number of messages sent in the current round to the node turnover record information corresponding to the identifier of the node in the token.
  • the node turnover record information is stored in the token, and the node turnover record information can be expressed in the form of a node turnover table.
  • S203 Pass the updated token to the corresponding next node according to the node turnover sequence indicated by the node turnover record information in the token.
  • the token in the embodiment of the present invention stores the node turnover record information. After each token obtains the token sending message in each round, the token records the number of messages sent in the current round in the node rotation sequence. So that the token initiating node within the logical token passing ring can arrange the node turnover order according to the quantity information recorded in the node's turnover sequence.
  • FIG. 3 is a schematic flowchart of the second embodiment of the method for controlling the token turnover of the present invention.
  • the method specifically includes:
  • the token initiating node records, as the logical token, the history of the message transmission status of each node in the ring, after the M round of token turnover ends, the number of accumulated messages corresponding to each node recorded in the history record information. information.
  • the value of the M is a preset value, which can be modified and adjusted according to user requirements.
  • the token initiating node saves and maintains the node turnover record information in the token while creating the token.
  • the node rotation distance information may be represented in the form of a node turnover table.
  • the node turnover table may include the node identifier and The node identifies the entry of the number of the current round of the corresponding node, and in this embodiment, the node turnover table further includes a turnaround flag, where the turnaround flag is used to indicate that the corresponding node sends a message or does not send a message.
  • the cold node, or the normal node that normally sends the message when the turnaround flag is set to 1, the corresponding node is the normal node, the token loops through the node, and when the turnaround flag is 0, the corresponding node is the cold node. , the node is skipped when the token loops.
  • the node turnover table indicates the node rotation order in the following manner: The node rotation order is determined according to the order of the node identifiers recorded in the entry and the turnaround flag thereof.
  • the token initiating node maintains the node turnover record information in the token, and locally maintains and maintains history record information, where the history record information includes a node identifier, a cumulative message sending quantity, and a turnaround flag bit, where the token originating node is in the The cumulative amount of messages sent by each node is recorded in the history information.
  • the history record information is represented in the form of a history record table.
  • each node obtains the token and sends a message, and records the number of the current sent message to the entry corresponding to the node on the node turnover record information. in.
  • FIG. 4 is A logical token is a schematic diagram of the delivery sequence of the tokens in the ring in the node.
  • the node A is the token initiator on the logical token delivery ring.
  • the S301 and the S302 may specifically include the following steps:
  • the turnaround identifier of each node is 1, and the node rotation sequence indicated by the node turnover record information is node A->node B->node C->node D- in FIG. >Node A.
  • node A After node A obtains the token, it sends 10 messages, and node A records 10 of the corresponding number of messages sent in the node turnover table shown in Table 1.
  • the node A delivers the token to the node B, and the node B sends 20 messages after acquiring the token, and the node B passes the token to the node C, and the node C does not send a message, and the record The number is 0, node C passes the token to the node D, the node D does not send a message, and the number of records is zero. Node D finally passes the token back to the node A.
  • the token turnover table in the token is as described in Table 1.
  • the node A which is the token initiating node, counts the historical transmission status of each node in the current round of the token turnover process, that is, the cumulative number of messages sent in the history table according to the record of the node turnover table, and obtains the number of messages sent as shown in Table 2 History table.
  • the node turnover table is initialized, and then the message is sent as needed and the number of current rounds of messages is recorded in the node turnover record information, and the next round of token turnover is entered, and after the token is obtained, the token is obtained. Update each node in the history information again based on the node turnover record information. The corresponding cumulative number of messages sent.
  • the token initiating node can determine the cumulative number of message transmissions corresponding to each node according to the history information as shown in Table 3 above, as described in Table 3: In the past M rounds In the card cycle, the token passes through each node, and the number of messages sent by the corresponding node is: Node A is 30, Node B is 45, and Node C and Node D are 0.
  • S303 Record a node whose number of accumulated message transmissions recorded in the history information after the end of the M round of token turnover is lower than or equal to a preset number threshold as a cold node.
  • the preset number threshold is set by the user according to the actual situation, and is set to 0 in this embodiment, that is, only the node that does not send the message is controlled.
  • the token initiating node A updates the locally stored node history information to obtain node history information as described in Table 4.
  • S304 Mark the cold node in the node turnover record information, and update the node rotation order indicated in the node turnover record information included in the token according to the marked cold node, and update the node rotation order.
  • the marked cold node is not included.
  • the S304 may specifically include:
  • the cold node is marked in the node turnover record information; the cold node is marked by setting the turnover identifier corresponding to the corresponding node identifier in the entry to the skip identifier 0.
  • the token initiating node A re-initializes the node turnover table, sets the turnover flag bits of the node C and the node D, marks the node C and the node D of the cold node in the node turnover table, and completes the node turnover record included in the token.
  • the node turnover sequence indicated in the information is updated, and the node turnover table as shown in Table 5 is obtained, and the subsequent node rotation order is node A->node B->node A.
  • the node rotation order indicated by the record order of the node identification and its turnover table is a token passing between the node A and the node B containing the node turnover table. That is, node A sends a message according to the token, and then passes the token to the node identifier of the node record of the next item in the table and the turnaround flag bit is 1 according to the turnaround flag bit. After the node B sends the message, according to the turnover The flag bit, skipping node C and node D marked as cold nodes, is passed directly to the node person.
  • Node A and Node D may not record their own current round of message transmission until the end of the N round of turnaround.
  • S306 Initialize the node turnover record information and the history record information to perform the above S301 to S305 again. The S301-S306 is repeated until each node in the logical token delivery ring is deactivated.
  • the node A After detecting that the token is cycled N times between the node A and the node B, the node A initializes the node turnover record information and the history record information, and sets the node turnover record information as a cold flag. The node's turnaround flag is set to the normal identity. The S301 to S305 are re-executed.
  • node A acts as a token initiating node, and the turnover flag remains as the normal identifier 1.
  • the present invention has the following advantages:
  • the node can send a message each time it needs to wait for the token, which causes the node to send a message with a certain delay, and the larger the node size, the higher the delay; Suddenly a large number of messages need to be sent, and the node needs to go through multiple token turnovers to send all the messages, and the later the message is sent later.
  • a master node master for example, DC in master ⁇ pacemaker in citrix XenServer, etc.
  • it usually needs to process more services (such as data synchronization, etc.).
  • the master node Compared with the non-master node, the master node has more messages to send, and the non-master node needs to send fewer messages. If a single-ring in-order protocol (totem) is used in a peer-to-peer architecture, based on the above scenario, most of the time, the non-master node receives the token without action, and a large number of messages on the master node are backlogged. In the case of the message delay, in the embodiment of the present invention, a method for improving the performance of the master node and improving the message throughput of the cluster system is proposed. Specifically, the method can selectively skip according to the historical transmission status of each node.
  • Some nodes with a small amount of messages are sent quickly, so that the tokens are quickly transmitted between the nodes that need to send the message, thereby enabling the nodes with the message sending requirements to send the messages faster, reducing the backlog of messages on these nodes, and reducing the message.
  • Delay improve node performance and availability, and improve message throughput for clustered systems.
  • the node with the message sending requirement in the embodiment, the master node is in this embodiment
  • the node with the message sending requirement can send the message faster, and the message is reduced.
  • the message is reduced.
  • a logical token delivery ring includes a plurality of nodes, and the token turnover control apparatus is responsible for The creation of the card and the maintenance of the token turnover mechanism are set in the token originating node. After the token initiating node issues the token, the token loops inside the ring, and each node on the ring sends when the token is received. Eliminate As shown in FIG. 5, the token turnover control apparatus includes:
  • the obtaining block 11 is configured to obtain historical record information of a message transmission situation of each node in the current logical token delivery ring during the past rounding of the token turn;
  • the manner in which the obtaining module 11 obtains the record information of the historical transmission status of each node may be: saving the node turnover record information in the token, and each node is acquired in the process of issuing the token for circulation.
  • the number of messages sent in the current round is recorded in the node turnover record information corresponding to the identifier of the node in the token.
  • the obtaining module 11 records the history of each node according to the node turnover record information obtained after the current round of loops. The status of the message is updated and recorded.
  • the statistics recorded for each node are the historical record records of each node.
  • each node records the number of messages sent in the current round to the node turnover record information corresponding to the node identifier in the token after acquiring the token and sending the message.
  • the obtaining module 11 accumulates the number of sent messages corresponding to each node recorded on the node turnover record information in the current round and the number of sent messages of the corresponding recorded nodes in the local. The total number of messages sent by each node after the end of the current round is obtained and recorded. After the preset round token loop, the total number of messages sent by each node recorded by the obtaining module 11 is each. The message sending status history information of the node.
  • the above-mentioned node turnover record information may be expressed in the form of a node turnover table, and may of course be expressed in the form of other script files, metadata, and the like.
  • the update module 12 is configured to update, according to the history information, a node rotation sequence indicated in the node turnover record information included in the token, where the node turnover record information is used to indicate that the token is transmitted in a logical token The node rotation order within the ring.
  • the update module 12 may determine, according to the history record information acquired by the obtaining module 11, a node of the logical information, in the updated node turnover sequence, skip the sending in the subsequent round-the-roll token turnover process.
  • a message or a node that does not send a message that is, does not pass a token to a node that sends less messages or does not send a message, exploits the right of these nodes to send messages, facilitates sending of data by other nodes that frequently send messages, and reduces the number of nodes that frequently send messages.
  • the message is sent with a delay.
  • the sending module 13 is configured to deliver the updated token to the corresponding next node according to the node rotation sequence indicated by the updated node rotation record information.
  • the order of the tokens in the nodes can be updated according to the historical transmission status of each node, which can enable the nodes with message sending requirements to send messages faster, reduce the backlog of messages on these nodes, reduce message delay, and improve node performance. And availability, improved message throughput.
  • the token rotation control device includes the acquisition module 11 in the first embodiment of the device.
  • the update module 12 and the sending module 13 further, the acquiring module 11 in the token rotation control device in the embodiment may specifically include:
  • the statistic unit 111 is configured to: after each token acquires the token, update the history information of the local maintenance of the node according to the number of messages sent by each node recorded in the node turnover record information in the token The number of accumulated messages corresponding to each node in the middle;
  • the obtaining unit 112 is configured to use, as the logical token, the history of the message transmission status of each node in the ring, after the end of the M-round token turnover, the number of accumulated messages corresponding to each node recorded in the history information.
  • the number of messages sent by each node in the logical token delivery ring during the token rounding process of the current round is recorded by the following method: After each node obtains the token and sends the message, the message sent by the current round is sent. The number of records is recorded in the node turnover record information corresponding to the node identifier in the token.
  • the update module 12 includes:
  • a determining unit 121 configured to record, as a cold node, a node whose number of accumulated message records recorded in the history information after the end of the M round of token rotation is lower than or equal to a preset number threshold;
  • the updating unit 122 is configured to mark the cold node in the node turnover record information, and update the node rotation order indicated in the node rotation record information included in the token according to the marked cold node, after the update The labeled cold node is not included in the node rotation sequence; wherein the N value is a preset value.
  • the updating unit 122 is specifically configured to: set, by the cold node determined by the determining unit, a corresponding turnaround flag in the node rotation record information in the token as a skip flag, to complete the inclusion of the token
  • the node turnover sequence indicated in the node turnover record information is updated.
  • the turnaround flag in the node turnover record information is used to record the normal node and the cold node.
  • the node A as the token originating node is provided with the token turnover control device in this embodiment, and Table 5 indicates that the token transfer order is from node A to node B to node C.
  • Node D after the number of the current round of messages is recorded in the corresponding entry, the node A determines that the turnaround flag of the node B is the normal flag 1, and then passes the token to the node B. After Node B records the number of rounds of this round of messages, it is determined that the turnaround flags of Node C and Node D are skip flags 0. Therefore, Node B skips Node C and Node D and passes the token directly to A.
  • the updating unit 122 of the token rotation control device is further configured to set a rotation flag of the cold node set as a skip flag in the node rotation record information to a normal identifier, In order to pass the token in the node turnover sequence on the node turnover record information during the next token turnover process, the token passes through the node.
  • the token rotation control device may further include an initialization module 14 for initializing and updating the node turnover information and the history information.
  • an initialization module 14 for initializing and updating the node turnover information and the history information.
  • the initialization module 14 is configured to initialize the node turnover record information and the history record information, and notify the acquisition module to continue to perform acquisition of each node in the current logical token delivery ring in the past M round after initialization is completed. History information of the message transmission status during the token turnover process.
  • the foregoing historical record information and the node turnover record information may be expressed in the form of a record table, and may also be identified in the form of a script file or metadata.
  • FIG. 7 is a schematic structural diagram of a node for transmitting a message based on a token turnover according to the present invention.
  • the node that sends a message based on the token turnover is used as a logical token delivery ring, except for the token.
  • the nodes include:
  • the communication module 21 is configured to send a message to be sent after receiving the token passed in the logical token delivery ring.
  • the node that sends a message based on the token turnover can send one or more messages.
  • the recording module 22 is configured to record, according to the number of messages sent by the node in the token rotation process of the current round, the number of messages sent in the current round to the node turnover record information corresponding to the node identifier in the token. in.
  • the node turnover record information is saved in the token, and the node turnover record information can be It is expressed in the form of a node turnover table.
  • the communication module 21 is further configured to deliver the updated token to the corresponding next node according to the node rotation sequence indicated by the node turnover record information in the token.
  • the token in the embodiment of the present invention stores the node turnover record information. After each token acquires the token sending message in each round, the token records the number of messages sent in the current round in the node rotation sequence.
  • the token originating node within the logical token passing ring can arrange the node turnover order according to the quantity information recorded in the node turnover sequence.
  • FIG. 8 is a schematic structural diagram of a cluster system using a token ring mechanism according to the present invention.
  • the system specifically includes: including multiple nodes, where the multiple nodes include: a master node (ie, the master in FIG. 8) a node and a slave node (ie, a slave node in FIG. 8).
  • the master node functions as a token initiating node for creating and maintaining a token
  • each slave node receives the token as a token.
  • sending a message to be sent and transmitting the token, where specifically,
  • the master node is configured to acquire history record information of a message sent by each node in the current logical token delivery ring during a past M round of token turnover; and according to the historical record information, the node included in the token
  • the node rotation sequence indicated in the turnover record information is updated, and the node rotation record information is used to indicate a node rotation order of the token in the logical token delivery ring; the node indicated according to the updated node turnover record information
  • the turnover sequence passes the updated token to the corresponding next node;
  • the slave node after receiving the token passed in the logical token delivery ring, sends a message to be sent (indicated by MSG1, MSG2, MSG3, ... in FIG. 8), based on the order of the node in the current round.
  • the number of messages sent during the card turnover process, and the number of messages sent in the current round is recorded in the node turnover record information corresponding to the node identifier in the token, according to the node turnover record information in the token.
  • the node turnover sequence passes the updated token to the corresponding next node.
  • the master node may be the token rotation control device in the first embodiment of the token rotation control device, or the token turnover in the second embodiment of the token rotation control device.
  • the control device, the slave node may be a node that sends a message based on token rotation in the above embodiment.
  • the node can send a message each time it needs to wait for the token to be sent, which causes the node to send a message.
  • the node size the higher the delay; if a node suddenly has a large number of messages to send, the node needs to go through multiple token turnovers to send all the messages, the more the message is behind The higher the transmission delay.
  • there will be a master node master for example, master in citrix XenServer, DC in pacemaker, etc.
  • the master node it usually needs to process more services (such as data synchronization, etc.).
  • the master node has more messages to send, and the non-master node needs to send fewer messages.
  • a single-ring in-order protocol (totem) is used in a peer-to-peer architecture, based on the above scenario, most of the time, the non-master node receives the token without action, and a large number of messages on the master node are backlogged.
  • the message delay in the embodiment of the present invention, a method for improving the performance of the master node and improving the message throughput of the cluster system is proposed. Specifically, the method can selectively skip according to the historical transmission status of each node.
  • Some nodes with a small amount of messages are sent quickly, so that the tokens are quickly transmitted between the nodes that need to send the message, thereby enabling the nodes with the message sending requirements to send the messages faster, reducing the backlog of messages on these nodes, and reducing the message.
  • Delay improve node performance and availability, and improve message throughput for clustered systems.
  • the node with the message sending requirement in the embodiment, the master node is in this embodiment
  • the node with the message sending requirement can send the message faster, and the message is reduced.
  • the message is reduced.
  • the node in the embodiment of the present invention is a physical node device, and may be, for example, a server or the like.
  • FIG. 9 is a schematic structural diagram of a node (ie, a physical node device) included in the cluster system using the token ring mechanism of the present invention shown in FIG. 8.
  • FIG. 9 shows a specific implementation of a node.
  • the node 100 includes a transmitting circuit 1002, a receiving circuit 1003, a power controller 1004, a processor 1006, a memory 1007, and an antenna 1001.
  • the processor 1006 controls the operation of the node 100.
  • the memory 1007 can include read only memory and random access memory and provides instructions and data to the processor 1006.
  • a portion of the memory 1007 may also include non-volatile line random access memory (NVRAM).
  • NVRAM non-volatile line random access memory
  • the node 100 may be embedded or may itself be a wireless communication device such as a mobile phone, and may also include a receiving circuit 1002 and The carrier of the receiving circuit 1003 is configured to allow data transmission and reception between the node 100 and the remote location.
  • the transmitting circuit 1002 and the receiving circuit 1003 can be coupled to the antenna 1001.
  • the various components of the node 100 are coupled by a bus system 3100.
  • the bus system 3100 includes a power bus, a control bus, and a status signal bus in addition to the data bus.
  • various buses are labeled as the bus system 3100 in the figure.
  • Node 100 can also include a decoding processor 1005.
  • Processor 1006 may be an integrated circuit chip that has the ability to execute instructions and data, as well as signal processing capabilities. In the implementation process, each step of the above method may be completed by an integrated logic circuit of hardware in the processor 1006 or an instruction in the form of software.
  • the above processor may be a general purpose processor (CPU), a digital signal processor (DSP), an application specific integrated circuit (ASIC ready-to-use programmable gate array (FPGA) or other programmable logic device, discrete gate or transistor logic device, discrete hardware).
  • the general purpose processor may be a microprocessor or the processor may be any conventional processor, etc.
  • the steps of the method disclosed in connection with the embodiments of the present invention may be directly implemented as hardware processor execution, or using hardware in the processor.
  • the software module combination is completed.
  • the software module can be located in a random storage medium, such as a random access memory, a flash memory, a read only memory, a programmable read only memory or an electrically erasable programmable memory, a register, etc.
  • the storage medium is located in the memory 1007.
  • the processor reads the information in the memory 1007 and completes the steps of the above method in combination with its hardware.
  • the storage medium may be a magnetic disk, an optical disk, a read-only memory (ROM), or a random access memory (RAM).

Landscapes

  • Small-Scale Networks (AREA)
  • Engineering & Computer Science (AREA)
  • Computer Networks & Wireless Communication (AREA)
  • Signal Processing (AREA)

Abstract

本发明实施例公开了一种令牌周转控制方法、装置及系统,其中,所述方法包括:获取当前逻辑令牌传递环内每一个节点在过去的M轮令牌周转过程中消息发送情况的历史记录信息;根据所述历史记录信息,对令牌中包含的节点周转记录信息中指示的节点周转顺序进行更新,所述节点周转记录信息用于指示所述令牌在逻辑令牌传递环内的节点周转顺序;按照所述更新后的节点周转记录信息指示的节点周转顺序将所述更新后的令牌传递到对应的下一个节点。可以让有消息发送需求的节点更快地发送消息,减少这些节点上消息的积压,降低消息延时,提高节点性能和可用性,提高了消息吞吐量。

Description

一种令牌周转控制方法、 装置及系统
本申请要求于 2012年 5月 18日提交中国专利局、申请号为 201210156854.X, 发明名称为 "一种令牌周转控制方法、 装置及系统" 的中国专利申请的优先权, 其全部内容通过引用结合在本申请中。 技术领域
本发明涉及令牌周转控制领域, 尤其涉及一种令牌周转控制方法、 装置及 系统。 背景技术
totem单环有序协议提供一个发送域内全局有序向上层服务递交消息的机 制。 这意味着消息需要可靠性和有序性保证, 施加于发送域之上的是一个逻辑 令牌传递环, 在该逻辑令牌传递环内传递的令牌用于控制对环的访问, 并且只 有持有令牌的节点能够发送消息。 逻辑令牌传递环内的每个节点每次得到令牌 后能够发送多个消息。 当所有节点都没有消息发送, 令牌还是持续在环内循环。
在 totem系统中, 节点每次需要等待拿到令牌之后才能发送消息, 这样造成 了节点发送消息有一定的延时, 节点规模越大, 延时越高; 假如某个节点突然 有大量的消息需要发送, 节点需要经过多次令牌周转才能将消息全部发送完, 则节点中越靠后的消息, 其发送延时越高。 发明内容
本发明实施例提供一种令牌周转控制方法、 装置及系统, 以降低消息延时, 提高消息吞吐量。
一方面, 本发明实施例提供了一种令牌周转控制方法, 包括:
获取当前逻辑令牌传递环内每一个节点在过去的 M轮令牌周转过程中消息 发送情况的历史记录信息;
根据所述历史记录信息, 对令牌中包含的节点周转记录信息中指示的节点 周转顺序进行更新, 所述节点周转记录信息用于指示所述令牌在逻辑令牌传递 环内的节点周转顺序;
按照所述更新后的节点周转记录信息指示的节点周转顺序将所述更新后的 令牌传递到对应的下一个节点。
另一方面, 本发明实施例还提供了一种令牌周转方法, 包括:
当前节点接收逻辑令牌传递环内传递过来的令牌后, 发送待发送的消息; 基于所述节点在本轮的令牌周转过程中发送消息的数量, 将本轮发送的消 息的数量记录到所述令牌中的与本节点标识对应的节点周转记录信息中;
根据所述令牌中的节点周转记录信息所指示的节点周转顺序将更新后的令 牌传递到对应的下一个节点。
再一方面, 本发明实施例还提供了一种令牌周转控制装置, 包括: 获耳 ^莫块, 用于获取当前逻辑令牌传递环内每一个节点在过去的 M轮令牌 周转过程中消息发送情况的历史记录信息;
更新模块, 用于根据所述历史记录信息, 对令牌中包含的节点周转记录信 息中指示的节点周转顺序进行更新, 所述节点周转记录信息用于指示所述令牌 在逻辑令牌传递环内的节点周转顺序;
发送模块, 用于按照所述更新后的节点周转记录信息指示的节点周转顺序 将所述更新后的令牌传递到对应的下一个节点。
再一方面, 本发明实施例还公开了一种基于令牌周转发送消息的节点, 包 括:
通信模块, 用于接收逻辑令牌传递环内传递过来的令牌后, 发送待发送的 消息;
记录模块, 用于基于所述节点在本轮的令牌周转过程中发送消息的数量, 将本轮发送的消息的数量记录到所述令牌中的与本节点标识对应的节点周转记 录信息中;
所述通信模块还用于根据所述令牌中的节点周转记录信息所指示的节点周 转顺序将更新后的令牌传递到对应的下一个节点。
最后, 本发明实施例还提供了一种使用令牌环机制的集群系统, 包括多个 节点, 所述多个节点包括: 主节点和从节点, 其中,
所述主节点, 用于获取当前逻辑令牌传递环内每一个节点在过去的 M轮令 牌周转过程中消息发送情况的历史记录信息; 根据所述历史记录信息, 对令牌 中包含的节点周转记录信息中指示的节点周转顺序进行更新, 所述节点周转记 录信息用于指示所述令牌在逻辑令牌传递环内的节点周转顺序; 按照所述更新 后的节点周转记录信息指示的节点周转顺序将所述更新后的令牌传递到对应的 下一个节点;
所述从节点, 用于接收逻辑令牌传递环内传递过来的令牌后, 发送待发送 的消息, 基于本节点在本轮的令牌周转过程中发送消息的数量, 将本轮发送的 消息的数量记录到所述令牌中的与本节点标识对应的节点周转记录信息中, 根 据所述令牌中的节点周转记录信息所指示的节点周转顺序将更新后的令牌传递 到对应的下一个节点。
实施本发明实施例, 具有如下有益效果:
能够根据每一个节点的历史发送情况来更新令牌在节点中的周转顺序, 可 以让有消息发送需求的节点更快地发送消息, 减少这些节点上消息的积压, 降 低消息延时, 提高节点性能和可用性, 提高了消息吞吐量。 附图说明 例或现有技术描述中所需要使用的附图作筒单地介绍, 显而易见地, 下面描述 中的附图仅仅是本发明的一些实施例, 对于本领域普通技术人员来讲, 在不付 出创造性劳动的前提下, 还可以根据这些附图获得其他的附图。
图 1是本发明的令牌周转控制方法的第一实施例流程示意图;
图 2是本发明实施例的令牌周转方法的实施例流程示意图;
图 3是本发明的令牌周转控制方法的第二实施例流程示意图;
图 4是本发明实施例的其中一种逻辑令牌传递环内令牌在节点中的传递顺 序示意图;
图 5是本发明的令牌周转控制装置的第一实施例结构组成示意图; 图 6是本发明的令牌周转控制装置的第二实施例结构组成示意图; 图 7是本发明的基于令牌周转发送消息的节点的结构组成示意图; 图 8是本发明的使用令牌环机制的集群系统的结构组成示意图;
图 9为图 8所示的本发明的使用令牌环机制的集群系统中包括的节点的结 构示意图。 具体实施方式 下面将结合本发明实施例中的附图, 对本发明实施例中的技术方案进行清 楚、 完整地描述, 显然, 所描述的实施例仅仅是本发明一部分实施例, 而不是 全部的实施例。 基于本发明中的实施例, 本领域普通技术人员在没有作出创造 性劳动前提下所获得的所有其他实施例, 都属于本发明保护的范围。
请参见图 1 ,是本发明的令牌周转控制方法的第一实施例流程示意图, 在本 实施例中, 逻辑令牌传递环内包括多个节点, 其中包括一个负责令牌创建令牌 发起节点, 该令牌发起节点还在令牌中保存一用于指示令牌在令牌传递环内的 节点周转顺序的节点周转记录信息。 令牌发起节点发出令牌后, 按照所述节点 周转记录信息, 令牌在所述逻辑令牌传递环内各个节点之间传递, 环上的每个 节点在接收到令牌时发送消息。 本实施例的所述方法具体包括:
S101 : 获取当前逻辑令牌传递环内每一个节点在过去的 M轮令牌周转过程 中消息发送情况的历史记录信息。
所述 S101获取逻辑令牌传递环内每一个节点的历史发送情况的记录信息的 方式可以为: 令牌发起节点可以在令牌中保存节点周转记录信息, 该节点周转 记录信息可以以节点周转表的形式表示, 在发出令牌进行周转传递的过程中, 每一个节点在获取到令牌并发送消息后, 将本轮发送的消息的数量记录到所述 令牌中的与本节点标识对应的节点周转记录信息中, 所述节点标识可以是相应 节点的名称, 或其他能唯一表示节点的其他标识。
在每一轮令牌周转结束后, 令牌回到令牌发起节点时, 令牌发起节点根据 本轮循环后得到的节点周转记录信息, 对每一个节点的历史发送情况进行统计 更新并记录, 在 M轮令牌周转结束后, 对每一个节点记录的统计结果即为每一 个节点的消息发送情况的历史记录信息。 所述 M的值可根据逻辑令牌传递环内 节点的具体情况进行调整更改, 以便更好地降低每一个节点的消息发送延时。
所述 S101中获取每一个节点的历史发送情况的记录信息的方式还可以为: 直接在令牌中设置保存节点周转记录信息, 在每一轮获取到令牌后, 根据令牌 中的所述节点周转记录信息中记录的每一个节点本轮的消息发送数量更新令牌 发起节点本地维护的所述历史记录信息中每一个节点对应的累计消息发送数 量; 在 M轮令牌周转结束后, 将所述历史记录信息中所记录的每一个节点对应 的累计消息发送数量作为逻辑令牌传递环内每一个节点的消息发送情况的历史 记录信息。 其中, 所述历史记录信息可以以历史记录表的形式表示, 当然也可 以用脚本文件或者元数据等形式表示。
S102: 根据所述历史记录信息, 对令牌中包含的节点周转记录信息中指示 的节点周转顺序进行更新。 如上述, 所述节点周转记录信息用于指示所述令牌 在逻辑令牌传递环内的节点周转顺序。
在所述 S102中, 根据所述历史记录信息, 确定逻辑令牌传递环中在历史的 点周转顺序中, 在后续 N轮令牌周转过程中, 跳过这些很少发送消息或者不发 送消息的节点, 即不向艮少发送消息或者不发送消息的节点传递令牌, 剥削这 些节点发送消息的权利, 方便其他常发送消息的节点发送数据, 降低常发送消 息的节点上的消息发送时延。
S 103: 按照所述更新后的节点周转记录信息指示的节点周转顺序将所述更 新后的令牌传递到对应的下一个节点。 即按照新的节点周转顺序, 跳过某些 ^艮 少发送消息或者不发送消息的节点传递令牌。
通过上述实施例的描述可知, 本发明具有以下优点:
能够根据每一个节点的历史发送情况来更新令牌在节点中的周转顺序, 可 以让有较大消息发送需求的节点更快地发送消息, 减少这些节点上消息的积压, 降低消息延时, 提高节点性能和可用性, 提高了消息吞吐量。
请参见图 2, 是本发明实施例的令牌周转方法的实施例流程示意图, 本实施 例中是在逻辑令牌传递环内, 每一个节点执行的令牌周转方法, 所述方法包括:
S201 : 当前节点接收逻辑令牌传递环内传递过来的令牌后, 发送待发送的 消息。 该节点可以发送一条或者多条消息。
S202: 基于所述节点在本轮的令牌周转过程中发送消息的数量, 将本轮发 送的消息的数量记录到所述令牌中的与本节点标识对应的节点周转记录信息 中。 所述节点周转记录信息保存在所述令牌中, 该节点周转记录信息可以以节 点周转表的形式表示。
S203: 根据所述令牌中的节点周转记录信息所指示的节点周转顺序将更新 后的令牌传递到对应的下一个节点。
本发明实施例中的令牌保存有节点周转记录信息, 每一个令牌在每一轮获 取到令牌发送消息后, 会将本轮发送的消息的数量记录在所述节点周转顺序中, 以便逻辑令牌传递环内的令牌发起节点能够根据该节点周转顺序中记录的数量 信息安排节点周转顺序。
下面结合图 3和图 4对本发明的令牌周转控制方法进行详细描述, 图 3是 本发明的令牌周转控制方法的第二实施例流程示意图, 本实施例中, 所述方法 具体包括:
S301 : 令牌发起节点在每一轮获取到令牌后, 根据令牌中的所述节点周转 记录信息中记录的每一个节点本轮的消息发送数量更新令牌发起节点本地维护 的历史记录信息中每一个节点对应的累计消息发送数量。
S302: 令牌发起节点在 M轮令牌周转结束后, 将所述历史记录信息所记录 的每一个节点对应的累计消息发送数量作为逻辑令牌传递环内每一个节点的消 息发送情况的历史记录信息。
所述 M的值为预设值, 其可根据用户需求进行修改调整。
令牌发起节点在创建令牌的同时, 在令牌中保存并维护节点周转记录信息, 该节点周转距离信息可以以节点周转表的形式表示, 具体的, 该节点周转表中 可以包括节点标识以及该节点标识对应的节点的本轮消息发送数量的表项, 并 且在本实施例中, 该节点周转表还包括周转标志, 该周转标志用来表示相应节 点为^艮少发送消息或者不发送消息的冷节点, 或者为常发送消息的正常节点, 当周转标志设置为 1 时则相应的节点为正常节点, 令牌循环时经过该节点, 当 周转标志位为 0时则相应的节点为冷节点, 令牌循环时跳过该节点。
当所述节点周转记录信息以节点周转表的形式表示时, 该节点周转表指示 节点周转顺序的方式为: 按照表项中记录的节点标识的顺序及其周转标志, 确 定节点周转顺序。
令牌发起节点在令牌中维护节点周转记录信息的同时, 在本地还设置维护 有历史记录信息, 该历史记录信息包括节点标识、 累计消息发送数量以及周转 标识位, 令牌发起节点在所述历史记录信息中记录每一个节点发送消息的累计 数量。 其中, 在本实施例中, 所述历史记录信息以历史记录表的形式表示。
其中, 记录每一个节点每一轮消息发送数量是通过: 每一个节点在获取到 令牌并发送消息后, 将本次发送消息的数量记录到所述节点周转记录信息上本 节点对应的表项中。
具体的, 下述的表 1为节点周转记录信息的其中一种表现形式, 图 4为其 中一种逻辑令牌传递环内令牌在节点中的传递顺序示意图, 节点 A为逻辑令牌 传递环上的令牌发起者, 所述 S301和所述 S302具体可包括以下步骤:
在第一轮的令牌周转过程中, 各个节点的周转标识均为 1 , 则节点周转记录 信息指示的节点周转顺序为图 4中的节点 A->节点 B->节点 C->节点 D->节点 A。
节点 A在获取到令牌后, 发送了 10条消息, 节点 A便在表 1所示的节点 周转表中对应的本轮消息发送数量中记录 10条。节点 A将令牌传递给所述节点 B, 所述节点 B在获取到令牌后发送了 20条消息, 节点 B将令牌传递给所述节 点 C, 所述节点 C没有发送消息, 记录的数量为 0, 节点 C将令牌传递给所述 节点 D, 所述节点 D没有发送消息, 记录的数量为 0。 节点 D最后将令牌传回 给所述节点 A。
所述节点 A在第一轮令牌周转结束后获取到令牌后, 令牌中的令牌周转表 如表 1 述。
Figure imgf000009_0001
作为令牌发起节点的节点 A根据该节点周转表的记录, 在历史记录表中统 计本轮令牌周转过程中每一个节点的历史发送情况即消息的累计发送数量, 得 到如表 2所述的历史记录表。
表 2:
Figure imgf000009_0002
在统计完成后, 初始化所述节点周转表, 然后再次根据需要发送消息并在 节点周转记录信息中记录本轮消息发送数量, 进入下一轮令牌周转, 并在周转 结束获取到令牌后, 再次根据节点周转记录信息更新历史记录信息中每个节点 对应的累计消息发送数量。
再执行下一轮令牌循环以记录累计消息发送数量, 在预设的 M轮令牌周转 结束 令牌发起节点更新得到最终的历史记录表, 以表 3为例。
Figure imgf000010_0001
在 M轮令牌周转结束后, 令牌发起节点根据如上述的表 3所示的历史记录 信息即可确定每一个节点对应的累计消息发送数量, 如表 3所述: 在过去的 M 轮令牌循环中, 令牌经过了每一个节点, 并且相应节点的消息发送数量为: 节 点 A为 30条, 节点 B为 45条, 节点 C和节点 D为 0条。
S303: 将 M轮令牌周转结束后的所述历史记录信息中记录的累计消息发送 数量低于或等于预设的数量阈值的节点记录为冷节点。
所述预设的数量阈值由用户根据实际情况设置, 本实施例中设置为 0, 即仅 对没有发送消息的节点进行控制。
所述令牌发起节点 A更新本地存储的节点历史记录信息, 得到如表 4所述 的节点历史记录信息。
表 4:
Figure imgf000010_0002
S304: 在所述节点周转记录信息中标记出所述冷节点, 并根据标记出的冷 节点对令牌中包含的节点周转记录信息中指示的节点周转顺序进行更新, 在更 新后的节点周转顺序中不包括所述被标记出的冷节点。 具体的, 所述 S304具体可以包括:
在所述节点周转记录信息中标记出所述冷节点; 标记冷节点的方式是通过 将表项中相应节点标识对应的周转标识置为跳过标识 0。
将所述节点周转记录信息中冷节点对应的周转标志置为跳过标志, 完成对 令牌中包含的节点周转记录信息中指示的节点周转顺序的更新, 以便于在接下 来的 N轮令牌周转过程中按照所述节点周转记录信息上的节点周转顺序传递令 牌时, 将跳过周转标志为跳过标志的冷节点传递令牌。
所述令牌发起节点 A再初始化节点周转表, 并设置节点 C和节点 D的周转 标志位, 在节点周转表中标记出冷节点的节点 C和节点 D, 完成令牌中包含的 节点周转记录信息中指示的节点周转顺序的更新, 得到如表 5 所示的节点周转 表, 后的节点周转顺序为节点 A->节点 B->节点 A。
Figure imgf000011_0001
根据初始化的表 5 所示的节点周转表中根据节点标识的记录顺序及其周转 表位指示的节点周转顺序是在节点 A和节点 B之间传递包含该节点周转表的令 牌。 即节点 A根据令牌发送消息, 然后根据周转标志位, 将令牌传递给表中下 一项记录的节点标识且周转标志位为 1的所述节点 B, 节点 B发送了消息后, 根据周转标志位, 跳过被标记为冷节点的节点 C和节点 D, 直接传递给节点人。
S305: 在接下来的 N轮令牌周转结束后, 将所述节点周转记录信息中置为 跳过标志的冷节点的周转标志设置为正常标识, 以便于在接下来的令牌周转过 程中按照所述节点周转记录信息上的节点周转顺序传递令牌时, 令牌经过该节 点。 所述 N值也可根据具体情况进行设置。
在接下来的 N轮令牌周转过程中,节点 A和节点 D可不用记录其自身的本 轮消息发送数量, 直至 N轮周转结束。
S306: 初始化所述节点周转记录信息以及所述历史记录信息, 以再次执行 上述的 S301至 S305。 重复所述 S301-S306直至逻辑令牌传递环中的各个节点停止使用。
节点 A在检测到令牌在节点 A和节点 B之间循环 N轮后,初始化所述节点 周转记录信息以及所述历史记录信息, 并将所述节点周转记录信息中置为跳过 标志的冷节点的周转标志设置为正常标识。 重新执行所述 S301至 S305。
需要说明的是, 节点 A作为令牌发起节点, 周转标志位保持为正常标识 1。 通过上述实施例的描述可知, 本发明具有以下优点:
一方面, 相对于现有技术中, 节点每次需要等待拿到令牌之后才能发送消 息, 这样造成了节点发送消息有一定的延时, 节点规模越大, 延时越高; 假如 某个节点突然有大量的消息需要发送, 节点需要经过多次令牌周转才能将消息 全部发送完, 则越位于后面的消息的发送延时越高。 在某些对等型架构中, 会 存在一个主控节点 master (例如 citrix XenServer中的 master ^ pacemaker中的 DC 等)。 作为主控节点, 通常需要处理更多的业务 (比如数据同步等), 相比于非主 控节点, 主控节点有更多消息需要发送, 非主控节点需要发送的消息则比较少。 如果在对等型架构中采用单环有序协议 (totem),基于上述场景, 在大多数时间 里, 非主控节点拿到令牌没有动作, 而主控节点上的大量消息则会积压, 有消 息延迟的情况, 本发明实施例中, 提出一种能够提升主控节点性能、 提高集群 系统的消息吞吐量的方法, 具体的, 能够根据每一个节点的历史发送情况, 选 择性地跳过一些消息发送量较少的节点, 使令牌在在消息发送需求的节点之间 快速传递, 从而实现了让有消息发送需求的节点更快地发送消息, 减少这些节 点上消息的积压, 降低消息延时, 提高节点性能和可用性, 提高了集群系统的 消息吞吐量。
另一方面, 本发明实施例中, 在使用令牌环机制的集群系统中, 通过本发 明可以让有消息发送需求的节点 (本实施例中以主控节点示意) 更快地发送消 息, 减少瞬间大量消息在节点上的积压, 降低消息延时, 提高节点性能和可用 性, 一定程度上提高了集群系统的消息吞吐量。
下面对本发明的令牌控制装置进行详细描述。
请参见图 5 , 是本发明的令牌周转控制装置的第一实施例结构组成示意图, 本实施例中, 逻辑令牌传递环内包括多个节点, 所述令牌周转控制装置用于负 责令牌的创建和对令牌周转机制进行维护, 其设置在令牌发起节点中, 令牌发 起节点发出令牌后, 令牌在环内循环, 环上的每个节点在接收到令牌时发送消 如图 5所示, 所述令牌周转控制装置包括:
获 莫块 11 , 用于获取当前逻辑令牌传递环内每一个节点在过去的 Μ轮令 牌周转过程中消息发送情况的历史记录信息;
具体的, 所述获取模块 11获取每一个节点的历史发送情况的记录信息的方 式可以为: 在令牌中保存节点周转记录信息, 在发出令牌进行循环的过程中, 每一个节点在获取到令牌并发送消息后, 将将本轮发送的消息的数量记录到所 述令牌中的与本节点标识对应的节点周转记录信息中。 在本轮循环结束后, 令 牌回到设置有所述令牌周转控制装置的令牌发起节点时, 所述获取模块 11根据 本轮循环后得到的节点周转记录信息, 对每一个节点的历史消息发送情况进行 统计更新并记录, 在达到预设的循环次数阈值时, 对每一个节点记录的统计结 果即为每一个节点的历史发送情况记录信息。
例如在本轮令牌循环中, 每一个节点在获取到令牌并发送消息后, 将本轮 发送的消息的数量记录到所述令牌中的与本节点标识对应的节点周转记录信息 中, 在本轮循环结束, 所述获取模块 11将本轮循环中节点周转记录信息上记录 的每一个节点所对应的已发送消息的数量与本地已记录的对应节点的已发送消 息的数量进行累加统计, 得到本轮循环结束后的每一个节点已发送消息的总数 并记录,在经过预设的 Μ轮令牌循环后,所述获取模块 11记录的每一个节点已 发送消息的总数即为每一个节点的消息发送情况历史记录信息。
上述的节点周转记录信息可以以节点周转表的形式表示, 当然也可以是其 他的例如脚本文件、 元数据等形式表示。
更新模块 12, 用于根据所述历史记录信息, 对令牌中包含的节点周转记录 信息中指示的节点周转顺序进行更新, 所述节点周转记录信息用于指示所述令 牌在逻辑令牌传递环内的节点周转顺序。
所述更新模块 12可以根据所述获取模块 11获取的历史记录信息, 确定逻 息的节点, 在更新后的节点周转顺序中, 在后续 Ν轮令牌周转过程中, 跳过这 些^艮少发送消息或者不发送消息的节点, 即不向艮少发送消息或者不发送消息 的节点传递令牌, 剥削这些节点发送消息的权利, 方便其他常发送消息的节点 发送数据, 降低常发送消息的节点上的消息发送时延。 发送模块 13 , 用于按照所述更新后的节点周转记录信息指示的节点周转顺 序将所述更新后的令牌传递到对应的下一个节点。
实施本发明实施例, 具有如下有益效果:
能够根据每一个节点的历史发送情况来更新令牌在节点中的周转顺序, 可 以让有消息发送需求的节点更快地发送消息, 减少这些节点上消息的积压, 降 低消息延时, 提高节点性能和可用性, 提高了消息吞吐量。
再请参见图 6, 是本发明的令牌周转控制装置的第二实施例结构组成示意 图, 在本实施例中, 所述令牌周转控制装置包括上述装置第一实施例中的获取 模块 11 , 更新模块 12以及发送模块 13 , 进一步的, 在本实施例中所述令牌周 转控制装置中的获取模块 11具体可以包括:
统计单元 111 , 用于在每一轮获取到令牌后, 根据令牌中的所述节点周转记 录信息中记录的每一个节点本轮的消息发送数量更新令牌发起节点本地维护的 历史记录信息中每一个节点对应的累计消息发送数量;
获取单元 112, 用于在 M轮令牌周转结束后, 将所述历史记录信息所记录 的每一个节点对应的累计消息发送数量作为逻辑令牌传递环内每一个节点的消 息发送情况的历史记录信息;
其中, 逻辑令牌传递环内每一个节点在本轮的令牌周转过程中的消息发送 数量是通过如下方法记录的: 每一个节点在获取到令牌并发送消息后, 将本轮 发送的消息的数量记录到所述令牌中的与本节点标识对应的节点周转记录信息 中
进一步具体的, 所述更新模块 12包括:
确定单元 121 , 用于将 M轮令牌周转结束后的所述历史记录信息中记录的 累计消息发送数量低于或等于预设的数量阈值的节点记录为冷节点;
更新单元 122, 用于在所述节点周转记录信息中标记出所述冷节点, 并根据 标记出的冷节点对令牌中包含的节点周转记录信息中指示的节点周转顺序进行 更新,在更新后的节点周转顺序中不包括所述被标记出的冷节点; 其中, 所述 N 值为预设值。
其中, 所述更新单元 122具体是用于将所述确定单元确定的冷节点在令牌 中的所述节点周转记录信息中对应的周转标志置为跳过标志, 以完成对令牌中 包含的节点周转记录信息中指示的节点周转顺序进行更新。 节点周转记录信息中的周转标志用于记录正常节点和冷节点, 在各节点传 递令牌时, 节点是根据周转表上所列的节点标识的顺序, 以及根据所述周转标 志的值, 来确定将令牌传递给哪一个节点。
具体以上述的表 5为例, 作为令牌发起节点的节点 A设置有本实施例中的 所述令牌周转控制装置, 表 5指示了令牌传递顺序为节点 A到节点 B到节点 C 到节点 D, 节点 A在对应表项中记录了本轮消息发送数量后, 确定节点 B的周 转标志为正常标志 1 , 则将令牌传递给所述节点 B。 而当节点 B记录完本轮消息 发送数量后, 确定节点 C以及节点 D的周转标志均为跳过标志 0, 因此, 节点 B会跳过节点 C和节点 D, 将令牌直接传递给 A。
最后在令牌周转 N轮结束后, 所述令牌周转控制装置的所述更新单元 122 还用于将所述节点周转记录信息中置为跳过标志的冷节点的周转标志设置为正 常标识, 以便于在接下来的令牌周转过程中按照所述节点周转记录信息上的节 点周转顺序传递令牌时, 令牌经过该节点。
同时, 所述令牌周转控制装置还可包括初始化模块 14用于对节点周转信息 和历史记录信息进行初始化更新, 具体的,
所述初始化模块 14, 用于初始化所述节点周转记录信息以及所述历史记录 信息, 并在初始化完成后通知所述获取模块继续执行获取当前逻辑令牌传递环 内每一个节点在过去的 M轮令牌周转过程中消息发送情况的历史记录信息。
在通过所述初始化模块 14初始化所述节点周转记录信息以及所述历史记录 信息后, 再进入下一轮的令牌周转控制。
具体的, 上述的历史记录信息和节点周转记录信息均可以以记录表的形式 表示, 当然也可以以脚本文件或者元数据等形式标识。
再请参见图 7, 是本发明的基于令牌周转发送消息的节点的结构组成示意 图, 本实施例中, 所述基于令牌周转发送消息的节点是作为逻辑令牌传递环内, 除令牌发起节点以外的其他节点, 所述节点包括:
通信模块 21 , 用于接收逻辑令牌传递环内传递过来的令牌后, 发送待发送 的消息。 该基于令牌周转发送消息的节点可以发送一条或者多条消息。
记录模块 22,用于基于所述节点在本轮的令牌周转过程中发送消息的数量, 将本轮发送的消息的数量记录到所述令牌中的与本节点标识对应的节点周转记 录信息中。 所述节点周转记录信息保存在所述令牌中, 该节点周转记录信息可 以以节点周转表的形式表示。
所述通信模块 21还用于根据所述令牌中的节点周转记录信息所指示的节点 周转顺序将更新后的令牌传递到对应的下一个节点。
本发明实施例中的令牌保存有节点周转记录信息, 每一个令牌在每一轮获 取到令牌发送消息后, 会将本轮发送的消息的数量记录在所述节点周转顺序中, 以便逻辑令牌传递环内的令牌发起节点能够根据该节点周转顺序中记录的数量 信息安排节点周转顺序。
在请参见图 8 , 是本发明的使用令牌环机制的集群系统的结构组成示意图, 所述系统具体包括: 包括多个节点, 所述多个节点包括: 主节点 (即图 8 中的 master node )和从节点(即图 8中的 slave node ), 在本实施例中, 所述主节点作 为令牌发起节点, 用于创建和维护令牌, 各个从节点在接收到所述作为令牌发 起节点的主节点发出的令牌后, 发送待发送的消息, 并传递所述令牌, 其中具 体的,
所述主节点, 用于获取当前逻辑令牌传递环内每一个节点在过去的 M轮令 牌周转过程中消息发送情况的历史记录信息; 根据所述历史记录信息, 对令牌 中包含的节点周转记录信息中指示的节点周转顺序进行更新, 所述节点周转记 录信息用于指示所述令牌在逻辑令牌传递环内的节点周转顺序; 按照所述更新 后的节点周转记录信息指示的节点周转顺序将所述更新后的令牌传递到对应的 下一个节点;
所述从节点, 用于接收逻辑令牌传递环内传递过来的令牌后, 发送待发送 的消息(图 8中以 MSG1 , MSG2 , MSG3...示意), 基于本节点在本轮的令牌周 转过程中发送消息的数量, 将本轮发送的消息的数量记录到所述令牌中的与本 节点标识对应的节点周转记录信息中, 根据所述令牌中的节点周转记录信息所 指示的节点周转顺序将更新后的令牌传递到对应的下一个节点。
具体的, 所述的主节点具体可以为上述的令牌周转控制装置的第一实施例 中的令牌周转控制装置, 或者为上述的令牌周转控制装置的第二实施例中的令 牌周转控制装置, 所述从节点可以为上述的实施例中的基于令牌周转发送消息 的节点。
对于本发明的所述使用令牌环机制的集群系统, 一方面, 相对于现有技术 中, 节点每次需要等待拿到令牌之后才能发送消息, 这样造成了节点发送消息 有一定的延时, 节点规模越大, 延时越高; 假如某个节点突然有大量的消息需 要发送, 节点需要经过多次令牌周转才能将消息全部发送完, 则越位于后面的 消息的发送延时越高。 在某些对等型架构中, 会存在一个主控节点 master (例如 citrix XenServer中的 master, pacemaker中的 DC等)。 作为主控节点, 通常需要 处理更多的业务 (比如数据同步等), 相比于非主控节点, 主控节点有更多消息需 要发送, 非主控节点需要发送的消息则比较少。
如果在对等型架构中采用单环有序协议 (totem),基于上述场景,在大多数时 间里, 非主控节点拿到令牌没有动作, 而主控节点上的大量消息则会积压, 有 消息延迟的情况, 本发明实施例中, 提出一种能够提升主控节点性能、 提高集 群系统的消息吞吐量的方法, 具体的, 能够根据每一个节点的历史发送情况, 选择性地跳过一些消息发送量较少的节点, 使令牌在在消息发送需求的节点之 间快速传递, 从而实现了让有消息发送需求的节点更快地发送消息, 减少这些 节点上消息的积压, 降低消息延时, 提高节点性能和可用性, 提高了集群系统 的消息吞吐量。
另一方面, 本发明实施例中, 在使用令牌环机制的集群系统中, 通过本发 明可以让有消息发送需求的节点 (本实施例中以主控节点示意) 更快地发送消 息, 减少瞬间大量消息在节点上的积压, 降低消息延时, 提高节点性能和可用 性, 一定程度上提高了集群系统的消息吞吐量。
应当理解的是, 本发明实施例中的节点为物理节点设备, 例如可以是服务 器等。
图 9为图 8所示的本发明的使用令牌环机制的集群系统中包括的节点 (即 物理节点设备) 的结构示意图, 如图 9所示, 图 9示出了一种节点的具体实施 例, 在该实施例中, 节点 100包括发射电路 1002、 接收电路 1003、 功率控制器 1004、 处理器 1006, 存储器 1007及天线 1001。 处理器 1006控制节点 100的操 作。 存储器 1007可以包括只读存储器和随机存取存储器, 并向处理器 1006提 供指令和数据。 存储器 1007 的一部分还可以包括非易失行随机存取存储器 ( NVRAM )„ 具体的应用中, 节点 100可以嵌入或者本身可以就是例如移动电 话之类的无线通信设备, 还可以包括容纳发射电路 1002和接收电路 1003的载 体, 以允许节点 100和远程位置之间进行数据发射和接收。发射电路 1002和接 收电路 1003可以耦合到天线 1001。节点 100的各个组件通过总线系统 3100耦 合在一起, 其中 总线系统 3100 除包括数据总线之外, 还包括电源总线、 控制 总线和状态信号总线。 但是为了清楚说明起见, 在图中将各种总线都标为总线 系统 3100。 节点 100还可以包括解码处理器 1005。
上述本发明实施例揭示的方法可以应用于处理器 1006中, 或者说由处理器 1006以实现, 上述本发明实施例中的节点可以具体为上述图 9所示的节点来实 现。 处理器 1006可能是一种集成电路芯片, 具有指令和数据的执行能力, 以及 信号的处理能力。 在实现过程中, 上述方法的各步骤可以通过处理器 1006中的 硬件的集成逻辑电路或者软件形式的指令完成。 上述的处理器可以是通用处理 器(CPU )、 数字信号处理器(DSP )、 专用集成电路(ASIC 现成可编程门阵 列 (FPGA )或者其他可编程逻辑器件、 分立门或者晶体管逻辑器件、 分立硬件 通用处理器可以是微处理器或者该处理器也可以是任何常规的处理器等。 结合 本发明实施例所公开的方法的步骤可以直接体现为硬件处理器执行完成, 或者 用处理器中的硬件及软件模块组合执行完成。 软件模块可以位于随机存储器, 闪存、 只读存储器, 可编程只读存储器或者电可擦写可编程存储器、 寄存器等 本领域成熟的存储介质中。该存储介质位于存储器 1007,处理器读取存储器 1007 中的信息, 结合其硬件完成上述方法的步骤。
本领域普通技术人员可以理解实现上述实施例方法中的全部或部分流程, 是可以通过计算机程序来指令相关的硬件来完成, 所述的程序可存储于一计算 机可读取存储介质中, 该程序在执行时, 可包括如上述各方法的实施例的流程。 其中, 所述的存储介质可为磁碟、 光盘、 只读存储记忆体(Read-Only Memory, ROM )或随机存储记忆体(Random Access Memory, RAM )等。
以上所揭露的仅为本发明较佳实施例而已, 当然不能以此来限定本发明之 权利范围, 因此依本发明权利要求所作的等同变化, 仍属本发明所涵盖的范围。

Claims

权 利 要 求
1、 一种令牌周转控制方法, 其特征在于, 包括:
获取当前逻辑令牌传递环内每一个节点在过去的 M轮令牌周转过程中消息 发送情况的历史记录信息;
根据所述历史记录信息, 对令牌中包含的节点周转记录信息中指示的节点 周转顺序进行更新, 所述节点周转记录信息用于指示所述令牌在逻辑令牌传递 环内的节点周转顺序;
按照所述更新后的节点周转记录信息指示的节点周转顺序将所述更新后的 令牌传递到对应的下一个节点。
2、 如权利要求 1所述的方法, 其特征在于, 所述节点周转记录信息还用于 记录逻辑令牌传递环内每一个节点在本轮的令牌周转过程中的消息发送数量; 所述获取当前逻辑令牌传递环内每一个节点在过去的 M轮令牌周转过程中 消息发送情况的历史记录信息包括:
在每一轮获取到令牌后, 根据令牌中的所述节点周转记录信息中记录的每 一个节点本轮的消息发送数量更新令牌发起节点本地维护的历史记录信息中每 一个节点对应的累计消息发送数量;
在 M轮令牌周转结束后, 将所述历史记录信息所记录的每一个节点对应的 累计消息发送数量作为逻辑令牌传递环内每一个节点的消息发送情况的历史记 录信息;
其中, 逻辑令牌传递环内每一个节点在本轮的令牌周转过程中的消息发送 数量是通过如下方法记录的: 每一个节点在获取到令牌并发送消息后, 将本轮 发送的消息的数量记录到所述令牌中的与本节点标识对应的节点周转记录信息 中。
3、 如权利要求 2所述的方法, 其特征在于, 所述根据所述历史记录信息, 对令牌中包含的节点周转记录信息中指示的节点周转顺序进行更新包括:
将 M轮令牌周转结束后的所述历史记录信息中记录的累计消息发送数量低 于或等于预设的数量阈值的节点记录为冷节点; 在所述节点周转记录信息中标记出所述冷节点, 并根据标记出的冷节点对 令牌中包含的节点周转记录信息中指示的节点周转顺序进行更新, 在更新后的 节点周转顺序中不包括所述被标记出的冷节点。
4、 如权利要求 3所述的方法, 其特征在于, 所述节点周转记录信息中还包 括用于记录每一个节点的周转标志;
所述在所述节点周转记录信息中标记出所述冷节点, 并根据标记出的冷节 点对令牌中包含的节点周转记录信息中指示的节点周转顺序进行更新包括: 在所述节点周转记录信息中标记出所述冷节点;
将所述节点周转记录信息中冷节点对应的周转标志置为跳过标志, 完成对 令牌中包含的节点周转记录信息中指示的节点周转顺序的更新, 以便于在接下 来的 N轮令牌周转过程中按照所述节点周转记录信息上的节点周转顺序传递令 牌时, 将跳过周转标志为跳过标志的冷节点传递令牌。
5、 如权利要求 4所述的方法, 其特征在于, 还包括:
在接下来的 N轮令牌周转结束后, 将所述节点周转记录信息中置为跳过标 志的冷节点的周转标志设置为正常标识, 以便于在接下来的令牌周转过程中按 照所述节点周转记录信息上的节点周转顺序传递令牌时, 令牌经过该节点; 初始化所述节点周转记录信息以及所述历史记录信息, 以再次执行所述获 取当前逻辑令牌传递环内每一个节点在过去的 M轮令牌周转过程中消息发送情 况的历史记录信息。
6、 一种令牌周转方法, 其特征在于, 包括:
当前节点接收逻辑令牌传递环内传递过来的令牌后, 发送待发送的消息; 基于所述节点在本轮的令牌周转过程中发送消息的数量, 将本轮发送的消 息的数量记录到所述令牌中的与本节点标识对应的节点周转记录信息中;
根据所述令牌中的节点周转记录信息所指示的节点周转顺序将更新后的令 牌传递到对应的下一个节点。
7、 一种令牌周转控制装置, 其特征在于, 包括: 获耳 莫块, 用于获取当前逻辑令牌传递环内每一个节点在过去的 M轮令牌 周转过程中消息发送情况的历史记录信息;
更新模块, 用于根据所述历史记录信息, 对令牌中包含的节点周转记录信 息中指示的节点周转顺序进行更新, 所述节点周转记录信息用于指示所述令牌 在逻辑令牌传递环内的节点周转顺序;
发送模块, 用于按照所述更新后的节点周转记录信息指示的节点周转顺序 将所述更新后的令牌传递到对应的下一个节点。
8、 如权利要求 7所述的装置, 其特征在于, 所述节点周转记录信息还用于 记录逻辑令牌传递环内每一个节点在本轮的令牌周转过程中的消息发送数量, 所述获取模块包括:
统计单元, 用于在每一轮获取到令牌后, 根据令牌中的所述节点周转记录 信息中记录的每一个节点本轮的消息发送数量更新令牌发起节点本地维护的历 史记录信息中每一个节点对应的累计消息发送数量;
获取单元, 用于在 M轮令牌周转结束后, 将所述历史记录信息所记录的每 一个节点对应的累计消息发送数量作为逻辑令牌传递环内每一个节点的消息发 送情况的历史记录信息;
其中, 逻辑令牌传递环内每一个节点在本轮的令牌周转过程中的消息发送 数量是通过如下方法记录的: 每一个节点在获取到令牌并发送消息后, 将本轮 发送的消息的数量记录到所述令牌中的与本节点标识对应的节点周转记录信息 中。
9、 如权利要求 8所述的装置, 其特征在于, 所述更新模块包括: 确定单元, 用于将 M轮令牌周转结束后的所述历史记录信息中记录的累计 消息发送数量低于或等于预设的数量阈值的节点记录为冷节点;
更新单元, 用于在所述节点周转记录信息中标记出所述冷节点, 并根据标 记出的冷节点对令牌中包含的节点周转记录信息中指示的节点周转顺序进行更 新, 在更新后的节点周转顺序中不包括所述被标记出的冷节点。
10、 如权利要求 9所述的装置, 其特征在于, 所述节点周转记录信息中还 包括用于记录每一个节点的周转标志;
所述更新单元具体是用于将所述确定单元确定的冷节点在令牌中的所述节 点周转记录信息中对应的周转标志置为跳过标志, 以完成对令牌中包含的节点 周转记录信息中指示的节点周转顺序进行更新。
11、 如权利要求 10所述的装置, 其特征在于,
所述更新单元还用于在接下来的 N轮令牌周转结束后, 将所述节点周转记 录信息中置为跳过标志的冷节点的周转标志设置为正常标识, 以便于在接下来 的令牌周转过程中按照所述节点周转记录信息上的节点周转顺序传递令牌时, 令牌经过该节点;
所述装置还包括:
初始化模块, 用于初始化所述节点周转记录信息以及所述历史记录信息, 并在初始化完成后通知所述获取模块继续执行获取当前逻辑令牌传递环内每一 个节点在过去的 M轮令牌周转过程中消息发送情况的历史记录信息。
12、 一种基于令牌周转发送消息的节点, 其特征在于, 包括:
通信模块, 用于接收逻辑令牌传递环内传递过来的令牌后, 发送待发送的 消息;
记录模块, 用于基于所述节点在本轮的令牌周转过程中发送消息的数量, 将本轮发送的消息的数量记录到所述令牌中的与本节点标识对应的节点周转记 录信息中;
所述通信模块还用于根据所述令牌中的节点周转记录信息所指示的节点周 转顺序将更新后的令牌传递到对应的下一个节点。
13、 一种使用令牌环机制的集群系统, 其特征在于, 包括多个节点, 所述 多个节点包括: 主节点和从节点, 其中,
所述主节点, 用于获取当前逻辑令牌传递环内每一个节点在过去的 M轮令 牌周转过程中消息发送情况的历史记录信息; 根据所述历史记录信息, 对令牌 中包含的节点周转记录信息中指示的节点周转顺序进行更新, 所述节点周转记 录信息用于指示所述令牌在逻辑令牌传递环内的节点周转顺序; 按照所述更新 后的节点周转记录信息指示的节点周转顺序将所述更新后的令牌传递到对应的 下一个节点;
所述从节点, 用于接收逻辑令牌传递环内传递过来的令牌后, 发送待发送 的消息, 基于本节点在本轮的令牌周转过程中发送消息的数量, 将本轮发送的 消息的数量记录到所述令牌中的与本节点标识对应的节点周转记录信息中, 根 据所述令牌中的节点周转记录信息所指示的节点周转顺序将更新后的令牌传递 到对应的下一个节点。
PCT/CN2012/084234 2012-05-18 2012-11-07 一种令牌周转控制方法、装置及系统 WO2013170592A1 (zh)

Applications Claiming Priority (2)

Application Number Priority Date Filing Date Title
CN201210156854.XA CN102710481B (zh) 2012-05-18 2012-05-18 一种令牌周转控制方法、装置及系统
CN201210156854.X 2012-05-18

Publications (1)

Publication Number Publication Date
WO2013170592A1 true WO2013170592A1 (zh) 2013-11-21

Family

ID=46903051

Family Applications (1)

Application Number Title Priority Date Filing Date
PCT/CN2012/084234 WO2013170592A1 (zh) 2012-05-18 2012-11-07 一种令牌周转控制方法、装置及系统

Country Status (2)

Country Link
CN (1) CN102710481B (zh)
WO (1) WO2013170592A1 (zh)

Families Citing this family (5)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN102710481B (zh) * 2012-05-18 2015-07-08 华为技术有限公司 一种令牌周转控制方法、装置及系统
CN104320427A (zh) * 2014-09-22 2015-01-28 国家电网公司 一种继电保护系统多智能体的分布式组通信方法
US10402360B2 (en) 2016-06-10 2019-09-03 Johnson Controls Technology Company Building management system with automatic equipment discovery and equipment model distribution
CN106851543A (zh) * 2017-02-23 2017-06-13 宇龙计算机通信科技(深圳)有限公司 一种消息广播方法和装置
CN110912794B (zh) * 2019-11-15 2021-07-16 国网安徽省电力有限公司安庆供电公司 基于令牌集的近似匹配策略

Citations (4)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US5365513A (en) * 1993-01-11 1994-11-15 John Fluke Mfg. Co. Token ring local area network testing apparatus for matching its speed to ring speed
CN102291309A (zh) * 2011-08-26 2011-12-21 华为技术有限公司 一种基于令牌环的网络流量控制方法、节点及系统
CN102387082A (zh) * 2011-11-25 2012-03-21 西安电子科技大学 基于流分类的分组流量控制系统及控制方法
CN102710481A (zh) * 2012-05-18 2012-10-03 华为技术有限公司 一种令牌周转控制方法、装置及系统

Family Cites Families (2)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US7965732B2 (en) * 2004-03-19 2011-06-21 Fujitsu Limited Scheduling token-controlled data transmissions in communication networks
CN101286917A (zh) * 2008-04-25 2008-10-15 南昌大学 一种基于休眠机制延长异构无线传感器网络寿命的方法

Patent Citations (4)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US5365513A (en) * 1993-01-11 1994-11-15 John Fluke Mfg. Co. Token ring local area network testing apparatus for matching its speed to ring speed
CN102291309A (zh) * 2011-08-26 2011-12-21 华为技术有限公司 一种基于令牌环的网络流量控制方法、节点及系统
CN102387082A (zh) * 2011-11-25 2012-03-21 西安电子科技大学 基于流分类的分组流量控制系统及控制方法
CN102710481A (zh) * 2012-05-18 2012-10-03 华为技术有限公司 一种令牌周转控制方法、装置及系统

Also Published As

Publication number Publication date
CN102710481B (zh) 2015-07-08
CN102710481A (zh) 2012-10-03

Similar Documents

Publication Publication Date Title
US9921998B2 (en) Sensors global bus
CN109690512B (zh) 具有触发操作的gpu远程通信
EP2263155B1 (en) Direct data transfer between slave devices
TWI464566B (zh) 半導體系統、半導體裝置及電子裝置初始化方法
JP6165342B2 (ja) エンベデッドメモリへのコマンドキューイングの提供
US10334047B2 (en) Remote direct memory access with reduced latency
US20150370735A1 (en) Dynamically adjustable multi-line bus shared by multi-protocol devices
CN111857838B (zh) 用于管理ufs装置与ufs主机之间的通信的方法和系统
US20210326290A1 (en) Unified systems and methods for interchip and intrachip node communication
WO2013170592A1 (zh) 一种令牌周转控制方法、装置及系统
WO2012031436A1 (zh) 数据存储与鉴权并行的处理方法和终端
WO2014146524A1 (en) Method and configuration center server for configuring server cluster
WO2021057666A1 (zh) 传输控制方法、网管服务器、基站及存储介质
JP2017519294A (ja) フラッシュメモリベースストレージデバイスのマルチホスト電力コントローラ(mhpc)
TW202014901A (zh) 從機間直接通訊
US8843972B2 (en) Method of requesting video data distributed across a plurality of video servers
CN108139955A (zh) 用于在计算设备中提供独立于操作系统的错误控制的系统和方法
US9490988B2 (en) Continuous information transfer with reduced latency
CN112702146B (zh) 数据处理方法及设备
CN107408092B (zh) 用于多端口物理层(phy)的锁相环(pll)的共享控制
US9509780B2 (en) Information processing system and control method of information processing system
US9405719B2 (en) Circuitry to generate and/or use at least one transmission time in at least one descriptor
CN116627869B (zh) 应用于电子设备的数据传输方法及装置
US20230275716A1 (en) System and method for assisting data transmission over virtual channels
JP6940283B2 (ja) Dma転送制御装置、dma転送制御方法、及び、dma転送制御プログラム

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: 12876995

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: 12876995

Country of ref document: EP

Kind code of ref document: A1