WO2015154483A1 - Procede d'equilibrage de trafic de reseau et super contrôleur - Google Patents

Procede d'equilibrage de trafic de reseau et super contrôleur Download PDF

Info

Publication number
WO2015154483A1
WO2015154483A1 PCT/CN2014/092213 CN2014092213W WO2015154483A1 WO 2015154483 A1 WO2015154483 A1 WO 2015154483A1 CN 2014092213 W CN2014092213 W CN 2014092213W WO 2015154483 A1 WO2015154483 A1 WO 2015154483A1
Authority
WO
WIPO (PCT)
Prior art keywords
traffic
switch
node controller
sum
transferred
Prior art date
Application number
PCT/CN2014/092213
Other languages
English (en)
Chinese (zh)
Inventor
喻敬海
王延松
吴少勇
吴春明
杜众
高文
周伯阳
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 WO2015154483A1 publication Critical patent/WO2015154483A1/fr

Links

Images

Classifications

    • H—ELECTRICITY
    • H04—ELECTRIC COMMUNICATION TECHNIQUE
    • H04L—TRANSMISSION OF DIGITAL INFORMATION, e.g. TELEGRAPHIC COMMUNICATION
    • H04L69/00—Network arrangements, protocols or services independent of the application payload and not provided for in the other groups of this subclass
    • H04L69/24—Negotiation of communication capabilities
    • H—ELECTRICITY
    • H04—ELECTRIC COMMUNICATION TECHNIQUE
    • H04L—TRANSMISSION OF DIGITAL INFORMATION, e.g. TELEGRAPHIC COMMUNICATION
    • H04L41/00—Arrangements for maintenance, administration or management of data switching networks, e.g. of packet switching networks
    • H04L41/08—Configuration management of networks or network elements
    • H04L41/0803—Configuration setting
    • H04L41/0823—Configuration setting characterised by the purposes of a change of settings, e.g. optimising configuration for enhancing reliability
    • H04L41/083—Configuration setting characterised by the purposes of a change of settings, e.g. optimising configuration for enhancing reliability for increasing network speed
    • H—ELECTRICITY
    • H04—ELECTRIC COMMUNICATION TECHNIQUE
    • H04L—TRANSMISSION OF DIGITAL INFORMATION, e.g. TELEGRAPHIC COMMUNICATION
    • H04L41/00—Arrangements for maintenance, administration or management of data switching networks, e.g. of packet switching networks
    • H04L41/08—Configuration management of networks or network elements
    • H04L41/0895—Configuration of virtualised networks or elements, e.g. virtualised network function or OpenFlow elements
    • H—ELECTRICITY
    • H04—ELECTRIC COMMUNICATION TECHNIQUE
    • H04L—TRANSMISSION OF DIGITAL INFORMATION, e.g. TELEGRAPHIC COMMUNICATION
    • H04L47/00—Traffic control in data switching networks
    • H04L47/10—Flow control; Congestion control
    • H04L47/12—Avoiding congestion; Recovering from congestion
    • H04L47/125—Avoiding congestion; Recovering from congestion by balancing the load, e.g. traffic engineering
    • H—ELECTRICITY
    • H04—ELECTRIC COMMUNICATION TECHNIQUE
    • H04L—TRANSMISSION OF DIGITAL INFORMATION, e.g. TELEGRAPHIC COMMUNICATION
    • H04L67/00—Network arrangements or protocols for supporting network services or applications
    • H04L67/01—Protocols
    • H04L67/10—Protocols in which an application is distributed across nodes in the network
    • H04L67/1001—Protocols in which an application is distributed across nodes in the network for accessing one among a plurality of replicated servers
    • H04L67/1004—Server selection for load balancing
    • H04L67/101—Server selection for load balancing based on network conditions
    • H—ELECTRICITY
    • H04—ELECTRIC COMMUNICATION TECHNIQUE
    • H04L—TRANSMISSION OF DIGITAL INFORMATION, e.g. TELEGRAPHIC COMMUNICATION
    • H04L67/00—Network arrangements or protocols for supporting network services or applications
    • H04L67/01—Protocols
    • H04L67/12—Protocols specially adapted for proprietary or special-purpose networking environments, e.g. medical networks, sensor networks, networks in vehicles or remote metering networks
    • H04L67/125—Protocols specially adapted for proprietary or special-purpose networking environments, e.g. medical networks, sensor networks, networks in vehicles or remote metering networks involving control of end-device applications over a network
    • H—ELECTRICITY
    • H04—ELECTRIC COMMUNICATION TECHNIQUE
    • H04L—TRANSMISSION OF DIGITAL INFORMATION, e.g. TELEGRAPHIC COMMUNICATION
    • H04L69/00—Network arrangements, protocols or services independent of the application payload and not provided for in the other groups of this subclass
    • H04L69/14—Multichannel or multilink protocols

Definitions

  • the present invention is not limited to the field of communications technologies, and in particular, to a method and a super controller for network traffic balancing.
  • the controller In a multi-controller Software Defined Network (SDN) cluster network architecture, the controller is designed according to the OpenFlow protocol, and is configured to manage data flows, configure network devices, and formulate flow tables. To undertake communication between network services and network equipment. Switches in software-defined networks also support the OpenFlow protocol. The switch is controlled by the controller, and the flow table defined by the controller is saved. The flow table entry is queried to determine the processing mode of the packet, including packet forwarding, buffering, submitting the controller, or discarding. In addition, the OpenFlow protocol provides a secure channel that is configured to connect the switch to the controller for communication.
  • SDN Software Defined Network
  • a software-defined network cluster network architecture if the total traffic of a controller controlled by a certain controller is large, a large network control flow is generated, and these controllers that control a large flow rate easily become a bottleneck of the entire controller cluster, that is, a certain Some controllers have high traffic load, while others may control less total traffic and are far less saturated. This will affect the cluster's control over the network and the performance of the cluster, especially if there is a large fluctuation in network traffic. Adjusting the traffic based on the real-time network status is likely to cause more adjustments, and the network is jittery, and the performance of the cluster is greatly reduced.
  • the embodiment of the invention provides a network traffic balancing method and a super controller, and the main purpose thereof is to solve the technical problem that the controllers control the traffic unevenness in the software-defined network cluster and reduce the performance of the cluster.
  • the embodiment of the present invention provides a method for network traffic balancing, and the method for balancing network traffic includes the following steps:
  • the super controller in the software-defined network cluster sends a collection instruction to the plurality of node controllers in each time interval, and obtains traffic statistics returned by the node controller according to the collection instruction.
  • Each of the time intervals is the same;
  • the super controller calculates, according to the traffic statistics, a sum of traffic of the node controller at a next time interval;
  • the super controller adjusts the switch managed by the node controller based on the sum of the traffic statistics and the traffic.
  • the step of calculating the sum of the traffic of the node controller in the next time interval according to the traffic statistics information includes:
  • the sum of the traffic controlled by the corresponding node controller at each time interval is obtained according to the traffic message;
  • the maximum value H i max of the sum of the traffic of each node controller of all time intervals is obtained, and the sum of the flows of each node controller of the next time interval T i next is calculated according to the maximum value.
  • the step of adjusting the switch controlled by the node controller based on the sum of the traffic statistics and the traffic includes:
  • the switch controlled by each node controller is adjusted according to the first ratio and the second ratio.
  • the step of adjusting the switch controlled by each node controller according to the first ratio and the second ratio includes:
  • the switch to be tuned to the corresponding node controller is obtained from the pool of the switch to be transferred.
  • the step of acquiring the switch to be invoked in the switch controlled by the corresponding node controller, and adding the switch to be transferred to the pool of the switch to be transferred includes:
  • the switches that are controlled by the corresponding node controller accumulate the traffic in the ascending order of the traffic
  • the switch that obtains the accumulated traffic is used as the switch to be called, and the switch to be transferred is added to the pool to be transferred.
  • an embodiment of the present invention provides a super controller, where the super controller includes:
  • a collecting module configured to send a collection instruction to the plurality of node controllers in each time interval, and obtain flow statistics information returned by the node controller according to the collection instruction, where each time interval is the same;
  • a calculation module configured to calculate, according to the traffic statistics, a sum of traffic of the node controller at a next time interval
  • the adjustment module is configured to adjust the switch controlled by the node controller based on the sum of the traffic statistics and the traffic.
  • the calculating module includes:
  • a first acquiring unit configured to: according to the traffic statistics information, try to obtain a traffic message of each node controller at a predetermined time;
  • a second acquiring unit configured to acquire, according to the traffic packet, a sum of traffic controlled by the corresponding node controller at each time interval according to the traffic packet;
  • the calculating unit is configured to obtain a maximum value H i max of the sum of the flows of each node controller of all time intervals, and calculate a sum T i next of the traffic of each node controller of the next time interval according to the maximum value.
  • the adjusting module includes:
  • a third obtaining unit configured to acquire a first ratio of a sum of the flows T i next to the maximum value H i max ;
  • a fourth obtaining unit configured to acquire a first sum of a sum of all the traffic of the node controller T i next and a second sum of all the maximum values H i max of the node controller, and obtain the first sum and the second sum Second ratio;
  • the adjusting unit is configured to adjust the switch controlled by each node controller according to the first ratio and the second ratio.
  • the adjusting unit includes:
  • a determining subunit configured to determine whether the first ratio is greater than the second ratio
  • the sub-unit is called, and if so, the switch to be called out of the switch controlled by the corresponding node controller is obtained, and the switch to be transferred is added to the pool to be transferred;
  • the sub-unit is called, and if not, the switch to be transferred by the corresponding node controller is obtained from the pool of the switch to be transferred.
  • the callout subunit is set to:
  • the switches that are controlled by the corresponding node controller accumulate the traffic in the ascending order of the traffic
  • the switch that obtains the accumulated traffic is used as the switch to be called, and the switch to be transferred is added to the pool to be transferred.
  • the call subunit is set to:
  • an embodiment of the present invention further provides a computer program, including program instructions, when the program instruction is executed by a super controller, so that the super controller can perform the method as described above.
  • an embodiment of the present invention further provides a carrier carrying the computer program as described above.
  • the network traffic balancing method and the super controller provided by the embodiments of the present invention create a manner in which controllers in the SDN cluster communicate with each other about traffic problems, and based on the predicted traffic of the node controller at the next time interval And the traffic statistics corresponding to the traffic carried by the node controller history to adjust the switch controlled by the node controller, thereby dynamically adjusting the traffic of the node controller in the SDN cluster; especially when the network traffic fluctuates greatly, The traffic load of the individual node controller is heavy, and the uneven traffic may cause bottlenecks in the node controller in the SDN cluster.
  • the present invention adjusts the switch controlled by the node controller to load the switch with smaller traffic to the traffic load.
  • the control of the node controller balances the traffic load of the node controller and can greatly improve the performance of the SDN cluster.
  • FIG. 1 is a schematic diagram of a method for network traffic balancing according to an embodiment of the present invention
  • FIG. 2 is a schematic diagram of a refinement process of step S102 in FIG. 1;
  • step S103 in FIG. 1 is a schematic diagram showing the refinement process of step S103 in FIG. 1;
  • step S1033 in FIG. 3 is a schematic diagram of a refinement process of step S1033 in FIG. 3;
  • FIG. 5 is a schematic diagram of functional modules of a super controller according to an embodiment of the present invention.
  • FIG. 6 is a schematic diagram of a refinement function module of the calculation 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 the adjustment unit of FIG. 7.
  • the embodiment of the invention provides a method for network traffic balancing, which is referred to FIG. 1 , in this embodiment,
  • the method for balancing network traffic includes:
  • Step S101 The super controller in the software-defined network cluster sends a collection instruction to the multiple node controllers in each time interval, and acquires traffic statistics information returned by the node controller according to the collection instruction, where each time interval the same;;
  • the software-defined network SDN cluster has a super controller and a plurality of node controllers, and each of the plurality of node controllers controls a plurality of switches.
  • the super controller in the SDN cluster initiates scheduling of the node controller traffic of the managed cluster, and issues a collection instruction.
  • the time interval is determined by the service in the SDN cluster. When the network traffic unevenness has little impact on the service, the time interval is long. When the network traffic is uneven, the time interval is short. All the node controllers in the SDN cluster start to report their traffic, including: the node controller sends collection commands to all the switches it controls through the secure channel. The switch obtains traffic statistics according to the collection command and sends the statistics to the node controller. Traffic statistics of all the switches that are controlled, and then send the collected traffic statistics to the cluster distributed shared storage system. The traffic statistics include: controller number, switch label, switch traffic y, and timestamp x. This timestamp is the time at which traffic statistics are collected for all managed switches.
  • Step S102 Calculate, according to the traffic statistics information, a sum of traffic of the node controller at a next time interval;
  • the traffic controlled by each node controller in the next time interval can be calculated, that is, the node can be accurately predicted in the next time interval. Controlled traffic.
  • the traffic information of the predetermined n times to the controller of each node is obtained from the traffic statistics as a complete traffic packet, and each time corresponds to the time in each time interval, if not obtained.
  • a complete traffic message of a node controller needs to continue to collect traffic statistics of the node controller, and is not used as a node controller for calculating the sum of the traffic of the next time interval.
  • the node controller For the node controller that can obtain the complete traffic message, obtain the sum of the traffic of the switch that is controlled by each of the predetermined time intervals, and then obtain the maximum value of the sum of the n traffic, and use the maximum according to the maximum value.
  • the regression analysis calculates the sum of the traffic of the node controllers at the next interval.
  • the maximum value of the sum of the flows of each node controller is set to zero.
  • the sum of the traffic of the node controller of the next time interval can be calculated, which is used as a traffic prediction mode to adjust the switch controlled by the node controller.
  • Step S103 adjusting the switch controlled by the node controller based on the sum of the traffic statistics and the traffic.
  • the switch that is controlled by the node controller can be adjusted by using a pool to be transferred, and the pool to be transferred is set to be empty when starting the adjustment.
  • the switch when a switch controlled by a node controller with a large traffic load is called out, the switch can be called out to the pool of the switch to be transferred; and the controller of a node with a small traffic load is controlled.
  • the switch in the pool of the switch to be transferred can be transferred to the node controller for control, thereby achieving network traffic balance in the SDN cluster.
  • the switch corresponding to the called node controller number and the switch corresponding to the transferred node controller number are transferred.
  • the switch command is then exchanged according to the slave-master mechanism to implement switch adjustment.
  • This embodiment creates a way for the controllers in the SDN cluster to communicate with each other on the traffic problem, and based on the predicted traffic flow of the node controller at the next time interval and the traffic corresponding to the traffic carried by the node controller history. Statistical information to adjust the switches managed by the node controller to dynamically adjust the traffic of the node controllers in the SDN cluster.
  • This embodiment can improve the performance of the SDN cluster, especially when the network traffic fluctuates greatly, the traffic load of the individual node controller is heavy, and the uneven traffic may cause a bottleneck of the node controller in the SDN cluster.
  • the switch controlled by the controller transfers the switch with smaller traffic to the node controller with heavy traffic load, and balances the traffic of the node controller. Load, can greatly improve the performance of SDN cluster.
  • the foregoing step S102 may include:
  • Step S1021 Try to obtain a traffic message of each node controller at a predetermined time according to the traffic statistics information.
  • step S1022 when the acquisition succeeds, the sum of the traffic controlled by the corresponding node controller at each time interval is obtained according to the traffic message;
  • Step S1023 Acquire a maximum value H i max of the sum of the traffic of each node controller of all time intervals, and calculate a sum T i next of the traffic of each node controller of the next time interval according to the maximum value.
  • the traffic statistics information ⁇ controller number, switch label, switch traffic Y, timestamp X ⁇ , try to obtain the traffic packets of each node controller in the predetermined n times, and each time corresponds to each time interval. If the traffic information of a node controller is not obtained, the traffic statistics of the node controller need to be collected continuously, and the node controller is not used as the sum of the traffic of the next time interval. Node controller.
  • the super controller reads the traffic statistics of each node controller from the cluster distributed shared storage system, and attempts to obtain a complete traffic message with n time stamps of each node controller.
  • the complete traffic packet refers to all the different switch traffic packets obtained by the same node controller in the preset n timestamps.
  • the value of n ranges from 5 to 30.
  • step S101 If less than n complete traffic messages are obtained, the process returns to step S101.
  • H i max max ⁇ H i max , T1, T2, T3, T4, T5 ... Tn ⁇ , ie Tm is taken as the maximum value H i max .
  • the value is:
  • X and Y in the equation are the time stamp and the sum of the flows in Tm, respectively.
  • the foregoing step S103 may include:
  • Step S1031 Acquire a first ratio of the sum of the flows T i next to the maximum value H i max ;
  • Step S1032 Acquire a first sum of the sum of all the traffic of the node controller T i next and a second sum of all the maximum values H i max of the node controller, and obtain the first sum and the second sum Second ratio;
  • Step S1033 Adjust the switch controlled by each node controller according to the first ratio and the second ratio.
  • the pool of the switch to be transferred is set to be empty, and the first ratio Pi of the sum of the flows of the node controllers T i next and the maximum value H i max is calculated:
  • step S1033 may include:
  • Step S10331 it is determined whether the first ratio is greater than the second ratio; if yes, then proceeds to step S10332, and if not, proceeds to step S10333;
  • Step S10332 Obtain a switch to be called out of the switch controlled by the corresponding node controller, and add the switch to be transferred to the pool to be transferred, and the process ends;
  • step S10333 the switch to be transferred by the corresponding node controller is obtained from the pool of the switch to be transferred, and the process ends.
  • the step S10332 may include: calculating, according to the maximum value H i max , the sum of the flows T i next and the second ratio, the callout traffic B osize of the corresponding node controller; and the switches controlled by the corresponding node controllers in ascending order of traffic When the accumulated traffic is greater than the outgoing traffic B osize , the switch that has acquired the accumulated traffic is used as the switch to be called, and the switch to be transferred is added to the pool to be transferred.
  • the callout traffic B osize T i next -H i max *Pv is calculated.
  • the traffic information of the latest timestamp of the corresponding node controller is read from the cluster distributed shared storage system, and the switch number and switch traffic are extracted from the traffic information.
  • the switch controlled by the node controller is sorted in ascending order by the switch traffic size, and the switch queue is obtained.
  • the switch is selected from the top of the switch queue and the traffic of the switch is accumulated.
  • the switch that has accumulated the traffic is obtained as the switch to be called out, and if the accumulated traffic is less than or equal to the called traffic B osize , the accumulated traffic is continued.
  • the node controller number to be transferred is empty.
  • Step S10333 includes: calculating, according to the maximum value H i max , the sum of the flows T i next and the second ratio, the incoming traffic B isize of the corresponding node controller; and ascending the traffic in the pool of the to-be-switched switch pool sequentially accumulating the flow; when the accumulated flow is greater than the flow rate transferred B isize, has acquired the integrated flow rate switch switches transferred as required, the switches need to be transferred into the corresponding node controller control.
  • the incoming traffic B isize H i max *Pv-T i next is calculated.
  • the switch In the pool to be mobilized, the switch is sorted in ascending order by the switch traffic size to obtain the switch queue. The switch is selected from the top of the switch queue and the traffic of the switch is accumulated.
  • the node controller includes A, B, and C, and the switch controlled by the node controller A is S1, S2, and S3; the controllers controlled by the node controller B are S4 and S5; the switches controlled by the node controller C are S6, S7, and S8.
  • the obtained traffic statistics are as follows:
  • Node controller number A B C The sum of flows T i next 39.7 16.3 15.2 Maximum value H i max 36 16 31 Pi 1.1 1.0 0.49
  • An embodiment of the present invention further provides a super controller.
  • the super controller includes:
  • the collecting module 101 is configured to send a collection instruction to the plurality of node controllers in each time interval, and obtain flow statistics information returned by the node controller according to the collection instruction, where each time interval is the same;
  • the software-defined network SDN cluster has a super controller and a plurality of node controllers, and each of the plurality of node controllers controls a plurality of switches.
  • the super controller in the SDN cluster initiates scheduling of the node controller traffic of the managed cluster, and issues a collection instruction.
  • the time interval is determined by the service in the SDN cluster.
  • the time interval in which the network traffic is unevenly affected to the service is long, and the time interval in which the network traffic is uneven is large.
  • All the node controllers in the SDN cluster start to report their traffic, including: the node controller sends collection commands to all the switches it controls through the secure channel.
  • the switch obtains traffic statistics according to the collection command and sends the statistics to the node controller. Traffic statistics of all the switches that are controlled, and then send the collected traffic statistics to the cluster distributed shared storage system.
  • the traffic statistics include: controller number, switch label, switch traffic y, and timestamp x. This timestamp is the time at which traffic statistics are collected for all managed switches.
  • the calculating module 102 is configured to calculate, according to the traffic statistics, a sum of traffic of the node controller at a next time interval;
  • the next time interval can be calculated.
  • the traffic controlled by the node controller can accurately predict the traffic controlled by each node controller in the next time interval.
  • the traffic information of the predetermined n times to the controller of each node is obtained from the traffic statistics as a complete traffic packet, and each time corresponds to the time in each time interval, if not obtained.
  • a complete traffic message of a node controller needs to continue to collect traffic statistics of the node controller, and is not used as a node controller for calculating the sum of the traffic of the next time interval.
  • the node controller For the node controller that can obtain the complete traffic message, obtain the sum of the traffic of the switch that is controlled by each of the predetermined time intervals, and then obtain the maximum value of the sum of the n traffic, and use the maximum according to the maximum value.
  • the regression analysis calculates the sum of the traffic of the node controllers at the next interval.
  • the maximum value of the sum of the flows of each node controller is set to zero.
  • the sum of the traffic of the node controller of the next time interval can be calculated, which is used as a traffic prediction mode to adjust the switch controlled by the node controller.
  • the adjustment module 103 is configured to adjust the switch controlled by the node controller based on the sum of the traffic statistics and the traffic.
  • the switch that is controlled by the node controller can be adjusted by using a pool to be transferred, and the pool to be transferred is set to be empty when starting the adjustment.
  • the switch when a switch controlled by a node controller with a large traffic load is called out, the switch can be called out to the pool of the switch to be transferred; and the controller of a node with a small traffic load is controlled.
  • the switch in the pool of the switch to be transferred can be transferred to the node controller for control, thereby achieving network traffic balance in the SDN cluster.
  • each switch in the switch pool is to be mobilized, if the node controller is called If the number is not empty, transfer the switch command to the switch corresponding to the node controller number that is called and the switch corresponding to the node controller number that is transferred, and then switch according to the slave-master mechanism to implement switch adjustment.
  • the super controller creates a way for the controllers in the SDN cluster to communicate with each other on the traffic problem, and based on the predicted traffic flow of the node controller at the next time interval, and the traffic carried by the node controller history. Corresponding traffic statistics are used to adjust the switches managed by the node controller to dynamically adjust the traffic of the node controllers in the SDN cluster.
  • This embodiment can improve the performance of the SDN cluster, especially when the network traffic fluctuates greatly, the traffic load of the individual node controller is heavy, and the uneven traffic may cause a bottleneck of the node controller in the SDN cluster.
  • the switch controlled by the controller transfers the switch with smaller traffic to the node controller with heavy traffic load, and balances the traffic load of the node controller, which can greatly improve the performance of the SDN cluster.
  • the computing module 102 can include:
  • the first obtaining unit 1021 is configured to, according to the traffic statistics information, attempt to acquire a traffic message of each node controller at a predetermined time;
  • the second obtaining unit 1022 is configured to: when the obtaining succeeds, acquire, according to the traffic message, a sum of traffic controlled by the corresponding node controller at each time interval;
  • the calculating unit 1023 is configured to acquire a maximum value H i max of the sum of the flows of each node controller of all time intervals, and calculate a sum T i next of the traffic of each node controller of the next time interval according to the maximum value.
  • the traffic statistics information ⁇ controller number, switch label, switch traffic Y, timestamp X ⁇ , try to obtain the traffic packets of each node controller in the predetermined n times, and each time corresponds to each time interval. If the traffic information of a node controller is not obtained, the traffic statistics of the node controller need to be collected continuously, and the node controller is not used as the sum of the traffic of the next time interval. Node controller.
  • the first obtaining unit 1021 reads the traffic statistics of each node controller from the cluster distributed shared storage system, and attempts to obtain a complete traffic report of each node controller with n timestamps. Text.
  • a complete traffic message refers to the same timestamp of the same node controller. All different switch traffic messages and the timestamp is the most recent. The above n value ranges from 5 to 30.
  • the collection module 101 continues to send a collection instruction to the multiple node controllers in each time interval to obtain the traffic statistics returned by the node controller according to the collection instruction.
  • the node controller that obtains the complete traffic message reaches n is the node controller that meets the condition, and the second acquiring unit 1022 obtains the traffic that is controlled at each time of the n times.
  • X and Y in the equation are the time stamp and the sum of the flows in Tm, respectively.
  • the adjustment module 103 includes:
  • the third obtaining unit 1031 is configured to acquire a first ratio of the sum of the traffic T i next and the maximum value H i max ;
  • the fourth obtaining unit 1032 is configured to acquire a first sum of the sum of all the traffic of the node controller T i next and a second sum of all the maximum values H i max of the node controller, and acquire the first a second ratio of the sum to the second sum;
  • the adjusting unit 1033 is configured to adjust the switch managed by each node controller according to the first ratio and the second ratio.
  • the pool to be transferred is set to be empty, and the third obtaining unit 1031 calculates a first ratio Pi of the sum of the traffic of the node controller T i next and the maximum value H i max :
  • the fourth obtaining unit 1032 calculates a first sum of the sum of all the traffic of the node controller T i next and a second sum of all the maximum values H i max of the node controller, and calculates the first sum and the second sum Two ratio P V :
  • the adjusting unit 1033 may include:
  • the determining subunit 10331 is configured to determine whether the first ratio is greater than the second ratio
  • the sub-unit 10332 is called, and if so, the switch to be called out of the switch controlled by the corresponding node controller is obtained, and the switch to be called is added to the pool to be transferred;
  • the call to the sub-unit 10333 is set to be, if not, the switch to be transferred from the corresponding node controller is obtained from the pool of the switch to be transferred.
  • the callout subunit 10332 is configured to: calculate a callout traffic B osize of the corresponding node controller according to the maximum value H i max , the sum of the flows T i next and the second ratio; and control the corresponding node
  • the switch that manages the load accumulates the traffic in the ascending order of the traffic; when the accumulated traffic is greater than the buffered traffic, the switch that obtains the accumulated traffic is used as the switch to be called, and the switch to be transferred is added to the pool to be transferred. in.
  • the callout traffic B osize T i next -H i max *Pv is calculated.
  • the traffic information of the latest timestamp of the corresponding node controller is read from the cluster distributed shared storage system, and the switch number and switch traffic are extracted from the traffic information.
  • the switch controlled by the node controller is sorted in ascending order by the switch traffic size, and the switch queue is obtained.
  • the switch is selected from the top of the switch queue and the traffic of the switch is accumulated.
  • the switch that obtains the accumulated traffic is used as the switch to be called out, and when the accumulated traffic is less than or equal to the called traffic B osize , the accumulated traffic is continued.
  • the node controller number to be transferred is empty.
  • the call-in sub-unit 10333 is configured to: calculate a call-in traffic Bizeize of the corresponding node controller according to the maximum value H i max , the sum of the flows T i next and the second ratio; switch switch pool in ascending order of cumulative traffic flow; and when the accumulated flow is greater than the flow rate transferred B isize, has acquired the integrated flow rate switch switches transferred as required, the switches need to be transferred into corresponding The node controller is in control.
  • the incoming traffic B isize H i max *Pv-T i next is calculated.
  • the switch In the pool to be mobilized, the switch is sorted in ascending order by the switch traffic size to obtain the switch queue. The switch is selected from the top of the switch queue and the traffic of the switch is accumulated.
  • the above technical solution creates a way for the controllers in the SDN cluster to communicate with each other on the traffic problem, and based on the predicted traffic interval of the node controller at the next time interval, the node controller
  • the traffic statistics corresponding to the traffic carried by the history adjust the switch controlled by the node controller to dynamically adjust the traffic of the node controller in the SDN cluster; especially when the network traffic fluctuates greatly, the traffic load of the individual node controller When the traffic is uneven, the node controller in the SDN cluster may be bottlenecked.
  • the present invention adjusts the switch controlled by the node controller to switch the switch with smaller traffic to the node controller with heavy traffic load. Balancing the traffic load of the node controller can greatly improve the performance of the SDN cluster.

Landscapes

  • Engineering & Computer Science (AREA)
  • Computer Networks & Wireless Communication (AREA)
  • Signal Processing (AREA)
  • Computer Security & Cryptography (AREA)
  • Health & Medical Sciences (AREA)
  • Computing Systems (AREA)
  • General Health & Medical Sciences (AREA)
  • Medical Informatics (AREA)
  • Data Exchanges In Wide-Area Networks (AREA)

Abstract

La présente invention concerne un procédé d'équilibrage de trafic de réseau et un super contrôleur. Le procédé d'équilibrage de trafic de réseau comprend les étapes suivantes: la transmission par un super contrôleur dans un faisceau de réseaux défini par logiciel d'une instruction de collecte à une pluralité de contrôleurs de nœuds à l'intérieur de chaque intervalle de temps, pour acquérir une information statistique de trafic renvoyée par les contrôleurs de nœuds selon l'instruction de collecte; en fonction de l'information statistique de trafic, le calcul de la somme de trafic des contrôleurs de nœuds à l'intérieur de l'intervalle de temps suivant; et sur la base de l'information statistique de trafic et la somme de trafic, l'ajustement d'un commutateur géré et commandé par les contrôleurs de nœuds. La solution technique peut équilibrer la charge de trafic de contrôleurs de nœuds, ce qui permet d'améliorer considérablement la performance d'un faisceau de réseaux partagés de transmission de données (RPTD).
PCT/CN2014/092213 2014-09-03 2014-11-25 Procede d'equilibrage de trafic de reseau et super contrôleur WO2015154483A1 (fr)

Applications Claiming Priority (2)

Application Number Priority Date Filing Date Title
CN201410447127.8A CN105471755B (zh) 2014-09-03 2014-09-03 网络流量均衡的方法及超级控制器
CN201410447127.8 2014-09-03

Publications (1)

Publication Number Publication Date
WO2015154483A1 true WO2015154483A1 (fr) 2015-10-15

Family

ID=54287235

Family Applications (1)

Application Number Title Priority Date Filing Date
PCT/CN2014/092213 WO2015154483A1 (fr) 2014-09-03 2014-11-25 Procede d'equilibrage de trafic de reseau et super contrôleur

Country Status (2)

Country Link
CN (1) CN105471755B (fr)
WO (1) WO2015154483A1 (fr)

Cited By (1)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US10656960B2 (en) 2017-12-01 2020-05-19 At&T Intellectual Property I, L.P. Flow management and flow modeling in network clouds

Families Citing this family (5)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN106059805B (zh) * 2016-05-27 2020-10-02 新华三技术有限公司 一种流量分布的分析方法及装置
CN106572020B (zh) * 2016-10-18 2020-05-22 上海斐讯数据通信技术有限公司 一种基于sdn的全网络整形方法、装置和系统
CN107425987B (zh) * 2017-09-12 2019-11-01 中国联合网络通信集团有限公司 一种交换机控制方法、装置及系统
CN111130945B (zh) * 2019-12-30 2021-12-28 江苏万佳科技开发股份有限公司 一种数据监测云平台及使用方法
CN113347107B (zh) * 2020-03-02 2022-10-14 中国移动通信集团浙江有限公司 基于上行报文的流量调度方法、装置及计算设备

Citations (6)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN101695050A (zh) * 2009-10-19 2010-04-14 浪潮电子信息产业股份有限公司 一种基于网络流量自适应预测的动态负载均衡方法
CN103457776A (zh) * 2013-09-04 2013-12-18 杭州华三通信技术有限公司 一种双向协议无关组播中动态部署汇聚点的方法及装置
CN103581036A (zh) * 2013-10-31 2014-02-12 华为技术有限公司 一种虚拟机网络流量的控制方法、装置和系统
CN103618621A (zh) * 2013-11-21 2014-03-05 华为技术有限公司 一种软件定义网络sdn的自动配置方法、设备及系统
US20140089500A1 (en) * 2012-09-25 2014-03-27 Swaminathan Sankar Load distribution in data networks
CN103825838A (zh) * 2014-02-24 2014-05-28 上海交通大学 一种数据中心去带宽碎片化流调度方法

Family Cites Families (4)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US7239608B2 (en) * 2002-04-26 2007-07-03 Samsung Electronics Co., Ltd. Router using measurement-based adaptable load traffic balancing system and method of operation
CN101212819B (zh) * 2006-12-27 2010-12-08 华为技术有限公司 一种基于流量预测的周期性自适应汇聚的方法及系统
CN101286937B (zh) * 2008-05-16 2011-01-05 成都市华为赛门铁克科技有限公司 一种网络流量控制方法、装置及系统
CN103179046B (zh) * 2013-04-15 2016-03-30 昆山天元昌电子有限公司 基于openflow的数据中心流量控制方法及系统

Patent Citations (6)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN101695050A (zh) * 2009-10-19 2010-04-14 浪潮电子信息产业股份有限公司 一种基于网络流量自适应预测的动态负载均衡方法
US20140089500A1 (en) * 2012-09-25 2014-03-27 Swaminathan Sankar Load distribution in data networks
CN103457776A (zh) * 2013-09-04 2013-12-18 杭州华三通信技术有限公司 一种双向协议无关组播中动态部署汇聚点的方法及装置
CN103581036A (zh) * 2013-10-31 2014-02-12 华为技术有限公司 一种虚拟机网络流量的控制方法、装置和系统
CN103618621A (zh) * 2013-11-21 2014-03-05 华为技术有限公司 一种软件定义网络sdn的自动配置方法、设备及系统
CN103825838A (zh) * 2014-02-24 2014-05-28 上海交通大学 一种数据中心去带宽碎片化流调度方法

Cited By (1)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US10656960B2 (en) 2017-12-01 2020-05-19 At&T Intellectual Property I, L.P. Flow management and flow modeling in network clouds

Also Published As

Publication number Publication date
CN105471755A (zh) 2016-04-06
CN105471755B (zh) 2019-11-05

Similar Documents

Publication Publication Date Title
WO2015154483A1 (fr) Procede d'equilibrage de trafic de reseau et super contrôleur
US10129043B2 (en) Apparatus and method for network flow scheduling
EP3047609B1 (fr) Systèmes et procédé pour la réconfiguration de routes
JP4796668B2 (ja) バス制御装置
US8670310B2 (en) Dynamic balancing priority queue assignments for quality-of-service network flows
Dong et al. Service systems with slowdowns: Potential failures and proposed solutions
US9712448B2 (en) Proxy server, hierarchical network system, and distributed workload management method
CN108476175B (zh) 使用对偶变量的传送sdn流量工程方法与系统
CN102823206B (zh) 用于降低分组处理线卡中的能量消耗的方法以及分组处理线卡
CN108293200B (zh) 设备吞吐量确定
WO2015131699A1 (fr) Procédé et dispositif d'équilibrage de charge
US9197444B2 (en) Multiple-interface network device and selection method for transmitting network packets
WO2016095410A1 (fr) Dispositif et procédé de distribution de trafic de liaison
KR101448413B1 (ko) Atca-기반 장비에서 통신 트래픽을 스케줄링하기 위한 방법 및 장치
Walton Concave switching in single and multihop networks
US20180007455A1 (en) Method and system for distributed control of large photonic switched networks
US9319325B2 (en) Adaptive method and system of regulation of yellow traffic in a network
CN106453114B (zh) 流量分配方法及装置
CN107995121A (zh) 基于通配符的软件定义网络中流量统计方法
JP2018098590A (ja) トラヒック需要予測装置、トラヒック需要予測方法、及びプログラム
KR101465884B1 (ko) 소프트웨어 정의 네트워크에서의 확률적 컨트롤러 선택 방법 및 장치
Lebedenko et al. Linear-quadratic model of optimal queue management on interface of telecommunication network router
Li et al. Dynamic overload balancing in server farms
US11924061B2 (en) Required communication quality estimation apparatus, required communication quality estimation method and program
KR101180997B1 (ko) 사이버 물리 시스템에서 부하균형을 고려한 동적 제어 관리 시스템 및 제어 방법

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: 14888761

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: 14888761

Country of ref document: EP

Kind code of ref document: A1