WO2024066623A1 - 调度策略的确定方法及相关装置 - Google Patents

调度策略的确定方法及相关装置 Download PDF

Info

Publication number
WO2024066623A1
WO2024066623A1 PCT/CN2023/105128 CN2023105128W WO2024066623A1 WO 2024066623 A1 WO2024066623 A1 WO 2024066623A1 CN 2023105128 W CN2023105128 W CN 2023105128W WO 2024066623 A1 WO2024066623 A1 WO 2024066623A1
Authority
WO
WIPO (PCT)
Prior art keywords
scheduling
strategy
link
application
policy
Prior art date
Application number
PCT/CN2023/105128
Other languages
English (en)
French (fr)
Inventor
夏斌
徐慧颖
张许宝
杜正贤
吕文祥
李粤琛
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 WO2024066623A1 publication Critical patent/WO2024066623A1/zh

Links

Classifications

    • HELECTRICITY
    • H04ELECTRIC COMMUNICATION TECHNIQUE
    • H04LTRANSMISSION OF DIGITAL INFORMATION, e.g. TELEGRAPHIC COMMUNICATION
    • H04L47/00Traffic control in data switching networks
    • H04L47/50Queue scheduling
    • H04L47/58Changing or combining different scheduling modes, e.g. multimode scheduling
    • HELECTRICITY
    • H04ELECTRIC COMMUNICATION TECHNIQUE
    • H04LTRANSMISSION OF DIGITAL INFORMATION, e.g. TELEGRAPHIC COMMUNICATION
    • H04L47/00Traffic control in data switching networks
    • H04L47/10Flow control; Congestion control
    • H04L47/12Avoiding congestion; Recovering from congestion
    • HELECTRICITY
    • H04ELECTRIC COMMUNICATION TECHNIQUE
    • H04LTRANSMISSION OF DIGITAL INFORMATION, e.g. TELEGRAPHIC COMMUNICATION
    • H04L47/00Traffic control in data switching networks
    • H04L47/50Queue scheduling

Definitions

  • the present application relates to the field of network technology, and in particular to a method for determining a scheduling strategy and a related device.
  • a link is a path established between two nodes for transmitting traffic of one or several applications. For example, a link established between node A and node B for transmitting traffic of office applications.
  • the present application provides a method for determining a scheduling strategy and a related device, which can avoid the problem in the related art of directly limiting the speed of low-priority applications, which will definitely damage the quality of low-priority services.
  • a method for determining a scheduling strategy is provided.
  • the method can be performed by a control device or a network device in a network.
  • the method includes: in response to congestion on a first link, determining a first scheduling strategy for traffic of at least one application on the first link according to scheduling strategies respectively supported by multiple applications to which traffic carried by the first link belongs; and applying the first scheduling strategy.
  • a scheduling strategy is determined based on the scheduling strategies supported by multiple applications on the link, so that the control device or network device can coordinate and schedule the traffic of each application based on the scheduling strategies supported by each application, and can make full use of network resources to transmit the traffic of each application, thereby ensuring the service quality of each application to the greatest extent possible, avoiding the problem in related technologies of directly limiting the speed of low-priority applications, which will definitely damage the quality of low-priority services.
  • the scheduling strategy includes a lossless scheduling strategy and a lossy scheduling strategy.
  • the lossy scheduling strategy refers to a scheduling strategy that causes application data packet loss
  • the lossless scheduling strategy refers to a scheduling strategy that does not cause application data packet loss.
  • the lossless scheduling strategy includes at least one of the following: path adjustment strategy, compression strategy.
  • the lossy scheduling strategy includes: speed limit strategy.
  • the scheduling strategy also includes pipeline mapping adjustment strategy, and the pipeline mapping adjustment strategy can be either a lossless scheduling strategy or a lossy scheduling strategy.
  • determining a first scheduling policy for traffic of at least one application on the first link according to the scheduling policies respectively supported by multiple applications to which traffic carried by the first link belongs includes: in response to the existence of an application supporting a lossless scheduling policy, determining the first scheduling policy according to the scheduling policy supported by the application supporting the lossless scheduling policy. In response to the absence of an application supporting the lossless scheduling policy, determining the first scheduling policy according to the scheduling policy supported by the application supporting the lossy scheduling policy.
  • the lossless scheduling strategy is given priority, so as to ensure the service quality of all applications in the link as much as possible.
  • determining the first scheduling strategy according to the scheduling strategy supported by the application supporting the lossless scheduling strategy includes:
  • determining that the first scheduling policy is a scheduling policy supported by the applications supporting the lossless scheduling policy; or, in response to the number of applications supporting the lossless scheduling policy being multiple, determining that the first scheduling policy includes a scheduling policy supported by at least one application among the applications supporting the lossless scheduling policy.
  • determining the first scheduling strategy as a scheduling strategy supported by at least one application among the applications supporting the lossless scheduling strategy includes: combining the applications supporting the lossless scheduling strategy and the scheduling strategies supported by the applications to obtain at least one strategy combination, each strategy combination corresponding to a scheduling strategy for at least one application; determining whether the congestion of the first link can be resolved when scheduling is performed using scheduling strategies corresponding to each strategy combination in at least one strategy combination; in response to the existence of a strategy combination that can resolve the congestion of the first link, selecting a strategy combination from the strategy combinations that can resolve the congestion of the congested first link, and adopting the scheduling strategy corresponding to the selected strategy combination as the first scheduling strategy.
  • the combination of applications and scheduling strategies is actually a combination based on the scheduling strategy as the precision.
  • application A supports two scheduling strategies a and b.
  • application A-scheduling strategy a and application A-scheduling strategy b.
  • strategy combination determines whether there is no combination (ie, strategy combination) that can solve the congestion of the first link. If there is no combination (ie, strategy combination) that can solve the congestion of the first link, continue to combine and select until all combinations are exhausted and there is still no combination that can solve the congestion, then determine the first scheduling strategy based on the scheduling strategies supported by the application that supports the lossy scheduling strategy.
  • determining whether the congestion of the first link can be resolved by using the scheduling strategy corresponding to the combination it is determined based on current traffic data and the above selected scheduling strategy whether the congestion problem of the first link can be resolved after scheduling.
  • determining whether the congestion of the first link can be resolved by using the scheduling strategy corresponding to the combination it is determined based on the predicted traffic data and the above selected scheduling strategy whether the congestion problem of the first link can be resolved after scheduling.
  • the scheduling policy corresponding to the first policy combination can solve the congestion of the first link, and the first policy combination is any one of at least one policy combination.
  • the scheduling strategy is determined based on the predicted traffic data, so that the scheduling strategy executed by the network device can better match the traffic carried by the link in the future, further alleviate the congestion that may occur in the link in the future, and also make the frequency of scheduling required lower, saving network resources.
  • the scheduling policy corresponding to the first policy combination can resolve the congestion of the first link, including: predicting the traffic data of each application based on the traffic profiles of multiple applications; determining the traffic value expected to be carried by the first link based on the scheduling policy corresponding to the first policy combination and the predicted traffic data of each application; and determining whether the first link is congested based on the traffic value expected to be carried by the first link.
  • using the traffic profile of each application to predict traffic data can improve prediction accuracy, thereby improving the accuracy of the judgment on whether the congestion of the first link can be resolved.
  • selecting a strategy combination from strategy combinations that can solve the congestion of the congested first link includes: selecting a combination from strategy combinations that can solve the congestion of the congested first link based on at least one of application priority and application requirement.
  • a scheduling policy corresponding to a policy combination with a lower priority is selected.
  • a scheduling policy corresponding to a policy combination that can meet the application requirements in the combination is selected.
  • a scheduling policy corresponding to a policy combination with a lower priority among the combinations that can meet the application requirements in the combination is selected.
  • the method also includes: determining the links involved in the first scheduling strategy; determining whether the links involved in the first scheduling strategy will be congested after the first scheduling strategy is executed based on the traffic data of the links involved in the first scheduling strategy; in response to the links involved in the first scheduling strategy not being congested, generating a first scheduling strategy instruction based on the first scheduling strategy.
  • the method also includes: in response to congestion on the second link, determining a second scheduling policy for traffic of at least one application on the second link based on scheduling policies respectively supported by multiple applications to which the traffic carried by the second link belongs; determining the links involved in the first scheduling policy and the second scheduling policy; determining whether the links involved will be congested after the execution of the first scheduling policy and the second scheduling policy based on traffic data of the links involved; and in response to the links involved not being congested, generating a first scheduling policy instruction based on the first scheduling policy, and generating a second scheduling policy instruction based on the second scheduling policy.
  • the method also includes: before applying the first scheduling strategy, determining the expected benefit after using the first scheduling strategy for traffic scheduling; after applying the first scheduling strategy, determining the actual benefit; in response to congestion on the third link, determining a third scheduling strategy based on the expected benefit and the actual benefit.
  • the method further includes: generating a traffic profile of each application based on traffic data of multiple applications; and adjusting a mapping relationship between each application and a pipeline based on the traffic profile of each application, wherein the pipeline includes at least one of a tunnel and a queue.
  • the mapping relationship between each application and the pipeline is adjusted, including: according to the traffic pattern of each application, applications with the same traffic pattern are mapped to the same pipeline, and applications with different traffic patterns are mapped to different pipelines.
  • the traffic of applications with the same traffic pattern is mapped to the same pipeline, so that the traffic characteristics of different pipelines are easier to distinguish, thereby facilitating the management and scheduling of each pipeline.
  • the method further includes: obtaining a traffic profile of each group under each grouping method after grouping multiple applications using multiple grouping methods, each group including at least one application; and adjusting a mapping relationship between multiple individuals and pipelines according to the traffic profile of each group, the pipeline including at least one of a tunnel and a queue.
  • the mapping relationship between multiple applications and pipelines involved in the traffic statistics data is adjusted, including: based on the smoothness of the traffic profile of each group under each grouping method, selecting a grouping method whose smoothness meets the smoothness condition; mapping applications in the same group under the selected grouping method to the same pipeline, and mapping applications in different groups to different pipelines.
  • the stability refers to the degree of stability, which can be evaluated according to the flow pattern and average-to-peak ratio of the flow.
  • the flow pattern can be selected as small-value oscillation and stable double-peak type, and then the average-to-peak ratio that meets the requirements is selected among these flow patterns.
  • a device for determining a scheduling strategy comprising:
  • a determining unit configured to determine, in response to congestion on the first link, a first scheduling strategy for traffic of at least one application on the first link according to scheduling strategies respectively supported by multiple applications to which traffic carried by the first link belongs;
  • An application unit configured to apply a first scheduling strategy.
  • the determining unit is configured to determine the first scheduling strategy in response to the existence of an application supporting the lossless scheduling strategy and according to scheduling strategies supported by the application supporting the lossless scheduling strategy.
  • the determination unit is used to determine, in response to the number of applications supporting the lossless scheduling policy being one, that the scheduling policy supported by the applications supporting the lossless scheduling policy is the first scheduling policy; or, in response to the number of applications supporting the lossless scheduling policy being at least two, determine that the scheduling policy supported by at least one application among the applications supporting the lossless scheduling policy is the first scheduling policy.
  • the determination unit is used to combine applications that support lossless scheduling strategies and scheduling strategies supported by the applications to obtain at least one strategy combination, each strategy combination corresponding to a scheduling strategy for at least one application; determine whether the congestion of the first link can be resolved when scheduling using scheduling strategies corresponding to each strategy combination in at least one strategy combination; in response to the existence of a strategy combination that can resolve the congestion of the first link, select a strategy combination from the strategy combinations that can resolve the congestion of the first link, and adopt the scheduling strategy corresponding to the selected strategy combination as the first scheduling strategy.
  • the determination unit is used to determine whether the scheduling policy corresponding to the first policy combination can solve the congestion of the first link based on the predicted traffic data and the scheduling policy corresponding to the first policy combination, and the first policy combination is any one of at least one policy combination.
  • the determination unit is used to predict the traffic data of each application based on the traffic portraits of multiple applications; determine the traffic value expected to be carried by the first link based on the scheduling strategy corresponding to the first strategy combination and the predicted traffic data of each application; and determine whether the first link is congested based on the traffic value expected to be carried by the first link.
  • the determining unit is configured to select a policy combination from policy combinations that can resolve congestion of the congested first link based on at least one of an application priority and an application requirement.
  • the determining unit is configured to determine the first scheduling strategy according to scheduling strategies supported by applications supporting the lossy scheduling strategy in response to the absence of an application supporting the lossless scheduling strategy.
  • the determination unit is further used to determine the link involved in the first scheduling strategy; and determine whether the link involved in the first scheduling strategy will be congested after the first scheduling strategy is executed according to the traffic data of the link involved in the first scheduling strategy;
  • the device also includes: a generating unit, configured to generate a first scheduling strategy instruction based on the first scheduling strategy in response to the link involved in the first scheduling strategy not being congested.
  • a generating unit configured to generate a first scheduling strategy instruction based on the first scheduling strategy in response to the link involved in the first scheduling strategy not being congested.
  • the determination unit is further configured to, in response to congestion on the second link, determine a second scheduling strategy for traffic of at least one application on the second link according to scheduling strategies respectively supported by multiple applications to which traffic carried by the second link belongs; determine the links involved in the first scheduling strategy and the second scheduling strategy; and determine, based on traffic data of the involved links, the links involved after the first scheduling strategy and the second scheduling strategy are executed. Will the road be congested?
  • the generating unit is configured to generate a first scheduling policy instruction based on the first scheduling policy and generate a second scheduling policy instruction based on the second scheduling policy in response to the involved link not being congested.
  • the determination unit is also used to determine the expected benefit after using the first scheduling strategy for traffic scheduling before applying the first scheduling strategy; determine the actual benefit after applying the first scheduling strategy; and determine the third scheduling strategy based on the expected benefit and the actual benefit in response to congestion on the third link.
  • a generating unit is used to generate a traffic profile of each application based on the traffic data of multiple applications;
  • the device also includes: an adjustment unit, which is used to adjust the mapping relationship between each application and the pipeline according to the traffic profile of each application, and the pipeline includes at least one of a tunnel and a queue.
  • a generating unit is used to obtain a traffic profile of each group under each grouping method after grouping multiple applications in multiple grouping methods, each group including at least one application;
  • the adjustment unit is used to adjust the mapping relationship between multiple groups and pipelines according to the traffic profile of each group, and the pipeline includes at least one of a tunnel and a queue.
  • an electronic device which is a control device or a network device.
  • the electronic device includes a processor and a memory.
  • the memory is used to store software programs and modules.
  • the processor implements the method in the first aspect or any possible implementation of the first aspect by running or executing the software program and/or module stored in the memory.
  • the number of the processors is one or more, and the number of the memories is one or more.
  • the memory may be integrated with the processor, or the memory may be provided separately from the processor.
  • the memory can be a non-transitory memory, such as a read-only memory (ROM), which can be integrated with the processor on the same chip or can be set on different chips.
  • ROM read-only memory
  • a computer program product comprising computer program code, and when the computer program code is executed by a computer, the computer executes the method in the first aspect or any possible implementation of the first aspect.
  • the present application provides a computer-readable storage medium, wherein the computer-readable storage medium is used to store program codes executed by a processor, wherein the program codes include a method for implementing any possible implementation of the first aspect above.
  • a chip comprising a processor, the processor being used to call and execute instructions stored in a memory from the memory, so that a communication device equipped with the chip executes a method in any possible implementation manner of the above-mentioned first aspect.
  • another chip in a seventh aspect, another chip is provided.
  • the another chip includes an input interface, an output interface, a processor, and a memory.
  • the input interface, the output interface, the processor, and the memory are connected via an internal connection path.
  • the processor is used to execute the code in the memory, and when the code is executed, the processor is used to execute the method in any possible implementation of the first aspect above.
  • FIG1 is a schematic diagram of an application scenario provided by an embodiment of the present application.
  • FIG2 is a schematic diagram of an application scenario provided by an embodiment of the present application.
  • FIG3 is a flow chart of a method for determining a scheduling strategy provided in an embodiment of the present application.
  • FIG4 is a flow chart of a method for determining a scheduling strategy provided in an embodiment of the present application.
  • FIG5 is a schematic diagram of a WFQ weight adjustment provided in an embodiment of the present application.
  • FIG6 is a flow chart of a method for determining a scheduling strategy provided in an embodiment of the present application.
  • FIG7 is a flow chart of a method for determining a scheduling strategy provided in an embodiment of the present application.
  • FIG8 is a flow chart of a method for determining a scheduling strategy provided in an embodiment of the present application.
  • FIG9 is a flow chart of a method for determining a scheduling strategy provided in an embodiment of the present application.
  • FIG10 is a flowchart of a method for determining a scheduling strategy provided in an embodiment of the present application.
  • FIG11 is a flowchart of a method for determining a scheduling strategy provided in an embodiment of the present application.
  • FIG12 is a schematic diagram of a network topology provided in an embodiment of the present application.
  • FIG13 is a schematic diagram of a scheduling process provided in an embodiment of the present application.
  • FIG14 is a schematic diagram of a module provided in an embodiment of the present application.
  • FIG15 is a block diagram of a device for determining a scheduling strategy provided in an embodiment of the present application.
  • FIG. 16 is a schematic diagram of the structure of a control device or a network device provided in an embodiment of the present application.
  • the application scenario can be a distributed storage system or a communication network such as a data center.
  • the following uses a distributed storage system as an example to illustrate the application scenario of the present application.
  • FIG1 is a schematic diagram of the structure of an application scenario provided by an embodiment of the present application.
  • the application scenario includes a control device 11 and a network device 12.
  • the control device 11 may be a management node in a network, for example, a network management device (such as a software defined network (SDN) controller).
  • the network device 12 may be a network node in a network, for example, a routing device.
  • the network device 12 includes a provider device, that is, a P device in the figure, and the network device 12 also includes a provider edge device, that is, a PE device in the figure.
  • control device 11 is connected to the PE device.
  • the control device 11 can be connected to each P device through the PE device, or the control device 11 can be directly connected to each P device (not directly shown to avoid overly complicated lines in the figure).
  • a link is established between any two network devices 12 for transmitting service data.
  • a tunnel is configured between two PE devices. The tunnel is used for transmission of a specific application and can transmit traffic of at least one application.
  • PE1 is the ingress device of the tunnel and PE2 is the egress device of the tunnel.
  • the tunnel can be carried on one path, such as the PE1-P1-PE2 path.
  • the tunnel can also be carried on multiple paths, such as the PE1-P1-PE2 and PE1-P2-P3-PE2 paths, where the two paths transmit the traffic of the tunnel in a certain proportion.
  • FIG2 is a schematic diagram of the structure of another application scenario provided by an embodiment of the present application.
  • the difference between this application scenario and the application scenario of FIG1 is that the network device 12 also includes PE3, and a tunnel may also be configured between PE3 and PE2, and the tunnel is carried on the PE3-P2-P3-PE2 path.
  • P2 only receives the traffic of PE1
  • P2 aggregates the traffic of PE1 and PE3.
  • the application scenario may also include a user device.
  • the user device may be an independent device connected to the control device, or the user device may be integrated in the control device.
  • FIG3 is a flow chart of a method for determining a scheduling strategy provided by an embodiment of the present application.
  • the method can be applied to a control device or a network device.
  • the following is an example of the method being applied to a control device.
  • the method includes the following steps.
  • the control device determines a scheduling policy for traffic of at least one application on the first link according to scheduling policies respectively supported by multiple applications to which the traffic carried by the first link belongs.
  • Link congestion refers to the traffic carried by the link exceeding the congestion threshold.
  • the congestion threshold is 80%. For a link with a bandwidth of 100 Mbps, when the traffic exceeds 80 Mbps, it is determined that the link is congested.
  • the scheduling policy supported by the application is usually a scheduling policy that has no effect or has a small effect on the business instructions of the application.
  • control device configures the scheduling policy instruction on an ingress device that carries traffic of the at least one application.
  • control device configures the scheduling policy instructions of each application to the ingress device that carries the traffic of the application.
  • the ingress device is, for example, an ingress device of a tunnel that carries traffic of each application function in the at least one application.
  • a scheduling strategy is determined according to the scheduling strategies supported by multiple applications on the link, so that the control device can coordinate and schedule the traffic of each application based on the scheduling strategies supported by each application, and can make full use of network resources to transmit the traffic of each application, thereby ensuring the service quality of each application to the greatest extent possible, avoiding the direct dispatch of low-priority applications in the related art.
  • FIG4 is a flow chart of another method for determining a scheduling strategy provided by an embodiment of the present application.
  • the method can be applied to a control device or a network device.
  • the following is an example of the method being applied to a control device.
  • the method includes the following steps.
  • S21 The control device determines that congestion exists on the first link.
  • control device receives traffic data sent by the network device, and determines that congestion exists on the first link according to the traffic data.
  • each network device e.g., each P device and each PE device in Figures 1 to 2 collects traffic from the outbound port and/or inbound port, and then sends the traffic data of the corresponding link to the control device.
  • the link identifier can be carried when sending the traffic data.
  • For a link it can be determined whether the link is congested based on the traffic sent by the egress port of the link's sending end or the traffic received by the ingress port of the receiving end.
  • the inlet device of the tunnel in the network collects the traffic of the tunnel and/or collects the traffic of the application, and then sends the traffic data of the tunnel and/or application to the control device.
  • the traffic of the application is collected based on the application flow identifier
  • the traffic of the tunnel is collected based on the tunnel identifier.
  • the tunnel identifier, application flow identifier, etc. can be carried when sending the traffic data for indication.
  • the network device can collect data at a predetermined time interval, such as collecting flow data once a second, and sending each collected flow data to the control device.
  • a predetermined time interval such as collecting flow data once a second
  • the network device can also collect for a period of time and then send it, such as collecting flow data once a second, and sending 60 collected flow data to the control device every minute.
  • control device After the control device obtains the traffic data, it determines whether each link in the network topology is congested based on the traffic data.
  • the method of determining whether a link is congested is to determine whether the traffic carried by the link exceeds the congestion threshold. Different links can have corresponding congestion thresholds.
  • the traffic carried by the link can be determined by the traffic data of the link, or by the traffic data of the application or tunnel carried on the link, which will not be elaborated herein.
  • control device receives a first link congestion message sent by the network device, where the first link congestion message is used to indicate that congestion exists on the first link.
  • the process of determining congestion is completed by the network device.
  • the network device can determine link congestion in the aforementioned manner.
  • the network device can also determine link congestion based on the performance data of the egress port, egress queue, or ingress queue associated with the link detected by the device. For example, when the packet loss rate of the egress port is greater than a threshold, the network device determines that link congestion occurs in the link associated with the egress port. For another example, when the queue occupancy of the egress queue exceeds a threshold, or when the queue occupancy of the egress queue exceeds the threshold for a duration that exceeds a duration threshold, the network device determines that link congestion occurs in the link associated with the egress queue.
  • S22 The control device obtains scheduling strategies supported by each application to which the traffic carried by the first link belongs.
  • the scheduling policies supported by each application may be configured by the user on the user equipment and then sent to the control device.
  • the scheduling strategies supported by each application may be pre-configured in the control device.
  • the control device determines a scheduling policy for traffic of at least one application on the first link according to scheduling policies supported by multiple applications to which the traffic carried by the first link belongs.
  • the control device sends a scheduling policy instruction to an ingress device that carries the traffic of the at least one application, so that the ingress device executes the scheduling policy instruction.
  • the control device may also generate a scheduling strategy instruction for each application in the at least one application, and configure the scheduling strategy instruction for each application. For example, the control device sends the scheduling strategy instruction to the inlet device carrying the application to instruct the inlet device to execute the scheduling strategy instruction.
  • the inlet device is, for example, a tunnel inlet device.
  • a network device eg, an entrance device of a tunnel
  • the network device may directly generate a scheduling policy instruction and execute the scheduling policy instruction.
  • the scheduling strategy includes a lossless scheduling strategy and a lossy scheduling strategy.
  • the lossy scheduling strategy refers to a scheduling strategy that causes application data packet loss
  • the lossless scheduling strategy refers to a scheduling strategy that does not cause application data packet loss.
  • the lossless scheduling strategy includes at least one of the following: path adjustment strategy, compression strategy.
  • the lossy scheduling strategy includes: speed limit strategy.
  • the scheduling strategy also includes pipeline mapping adjustment strategy, and the pipeline mapping adjustment strategy can be either a lossless scheduling strategy or a lossy scheduling strategy.
  • the pipeline mapping adjustment strategy is lossy depends on the type of the adjusted pipeline. For example, when there is a rate-limited pipeline (tunnel or queue), the pipeline mapping adjustment will map the application from the non-rate-limited pipeline to the rate-limited pipeline, which is a lossy scheduling strategy; if the adjusted pipeline is a non-rate-limited pipeline, the pipeline mapping adjustment is a lossless scheduling strategy.
  • the path adjustment strategy includes a routing adjustment strategy, a tunnel diversion path adjustment strategy, a tunnel diversion ratio adjustment strategy, and the like.
  • route adjustment strategy refers to a strategy for changing the transmission path of application traffic. Taking Figure 1 as an example, route adjustment may refer to adjusting the application route from PE1-P1-PE2 to PE1-P2-P3-PE2.
  • Tunnel diversion path adjustment refers to adjusting the application traffic from one path of the tunnel to another path, or creating a new path for the tunnel and adjusting the application traffic from one path of the tunnel to the newly created path.
  • tunnel diversion path adjustment can refer to adjusting the application from the path PE1-P2-P3-PE2 to the path PE1-P1-PE2.
  • Tunnel traffic splitting ratio adjustment refers to the strategy of adjusting the traffic splitting ratio of application traffic on multiple paths of the tunnel. Taking Figure 2 as an example, the traffic splitting ratio of application traffic on the two paths of the tunnel, PE1-P2-P3-PE2 and PE1-P1-PE2, is 1:1. Through tunnel traffic splitting ratio adjustment, the above splitting ratio can be adjusted from 1:1 to 2:1.
  • the compression strategy refers to a strategy of compressing application data packets using a compression algorithm, thereby reducing the bandwidth required for transmitting the data packets by compressing the data packets.
  • the speed limit policy includes application traffic speed limit policy, application traffic tunnel speed limit policy, application mapped queue shaping policy, application mapped weighted fair queue (WFQ) weight adjustment policy, etc.
  • the application traffic rate limiting strategy refers to a strategy for limiting the traffic rate of the application, for example, limiting the rate of data packets with the application flow identifier of the application.
  • the speed limit strategy of the tunnel where the application traffic is located refers to the strategy for limiting the speed of the traffic of the entire tunnel.
  • the speed limit is set for the data packets carrying the tunnel identifier.
  • the speed limit of the tunnel where the application traffic is located can be configured by adjusting the committed access rate (CAR) value, usually at the ingress port of the tunnel entrance device, such as limiting the bandwidth of the tunnel where the virus library download application is located from 100Mbps to 50Mbps.
  • CAR committed access rate
  • the queue shaping strategy of application mapping refers to the strategy of limiting the speed of the queue mapped to the application.
  • the queue here refers to the sending queue of the network device. It is usually operated at the egress port of the tunnel inlet device, for example, reducing the shaping value of queue 1 from 100Mbps to 50Mbps.
  • the WFQ weight adjustment strategy of the queue to which the mapping is applied refers to a strategy for adjusting the weight of the WFQ queue, so that the bandwidth obtained by the queue changes.
  • the WFQ weight adjustment is exemplarily illustrated below in conjunction with Figure 5, which is a schematic diagram of a WFQ weight adjustment provided in an embodiment of the present application.
  • Q represents the queue type
  • PQ represents the queue type as a priority queue
  • WFQ represents the queue type as a weighted fair queue, wherein the PQ queue uses the link bandwidth first and the WFQ queue uses the remaining bandwidth.
  • W represents the weight.
  • the weight ratio of the four WFQ queues was 1:2:3:3, and after the adjustment, the weight ratio of the four WFQ queues was 2:2:3:3. Increasing the weight of the first WFQ queue is equivalent to reducing the weights of the other queues in the WFQ queue.
  • the pipeline mapping adjustment strategy includes a tunnel mapping adjustment strategy, a queue mapping adjustment strategy, and the like.
  • the tunnel mapping adjustment strategy refers to the strategy for adjusting the tunnel mapped by the application, and adjusting the tunnel mapped by the application from one tunnel to another. For example, there are tunnels A and B. Application A is originally mapped to tunnel A, and application A is mapped to tunnel B through adjustment. If tunnel B is not speed-limited, the above adjustment belongs to the lossless scheduling strategy; if tunnel B is speed-limited, the above adjustment belongs to the lossy scheduling strategy.
  • the queue mapping adjustment strategy refers to the strategy for adjusting the queue mapped by the application, adjusting the queue mapped by the application from one queue to another.
  • the network device has queue A and queue B.
  • Application B was originally mapped to queue A, and application B is mapped to queue B through adjustment. If queue B is not speed-limited, the above adjustment belongs to a lossless scheduling strategy; if queue B is speed-limited, the above adjustment belongs to a lossy scheduling strategy.
  • the virus library download application was originally mapped to queue A, queue A is a non-speed-limited queue, and queue B is a speed-limited queue configured with queue shaping. In order to protect the bandwidth of other applications in queue A, the virus library download application is remapped to queue B.
  • the scheduling strategies supported by the applications may be configured for each application respectively, so that the control device can learn the scheduling strategies supported by each application.
  • the scheduling policies supported by the application by default are configured. In this case, each application supports these scheduling policies by default.
  • some applications are configured to support scheduling policies separately, while other applications are not configured to support scheduling policies separately. These applications support the scheduling policies supported by the aforementioned applications by default.
  • FIG6 is a flow chart of another method for determining a scheduling strategy provided by an embodiment of the present application.
  • the method can be applied to a control device or a network device. The following description is made by taking the application of the method to a control device as an example. As shown in FIG6, the method includes the following steps.
  • the control device determines an application that supports a lossless scheduling strategy from among multiple applications to which the traffic carried by the first link belongs.
  • step S32 When there is an application supporting the lossless scheduling strategy, step S32 is executed; when there is no application supporting the lossless scheduling strategy, step S33 is executed.
  • control device determines the scheduling strategy according to the scheduling strategy supported by the application supporting the lossless scheduling strategy, and generates a scheduling strategy instruction.
  • the scheduling policy instruction includes the identifier of the application to be scheduled, the identifier of the scheduling policy to be used, and scheduling related parameters.
  • the scheduling related parameters are parameter values involved in scheduling execution, such as the identifier of the adjusted link during path adjustment.
  • step S32 in response to the number of applications supporting the lossless scheduling policy being one, determining that the scheduling policy includes a scheduling policy supported by the application supporting the lossless scheduling policy. In response to the number of applications supporting the lossless scheduling policy being multiple, determining that the scheduling policy includes a scheduling policy supported by at least one application among the multiple applications supporting the lossless scheduling policy.
  • control device may determine the scheduling strategy in the following manner:
  • the combination of applications and scheduling policies is actually to combine them with the scheduling policy as the precision.
  • application A supports two scheduling policies a and b.
  • application A-scheduling policy a and application A-scheduling policy b. That is, one policy combination includes application A-scheduling policy a, and the other policy combination includes application A-scheduling policy b.
  • the combination can also be based on certain conditions, such as selecting applications with priorities lower than a certain threshold for combination, etc.
  • the conditions here are not limited in this application.
  • step S33 If there is no combination that can solve the congestion of the first link, continue to combine and select until all combinations are exhausted and there is still no combination that can solve the congestion, then the control device executes step S33.
  • control device selects a policy combination from policy combinations that can resolve congestion of the congested first link based on at least one of a priority of the application and an application requirement.
  • the application priority may be represented by numbers, such as 1, 2, 3, etc. Generally, the larger the number, the lower the priority.
  • the application requirements include at least one of the following: bandwidth requirements, latency requirements, packet loss requirements, jitter requirements, service level agreement requirements, and customer experience requirements.
  • bandwidth requirements may include collection bandwidth requirements and guaranteed bandwidth requirements.
  • Collection bandwidth requirements refer to the bandwidth occupied by the traffic data collected when collecting traffic for the application. By limiting the collection bandwidth requirements, it is ensured that the traffic data collected for the application is sufficient as data support for determining the scheduling strategy.
  • Guaranteed bandwidth requirements are used to limit the minimum guaranteed bandwidth, maximum guaranteed bandwidth and/or bandwidth guarantee interval required by the application.
  • the minimum guaranteed bandwidth is also the minimum bandwidth that needs to be guaranteed, which can be limited by the committed information rate (CIR), such as 50Mbps; the maximum guaranteed bandwidth is the upper limit of the guaranteed bandwidth, which can be limited by the peak information rate (PIR), such as 100Mbps; the bandwidth guarantee interval includes the minimum bandwidth and maximum bandwidth that need to be guaranteed, such as 50Mbps ⁇ 100Mbps.
  • CIR committed information rate
  • PIR peak information rate
  • the bandwidth guarantee interval includes the minimum bandwidth and maximum bandwidth that need to be guaranteed, such as 50Mbps ⁇ 100Mbps.
  • control device selects a scheduling policy corresponding to a policy combination with a lower priority based on the priority of each application.
  • the control device selects the combination of application C and application D for scheduling.
  • the priority can consider the average priority or the highest priority.
  • This implementation method can ensure that the services of high-priority applications are not affected by scheduling low-priority services.
  • control device selects, based on application requirements of each application, a scheduling policy corresponding to a policy combination that can meet the application requirements in the policy combination.
  • the default supported scheduling policies are selected and combined. At this time, the selected scheduling policy may not match the application requirements. Therefore, this situation needs to be excluded to ensure the service quality of the application.
  • control device selects, based on the priorities of the respective applications and the application requirements of the respective applications, a scheduling strategy corresponding to a combination with a lower priority among the combinations that can meet the application requirements in the combination.
  • the process of determining the scheduling strategy it is necessary to consider whether the determined scheduling strategy can solve the congestion problem of the first link.
  • the control device determines whether the congestion problem of the first link can be solved after scheduling based on the current traffic data and the above selected scheduling strategy. For example, the control device calculates whether the first link will still be congested after scheduling is completed based on the current traffic value of the first link and the traffic reduced by scheduling.
  • the control device determines whether the congestion problem of the first link can be solved after scheduling based on the predicted traffic data and the above selected scheduling strategy. For example, the control device can predict the traffic data of each application based on the traffic profile of the application, and then determine whether the first link will still be congested based on the traffic value expected to be carried by the first link and the congestion threshold based on the selected scheduling strategy.
  • the traffic data of each application is predicted respectively; based on the scheduling policy corresponding to the first policy combination and the predicted traffic data of each application, the traffic value expected to be carried by the first link is determined; based on the traffic value expected to be carried by the first link, whether the first link is congested is determined.
  • the first policy combination is any one of at least one policy combination.
  • whether there is congestion is determined according to the expected carried traffic value and the congestion threshold.
  • the control device generates a traffic profile of each application based on the traffic sequence of each application.
  • the traffic sequence usually corresponds to a longer time, such as one or more days, so that the traffic profile can be accurately portrayed.
  • traffic profile refers to the traffic characteristics obtained based on the traffic sequence, such as traffic pattern, traffic peak value, average-peak ratio and other characteristics.
  • the traffic data (e.g., traffic peak) of the application within a period of time can be predicted.
  • the application carried by the first link can be determined if the scheduling strategy corresponding to the first strategy combination is used.
  • the predicted traffic data of the application carried by the first link the traffic value expected to be carried by the first link is determined, and then it is predicted whether there will be congestion, that is, whether the congestion problem has been solved.
  • the scheduling strategy is determined based on the predicted traffic data, so that the scheduling strategy executed by the network device can better match the traffic carried by the link in the future, further alleviate the congestion that may occur in the link in the future, and also make the frequency of scheduling required lower, saving network resources.
  • control device determines the scheduling strategy according to the scheduling strategy supported by the application supporting the lossy scheduling strategy, and generates a scheduling strategy instruction.
  • the parameters for lossy scheduling are also determined.
  • a combination method can also be used to list the combination of applications and strategies, and then select the strategy combination that can solve the congestion of the first link, until all the listed strategy combinations cannot solve the congestion, then select the strategy combination that cannot solve the congestion.
  • the parameters involved in the scheduling use the default values or the values corresponding to the application requirements. For example, if a speed limit scheduling strategy is involved, the speed limit value is set according to the minimum guaranteed bandwidth supported by the application.
  • the application requirements of key applications include bandwidth requirements, and the minimum guaranteed bandwidth is relatively high.
  • the policy combination in which the scheduling policy including the key applications is a rate limit policy can be excluded.
  • the application requirements of the disaster recovery application include bandwidth requirements, and the minimum guaranteed bandwidth is low.
  • a policy combination in which the scheduling policy of the disaster recovery application is a speed limit policy can be included in the selection range.
  • the speed limit policy can be executed for the application or for the tunnel.
  • the speed of the tunnel can be directly limited.
  • the speed of the tunnel can also be directly limited.
  • the solution provided in the present application can realize scheduling of different granularities, such as application granularity, tunnel granularity, etc.
  • the scheduling schemes of different granularities make the traffic scheduling of the present application more refined and easier to alleviate link congestion.
  • the scheduling strategy is first determined based on at least one of the application priority and application requirements; if the scheduling strategy includes one of the path adjustment strategy, the tunnel diversion path adjustment strategy, and the pipeline mapping adjustment strategy, the links involved in the scheduling strategy (links other than the first link) are determined; according to the traffic data of the links involved in the scheduling strategy, it is determined whether the links involved in the scheduling strategy will be congested after the scheduling strategy is executed. If there will be no congestion, the scheduling strategy instruction is generated using the scheduling strategy. If there will be congestion, the scheduling strategy is re-determined according to step S32.
  • step S33 is executed.
  • step S33 similarly, if the scheduling strategy includes a pipeline mapping adjustment strategy, it is also necessary to consider whether the link involved in the scheduling strategy will be congested. If not, the scheduling strategy is used to generate a scheduling strategy instruction. If congestion occurs, the scheduling strategy is re-determined according to step S33.
  • the above-mentioned embodiments are mainly directed to the method for determining the scheduling strategy when a single link is congested.
  • the scheduling strategy is determined for the two links respectively, and the scheduling goal is that the links involved in the scheduling strategy will not be congested.
  • the following is an exemplary description of the detailed process of how to determine the scheduling strategy when there are two or more congested links in conjunction with Figure 7.
  • Figure 7 is a flow chart of another method for determining the scheduling strategy provided by an embodiment of the present application. The method can be applied to a control device or to a network device. The following description will be given by taking the application of the method to a control device as an example. As shown in Figure 7, the method comprises the following steps.
  • S41 The control device determines that the first link and the second link are congested.
  • the first link and the second link are any two of the congested links. Taking FIG2 as an example, the first link is P1-PE2 (belonging to the path PE1-P1-PE2), and the second link is P2-P3 (belonging to the path PE1-P2-P3-PE2).
  • the control device obtains scheduling policies supported by each application to which the traffic carried by the first link belongs, and obtains scheduling policies supported by each application to which the traffic carried by the second link belongs.
  • the control device determines a first scheduling policy for the traffic of at least one application on the first link according to the scheduling policies supported by each application to which the traffic carried by the first link belongs; and determines a second scheduling policy for the traffic of at least one application on the second link according to the scheduling policies supported by each application to which the traffic carried by the second link belongs.
  • control device determines whether the links involved in the first scheduling strategy and the second scheduling strategy will be congested after the first scheduling strategy and the second scheduling strategy are executed according to the traffic data of the links involved in the first scheduling strategy and the second scheduling strategy. If not, execute S45. If congested, repeat step S43.
  • step S32 For each link involved, it is determined based on the collected traffic data whether congestion will occur after scheduling. The detailed steps of the determination can be found in step S32.
  • control device adopts the first scheduling strategy to generate a first scheduling strategy instruction, and adopts the second scheduling strategy to generate a second scheduling strategy instruction.
  • the scheduling strategy for each link congestion will not cause congestion on other links. That is, the solution takes the entire network as the scheduling object, and can globally schedule traffic to solve the congestion problem in the network.
  • the benefit after scheduling can be predicted, which is used to indicate the improvement of congested links.
  • the actual benefit of scheduling is determined.
  • the actual benefit can be fed back to the user equipment for visual display, and on the other hand, based on the expected The deviation between the revenue and the actual revenue can provide guidance for the determination of subsequent scheduling strategies, thereby achieving better results after scheduling.
  • FIG8 is a flow chart of another method for determining a scheduling strategy provided by an embodiment of the present application.
  • the method can be applied to a control device or a network device.
  • the following is an example of the method being applied to a control device. As shown in FIG8, the method includes the following steps.
  • control device determines, based on the determined scheduling strategy, the expected benefit after using the scheduling strategy to perform traffic scheduling.
  • Step S51 is performed before applying the scheduling strategy, that is, predicting the benefits generated by applying the scheduling strategy.
  • the flow data of the first link after the scheduling strategy is executed is predicted, and the expected benefit after scheduling is determined according to the predicted flow data of the first link.
  • the expected benefits may include: the number of congested links is reduced, the proportion of traffic occupancy of congested links is reduced, etc.
  • the control device determines the actual benefit.
  • control device receives scheduled traffic data sent by the network device, and determines the actual revenue of the first link according to the traffic data.
  • control device receives the actual revenue sent by the network device.
  • control device determines a scheduling strategy based on the expected benefit and the actual benefit.
  • the third link may be the first link or the second link mentioned above, or may be a link other than the first link and the second link.
  • the control device determines the scheduling strategy, in addition to the factors considered in the previous article, the deviation between the expected benefit and the actual benefit is additionally considered. For example, when the actual benefit does not reach the expected benefit, when determining the scheduling strategy for the third link, a threshold lower than the congestion threshold of the third link can be used as the target for scheduling, so that the traffic scheduling is more powerful, so that the final actual benefit can meet the requirements. On the contrary, when the actual benefit exceeds the expected benefit, when determining the scheduling strategy for the third link, a threshold higher than the congestion threshold of the third link can be used as the target for scheduling. If the actual benefit is equivalent to the expected benefit, scheduling can be performed with the congestion threshold of the third link as the target.
  • Figure 9 is a flow chart of another method for determining a scheduling strategy provided by an embodiment of the present application.
  • the method can be applied to a control device or to a network device.
  • the following is an example of the application of the method to a control device. As shown in Figure 9, the method includes the following steps.
  • control device generates a traffic profile of each application based on the traffic data of each application.
  • control device adjusts the mapping relationship between each application and the pipeline according to the traffic profile of each application, and the pipeline includes at least one of a tunnel and a queue.
  • control device adjusts the mapping relationship between each application and the pipeline according to the traffic profile of each application, including: the control device maps applications with the same traffic pattern to the same pipeline and maps applications with different traffic patterns to different pipelines according to the traffic pattern of each application.
  • the traffic of applications with the same traffic pattern is mapped to the same pipeline, so that the traffic characteristics of different pipelines are easier to distinguish, thereby facilitating the management and scheduling of each pipeline.
  • the traffic pattern refers to the traffic change trend reflected by the traffic profile, including small value oscillation type, sudden burr type, stable double peak type and oscillating double peak type. Determining the traffic pattern based on traffic data is actually to determine the traffic pattern based on the change trend of the waveform formed by the traffic data. The detailed process of determining the traffic pattern is not described here.
  • the traffic pattern is used to indicate the changing trend of the traffic value of the application within the first time period.
  • the first time period can refer to the collection time period of traffic data.
  • the traffic is a relatively low traffic value (for example, 1 to 3 Mb/s) during most of the time (for example, more than 50%) within the collection time period.
  • the burst burr type traffic pattern the span of the traffic value in the traffic sequence is very large, ranging from a few Mb/s to hundreds of Mb/s, and the traffic change is irregular, that is, no obvious peaks and troughs are formed.
  • the traffic value in the traffic sequence changes in a peak-valley-peak waveform, and there are few traffic glitches in the peaks and valleys.
  • the traffic sequence The flow rate changes in a peak-valley-peak waveform, and there are many flow glitches in the peaks and valleys (more than the smooth double peak type).
  • Figure 10 is a flow chart of another method for determining a scheduling strategy provided by an embodiment of the present application.
  • the method can be applied to a control device or a network device.
  • the following is an example of the method being applied to a control device.
  • the method includes the following steps.
  • control device obtains the traffic profile of each group under each grouping method after grouping multiple applications using multiple grouping methods, and each group includes at least one application.
  • the multiple applications are applications involved in the first traffic statistical data.
  • control device adjusts the mapping relationship between multiple applications and pipelines according to the traffic profile of each group, and the pipeline includes at least one of a tunnel and a queue.
  • control device adjusts the mapping relationship between multiple applications and pipelines involved in the traffic statistics data according to the traffic profile of each group, including: the control device selects a grouping method whose smoothness meets the smoothness condition according to the smoothness of the traffic profile of each group under each grouping method; the control device maps the applications of the same group under the selected grouping method to the same pipeline, and maps applications of different groups to different pipelines.
  • the stability refers to the degree of stability, which can be evaluated according to the flow pattern and average-to-peak ratio of the flow.
  • the flow pattern can be selected as small-value oscillation and stable double-peak type, and then the average-to-peak ratio that meets the requirements is selected among these flow patterns.
  • FIG11 is a flow chart of another method for determining a scheduling strategy provided by an embodiment of the present application.
  • the method can be jointly performed by a control device and a network device (tunnel entrance device) in the application scenario shown in FIG1 or FIG2, and a user device (the user device can be a server device) not shown in the figure.
  • the method includes the following steps.
  • the user equipment sends user configuration information to the control device.
  • the control device receives the user configuration information sent by the user equipment.
  • the user configuration information includes at least one of application priority, application requirements, and application support policies.
  • APP application
  • priority 6 priority 6
  • minimum guaranteed bandwidth 10 Mbps
  • application support policy application speed limit policy
  • APP2 priority 5
  • application support policy tunnel mapping adjustment policy, tunnel diversion path adjustment policy, application speed limit policy.
  • application support policy tunnel mapping adjustment policy, tunnel diversion path adjustment policy, application speed limit policy.
  • APP4 priority 3
  • application support policy tunnel mapping adjustment policy
  • application speed limit policy 80 Mbps
  • the definition of each application may be included in the user configuration information.
  • the definition of the application can be distinguished by application flow identification information, wherein the application flow identification information is based on the Internet Protocol (IP) segment, quintuple, or application aware networking (APN) identification.
  • IP Internet Protocol
  • API application aware networking
  • the application flow identification information may be one of the following:
  • a tuple consisting of the source IP segment and the destination IP segment
  • a five-tuple consisting of source IP address, destination IP address, source port, destination port, and protocol number
  • the APN 6 identification is usually composed of the source IP segment-destination IP segment and service level agreement (SLA).
  • SLA service level agreement
  • Step S81 is an optional step.
  • the user configuration information can also be directly configured in the control device without the participation of the user device.
  • the network Before step S81, the network has completed the mapping configuration between the application and the tunnel, and the mapping configuration between the application and the queue.
  • the mapping configuration between the application and the tunnel, and the mapping configuration between the application and the queue are completed according to the priority of the application.
  • control device sends access control list (ACL) configuration information to the network device.
  • ACL access control list
  • the ACL configuration information is used to instruct network devices which traffic can pass and which traffic needs to be blocked.
  • Step S82 is an optional step.
  • the ACL configuration information can also be directly configured in the network device without the participation of the control device.
  • S83 The network device performs ACL configuration according to the ACL configuration information.
  • Access control of the traffic is performed according to the ACL configuration information received in step S82, which will not be elaborated in this application.
  • the network device collects flow data and sends the collected flow data to the control device.
  • the control device receives the flow data sent by the network device.
  • network devices can also send traffic without distinguishing applications to the control device, which will then distinguish the traffic.
  • the control device aggregates the traffic of the same application based on the source IP address and destination IP address of the traffic, thereby determining the traffic data of each application.
  • control device determines whether each link in the network is congested according to the traffic data.
  • the causes of link congestion include link failure and sudden increase in service traffic corresponding to the application (such as sudden increase in service traffic caused by a temporary video conference).
  • Link failure includes line failure, equipment failure, etc.
  • step S85 may include: the control device determines the flow data of each link in the network based on the flow data of each tunnel and the mapping relationship between the tunnel and the link; the control device determines whether there is congestion in each link in the network based on the flow data of each link in the network.
  • the network device sends the flow data of each application, and the flow data of each tunnel can be determined based on the flow data of each application. In another case, the network device sends the flow data of each tunnel.
  • the traffic data of the P2-P3 link can be determined based on the traffic of the tunnel passing through the PE1-P2-P3-PE2 path and the traffic of the tunnel passing through the PE3-P2-P3-PE2 path. If the two tunnels only pass through the above two paths, the traffic data of the P2-P3 link can be obtained by directly adding the traffic of the two tunnels. If the tunnel passing through the PE1-P2-P3-PE2 path also passes through the PE1-P1-PE2 path, the traffic splitting ratio of the tunnel on the two paths needs to be considered. The traffic passing through the PE1-P2-P3-PE2 path is determined according to the splitting ratio, and it is added to the traffic passing through the PE3-P2-P3-PE2 path to obtain the traffic data of the P2-P3 link.
  • Figure 12 is a schematic diagram of a network topology provided by an embodiment of the present application.
  • Node S is a tunnel entrance device, and the tunnel includes three paths (S-M-C-D-T, S-M-D-T, S-M-A-B-D-T) in the figure.
  • the flow of the middle path is determined according to the diversion ratio of the three paths in the tunnel, that is, the flow data of the M-D link.
  • control device determines whether there is network congestion based on the traffic data sent by each network device.
  • the control device conducts a comprehensive analysis based on the traffic situation of the entire network to avoid situations where it is difficult to make a comprehensive judgment based on only local analysis.
  • control device When congestion exists on the first link, the control device generates a scheduling policy instruction for traffic of at least one application on the first link according to scheduling policies supported by each application to which the traffic carried by the first link belongs.
  • control device sends the scheduling strategy instruction to the user equipment.
  • the user equipment receives the scheduling strategy instruction.
  • the control device can also calculate the expected benefits that can be obtained if the scheduling strategy instruction is executed.
  • the benefit can be the congestion improvement, such as the number of congested links reduced, the proportion of congested link traffic occupancy reduced, etc.
  • the expected benefit is sent to the user device for display, so that the human can more clearly perceive the expected effect before making a decision, making the decision simpler, and seeing the effect of the execution after the decision is made, which not only visualizes the benefit of the decision, but also can serve as experience feedback for the next decision.
  • the user equipment receives manual confirmation or modification information for the scheduling policy instruction.
  • the user equipment sends confirmation or modification information to the control device.
  • the control device receives the confirmation or modification information sent by the user equipment and determines the scheduling policy instruction to be sent.
  • Steps S87 to S89 are optional steps, and the control device may also directly issue the scheduling policy instruction without confirmation or modification by the user equipment.
  • the control device sends a scheduling policy instruction to a network device.
  • the network device receives the scheduling policy instruction.
  • the network device is a tunnel entry device that carries the at least one application mentioned above.
  • the network device executes the scheduling policy instruction.
  • steps S84 to S811 may be executed in a loop. For example, after executing a scheduling policy instruction once, traffic data is collected again, and when there is still link congestion, the scheduling policy instruction is regenerated and executed.
  • the scheduling strategy selected by the scheduling strategy instruction may be implemented in the order of first selecting the lossless scheduling strategy and then selecting the lossy scheduling strategy.
  • FIG13 is a schematic diagram of a scheduling process provided by an embodiment of the present application.
  • step A first executes a lossless scheduling strategy.
  • step B executes a lossy scheduling strategy.
  • step C gives a capacity expansion suggestion.
  • both step A and step B can be executed once or multiple times.
  • the method may further include: controlling the device to output a capacity expansion suggestion for the first link. That is, when the current network capacity cannot meet the bandwidth demand of the application on the first link, generating a capacity expansion suggestion for the first link.
  • the above path adjustment corresponds to the lossless scheduling strategy, and the speed limit corresponds to the lossy scheduling strategy.
  • the lossless scheduling strategy is preferred to ensure the service quality of each application.
  • lossy scheduling is used to prioritize the service quality of core applications.
  • control device may also output a capacity reduction suggestion for the first link.
  • the network device collects flow data again and sends the collected flow data to the control device.
  • the control device receives the flow data sent by the network device.
  • control device determines the actual revenue according to the flow data.
  • the actual benefits may include a reduction in the number of congested links, a reduction in the traffic occupancy rate of congested links, etc.
  • step S812 and step S813 does not require the network device to perform calculations, but is calculated by the control device, so the resource occupation of the network device is low.
  • step S812 In addition to the actual revenue determination method in step S812 and step S813, another method is for the network device to determine the actual revenue and then send the actual revenue to the control device.
  • S814 The control device sends the actual revenue to the user equipment.
  • the user equipment receives the actual revenue.
  • S815 The user equipment displays the actual revenue.
  • Steps S812 to S815 are optional steps.
  • control device receives the traffic data after scheduling sent by the network device, and can determine the benefits generated by the scheduling based on these traffic data, and send the actual benefits to the user device for display, thereby providing feedback to the user on the effect of the above-mentioned scheduling strategy determination plan.
  • control device adjusts the mapping relationship between each application and the pipeline, where the pipeline includes at least one of a tunnel and a queue.
  • control device adjusts the mapping relationship between each application and the pipeline, and sends the adjusted mapping relationship to the network device so that the network device also adjusts accordingly.
  • step S816 and the aforementioned steps S812 to S815 may not be fixed, for example, step S816 may be executed before step S812.
  • the network transmission can also be improved by adjusting the pipeline mapping relationship.
  • This application supports the network's end-to-end refined management demands based on applications from configuration to scheduling; based on traffic profiling and feature analysis, it provides a basis for application feature visualization, anomaly detection, intelligent scheduling, management decisions, etc.; based on scheduling management at different granularities such as tunnels, paths, and applications, it flexibly combines to achieve customers' differentiated management demands, and based on multi-granularity scheduling, it realizes customers' automatic tuning demands and minimum lossy optimization goals.
  • Fig. 14 is a module diagram of a system for executing the above method provided by an embodiment of the present application.
  • the user equipment includes a user configuration module, and the user configuration module generates user configuration information based on user input, that is, executes step S81.
  • the network device includes a data acquisition module, which is used to collect traffic data and send it to the control device, that is, to execute steps S84 and S812 .
  • control device includes a service management module, and the service management module is used to determine whether congestion occurs in each link based on traffic data, that is, to execute step S85 .
  • the control device also includes a control scheduling module, a pipeline management module, a lossless optimization module, a lossy optimization module and a link management module.
  • the business management module is used to generate a traffic portrait based on the traffic data, thereby realizing traffic trend prediction.
  • the control scheduling module is responsible for policy scheduling, and generates scheduling policy instructions by calling the pipeline management module, the lossless optimization module, the lossy optimization module and the link management module. In addition, it is also used to execute pipeline deletion and output network expansion and contraction recommendations.
  • the pipeline management module is used to implement the mapping configuration of the initial application and the pipeline. For example, a single core application is mapped to a single queue or tunnel, while general priority applications or low priority applications are mapped to one queue or tunnel.
  • the lossless optimization module is used to provide lossless optimization scheduling strategies.
  • the lossy optimization module is used to provide lossy optimization scheduling strategies.
  • the link management module is used to provide link expansion and contraction suggestions, network topology design suggestions, etc.
  • Step S86 is implemented by controlling the scheduling module, the pipeline management module, the lossless optimization module, the lossy optimization module and the link management module together.
  • the network device further includes a policy management module for executing the scheduling policy instruction issued by the control device, that is, executing step S811 .
  • the user equipment further includes a visualization module for displaying expected benefits and/or actual benefits, that is, executing step S815 .
  • the solution provided by this application realizes the visualization and sensing of scheduling through the data acquisition module and the visualization module, that is, based on traffic statistics, after analysis and processing, the scheduling strategy is completed, and finally the visualization of benefits is realized.
  • the analysis and recommendation of the scheduling strategy are realized by controlling the scheduling module, the pipeline management module, the lossless optimization module, the lossy optimization module and the link management module.
  • the scheduling strategy is manually confirmed/modified through the user configuration module and the visualization module, thereby realizing the decision on the scheduling strategy.
  • the execution of the scheduling decision is realized through the policy management module.
  • the visualization module can also cooperate with the control equipment to output abnormal alarms and prompts when the traffic pattern is abnormal.
  • FIG15 is a block diagram of a scheduling strategy determination device provided in an embodiment of the present application.
  • the scheduling strategy determination device can be implemented as all or part of a control device through software, hardware, or a combination of both.
  • the scheduling strategy determination device can include: a determination unit 901 and an application unit 902.
  • a determining unit 901 is configured to determine, in response to congestion on the first link, a first scheduling strategy for traffic of at least one application on the first link according to scheduling strategies respectively supported by multiple applications to which traffic carried by the first link belongs;
  • the application unit 902 is configured to apply a first scheduling strategy.
  • the determining unit 901 is configured to determine the first scheduling strategy according to scheduling strategies supported by the application supporting the lossless scheduling strategy in response to the existence of the application supporting the lossless scheduling strategy.
  • the determination unit 901 is used to determine that the first scheduling policy is a scheduling policy supported by applications supporting the lossless scheduling policy in response to the number of applications supporting the lossless scheduling policy being one; or, in response to the number of applications supporting the lossless scheduling policy being multiple, determine that the first scheduling policy includes scheduling policies supported by at least one application among the applications supporting the lossless scheduling policy.
  • the determination unit 901 is used to combine applications that support lossless scheduling strategies and scheduling strategies supported by the applications to obtain at least one strategy combination, each strategy combination corresponding to a scheduling strategy for at least one application; determine whether the congestion of the first link can be resolved when scheduling using scheduling strategies corresponding to each strategy combination in at least one strategy combination; in response to the existence of a strategy combination that can resolve the congestion of the first link, select a strategy combination from the strategy combinations that can resolve the congestion of the congested first link, and adopt the scheduling strategy corresponding to the selected strategy combination as the first scheduling strategy.
  • the determination unit 901 is used to determine whether the scheduling policy corresponding to the first policy combination can solve the congestion of the first link based on the predicted traffic data and the scheduling policy corresponding to the first policy combination, and the first policy combination is any one of at least one policy combination.
  • the determination unit 901 is used to predict the traffic data of each application based on the traffic portraits of multiple applications; determine the traffic value expected to be carried by the first link based on the scheduling strategy corresponding to the first strategy combination and the predicted traffic data of each application; and determine whether the first link is congested based on the traffic value expected to be carried by the first link.
  • the determining unit 901 is configured to select a policy combination from policy combinations that can resolve congestion of the congested first link based on at least one of an application priority and an application requirement.
  • the determining unit 901 is configured to determine the first scheduling strategy according to scheduling strategies supported by applications supporting the lossy scheduling strategy in response to the absence of an application supporting the lossless scheduling strategy.
  • the determining unit 901 is further configured to determine a link involved in the first scheduling strategy; and determine, based on traffic data of the link involved in the first scheduling strategy, whether the link involved in the first scheduling strategy will be congested after the first scheduling strategy is executed;
  • the device also includes: a generating unit 903, configured to generate a first scheduling strategy instruction based on the first scheduling strategy in response to the link involved in the first scheduling strategy not being congested.
  • the determination unit 901 is further configured to, in response to congestion on the second link, determine a second scheduling strategy for traffic of at least one application on the second link according to scheduling strategies respectively supported by multiple applications to which traffic carried by the second link belongs; determine links involved in the first scheduling strategy and the second scheduling strategy; and determine, according to traffic data of the involved links, whether the involved links will be congested after the first scheduling strategy and the second scheduling strategy are executed;
  • a generating unit 903 is configured to generate a first scheduling strategy instruction based on the first scheduling strategy in response to the involved link not being congested, and The second scheduling strategy generates a second scheduling strategy instruction.
  • the determination unit 901 is also used to determine the expected benefit after using the first scheduling strategy for traffic scheduling before applying the first scheduling strategy; determine the actual benefit after applying the first scheduling strategy; and determine the third scheduling strategy based on the expected benefit and the actual benefit in response to congestion on the third link.
  • the generating unit 903 is used to generate a traffic profile of each application according to the traffic data of multiple applications;
  • the device also includes: an adjustment unit 904, which is used to adjust the mapping relationship between each application and the pipeline according to the traffic profile of each application, and the pipeline includes at least one of a tunnel and a queue.
  • the generating unit 903 is configured to obtain a traffic profile of each group in each grouping mode after grouping multiple applications in multiple grouping modes, where each group includes at least one application;
  • the adjustment unit 904 is used to adjust the mapping relationship between multiple groups and pipelines according to the traffic profile of each group, and the pipeline includes at least one of a tunnel and a queue.
  • the determination unit is implemented by the control scheduling module, pipeline management module, lossless optimization module, lossy optimization module and link management module in FIG. 14 , and the application unit can be integrated in the control scheduling module.
  • the device for determining the scheduling strategy provided in the above embodiment only uses the division of the above functional units as an example when determining the scheduling strategy.
  • the above functions can be assigned to different functional units as needed, that is, the internal structure of the device is divided into different functional units to complete all or part of the functions described above.
  • the device for determining the scheduling strategy provided in the above embodiment and the method for determining the scheduling strategy belong to the same concept. The specific implementation process is detailed in the method embodiment and will not be repeated here.
  • FIG16 shows a schematic diagram of the structure of a control device or network device 150 provided in an embodiment of the present application.
  • the control device or network device 150 shown in FIG16 is used to perform the operations involved in the method for determining the scheduling strategy shown in any of FIG3 to FIG14.
  • the control device or network device 150 can be implemented by a general bus architecture.
  • control device or network device 150 includes at least one processor 151 , a memory 153 , and at least one communication interface 154 .
  • the processor 151 is, for example, a general-purpose central processing unit (CPU), a digital signal processor (DSP), a network processor (NP), a data processing unit (DPU), a microprocessor, or one or more integrated circuits for implementing the solution of the present application.
  • the processor 151 includes an application-specific integrated circuit (ASIC), a programmable logic device (PLD) or other programmable logic devices, transistor logic devices, hardware components, or any combination thereof.
  • ASIC application-specific integrated circuit
  • PLD programmable logic device
  • the PLD is, for example, a complex programmable logic device (CPLD), a field-programmable gate array (FPGA), a generic array logic (GAL), or any combination thereof.
  • CPLD complex programmable logic device
  • FPGA field-programmable gate array
  • GAL generic array logic
  • the processor may also be a combination that implements a computing function, such as a combination of one or more microprocessors, a combination of a DSP and a microprocessor, and the like.
  • control device or network device 150 further includes a bus.
  • the bus is used to transmit information between the components of the control device or network device 150.
  • the bus may be a peripheral component interconnect (PCI) bus or an extended industry standard architecture (EISA) bus.
  • PCI peripheral component interconnect
  • EISA extended industry standard architecture
  • the bus may be divided into an address bus, a data bus, a control bus, etc.
  • FIG16 is represented by only one thick line, but it does not mean that there is only one bus or one type of bus.
  • the memory 153 is, for example, a read-only memory (ROM) or other types of static storage devices that can store static information and instructions, or a random access memory (RAM) or other types of dynamic storage devices that can store information and instructions, or an electrically erasable programmable read-only memory (EEPROM), a compact disc read-only memory (CD-ROM) or other optical disk storage, optical disk storage (including compressed optical disk, laser disk, optical disk, digital versatile disk, Blu-ray disk, etc.), a magnetic disk storage medium or other magnetic storage device, or any other medium that can be used to carry or store the desired program code in the form of instructions or data structures and can be accessed by a computer, but is not limited thereto.
  • the memory 153 is, for example, independent and connected to the processor 151 through a bus.
  • the memory 153 can also be integrated with the processor 151.
  • the communication interface 154 uses any transceiver or other device to communicate with other devices or communication networks, such as Ethernet, a radio access network (RAN), or a wireless local area network (WLAN).
  • the communication interface 154 may include a wired
  • the communication interface may also include a wireless communication interface.
  • the communication interface 154 may be an Ethernet interface, a Fast Ethernet (FE) interface, a Gigabit Ethernet (GE) interface, an Asynchronous Transfer Mode (ATM) interface, a wireless local area network (WLAN) interface, a cellular network communication interface or a combination thereof.
  • the Ethernet interface may be an optical interface, an electrical interface or a combination thereof.
  • the communication interface 154 may be used to control the device or the network device 150 to communicate with other devices.
  • the processor 151 may include one or more CPUs, such as CPU0 and CPU1 shown in FIG16 . Each of these processors may be a single-core (single-CPU) processor or a multi-core (multi-CPU) processor.
  • the processor here may refer to one or more devices, circuits, and/or processing cores for processing data (e.g., computer program instructions).
  • control device or network device 150 may include multiple processors, such as the processor 151 and the processor 155 shown in Figure 16. Each of these processors may be a single-core processor (single-CPU) or a multi-core processor (multi-CPU).
  • the processor here may refer to one or more devices, circuits, and/or processing cores for processing data (such as computer program instructions).
  • the control device or network device 150 may also include an output device and an input device.
  • the output device communicates with the processor 151 and can display information in a variety of ways.
  • the output device may be a liquid crystal display (LCD), a light emitting diode (LED) display device, a cathode ray tube (CRT) display device, or a projector.
  • the input device communicates with the processor 151 and can receive user input in a variety of ways.
  • the input device may be a mouse, a keyboard, a touch screen device, or a sensor device.
  • the memory 153 is used to store the program code 1510 for executing the solution of the present application, and the processor 151 can execute the program code 1510 stored in the memory 153. That is, the control device or network device 150 can execute the program code 1510 in the memory 153 through the processor 151 to implement the data processing method provided by the method embodiment.
  • the program code 1510 may include one or more software modules.
  • the processor 151 itself may also store the program code or instruction for executing the solution of the present application.
  • control device or network device 150 of the embodiment of the present application may correspond to the controller in the above-mentioned method embodiments, and the processor 151 in the control device or network device 150 reads the instructions in the memory 153, so that the control device or network device 150 shown in Figure 16 can execute all or part of the operations performed by the controller.
  • the processor 151 is used to determine, in response to congestion on the first link, a first scheduling policy for traffic of at least one application on the first link according to scheduling policies respectively supported by multiple applications to which the traffic carried by the first link belongs; and apply the first scheduling policy.
  • each step of the method for determining the scheduling strategy shown in any of Figures 3 to 14 is completed by an integrated logic circuit of hardware or software instructions in the processor of the control device or network device 150.
  • the steps of the method disclosed in conjunction with the embodiment of the present application can be directly embodied as a hardware processor, or a combination of hardware and software modules in the processor.
  • the software module can be located in a mature storage medium in the field 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 a memory, and the processor reads the information in the memory, and completes the steps of the above method in conjunction with its hardware. To avoid repetition, it is not described in detail here.
  • the embodiment of the present application also provides a chip, including: an input interface, an output interface, a processor and a memory.
  • the input interface, the output interface, the processor and the memory are connected through an internal connection path.
  • the processor is used to execute the code in the memory.
  • the processor is used to execute any of the above-mentioned methods for determining the scheduling strategy.
  • the processor may be a CPU, or other general-purpose processors, DSPs, ASICs, FPGAs or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components, etc.
  • a general-purpose processor may be a microprocessor or any conventional processor, etc. It is worth noting that the processor may be a processor supporting an ARM architecture.
  • the processor is one or more and the memory is one or more.
  • the memory can be integrated with the processor, or the memory is separately arranged with the processor.
  • the memory can include a read-only memory and a random access memory, and provide instructions and data to the processor.
  • the memory can also include a non-volatile random access memory.
  • the memory can also store a reference block and a target block.
  • the memory may be a volatile memory or a nonvolatile memory, or may include both volatile and nonvolatile memories.
  • the nonvolatile memory may be a ROM, a PROM, an EPROM, an EEPROM, or a flash memory.
  • the volatile memory may be a RAM, which is used as an external By way of example and not limitation, many forms of RAM are available, such as SRAM, DRAM, SDRAM, DDR SDRAM, ESDRAM, SLDRAM, and DR RAM.
  • a computer-readable storage medium which stores computer instructions.
  • the control device executes the method for determining the scheduling strategy provided above.
  • a computer program product including instructions is also provided, which, when executed on a control device, enables the control device to execute the method for determining the scheduling strategy provided above.
  • the computer program product includes one or more computer instructions.
  • the computer can be a general-purpose computer, a special-purpose computer, a computer network, or other programmable device.
  • the computer instructions can be stored in a computer-readable storage medium, or transmitted from one computer-readable storage medium to another computer-readable storage medium.
  • the computer instructions can be transmitted from one website, computer, server or data center to another website, computer, server or data center by wired (e.g., coaxial cable, optical fiber, digital subscriber line) or wireless (e.g., infrared, wireless, microwave, etc.) means.
  • the computer-readable storage medium can be any available medium that can be accessed by a computer or a data storage device such as a server or data center that includes one or more available media integrated.
  • the available medium can be a magnetic medium (e.g., a floppy disk, a hard disk, a tape), an optical medium (e.g., a DVD), or a semiconductor medium (e.g., a solid-state drive Solid State Disk), etc.

Landscapes

  • Engineering & Computer Science (AREA)
  • Computer Networks & Wireless Communication (AREA)
  • Signal Processing (AREA)
  • Data Exchanges In Wide-Area Networks (AREA)

Abstract

公开了一种调度策略的确定方法及相关装置。该方法包括:响应于第一链路存在拥塞,根据所述第一链路承载的流量所属的多个应用分别支持的调度策略,确定针对所述第一链路上的至少一个应用的流量的第一调度策略;应用所述第一调度策略。在第一链路存在拥塞时,根据该链路上的多个应用支持的调度策略确定调度策略,使得控制设备或网络设备可以基于各个应用分别支持的调度策略统筹调度各个应用的流量,可以充分利用网络资源以传输各个应用的流量,进而可以最大可能地保证各个应用的业务质量,避免相关技术中直接对低优先级应用进行限速,造成的低优先级业务质量肯定受损的问题。

Description

调度策略的确定方法及相关装置
本申请要求于2022年9月30日提交的申请号为202211215235.3、申请名称为“调度策略的确定方法及相关装置”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
本申请涉及网络技术领域,特别涉及一种调度策略的确定方法及相关装置。
背景技术
链路是指在两个节点之间建立的用于传输某种或某几种应用的流量的通路。例如,在节点A和节点B之间建立的用于传输办公类应用的流量的链路。
当链路同时承载多个应用的流量,在链路出现拥塞时,为了保障关键应用的带宽,直接对低优先级应用的流量进行限速,造成低优先级应用的业务质量肯定受损。
发明内容
本申请提供了一种调度策略的确定方法及相关装置,能够避免相关技术中直接对低优先级应用进行限速,造成的低优先级业务质量肯定受损的问题。
第一方面,提供了一种调度策略的确定方法。该方法可以由网络中的控制设备或网络设备执行。该方法包括:响应于第一链路存在拥塞,根据第一链路承载的流量所属的多个应用分别支持的调度策略,确定针对第一链路上的至少一个应用的流量的第一调度策略;应用第一调度策略。
在第一链路存在拥塞时,根据该链路上的多个应用支持的调度策略确定调度策略,使得控制设备或网络设备可以基于各个应用分别支持的调度策略统筹调度各个应用的流量,可以充分利用网络资源以传输各个应用的流量,进而可以最大可能地保证各个应用的业务质量,避免相关技术中直接对低优先级应用进行限速,造成的低优先级业务质量肯定受损的问题。
在本申请的实现方式中,调度策略包括无损调度策略和有损调度策略。有损调度策略是指会造成应用数据包丢失的调度策略,无损调度策略是指不会造成应用数据包丢失的调度策略。
其中,无损调度策略包括如下至少一种:路径调整策略、压缩策略。有损调度策略包括:限速策略。除此之外,调度策略还包括管道映射调整策略,而管道映射调整策略既可以是无损调度策略,也可以是有损调度策略。
在本申请的实现方式中,根据第一链路承载的流量所属的多个应用分别支持的调度策略,确定针对第一链路上的至少一个应用的流量的第一调度策略,包括:响应于存在支持无损调度策略的应用,根据支持无损调度策略的应用所支持的调度策略,确定第一调度策略。响应于不存在支持无损调度策略的应用,根据支持有损调度策略的应用所支持的调度策略,确定第一调度策略。
在该实现方式中,在确定调度策略时优先确定无损调度策略,从而尽量保证链路中所有应用的业务质量。
示例性地,根据支持无损调度策略的应用所支持的调度策略,确定第一调度策略,包括:
响应于支持无损调度策略的应用的数量为一个,确定第一调度策略为支持无损调度策略的应用所支持的调度策略;或,响应于支持无损调度策略的应用的数量为多个,确定第一调度策略包括支持无损调度策略的应用中至少一个应用所支持的调度策略。
其中,确定第一调度策略为支持无损调度策略的应用中至少一个应用所支持的调度策略,包括:对支持无损调度策略的应用及应用支持的调度策略进行组合,得到至少一种策略组合,每种策略组合对应针对至少一个应用的调度策略;确定采用至少一种策略组合中的各个策略组合对应的调度策略进行调度时是否能够解决第一链路的拥塞;响应于存在能解决第一链路的拥塞的策略组合,从能解决拥塞第一链路的拥塞的策略组合中选择策略组合,采用选出的策略组合对应的调度策略作为第一调度策略。
其中,对应用及调度策略进行组合,实际是以调度策略作为精度进行组合,例如,应用A支持两种调度策略a、b,组合时应用A有两种加入组合的方式,分别是应用A-调度策略a、应用A-调度策略b。
如果不存在能够解决第一链路的拥塞的组合(也即策略组合),则继续组合并选择,直到穷举了所有组合方式仍不存在能够解决拥塞的组合,则根据支持有损调度策略的应用所支持的调度策略,确定第一调度策略。
在该实现方式中,通过对应用及调度策略进行组合,并从这些组合中选择能够解决第一链路拥塞的组合,保证调度策略执行的效果。
在一种可能的实现方式中,在确定采用组合对应的调度策略进行调度时是否能够解决第一链路的拥塞时,基于当前的流量数据和上述选择的调度策略判断调度后是否能够解决第一链路的拥塞问题。
在另一种可能的实现方式中,在确定采用组合对应的调度策略进行调度时是否能够解决第一链路的拥塞时,基于预测的流量数据和上述选择的调度策略判断调度后是否能够解决第一链路的拥塞问题。
例如,基于预测的流量数据和第一策略组合对应的调度策略,判断第一组合对应的调度策略是否能够解决第一链路的拥塞,第一策略组合为至少一种策略组合中的任一个。
在该实现方式中,基于预测的流量数据确定调度策略,使得网络设备执行的调度策略更能匹配链路未来承载的流量,能进一步缓解链路未来可能出现的拥塞,也使得需要进行调度的频率更低,节省了网络资源。
示例性地,基于预测的流量数据和第一策略组合对应的调度策略,判断第一策略组合对应的调度策略是否能够解决第一链路的拥塞,包括:基于多个应用的流量画像预测各个应用的流量数据;基于第一策略组合对应的调度策略和预测的各个应用的流量数据,确定第一链路预期承载的流量值;基于第一链路预期承载的流量值判断第一链路是否拥塞。
在该实现方式中,采用各个应用的流量画像进行流量数据的预测,能够提高预测精度,进而提高对于能否解决第一链路拥塞的判断的准确性。
在本申请的实现方式中,从能解决拥塞第一链路的拥塞的策略组合中选择策略组合,包括:基于应用的优先级和应用需求中的至少一项,从能解决拥塞第一链路的拥塞的策略组合中选择组合。
在一种可能的实现方式中,基于各个应用的优先级,选择优先级较低的策略组合对应的调度策略。
在另一种可能的实现方式中,基于各个应用的应用需求,选择能够满足组合中应用需求的策略组合对应的调度策略。
在又一种可能的实现方式中,基于各个应用的优先级和各个应用的应用需求,选择能够满足组合中应用需求的组合中优先级较低的策略组合对应的调度策略。
可选地,该方法还包括:确定第一调度策略涉及的链路;根据第一调度策略涉及的链路的流量数据,确定第一调度策略执行后第一调度策略涉及的链路是否会拥塞;响应于第一调度策略涉及的链路不会拥塞,基于第一调度策略生成第一调度策略指令。
在该实现方式中,考虑第一调度策略执行后涉及到的其他链路是否拥塞,避免解决第一链路拥塞时造成其他链路拥塞的问题,从而保证整个网络传输的业务质量。
可选地,该方法还包括:响应于第二链路存在拥塞,根据第二链路承载的流量所属的多个应用分别支持的调度策略,确定针对第二链路上的至少一个应用的流量的第二调度策略;确定第一调度策略和第二调度策略涉及的链路;根据涉及的链路的流量数据,确定第一调度策略和第二调度策略执行后涉及的链路是否会拥塞;响应于涉及的链路不会拥塞,基于第一调度策略生成第一调度策略指令,基于第二调度策略生成第二调度策略指令。
在该实现方式中,在对多条链路拥塞进行调度时,使得每个链路拥塞的调度策略都不会造成其他链路的拥塞。该方案以整个网络为调度对象,避免了相关技术中只着眼于单个设备或链路,带来的额外网络问题。
可选地,该方法还包括:在应用第一调度策略前,确定采用第一调度策略进行流量调度后的预期收益;在应用第一调度策略后,确定实际收益;响应于第三链路存在拥塞,基于预期收益和实际收益确定第三调度策略。
在该实现方式中,基于预期收益和实际收益的偏差,可以为后续的调度策略的确定提供指导,从而使得调度后的效果更好。
在一种可能的实现方式中,该方法还包括:根据多个应用的流量数据生成各个应用的流量画像;根据各个应用的流量画像,调整各个应用和管道的映射关系,管道包括隧道和队列中的至少一个。
示例性地,根据各个应用的流量画像,调整各个应用和管道的映射关系,包括:根据各个应用的流量模式,将流量模式相同的应用映射到同一个管道,将流量模式不同的应用映射到不同的管道。
在该实现方式中,将相同流量模式的应用的流量映射到相同的管道,从而使得不同管道的流量特征更容易区分,从而方便对于各个管道的管控和调度。
在另一种可能的实现方式中,该方法还包括:获取采用多种分组方式对多个应用进行分组之后得到的每种分组方式下各组的流量画像,每组包括至少一个应用;根据各组的流量画像,调整多个各个和管道的映射关系,管道包括隧道和队列中的至少一个。
示例性地,根据各组的流量画像,调整流量统计数据中涉及的多个应用和管道的映射关系,包括:根据每种分组方式下各组的流量画像的平稳度,选择平稳度满足平稳度条件的分组方式;将选出的分组方式下的同一组的应用映射到同一个管道,将不同组的应用映射到不同的管道。
其中,平稳度是指平稳的程度,可以根据流量的流量模式及均峰比等参数来评价。在上述选择过程中,可以选择流量模式为小值振荡、平稳双峰型,然后选择这些流量模式中均峰比满足要求的。
在该实现方式中,将不同画像类型组合到一起进行互补,比如双峰、毛刺这些组合到一起,使得组合出的管道的流量画像更平稳型。这种组合使得得到的管道的流量平稳,可以节省链路带宽调整的次数。
第二方面,提供了一种调度策略的确定装置,该装置包括:
确定单元,用于响应于第一链路存在拥塞,根据第一链路承载的流量所属的多个应用分别支持的调度策略,确定针对第一链路上的至少一个应用的流量的第一调度策略;
应用单元,用于应用第一调度策略。
可选地,该确定单元,用于响应于存在支持无损调度策略的应用,根据支持无损调度策略的应用所支持的调度策略,确定第一调度策略。
可选地,该确定单元,用于响应于支持无损调度策略的应用的数量为一个,确定支持无损调度策略的应用所支持的调度策略为第一调度策略;或,响应于支持无损调度策略的应用的数量为至少两个,确定支持无损调度策略的应用中至少一个应用所支持的调度策略为第一调度策略。
可选地,该确定单元,用于对支持无损调度策略的应用及应用支持的调度策略进行组合,得到至少一种策略组合,每种策略组合对应针对至少一个应用的调度策略;确定采用至少一种策略组合中的各个策略组合对应的调度策略进行调度时是否能够解决第一链路的拥塞;响应于存在能解决第一链路的拥塞的策略组合,从能解决拥塞第一链路的拥塞的策略组合中选择策略组合,采用选出的策略组合对应的调度策略作为第一调度策略。
可选地,该确定单元,用于基于预测的流量数据和第一策略组合对应的调度策略,判断第一策略组合对应的调度策略是否能够解决第一链路的拥塞,第一策略组合为至少一种策略组合中的任一个。
可选地,该确定单元,用于基于多个应用的流量画像预测各个应用的流量数据;基于第一策略组合对应的调度策略和预测的各个应用的流量数据,确定第一链路预期承载的流量值;基于第一链路预期承载的流量值判断第一链路是否拥塞。
可选地,该确定单元,用于基于应用的优先级和应用需求中的至少一项,从能解决拥塞第一链路的拥塞的策略组合中选择策略组合。
可选地,该确定单元,用于响应于不存在支持无损调度策略的应用,根据支持有损调度策略的应用所支持的调度策略,确定第一调度策略。
可选地,该确定单元,还用于确定第一调度策略涉及的链路;根据第一调度策略涉及的链路的流量数据,确定第一调度策略执行后第一调度策略涉及的链路是否会拥塞;
该装置还包括:生成单元,用于响应于第一调度策略涉及的链路不会拥塞,基于第一调度策略生成第一调度策略指令。
可选地,该确定单元,还用于响应于第二链路存在拥塞,根据第二链路承载的流量所属的多个应用分别支持的调度策略,确定针对第二链路上的至少一个应用的流量的第二调度策略;确定第一调度策略和第二调度策略涉及的链路;根据涉及的链路的流量数据,确定第一调度策略和第二调度策略执行后涉及的链 路是否会拥塞;
生成单元,用于响应于涉及的链路不会拥塞,基于第一调度策略生成第一调度策略指令,基于第二调度策略生成第二调度策略指令。
可选地,该确定单元,还用于在应用第一调度策略前,确定采用第一调度策略进行流量调度后的预期收益;在应用第一调度策略后,确定实际收益;响应于第三链路存在拥塞,基于预期收益和实际收益确定第三调度策略。
可选地,生成单元,用于根据多个应用的流量数据生成各个应用的流量画像;
该装置还包括:调整单元,用于根据各个应用的流量画像,调整各个应用和管道的映射关系,管道包括隧道和队列中的至少一个。
可选地,生成单元,用于获取采用多种分组方式对多个应用进行分组之后得到的每种分组方式下各组的流量画像,每组包括至少一个应用;
调整单元,用于根据各组的流量画像,调整多个各个和管道的映射关系,管道包括隧道和队列中的至少一个。
第三方面,提供了一种电子设备,该电子设备为控制设备或网络设备。所述电子设备包括处理器和存储器。所述存储器用于存储软件程序以及模块。所述处理器通过运行或执行存储在所述存储器内的软件程序和/或模块实现上述第一方面或第一方面的任一种可能的实施方式中的方法。
可选地,所述处理器为一个或多个,所述存储器为一个或多个。
可选地,所述存储器可以与所述处理器集成在一起,或者所述存储器与处理器分离设置。
在具体实现过程中,存储器可以为非瞬时性(non-transitory)存储器,例如只读存储器(read only memory,ROM),其可以与处理器集成在同一块芯片上,也可以分别设置在不同的芯片上,本申请对存储器的类型以及存储器与处理器的设置方式不做限定。
第四方面,提供了一种计算机程序产品。所述计算机程序产品包括计算机程序代码,当所述计算机程序代码被计算机运行时,使得所述计算机执行上述第一方面或第一方面的任一种可能的实施方式中的方法。
第五方面,本申请提供了一种计算机可读存储介质,所述计算机可读存储介质用于存储处理器所执行的程序代码,所述程序代码包括用于实现上述第一方面任一种可能的实施方式中的方法。
第六方面,提供了一种芯片,包括处理器,处理器用于从存储器中调用并运行所述存储器中存储的指令,使得安装有所述芯片的通信设备执行上述第一方面任一种可能的实施方式中的方法。
第七方面,提供另一种芯片。所述另一种芯片包括输入接口、输出接口、处理器和存储器。所述输入接口、输出接口、所述处理器以及所述存储器之间通过内部连接通路相连。所述处理器用于执行所述存储器中的代码,当所述代码被执行时,所述处理器用于执行上述第一方面任一种可能的实施方式中的方法。
附图说明
图1是本申请实施例提供的一种应用场景的示意图;
图2是本申请实施例提供的一种应用场景的示意图;
图3是本申请实施例提供的一种调度策略的确定方法的流程图;
图4是本申请实施例提供的一种调度策略的确定方法的流程图;
图5是本申请实施例提供的一种WFQ权重调整的示意图;
图6是本申请实施例提供的一种调度策略的确定方法的流程图;
图7是本申请实施例提供的一种调度策略的确定方法的流程图;
图8是本申请实施例提供的一种调度策略的确定方法的流程图;
图9是本申请实施例提供的一种调度策略的确定方法的流程图;
图10是本申请实施例提供的一种调度策略的确定方法的流程图;
图11是本申请实施例提供的一种调度策略的确定方法的流程图;
图12是本申请实施例提供的一种网络拓扑示意图;
图13是本申请实施例提供的一种调度过程示意图;
图14是本申请实施例提供的一种模块示意图;
图15是本申请实施例提供的一种调度策略的确定装置的框图;
图16是本申请实施例提供的一种控制设备或网络设备的结构示意图。
具体实施方式
为使本申请的目的、技术方案和优点更加清楚,下面将结合附图对本申请实施方式作进一步地详细描述。
为便于对本申请实施例提供的技术方案的理解,首先介绍一下本申请的应用场景。该应用场景可以是分布式存储系统或者数据中心等通信网络。下面以分布式存储系统为例,对本申请应用场景进行说明。
图1是本申请实施例提供的一种应用场景的结构示意图。参见图1,该应用场景包括控制设备11和网络设备12。控制设备11可以是网络中的管理节点,例如,网管设备(如软件定义网络(software defined network,SDN)控制器)。网络设备12可以是网络中的网络节点,例如路由设备。如图1所示,网络设备12包括运营商(provider)设备,也即图中P设备,网络设备12还包括运营商边缘(provider edge)设备,也即图中PE设备。
如图1所示,控制设备11和PE设备连接。控制设备11可以通过PE设备和各个P设备连接,或者,控制设备11也可以直接和各个P设备连接(为了避免附图中线条过于复杂并未直接示出)。
图1中任意两个网络设备12之间建立链路,用于传输业务数据。两个PE设备之间配置有隧道(tunnel)。该隧道用于特定应用的传输,可以传输至少一个应用的流量。
例如,从PE1到PE2的隧道,PE1为隧道的入口设备,PE2为隧道的出口设备。该隧道既可以承载在一个路径上,比如PE1-P1-PE2路径。该隧道也可以承载在多个路径上,比如PE1-P1-PE2和PE1-P2-P3-PE2两条路径,两条路径按照一定比例传输该隧道的流量。
图2是本申请实施例提供的另一种应用场景的结构示意图。参见图2,该应用场景和图1的应用场景相比,区别在于网络设备12还包括PE3,PE3和PE2之间也可以配置有隧道,该隧道承载在PE3-P2-P3-PE2路径上。图1中P2仅接收PE1的流量,而图2中P2汇聚了PE1和PE3的流量。
由于P2-P3链路同时承载2个隧道的流量,因此对于P2-P3链路的流量调度要同时考虑两个隧道的流量。
除了控制设备和网络设备,该应用场景还可以包括用户设备,用户设备可以是独立的设备,与控制设备相连,用户设备也可以集成在控制设备中。
图3是本申请实施例提供的一种调度策略的确定方法的流程图。该方法可以应用在控制设备,也可以应用在网络设备,下面以该方法应用在控制设备为例进行说明。如图3所示,该方法包括如下步骤。
S11,响应于第一链路存在拥塞,控制设备根据第一链路承载的流量所属的多个应用分别支持的调度策略,确定针对第一链路上的至少一个应用的流量的调度策略。
其中,链路拥塞是指链路承载的流量超过拥塞阈值。
例如,拥塞阈值为80%,对于带宽为100Mbps的链路,当流量超过80Mbps时,确定该链路存在拥塞。
其中,应用支持的调度策略通常是对应用的业务指令没有影响或者影响较小的调度策略。
S12,控制设备应用该调度策略。
例如,控制设备将该调度策略指令配置于在承载上述至少一个应用的流量的入口设备。
例如,控制设备将每个应用的调度策略指令配置于承载该应用的流量的入口设备。
入口设备例如为承载上述至少一个应用中的每个应用功能的流量的隧道的入口设备。
在本申请实施例中,在第一链路存在拥塞时,根据该链路上的多个应用支持的调度策略确定调度策略,使得控制设备可以基于各个应用分别支持的调度策略统筹调度各个应用的流量,可以充分利用网络资源以传输各个应用的流量,进而可以最大可能地保证各个应用的业务质量,避免相关技术中直接对低优先级应 用进行限速,造成的低优先级业务质量肯定受损的问题。
图4是本申请实施例提供的另一种调度策略的确定方法的流程图。该方法可以应用在控制设备,也可以应用在网络设备,下面以该方法应用在控制设备为例进行说明。如图4所示,该方法包括如下步骤。
S21,控制设备确定第一链路存在拥塞。
在一种可能的实现方式中,控制设备接收网络设备发送的流量数据,并根据该流量数据确定第一链路存在拥塞。
例如,每个网络设备(例如,图1~2中的各个P设备和各个PE设备)采集出端口和/或入口端的流量,然后将对应的链路的流量数据发送给控制设备。为了让控制设备知道采集的流量数据的归属,可以在发送流量数据时携带链路标识。
对于一条链路,可以基于链路的发送端的出端口发送的流量,或接收端的入端口的接收流量判断该链路是否存在拥塞。
再例如,网络中的隧道的入口设备(例如,图1~2中的PE设备)采集隧道的流量和/或采集应用的流量,然后将隧道和/或应用的流量数据发送给控制设备。采集时,基于应用流标识来采集应用的流量,基于隧道标识来采集隧道的流量。同样,为了让控制设备知道采集的流量数据的归属,可以在发送流量数据时携带隧道标识、应用流标识等进行指示。
无论是以链路、隧道还是应用作为采集粒度,网络设备都可以按照预定时间间隔采集,如每秒钟采集一次流量数据,每采集一个流量数据即发送给控制设备。当然,网络设备也可以采集一段时间然后再发送,如每秒钟采集一次流量数据,每分钟将采集到的60个流量数据发送给控制设备。
控制设备获取到流量数据后,基于流量数据确定网络拓扑中的各条链路是否存在拥塞。其中,确定链路是否拥塞的方式是判断链路承载的流量是否超过拥塞阈值。不同链路可以各自有对应的拥塞阈值。
对于一条链路,链路承载的流量可以通过该链路的流量数据确定,也可以通过该链路上承载的应用或隧道的流量数据确定,对此不做赘述。
在另一种可能的实现方式中,控制设备接收网络设备发送的第一链路拥塞消息,第一链路拥塞消息用于指示第一链路存在拥塞。
在该实现方式中,确定拥塞的过程由网络设备完成。网络设备可以采用前述方式确定链路拥塞。网络设备也可以根据该设备检测到与该链路关联的出端口、出队列或入队列的性能数据确定链路拥塞。例如,当出端口的丢包率大于阈值时,网络设备确定与该出端口关联的链路发生链路拥塞。又例如,当出队列的队列占用率超过阈值时,或者当出队列的队列占用率超过阈值的持续时长超过时长阈值时,网络设备确定与该出队列关联的链路发生链路拥塞。
S22,控制设备获取第一链路承载的流量所属的各个应用支持的调度策略。
在一种可能的实现方式中,各个应用支持的调度策略可以由用户在用户设备上进行配置,然后发送给控制设备。
在另一种可能的实现方式中,各个应用支持的调度策略可以预先配置在控制设备中。
S23,控制设备根据第一链路承载的流量所属的多个应用支持的调度策略,确定针对第一链路上的至少一个应用的流量的调度策略。
S24,控制设备向承载上述至少一个应用的流量的入口设备发送调度策略指令,以使该入口设备执行该调度策略指令。
在控制设备确定了至少一个应用的流量的调度策略后,控制设备还可以为该至少一个应用中的每个应用生成调度策略指令,并为每个应用配置调度策略指令。例如,控制设备向承载应用的入口设备发送调度策略指令以指示入口设备执行该调度策略指令。入口设备例如为隧道入口设备。
当然,若网络设备(例如,隧道的入口设备)执行该方法,则网络设备可以直接生成调度策略指令并执行该调度策略指令。
以图2为例,第一链路P2-P3承载有PE1-PE2隧道的流量以及PE3-PE2隧道的流量。在上述过程中,如果生成了针对PE1-PE2隧道所承载的应用的调度策略,则将该调度策略指令发送给PE1;如果生成了针对PE3-PE2隧道所承载的应用的调度策略指令,则将该调度策略指令发送给PE3;如果同时生成了针对两个隧道所承载的应用的调度策略指令,则同时给PE1和PE3发送调度策略指令。
在本申请实施例中,调度策略包括无损调度策略和有损调度策略。有损调度策略是指会造成应用数据包丢失的调度策略,无损调度策略是指不会造成应用数据包丢失的调度策略。
其中,无损调度策略包括如下至少一种:路径调整策略、压缩策略。有损调度策略包括:限速策略。除此之外,调度策略还包括管道映射调整策略,而管道映射调整策略既可以是无损调度策略,也可以是有损调度策略。
管道映射调整策略是否有损与调整后的管道的类型相关。例如,存在限速管道(隧道或队列)时,管道映射调整将应用从非限速管道映射到限速管道中,则属于有损调度策略;如果调整后的管道为非限速管道,则管道映射调整属于无损调度策略。
示例性地,路径调整策略包括路由调整策略、隧道分流路径调整策略、隧道分流比例调整策略等。
其中,路由调整策略是指改变应用流量的传输路径的策略。以图1为例,路由调整可以是指将应用的路由从PE1-P1-PE2调整到PE1-P2-P3-PE2。
隧道分流路径调整是指将应用的流量从隧道的一条路径调整到另一条路径上,或者,为隧道新建一条路径,将应用的流量从隧道的一条路径调整到新建的路径上的策略。以图2为例,隧道分流路径调整可以是指将应用从路径PE1-P2-P3-PE2调整到路径PE1-P1-PE2。
隧道分流比例调整是指调整应用的流量在隧道的多条路径上的分流比例的策略。以图2为例,应用的流量在隧道的两条路径PE1-P2-P3-PE2和PE1-P1-PE2上的分流比例为1:1,通过隧道分流比例调整可以将上述分流比例从1:1调整为2:1。
示例性地,压缩策略是指采用压缩算法对应用的数据包进行压缩的策略,通过对数据包进行压缩,减小传输数据包所需的带宽。
示例性地,限速策略包括应用流量限速策略、应用的流量所在隧道限速策略、应用映射的队列整形策略、应用映射的权重公平队列(weighted fair queuing,WFQ)权重调整策略等。
其中,应用流量限速策略是指针对该应用的流量进行限速的策略。例如针对具有该应用的应用流标识的数据包进行限速。
应用的流量所在隧道限速策略是指对整个隧道的流量进行限速的策略。例如针对携带有该隧道标识的数据包进行限速。应用的流量所在隧道限速可以通过调整承诺访问速率(committed access rate,CAR)值配置,通常在隧道入口设备的入端口处操作,如将病毒库下载应用所在隧道的带宽由100Mbps限速至50Mbps。
应用映射的队列整形(shaping)策略是指对该应用所映射到的队列进行限速的策略。这里的队列是指网络设备的发送队列。通常在隧道入口设备的出端口处操作,例如将队列1的shaping值由100Mbps降低至50Mbps。
应用映射的队列的WFQ权重调整策略是指对WFQ队列的权重进行调整的策略,从而使得队列获得的带宽发生变化。下面结合图5对WFQ权重调整进行示例性说明,图5是本申请实施例提供的一种WFQ权重调整的示意图。参见图5,Q表示队列类型,PQ表示队列类型为优先队列(priority queue),WFQ表示队列类型为权重公平队列,其中,PQ队列优先使用链路带宽,WFQ队列使用剩余带宽。W表示权重。在WFQ权重调整时,调整的是队列类型为WFQ队列的权重。在调整前,4个WFQ队列的权重比为1:2:3:3,在调整后,4个WFQ队列的权重比为2:2:3:3。增加了第一个WFQ队列的权重,相当于降低了WFQ队列中其他几个队列的权重。
示例性地,管道映射调整策略包括隧道映射调整策略、队列映射调整策略等。
其中,隧道映射调整策略是指调整应用映射的隧道的策略,将应用映射的隧道从一个隧道调整到另一个隧道。例如,存在隧道A和隧道B,原本应用A映射到隧道A,通过调整将应用A映射到隧道B。如果隧道B不限速,则上述调整属于无损调度策略;如果隧道B限速,则上述调整属于有损调度策略。
其中,队列映射调整策略是指调整应用映射的队列的策略,将应用映射的队列从一个队列调整到另一个队列。例如,网络设备存在队列A和队列B,原本应用B映射到队列A,通过调整将应用B映射到队列B。如果队列B不限速,则上述调整属于无损调度策略;如果队列B限速,则上述调整属于有损调度策略。如,原本病毒库下载应用映射到队列A,队列A为非限速队列,队列B为配置有队列整形的限速队列,为了保护队列A中其他应用的带宽,将病毒库下载应用重新映射到队列B。
在一种可能的实现方式中,应用支持的调度策略可以分别针对各个应用进行配置,从而使得控制设备能够获知每个应用所支持的调度策略。
在另一种可能的实现方式中,不针对每个应用分别进行配置,而是配置应用默认支持的调度策略,这种情况下,每个应用都默认支持这些调度策略。
在又一种可能的实现方式中,部分应用单独配置支持的调度策略,另一部分应用不单独配置支持的调度策略,这些应用支持前述应用默认支持的调度策略。
下面结合图6对本申请如何确定调度策略的详细过程进行示例性说明。图6是本申请实施例提供的另一种调度策略的确定方法的流程图。该方法可以应用在控制设备,也可以应用在网络设备,下面以该方法应用在控制设备为例进行说明。如图6所示,该方法包括如下步骤。
S31,控制设备在第一链路承载的流量所属的多个应用中确定支持无损调度策略的应用。
当存在支持无损调度策略的应用时,执行步骤S32;当不存在支持无损调度策略的应用时,执行步骤S33。
S32,控制设备根据支持无损调度策略的应用所支持的调度策略,确定调度策略,并生成调度策略指令。
示例性地,调度策略指令包括待调度的应用的标识、待使用的调度策略的标识以及调度的相关参数。调度的相关参数是调度执行时涉及的参数值,比如路径调整时调整后的链路的标识等。
在步骤S32中,响应于支持无损调度策略的应用的数量为一个,确定调度策略包括支持无损调度策略的应用所支持的调度策略。响应于支持无损调度策略的应用的数量为多个,确定调度策略包括支持无损调度策略的多个应用中的至少一个应用所支持的调度策略。
对于支持无损调度策略的应用的数量为多个的情况而言,控制设备可以按照如下方式确定调度策略:
对支持无损调度策略的应用及应用支持的调度策略进行组合,得到至少一种策略组合(可能的部分或全部组合),每种策略组合对应针对至少一个应用的调度策略;确定采用至少一种策略组合中的各个策略组合对应的调度策略进行调度时是否能够解决第一链路的拥塞;响应于存在能解决第一链路的拥塞的策略组合,从能解决拥塞第一链路的拥塞的策略组合中选择组合,采用选出的策略组合对应的调度策略作为确定出的调度策略。
其中,对应用及调度策略进行组合,实际是以调度策略作为精度进行组合,例如,应用A支持两种调度策略a、b,组合时应用A有两种加入组合的方式,分别是应用A-调度策略a、应用A-调度策略b。即,一个策略组合包括应用A-调度策略a,另一个策略组合包括应用A-调度策略b。
另外,在进行组合时,还可以基于某种条件进行组合,例如选择优先级低于某个阈值的应用进行组合等,这里的条件本申请不做限制。
如果不存在能够解决第一链路的拥塞的组合,则继续组合并选择,直到穷举了所有组合方式仍不存在能够解决拥塞的组合,则控制设备执行步骤S33。
示例性地,控制设备基于应用的优先级和应用需求中的至少一项,从能解决拥塞第一链路的拥塞的策略组合中选择一种策略组合。
下面先对应用的优先级和应用需求进行介绍,然后再说明如何选择。
其中,应用优先级可以采用数字表示,如1、2、3等,通常数字越大,优先级越低。
其中,应用需求包括如下至少一项:带宽需求、时延需求、丢包需求、抖动需求、服务水平协议需求、客户体验需求。
其中,带宽需求可以包括采集带宽需求和保障带宽需求,采集带宽需求是指对于该应用进行流量采集时采集到的流量数据所占用的带宽,通过限定采集带宽需求,保证针对应用采集的流量数据足够多,作为确定调度策略的数据支撑。保障带宽需求是用来限定应用所需的最小保障带宽、最大保障带宽和/或带宽保障区间。最小保障带宽也即需要保障的最小带宽,可以通过承诺信息速率(committed information rate,CIR)限定,例如50Mbps;最大保障带宽是保障的带宽上限,可以通过峰值信息速率(peak information rate,PIR)限定,例如100Mbps;带宽保障区间包括了需要保障的最小带宽和最大带宽,例如50Mbps~100Mbps。
在一种可能的实现方式中,控制设备基于各个应用的优先级,选择优先级较低的策略组合对应的调度策略。
例如,存在两个组合能够解决第一链路的拥塞问题,一个组合为应用A和应用B(为了方便说明,这里省略了各个应用支持的调度策略),一个组合为应用C和应用D,应用A~D的优先级分别为1、2、3、4,则应用C和应用D组合的优先级更低,控制设备选择应用C和应用D的组合进行调度。通常,组合中包括多个应用时,优先级可以考虑平均优先级,也可以考虑最高优先级。
这种实现方式,能够通过对低优先级业务进行调度,来保证高优先级应用的业务不受影响。
在另一种可能的实现方式中,控制设备基于各个应用的应用需求,选择能够满足策略组合中应用需求的策略组合对应的调度策略。
对于未单独配置支持的调度策略的应用,组合时从默认支持的调度策略中选择并组合。此时,有可能选择的调度策略与应用需求不匹配。因此,需要将这种情况排除出去,以保证应用的业务质量。
例如,应用A存在较低的时延需求,那么若对应用A执行压缩策略就与应用A的需求并不匹配。因为压缩策略涉及到数据发送方的压缩过程和数据接收方的解压缩过程,这两个过程均会带来额外的时延。在选择时,筛选掉包括应用A的压缩策略的组合。
在又一种可能的实现方式中,控制设备基于各个应用的优先级和各个应用的应用需求,选择能够满足组合中应用需求的组合中优先级较低的组合对应的调度策略。
其中,如何判断优先级,以及如何判断是否满足应用需求可以参见上述描述,这里不做赘述。
无论支持无损调度策略的应用的数量为一个还是多个,在确定调度策略的过程中,都要考虑确定的调度策略能否解决第一链路的拥塞问题。
在一种可能的实现方式中,控制设备基于当前的流量数据和上述选择的调度策略判断调度后是否能够解决第一链路的拥塞问题。例如,控制设备根据当前第一链路的流量值和调度减少的流量,计算调度完成后第一链路是否还会拥塞。
在另一种可能的实现方式中,控制设备基于预测的流量数据和上述选择的调度策略判断调度后是否能够解决第一链路的拥塞问题。例如,控制设备可以基于应用的流量画像预测各应用的流量数据,再基于选择的调度策略根据第一链路预期承载的流量值和拥塞阈值判断第一链路是否还会拥塞。
例如,基于各个应用的流量画像分别预测各个应用的流量数据;基于第一策略组合对应的调度策略和预测的各个应用的流量数据,确定第一链路预期承载的流量值;基于第一链路预期承载的流量值判断第一链路是否拥塞。其中,第一策略组合为至少一种策略组合中的任一个。
示例性地,根据预期承载的流量值和拥塞阈值判断是否拥塞。
其中,控制设备根据各个应用的流量序列生成各个应用的流量画像,该流量序列通常对应的时间较长,例如一天或多天,从而能够准确刻画出流量画像。
其中,流量画像是指基于流量序列获取的流量特征,例如包括流量模式、流量峰值、均峰比等特征。
在确定出应用的流量画像后,可以预测应用在一段时间内的流量数据(例如流量峰值)。根据第一策略组合的调度策略可以确定若使用该第一策略组合对应的调度策略后,第一链路承载的应用。根据预测出的第一链路承载的应用的流量数据确定第一链路预期承载的流量值,进而预测是否还会拥塞,也即是否解决了拥塞问题。
在该实现方式中,基于预测的流量数据确定调度策略,使得网络设备执行的调度策略更能匹配链路未来承载的流量,能进一步缓解链路未来可能出现的拥塞,也使得需要进行调度的频率更低,节省了网络资源。
S33,控制设备根据支持有损调度策略的应用所支持的调度策略,确定调度策略,生成调度策略指令。
除了步骤S31中确定出的支持无损调度策略的应用,其他应用即为支持有损调度策略的应用。
在确定有损调度策略时,除了像无损调度策略一样确定待调度的应用、待使用的策略,还确定有损调度时的参数,例如限速值。
同样地,在进行有损调度策略确定时,也可以采用组合的方式,列举出应用和策略的组合,然后从中选择能够解决第一链路拥塞的策略组合,直到列举所有策略组合均无法解决拥塞,则选择无法解决拥塞的策略组合。其中,调度涉及的参数采用默认值或者应用需要对应的值,例如涉及限速调度策略,则根据应用支持的最小保障带宽设置限速值。
选择时,仍然可以考虑应用优先级和应用需求中的至少一个。选择方法和无损调度相同,这里不再赘述。
例如,关键应用的应用需求包括带宽需求,且最小保障带宽较高,这种情况下可以排除包含关键应用的调度策略为限速策略的策略组合。
再例如,灾备应用的应用需求包括带宽需求,且最小保障带宽较低,这种情况下可以将包含灾备应用的调度策略为限速策略的策略组合纳入选择范围。
在本申请实施例中,限速策略可以针对应用执行,也可以针对隧道执行。例如,隧道中仅包括单个应用时,若单个应用支持限速策略,可以直接对隧道进行限速。再例如,隧道承载的多个应用均支持限速策略,也可以直接对隧道进行限速。
由于本申请提供的方案能够实现不同粒度的调度,例如应用粒度、隧道粒度等,通过不同粒度的调度方案使得本申请的流量调度更精细,更容易缓解链路拥塞。
在上述调度策略确定过程中,仅考虑了对第一链路上的应用的流量进行调度,从而使得第一链路不再拥塞。但如果是路径调整策略、管道映射调整策略将第一链路上的流量转移到其他链路,有可能导致其他链路出现拥塞。为了避免这种情况的发生,在采用路径调整策略、管道映射调整策略时,还需要考虑调度后,受影响的其他链路的拥塞情况。
比如,在步骤S32中,先基于应用的优先级和应用需求中的至少一项确定调度策略;如果该调度策略包括路径调整策略、隧道分流路径调整策略、管道映射调整策略中的一种,则确定该调度策略涉及的链路(第一链路之外的链路);根据调度策略涉及的链路的流量数据,确定调度策略执行后该调度策略涉及的链路是否会拥塞。如果不会拥塞,则采用该调度策略生成调度策略指令。如果会拥塞,则按照步骤S32重新确定调度策略。
如果步骤S32中所有可能的调度方式均会造成调度策略涉及的链路出现拥塞,则执行步骤S33。
在步骤S33中,同样地,如果该调度策略包括管道映射调整策略,也需要考虑调度策略涉及的链路是否会出现拥塞。如果不会拥塞,则采用该调度策略生成调度策略指令。如果会拥塞,则按照步骤S33重新确定调度策略。
前述实施例主要针对的是单条链路拥塞时的调度策略确定的方法,而当网络同时存在两条或多条链路拥塞时,以两条为例,分别针对两条链路确定调度策略,并且调度的目标是调度策略涉及的链路不会拥塞。下面结合图7对存在两条或多条拥塞链路时,如何确定调度策略的详细过程进行示例性说明。图7是本申请实施例提供的另一种调度策略的确定方法的流程图。该方法可以应用在控制设备,也可以应用在网络设备,下面以该方法应用在控制设备为例进行说明。如图7所示,该方法包括如下步骤。
S41,控制设备确定第一链路和第二链路拥塞。
第一链路和第二链路是拥塞链路中的任意两条。以图2为例,第一链路为P1-PE2(属于路径PE1-P1-PE2)、第二链路为P2-P3(属于路径PE1-P2-P3-PE2)。
这里仅仅是以两条链路的情况进行示例性说明,多条链路拥塞的实现方案和两条本质相同。
S42,控制设备获取第一链路承载的流量所属的各个应用支持的调度策略,获取第二链路承载的流量所属的各个应用支持的调度策略。
S43,控制设备根据第一链路承载的流量所属的各个应用支持的调度策略,确定针对第一链路上的至少一个应用的流量的第一调度策略;根据第二链路承载的流量所属的各个应用支持的调度策略,确定针对第二链路上的至少一个应用的流量的第二调度策略。
S44,控制设备根据第一调度策略和第二调度策略涉及的链路的流量数据,确定第一调度策略和第二调度策略执行后涉及的链路是否会拥塞。如果不会拥塞,则执行S45。如果会拥塞,则重新步骤S43。
针对涉及的每个链路,根据采集到的流量数据确定是否会在调度后发生拥塞,确定的详细步骤可以参见步骤S32。
S45,控制设备采用第一调度策略生成第一调度策略指令,采用第二调度策略生成第二调度策略指令。
在对多条链路拥塞进行调度时,使得每个链路拥塞的调度策略都不会造成其他链路的拥塞。即,该方案以整个网络为调度对象,可以从全局调度流量,解决网络中的拥塞问题。
在上述调度策略确定过程中,可以预测调度后的收益,该收益用于指示拥塞链路改善情况。在调度完成后,确定调度的实际收益。一方面实际收益可以反馈给用户设备进行可视化展示,另一方面,基于预期 收益和实际收益的偏差,可以为后续的调度策略的确定提供指导,从而使得调度后的效果更好。
下面结合图8示例性说明如何基于收益调整调度策略的确定过程。图8是本申请实施例提供的另一种调度策略的确定方法的流程图。该方法可以应用在控制设备,也可以应用在网络设备,下面以该方法应用在控制设备为例进行说明。如图8所示,该方法包括如下步骤。
S51,控制设备基于确定出的调度策略确定采用该调度策略进行流量调度后的预期收益。
步骤S51在应用该调度策略之前执行,也即预测应用该调度策略产生的收益。
根据第一链路的流量数据,预测调度策略执行后第一链路的流量数据。根据预测的第一链路的流量数据确定调度后的预期收益。
其中,预期收益可以包括:拥塞链路减少数量,拥塞链路流量占用降低比例等。
S52,控制设备在调度策略指令下发后,确定实际收益。
在一种可能的实现方式中,控制设备接收网络设备发送的调度后的流量数据,根据上述流量数据确定第一链路的实际收益。
在另一种可能的实现方式中,控制设备接收网络设备发送的实际收益。
在该步骤中,实际收益与预期收益包含的参数类型是相同的,不再赘述。
S53,控制设备在第三链路出现拥塞时,基于预期收益和实际收益进行调度策略的确定。
其中,第三链路可以是前文的第一链路或第二链路,也可以是第一链路和第二链路之外的链路。
在这里,控制设备进行调度策略确定时,除了前文考虑的因素外,再额外考虑预期收益和实际收益的偏差。例如,当实际收益未达到预期收益时,在针对第三链路进行调度策略确定时,可以以一个比第三链路拥塞阈值更低的阈值作为目标进行调度,这样流量调度的力度更高,从而使得最终的实际收益能够满足要求。相反,当实际收益超出预期收益时,在针对第三链路进行调度策略确定时,可以以一个比第三链路拥塞阈值更高的阈值作为目标进行调度。而如果实际收益与预期收益相当,则以第三链路拥塞阈值为目标进行调度即可。
前文中对于流量的调度主要是为了解决链路拥塞的问题,除此之外,本申请也可以在不考虑链路拥塞的情况下,对部分应用进行管道映射的调整,相当于执行管道映射调整策略。下面结合图9和图10进行示例性说明。图9是本申请实施例提供的另一种调度策略的确定方法的流程图。该方法可以应用在控制设备,也可以应用在网络设备,下面以该方法应用在控制设备为例进行说明。如图9所示,该方法包括如下步骤。
S61,控制设备根据各个应用的流量数据生成各个应用的流量画像。
S62,控制设备根据各个应用的流量画像,调整各个应用和管道的映射关系,管道包括隧道和队列中的至少一个。
示例性地,控制设备根据各个应用的流量画像,调整各个应用和管道的映射关系,包括:控制设备根据各个应用的流量模式,将流量模式相同的应用映射到同一个管道,将流量模式不同的应用映射到不同的管道。
在该实现方式中,将相同流量模式的应用的流量映射到相同的管道,从而使得不同管道的流量特征更容易区分,从而方便对于各个管道的管控和调度。
将相同流量模式的应用的流量映射到同一个管道,也即对各种流量模式的应用进行分类,从而使得分类后的管道的流量具有包含的应用对应的流量模式的特征,由于各个流量模式的变化趋势是已知的,因此,这种映射后各个管道的流量变化趋势已知,有利于调度和管理。
这里,流量模式是指流量画像所反映出的流量变化趋势,包括小值震荡型、突发毛刺型、平稳双峰型和震荡双峰型等。根据流量数据确定流量模式实际是根据流量数据形成的波形图的变化趋势来确定流量模式,这里对确定流量模式的详细过程不做赘述。
其中,流量模式用于指示应用在第一时间段内的流量值的变化趋势。其中,第一时间段可以是指流量数据的采集时间段。在小值震荡型流量模式中,在采集时间段内的大部分时间(例如50%以上)中流量均为一较低流量值(例如1~3Mb/s)。在突发毛刺型流量模式中,流量序列中流量值的跨度很大,从几Mb/s到上百Mb/s,流量变化也没有规律,也即没有形成明显的波峰和波谷。在平稳双峰型流量模式中,流量序列中流量值按峰-谷-峰波形变化,且在山峰和山谷中流量毛刺少。在震荡双峰型流量模式中,流量序列中 流量按峰-谷-峰波形变化,且在山峰和山谷中流量毛刺多(多于平稳双峰型)。
图10是本申请实施例提供的另一种调度策略的确定方法的流程图。该方法可以应用在控制设备,也可以应用在网络设备,下面以该方法应用在控制设备为例进行说明。如图10所示,该方法包括如下步骤。
S71,控制设备获取采用多种分组方式对多个应用进行分组之后得到的每种分组方式下各组的流量画像,每组包括至少一个应用。
例如,该多个应用为第一流量统计数据中涉及的应用。
S72,控制设备根据各组的流量画像,调整多个应用和管道的映射关系,管道包括隧道和队列中的至少一个。
示例性地,控制设备根据各组的流量画像,调整流量统计数据中涉及的多个应用和管道的映射关系,包括:控制设备根据每种分组方式下各组的流量画像的平稳度,选择平稳度满足平稳度条件的分组方式;控制设备将选出的分组方式下的同一组的应用映射到同一个管道,将不同组的应用映射到不同的管道。
其中,平稳度是指平稳的程度,可以根据流量的流量模式及均峰比等参数来评价。在上述选择过程中,可以选择流量模式为小值振荡、平稳双峰型,然后选择这些流量模式中均峰比满足要求的。
在该实现方式中,将不同画像类型组合到一起进行互补,比如双峰、毛刺这些组合到一起,使得组合出的管道的流量画像更平稳型。这种组合使得得到的管道的流量平稳,可以节省链路带宽调整的次数。
图11是本申请实施例提供的另一种调度策略的确定方法的流程图。该方法可以由图1或图2所示的应用场景中控制设备及网络设备(隧道入口设备)、以及图中未示出的用户设备(用户设备可以是服务器设备)共同执行。如图11所示,该方法包括如下步骤。
S81,用户设备向控制设备发送用户配置信息。控制设备接收用户设备发送的用户配置信息。
其中,用户配置信息包括应用的优先级、应用需求和应用支持策略中的至少一个。
下面通过示例对用户配置信息进行说明:
如应用(application,APP)1,优先级6,应用需求:最小保障带宽=10Mbps,应用支持策略:应用限速策略。
APP2,优先级5,应用需求:最小保障带宽=50Mbps,应用支持策略:隧道映射调整策略、隧道分流路径调整策略、应用限速策略。
APP3,优先级4,应用需求:最小保障带宽=80Mbps,应用支持策略:隧道映射调整策略、隧道分流路径调整策略、应用限速策略。
APP4,优先级3,应用需求:最小保障带宽=80Mbps,应用支持策略:隧道映射调整策略、应用限速策略。
APP5,优先级2,应用支持策略:路由调整策略。
为了保证网络设备能够区分各个应用,在用户配置信息中可以包含对各个应用的定义,对于应用的定义可以通过应用流标识信息来区分,其中,应用流标识信息基于互联网协议(Internet Protocol,IP)网段、五元组、或应用感知网络(application aware networking,APN)标识等来实现。
例如,应用流标识信息可以是如下有一种:
源IP段和目的IP段构成的二元组;
源IP地址、目的IP地址、源端口、目的端口和协议号构成的五元组;
APN 6标识。其中,APN标识通常由源IP段-目的IP段、服务水平协议(service level agreement,SLA)构成。
步骤S81为可选步骤,用户配置信息除了由用户设备配置到控制设备外,也可以直接在控制设备中配置,而无需用户设备的参与。
在步骤S81之前,网络已经完成应用和隧道的映射配置,应用和队列的映射配置。通常根据应用的优先级完成应用和隧道的映射配置,应用和队列的映射配置。
S82,控制设备向网络设备发送访问控制列表(access control list,ACL)配置信息。网络设备接收控制设备发送的ACL配置信息。
其中,ACL配置信息用于指示网络设备哪些流量可以通过,哪些流量需要阻止。
步骤S82为可选步骤,ACL配置信息除了由控制设备配置到网络设备外,也可以直接在网络设备中配置,而无需控制设备的参与。
S83,网络设备根据ACL配置信息进行ACL配置。
根据步骤S82接收到的ACL配置信息进行流量的访问控制,对此,本申请不做赘述。
S84,网络设备采集流量数据,向控制设备发送采集的流量数据。控制设备接收网络设备发送的流量数据。
在进行采集流量数据时,可以单个维度采集,例如只采集各个应用的流量数据,也可以多维度采集,例如同时对各个应用、各个隧道的流量数据进行采集。
除了网络设备按照上述方式采集流量外,网络设备还可以不区分应用,发送给控制设备后由控制设备进行区分,例如,控制设备根据流量的源IP地址和目的IP地址,将相同应用的流量的聚合,从而确定各个应用流量数据等。
S85,控制设备根据流量数据确定网络中的各条链路是否存在拥塞。
通常,链路拥塞的原因包括链路故障、应用对应的业务流量突增(如临时视频会议导致业务流量突增)。其中,链路故障包括线路故障、设备故障等。
除了直接根据链路的流量数据确定链路是否拥塞外,也可以根据链路承载的应用或者链路映射的隧道的流量数据确定是否拥塞。
示例性地,步骤S85可以包括:控制设备根据各个隧道的流量数据、隧道和链路的映射关系,确定网络中的各条链路的流量数据;控制设备根据网络中的各条链路的流量数据确定网络中的各条链路是否存在拥塞。
一种情况下,网络设备发送各个应用的流量数据,各个隧道的流量数据可以根据各个应用的流量数据确定。另一种情况下,网络设备发送的就是各个隧道的流量数据。
如图2所示,图中对于P2-P3链路的流量数据,可以根据经过PE1-P2-P3-PE2路径的隧道和经过PE3-P2-P3-PE2路径的隧道两部分的流量确定。如果两条隧道只经过上述两条路,则直接将两个隧道流量相加即可得到P2-P3链路的流量数据。而如果经过PE1-P2-P3-PE2路径的隧道还经过PE1-P1-PE2路径,此时需要考虑该隧道在两个路径上的分流比,按照分流比确定经过PE1-P2-P3-PE2路径的流量,将其和经过PE3-P2-P3-PE2路径的流量相加,得到P2-P3链路的流量数据。
图12是本申请实施例提供的一种网络拓扑示意图。节点S是隧道入口设备,隧道同时包括图中3条路径(S-M-C-D-T、S-M-D-T、S-M-A-B-D-T),在确定M-D链路的流量时,根据隧道中3条路径的分流比,确定出中间路径的流量,也即M-D链路的流量数据。
在该实现方式中,控制设备会基于各个网络设备发送的流量数据来判断是否存在网络拥塞,控制设备根据整个网络的流量情况综合分析,避免只局部分析难以判断全面的情况出现。
S86,在第一链路存在拥塞时,控制设备根据第一链路承载的流量所属的各个应用支持的调度策略,生成针对第一链路上的至少一个应用的流量的调度策略指令。
S87,控制设备将调度策略指令发送给用户设备。用户设备接收该调度策略指令。
可选地,控制设备除了将待执行的调度策略指令发送给用户设备,供人工进行确认外,控制设备还可以计算若执行该调度策略指令能够获得的预期收益,该收益可以是拥塞改善情况,例如拥塞链路减少数量,拥塞链路流量占用降低比例等。将该预期收益发送给用户设备进行展示,从而使得人工在决策前能够更清晰的感知预期效果,使得决策更简化,决策后看到执行的效果,不仅使得决策的收益可视化,而且可以作为下一次决策的经验反馈。
S88,用户设备接收人工针对该调度策略指令的确认或修改信息。
S89,用户设备向控制设备发送确认或修改信息。控制设备接收用户设备发送的确认或修改信息,并确定待发送的调度策略指令。
步骤S87至步骤S89为可选步骤,控制设备也可以不经过用户设备的确认或修改,直接下发调度策略指令。
S810,控制设备向至网络设备发送调度策略指令。该网络设备接收该调度策略指令。
该网络设备为承载上述至少一个应用的隧道入口设备。
S811,网络设备执行该调度策略指令。
在本申请实施例中,步骤S84至步骤S811可以循环执行,例如在执行一次调度策略指令后,重新采集流量数据,并在仍然有链路拥塞时,重新产生调度策略指令并执行。
在循环执行步骤S84至步骤S811的过程中,调度策略指令所选择的调度策略可以按照优先选择无损调度策略然后再选择有损调度策略的顺序实现。
图13是本申请实施例提供的一种调度过程示意图。如图13所示,步骤A先执行无损调度策略,在无损调度策略无法解除拥塞风险时,步骤B执行有损调度策略,在有损调度策略无法解除拥塞风险时,步骤C给出扩容建议。这里,步骤A和步骤B均可以执行一次或多次。
也即,该方法还可以包括:控制设备输出针对第一链路的扩容建议。也即在当前网络容量无法满足第一链路上的应用带宽诉求时,产生针对第一链路的扩容建议。
前述路径调整对应无损调度策略,限速对应有损调度策略。优先选择无损调度策略,保证各个应用的业务质量。而在无损调度难以解决链路拥塞时,则通过有损调度优先保证核心应用的业务质量。
除了给出扩容建议外,响应于第一链路处于轻载状态时,控制设备还可以输出第一链路的缩容建议。
S812,网络设备再次采集流量数据,向控制设备发送采集的流量数据。控制设备接收网络设备发送的流量数据。
S813,控制设备根据流量数据确定实际收益。
示例性地,实际收益可以包括拥塞链路减少数量,拥塞链路流量占用降低比例等。
步骤S812和步骤S813的实际收益确定方式无需网络设备进行计算,由控制设备计算,网络设备的资源占用低。
除了步骤S812和步骤S813的实际收益确定方式外,另一种方式是由网络设备确定出实际收益,然后将实际收益发送给控制设备。
S814,控制设备将实际收益发送给用户设备。用户设备接收该实际收益。
S815,用户设备展示该实际收益。
步骤S812至步骤S815为可选步骤。
在该实现方式中,控制设备接收网络设备发送的执行调度后的流量数据,根据这些流量数据能够确定调度产生的收益,将实际收益发送给用户设备进行展示,从而向用户反馈上述调度策略的确定方案的效果。
S816,控制设备调整各个应用和管道的映射关系,管道包括隧道和队列中的至少一个。
这里,控制设备调整各个应用和管道的映射关系,将调整后的映射关系发送给网络设备,以使网络设备也相应调整。
步骤S816与前述步骤S812~S815的先后关系可以不定,例如在步骤S812之前执行。
除了在链路发生拥塞时执行的调度策略外,在链路不发生拥塞时,也可以通过管道映射关系的调整,改善网络传输。
本申请通过从配置到调度,支撑网络基于应用的端到端精细化管理诉求;基于对流量进行画像和特征分析,为应用特征可视、异常发现、智能调度、管理决策等提供基础;基于隧道、路径、应用等不同粒度的调度管理,灵活组合达成客户差异化管理诉求,同时基于多粒度调度,实现客户自动调优诉求和最小有损优化目标。
图14是本申请实施例提供的一种执行上述方法的系统的模块示意图。参见图14,用户设备包括用户配置模块,用户配置模块基于用户输入生成用户配置信息,也即执行步骤S81。
参见图14,网络设备包括数据采集模块,数据采集模块用于采集流量数据,并发送给控制设备,也即执行步骤S84和S812。
参见图14,控制设备包括业务管理模块,业务管理模块用于基于流量数据确定各条链路是否发生拥塞,也即执行步骤S85。
参见图14,控制设备还包括控制调度模块、管道管理模块、无损优化模块、有损优化模块和链路管理模块。业务管理模块用于基于流量数据生成流量画像,进而实现流量趋势预测。控制调度模块负责策略调度,通过调用管道管理模块、无损优化模块、有损优化模块和链路管理模块实现调度策略指令的产生,除此之外还用于执行管道删减、网络扩缩容建议的输出。管道管理模块用于实现初始应用和管道的映射配置,如单个核心应用映射到单个队列或隧道,而一般优先级应用或低优先级应用,则多个映射到一个队列或隧道,这样在保障核心应用带宽时,在隧道粒度即可达到精准有损优化调度目的;此外,此方式可以减少单 应用单隧道映射时的海量隧道开销。无损优化模块用于提供无损优化调度策略。有损优化模块用于提供有损优化调度策略。链路管理模块用于提供链路扩缩容建议、网络拓扑设计建议等。
通过控制调度模块、管道管理模块、无损优化模块、有损优化模块和链路管理模块一起实现步骤S86。
参见图14,网络设备还包括策略管理模块,用于执行控制设备下发的调度策略指令,也即执行步骤S811。
参见图14,用户设备还包括可视化模块,用于展示预期收益和/或实际收益,也即执行步骤S815。
基于图14提供的系统模块结构可知,本申请提供的方案通过数据采集模块和可视化模块,实现了调度的可视化和传感化,也即基于流量统计,经过分析和处理,完成调度策略,最终实现了收益的可视化。通过控制调度模块、管道管理模块、无损优化模块、有损优化模块和链路管理模块实现了对于调度策略的分析和推荐。通过用户配置模块和可视化模块对调度策略进行人工确认/修改,从而实现了对于调度策略的决策。通过策略管理模块实现了对调度决策的执行。
除此之外,可视化模块还可以配合控制设备,在流量模式异常等情况下输出异常告警和提示等。
图15是本申请实施例提供的一种调度策略的确定装置的框图。该调度策略的确定装置可以通过软件、硬件或者两者的结合实现成为控制设备的全部或者一部分。该调度策略的确定装置可以包括:确定单元901和应用单元902。
确定单元901,用于响应于第一链路存在拥塞,根据第一链路承载的流量所属的多个应用分别支持的调度策略,确定针对第一链路上的至少一个应用的流量的第一调度策略;
应用单元902,用于应用第一调度策略。
可选地,该确定单元901,用于响应于存在支持无损调度策略的应用,根据支持无损调度策略的应用所支持的调度策略,确定第一调度策略。
可选地,该确定单元901,用于响应于支持无损调度策略的应用的数量为一个,确定第一调度策略为支持无损调度策略的应用所支持的调度策略;或,响应于支持无损调度策略的应用的数量为多个,确定第一调度策略包括支持无损调度策略的应用中的至少一个应用所支持的调度策略。
可选地,该确定单元901,用于对支持无损调度策略的应用及应用支持的调度策略进行组合,得到至少一种策略组合,每种策略组合对应针对至少一个应用的调度策略;确定采用至少一种策略组合中的各个策略组合对应的调度策略进行调度时是否能够解决第一链路的拥塞;响应于存在能解决第一链路的拥塞的策略组合,从能解决拥塞第一链路的拥塞的策略组合中选择策略组合,采用选出的策略组合对应的调度策略作为第一调度策略。
可选地,该确定单元901,用于基于预测的流量数据和第一策略组合对应的调度策略,判断第一策略组合对应的调度策略是否能够解决第一链路的拥塞,第一策略组合为至少一种策略组合中的任一个。
可选地,该确定单元901,用于基于多个应用的流量画像预测各个应用的流量数据;基于第一策略组合对应的调度策略和预测的各个应用的流量数据,确定第一链路预期承载的流量值;基于第一链路预期承载的流量值判断第一链路是否拥塞。
可选地,该确定单元901,用于基于应用的优先级和应用需求中的至少一项,从能解决拥塞第一链路的拥塞的策略组合中选择策略组合。
可选地,该确定单元901,用于响应于不存在支持无损调度策略的应用,根据支持有损调度策略的应用所支持的调度策略,确定第一调度策略。
可选地,该确定单元901,还用于确定第一调度策略涉及的链路;根据第一调度策略涉及的链路的流量数据,确定第一调度策略执行后第一调度策略涉及的链路是否会拥塞;
该装置还包括:生成单元903,用于响应于第一调度策略涉及的链路不会拥塞,基于第一调度策略生成第一调度策略指令。
可选地,该确定单元901,还用于响应于第二链路存在拥塞,根据第二链路承载的流量所属的多个应用分别支持的调度策略,确定针对第二链路上的至少一个应用的流量的第二调度策略;确定第一调度策略和第二调度策略涉及的链路;根据涉及的链路的流量数据,确定第一调度策略和第二调度策略执行后涉及的链路是否会拥塞;
生成单元903,用于响应于涉及的链路不会拥塞,基于第一调度策略生成第一调度策略指令,基于第 二调度策略生成第二调度策略指令。
可选地,该确定单元901,还用于在应用第一调度策略前,确定采用第一调度策略进行流量调度后的预期收益;在应用第一调度策略后,确定实际收益;响应于第三链路存在拥塞,基于预期收益和实际收益确定第三调度策略。
可选地,生成单元903,用于根据多个应用的流量数据生成各个应用的流量画像;
该装置还包括:调整单元904,用于根据各个应用的流量画像,调整各个应用和管道的映射关系,管道包括隧道和队列中的至少一个。
可选地,生成单元903,用于获取采用多种分组方式对多个应用进行分组之后得到的每种分组方式下各组的流量画像,每组包括至少一个应用;
调整单元904,用于根据各组的流量画像,调整多个各个和管道的映射关系,管道包括隧道和队列中的至少一个。
示例性地,确定单元由图14中的控制调度模块、管道管理模块、无损优化模块、有损优化模块和链路管理模块共同实现,应用单元可以集成在控制调度模块中。
需要说明的是,上述实施例提供的调度策略的确定装置在进行调度策略确定时,仅以上述各功能单元的划分进行举例说明,实际应用中,可以根据需要而将上述功能分配由不同的功能单元完成,即将设备的内部结构划分成不同的功能单元,以完成以上描述的全部或者部分功能。另外,上述实施例提供的调度策略的确定装置与调度策略的确定方法实施例属于同一构思,其具体实现过程详见方法实施例,这里不再赘述。
图16示出了本申请实施例提供的控制设备或网络设备150的结构示意图。图16所示的控制设备或网络设备150用于执行上述图3至图14任一幅所示的调度策略的确定方法所涉及的操作。该控制设备或网络设备150可以由一般性的总线体系结构来实现。
如图16所示,控制设备或网络设备150包括至少一个处理器151、存储器153以及至少一个通信接口154。
处理器151例如是通用中央处理器(central processing unit,CPU)、数字信号处理器(digital signal processor,DSP)、网络处理器(network processer,NP)、数据处理单元(Data Processing Unit,DPU)、微处理器或者一个或多个用于实现本申请方案的集成电路。例如,处理器151包括专用集成电路(application-specific integrated circuit,ASIC),可编程逻辑器件(programmable logic device,PLD)或者其他可编程逻辑器件、晶体管逻辑器件、硬件部件或者其任意组合。PLD例如是复杂可编程逻辑器件(complex programmable logic device,CPLD)、现场可编程逻辑门阵列(field-programmable gate array,FPGA)、通用阵列逻辑(generic array logic,GAL)或其任意组合。其可以实现或执行结合本申请实施例公开内容所描述的各种逻辑方框、模块和电路。所述处理器也可以是实现计算功能的组合,例如包括一个或多个微处理器组合,DSP和微处理器的组合等等。
可选的,控制设备或网络设备150还包括总线。总线用于在控制设备或网络设备150的各组件之间传送信息。总线可以是外设部件互连标准(peripheral component interconnect,简称PCI)总线或扩展工业标准结构(extended industry standard architecture,简称EISA)总线等。总线可以分为地址总线、数据总线、控制总线等。为便于表示,图16中仅用一条粗线表示,但并不表示仅有一根总线或一种类型的总线。
存储器153例如是只读存储器(read-only memory,ROM)或可存储静态信息和指令的其它类型的静态存储设备,又如是随机存取存储器(random access memory,RAM)或者可存储信息和指令的其它类型的动态存储设备,又如是电可擦可编程只读存储器(electrically erasable programmable read-only Memory,EEPROM)、只读光盘(compact disc read-only memory,CD-ROM)或其它光盘存储、光碟存储(包括压缩光碟、激光碟、光碟、数字通用光碟、蓝光光碟等)、磁盘存储介质或者其它磁存储设备,或者是能够用于携带或存储具有指令或数据结构形式的期望的程序代码并能够由计算机存取的任何其它介质,但不限于此。存储器153例如是独立存在,并通过总线与处理器151相连接。存储器153也可以和处理器151集成在一起。
通信接口154使用任何收发器一类的装置,用于与其它设备或通信网络通信,通信网络可以为以太网、无线接入网(RAN)或无线局域网(wireless local area networks,WLAN)等。通信接口154可以包括有线 通信接口,还可以包括无线通信接口。具体的,通信接口154可以为以太(Ethernet)接口、快速以太(Fast Ethernet,FE)接口、千兆以太(Gigabit Ethernet,GE)接口,异步传输模式(Asynchronous Transfer Mode,ATM)接口,无线局域网(wireless local area networks,WLAN)接口,蜂窝网络通信接口或其组合。以太网接口可以是光接口,电接口或其组合。在本申请实施例中,通信接口154可以用于控制设备或网络设备150与其他设备进行通信。
在具体实现中,作为一种实施例,处理器151可以包括一个或多个CPU,如图16中所示的CPU0和CPU1。这些处理器中的每一个可以是一个单核(single-CPU)处理器,也可以是一个多核(multi-CPU)处理器。这里的处理器可以指一个或多个设备、电路、和/或用于处理数据(例如计算机程序指令)的处理核。
在具体实现中,作为一种实施例,控制设备或网络设备150可以包括多个处理器,如图16中所示的处理器151和处理器155。这些处理器中的每一个可以是一个单核处理器(single-CPU),也可以是一个多核处理器(multi-CPU)。这里的处理器可以指一个或多个设备、电路、和/或用于处理数据(如计算机程序指令)的处理核。
在具体实现中,作为一种实施例,控制设备或网络设备150还可以包括输出设备和输入设备。输出设备和处理器151通信,可以以多种方式来显示信息。例如,输出设备可以是液晶显示器(liquid crystal display,LCD)、发光二级管(light emitting diode,LED)显示设备、阴极射线管(cathode ray tube,CRT)显示设备或投影仪(projector)等。输入设备和处理器151通信,可以以多种方式接收用户的输入。例如,输入设备可以是鼠标、键盘、触摸屏设备或传感设备等。
在一些实施例中,存储器153用于存储执行本申请方案的程序代码1510,处理器151可以执行存储器153中存储的程序代码1510。也即是,控制设备或网络设备150可以通过处理器151执行存储器153中的程序代码1510,来实现方法实施例提供的数据处理方法。程序代码1510中可以包括一个或多个软件模块。可选地,处理器151自身也可以存储执行本申请方案的程序代码或指令。
在具体实施例中,本申请实施例的控制设备或网络设备150可对应于上述各个方法实施例中的控制器,控制设备或网络设备150中的处理器151读取存储器153中的指令,使图16所示的控制设备或网络设备150能够执行控制器所执行的全部或部分操作。
具体的,处理器151用于响应于第一链路存在拥塞,根据所述第一链路承载的流量所属的多个应用分别支持的调度策略,确定针对所述第一链路上的至少一个应用的流量的第一调度策略;应用所述第一调度策略。
其他可选的实施方式,为了简洁,在此不再赘述。
其中,图3至图14任一幅所示的调度策略的确定方法的各步骤通过控制设备或网络设备150的处理器中的硬件的集成逻辑电路或者软件形式的指令完成。结合本申请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,处理器读取存储器中的信息,结合其硬件完成上述方法的步骤,为避免重复,这里不再详细描述。
本申请实施例还提供了一种芯片,包括:输入接口、输出接口、处理器和存储器。输入接口、输出接口、处理器以及存储器之间通过内部连接通路相连。处理器用于执行存储器中的代码,当代码被执行时,处理器用于执行上述任一种的调度策略的确定方法。
应理解的是,上述处理器可以是CPU,还可以是其他通用处理器、DSP、ASIC、FPGA或者其他可编程逻辑器件、分立门或者晶体管逻辑器件、分立硬件组件等。通用处理器可以是微处理器或者是任何常规的处理器等。值得说明的是,处理器可以是支持ARM架构的处理器。
进一步地,在一种可选的实施例中,上述处理器为一个或多个,存储器为一个或多个。可选地,存储器可以与处理器集成在一起,或者存储器与处理器分离设置。上述存储器可以包括只读存储器和随机存取存储器,并向处理器提供指令和数据。存储器还可以包括非易失性随机存取存储器。例如,存储器还可以存储参考块和目标块。
该存储器可以是易失性存储器或非易失性存储器,或可包括易失性和非易失性存储器两者。其中,非易失性存储器可以是ROM、PROM、EPROM、EEPROM或闪存。易失性存储器可以是RAM,其用作外 部高速缓存。通过示例性但不是限制性说明,许多形式的RAM可用。例如,SRAM、DRAM、SDRAM、DDR SDRAM、ESDRAM、SLDRAM和DR RAM。
本申请实施例中,还提供了一种计算机可读存储介质,计算机可读存储介质存储有计算机指令,当计算机可读存储介质中存储的计算机指令被控制设备执行时,使得控制设备执行上述所提供的调度策略的确定方法。
本申请实施例中,还提供了一种包含指令的计算机程序产品,当其在控制设备上运行时,使得控制设备执行上述所提供的调度策略的确定方法。
在上述实施例中,可以全部或部分地通过软件、硬件、固件或者其任意组合来实现。当使用软件实现时,可以全部或部分地以计算机程序产品的形式实现。所述计算机程序产品包括一个或多个计算机指令。在计算机上加载和执行所述计算机程序指令时,全部或部分地产生按照本申请所述的流程或功能。所述计算机可以是通用计算机、专用计算机、计算机网络、或者其他可编程装置。所述计算机指令可以存储在计算机可读存储介质中,或者从一个计算机可读存储介质向另一个计算机可读存储介质传输,例如,所述计算机指令可以从一个网站站点、计算机、服务器或数据中心通过有线(例如同轴电缆、光纤、数字用户线)或无线(例如红外、无线、微波等)方式向另一个网站站点、计算机、服务器或数据中心进行传输。所述计算机可读存储介质可以是计算机能够存取的任何可用介质或者是包含一个或多个可用介质集成的服务器、数据中心等数据存储设备。所述可用介质可以是磁性介质,(例如,软盘、硬盘、磁带)、光介质(例如,DVD)、或者半导体介质(例如固态硬盘Solid State Disk)等。
本领域普通技术人员可以理解实现上述实施例的全部或部分步骤可以通过硬件来完成,也可以通过程序来指令相关的硬件完成,所述的程序可以存储于一种计算机可读存储介质中,上述提到的存储介质可以是只读存储器,磁盘或光盘等。
以上所述仅为本申请的可选实施例,但本申请的保护范围并不局限于此,任何熟悉本技术领域的技术人员在本申请揭露的技术范围内,可轻易想到的变化或替换,都应涵盖在本申请的保护范围之内。因此,本申请的保护范围应该以权利要求的保护范围为准。
除非另作定义,此处使用的技术术语或者科学术语应当为本申请所属领域内具有一般技能的人士所理解的通常意义。本申请专利申请说明书以及权利要求书中使用的“第一”、“第二”、“第三”以及类似的词语并不表示任何顺序、数量或者重要性,而只是用来区分不同的组成部分。同样,“一个”或者“一”等类似词语也不表示数量限制,而是表示存在至少一个。“包括”或者“包含”等类似的词语意指出现在“包括”或者“包含”前面的元件或者物件涵盖出现在“包括”或者“包含”后面列举的元件或者物件及其等同,并不排除其他元件或者物件。
以上仅为本申请一个实施例,并不用以限制本申请,凡在本申请的精神和原则之内,所作的任何修改、等同替换、改进等,均应包含在本申请的保护范围之内。

Claims (25)

  1. 一种调度策略的确定方法,其特征在于,所述方法包括:
    响应于第一链路存在拥塞,根据所述第一链路承载的流量所属的多个应用分别支持的调度策略,确定针对所述多个应用中的至少一个应用的流量的第一调度策略;
    应用所述第一调度策略。
  2. 根据权利要求1所述的方法,其特征在于,所述根据所述第一链路承载的流量所属的多个应用分别支持的调度策略,确定针对所述第一链路上的至少一个应用的流量的第一调度策略,包括:
    响应于存在支持无损调度策略的应用,根据支持所述无损调度策略的应用所支持的调度策略,确定所述第一调度策略。
  3. 根据权利要求2所述的方法,其特征在于,所述根据支持所述无损调度策略的应用所支持的调度策略,确定所述第一调度策略,包括:
    响应于支持所述无损调度策略的应用的数量为一个,确定所述第一调度策略为支持所述无损调度策略的应用所支持的调度策略;或,响应于支持所述无损调度策略的应用的数量为多个,确定所述第一调度策略包括支持所述无损调度策略的应用中的至少一个应用所支持的调度策略。
  4. 根据权利要求3所述的方法,其特征在于,所述确定所述第一调度策略包括支持所述无损调度策略的应用中的至少一个应用所支持的调度策略,包括:
    对支持所述无损调度策略的应用及应用支持的调度策略进行组合,得到至少一种策略组合,每种策略组合对应针对至少一个应用的调度策略;
    确定采用所述至少一种策略组合中的各个策略组合对应的调度策略进行调度时是否能够解决所述第一链路的拥塞;
    响应于存在能解决所述第一链路的拥塞的策略组合,从能解决所述第一链路的拥塞的策略组合中选择策略组合,所述第一调度策略包括选出的策略组合对应的调度策略。
  5. 根据权利要求4所述的方法,其特征在于,所述确定采用所述至少一种策略组合中的各个策略组合对应的调度策略进行调度时是否能够解决所述第一链路的拥塞,包括:
    基于预测的流量数据和第一策略组合对应的调度策略,判断所述第一策略组合对应的调度策略是否能够解决所述第一链路的拥塞,所述第一策略组合为所述至少一种策略组合中的任一个。
  6. 根据权利要求5所述的方法,其特征在于,所述基于预测的流量数据和第一策略组合对应的调度策略,判断所述第一策略组合对应的调度策略是否能够解决所述第一链路的拥塞,包括:
    基于所述多个应用的流量画像预测各个应用的流量数据;
    基于所述第一策略组合对应的调度策略和预测的所述各个应用的流量数据,确定所述第一链路预期承载的流量值;
    基于所述第一链路预期承载的流量值判断所述第一链路是否拥塞。
  7. 根据权利要求4至6任一项所述的方法,其特征在于,所述从能解决拥塞所述第一链路的拥塞的策略组合中选择策略组合,包括:
    基于应用的优先级和应用需求中的至少一项,从能解决拥塞所述第一链路的拥塞的策略组合中选择策略组合。
  8. 根据权利要求2至7任一项所述的方法,其特征在于,所述根据所述第一链路承载的流量所属的多个应用分别支持的调度策略,确定针对所述第一链路上的至少一个应用的流量的第一调度策略,还包括:
    响应于不存在支持所述无损调度策略的应用,根据支持有损调度策略的应用所支持的调度策略,确定所述第一调度策略。
  9. 根据权利要求1至8任一项所述的方法,其特征在于,所述方法还包括:
    确定所述第一调度策略涉及的链路;
    根据所述第一调度策略涉及的链路的流量数据,确定所述第一调度策略执行后所述第一调度策略涉及的链路是否会拥塞;
    响应于所述第一调度策略涉及的链路不会拥塞,基于所述第一调度策略生成第一调度策略指令。
  10. 根据权利要求1至8任一项所述的方法,其特征在于,所述方法还包括:
    响应于第二链路存在拥塞,根据所述第二链路承载的流量所属的多个应用分别支持的调度策略,确定针对所述第二链路上的至少一个应用的流量的第二调度策略;
    确定所述第一调度策略和所述第二调度策略涉及的链路;
    根据所述涉及的链路的流量数据,确定所述第一调度策略和所述第二调度策略执行后所述涉及的链路是否会拥塞;
    响应于所述涉及的链路不会拥塞,基于所述第一调度策略生成第一调度策略指令,基于所述第二调度策略生成第二调度策略指令。
  11. 根据权利要求1至10任一项所述的方法,其特征在于,所述方法还包括:
    在应用所述第一调度策略前,确定采用所述第一调度策略进行流量调度后的预期收益;
    在应用所述第一调度策略后,确定实际收益;
    响应于第三链路存在拥塞,基于所述预期收益和所述实际收益确定第三调度策略。
  12. 根据权利要求1至11任一项所述的方法,其特征在于,所述方法还包括:
    根据所述多个应用的流量数据生成各个应用的流量画像;
    根据所述各个应用的流量画像,调整所述各个应用和管道的映射关系,所述管道包括隧道和队列中的至少一个。
  13. 根据权利要求1至11任一项所述的方法,其特征在于,所述方法还包括:
    获取采用多种分组方式对所述多个应用进行分组之后得到的每种分组方式下各组的流量画像,每组包括至少一个应用;
    根据所述各组的流量画像,调整所述多个应用和管道的映射关系,所述管道包括隧道和队列中的至少一个。
  14. 一种调度策略的确定装置,其特征在于,所述装置包括:
    确定单元,用于响应于第一链路存在拥塞,根据所述第一链路承载的流量所属的多个应用分别支持的调度策略,确定针对所述第一链路上的至少一个应用的流量的第一调度策略;
    应用单元,用于应用所述第一调度策略。
  15. 根据权利要求14所述的装置,其特征在于,所述确定单元用于:
    响应于存在支持无损调度策略的应用,根据支持所述无损调度策略的应用所支持的调度策略,确定所述第一调度策略。
  16. 根据权利要求15所述的装置,其特征在于,所述确定单元用于:
    响应于支持所述无损调度策略的应用的数量为一个,确定所述第一调度策略为支持所述无损调度策略的应用所支持的调度策略;或,响应于支持所述无损调度策略的应用的数量为多个,确定所述第一调度策略包括支持所述无损调度策略的应用中的至少一个应用所支持的调度策略。
  17. 根据权利要求16所述的装置,其特征在于,所述确定单元用于:
    对支持所述无损调度策略的应用及应用支持的调度策略进行组合,得到至少一种策略组合,每种策略组合对应针对至少一个应用的调度策略;
    确定采用所述至少一种策略组合中的各个策略组合对应的调度策略进行调度时是否能够解决所述第一链路的拥塞;
    响应于存在能解决所述第一链路的拥塞的策略组合,从能解决所述第一链路的拥塞的策略组合中选择策略组合,所述第一调度策略包括选出的策略组合对应的调度策略。
  18. 根据权利要求14至17任一项所述的装置,其特征在于,所述确定单元用于:
    确定所述第一调度策略涉及的链路;
    根据所述第一调度策略涉及的链路的流量数据,确定所述第一调度策略执行后所述第一调度策略涉及的链路是否会拥塞;
    响应于所述第一调度策略涉及的链路不会拥塞,基于所述第一调度策略生成第一调度策略指令。
  19. 根据权利要求14至17任一项所述的装置,其特征在于,所述确定单元用于:
    响应于第二链路存在拥塞,根据所述第二链路承载的流量所属的多个应用分别支持的调度策略,确定针对所述第二链路上的至少一个应用的流量的第二调度策略;
    确定所述第一调度策略和所述第二调度策略涉及的链路;
    根据所述涉及的链路的流量数据,确定所述第一调度策略和所述第二调度策略执行后所述涉及的链路是否会拥塞;
    响应于所述涉及的链路不会拥塞,基于所述第一调度策略生成第一调度策略指令,基于所述第二调度策略生成第二调度策略指令。
  20. 根据权利要求14至19任一项所述的装置,其特征在于,所述确定单元用于:
    在应用所述第一调度策略前,确定采用所述第一调度策略进行流量调度后的预期收益;
    在应用所述第一调度策略后,确定实际收益;
    响应于第三链路存在拥塞,基于所述预期收益和所述实际收益确定第三调度策略。
  21. 根据权利要求14至20任一项所述的装置,其特征在于,所述装置还包括生成单元和调整单元,
    所述生成单元,用于根据所述多个应用的流量数据生成各个应用的流量画像;
    所述调整单元,用于根据所述各个应用的流量画像,调整所述各个应用和管道的映射关系,所述管道包括隧道和队列中的至少一个。
  22. 根据权利要求14至20任一项所述的装置,其特征在于,所述装置还包括生成单元和调整单元,
    所述生成单元,用于获取采用多种分组方式对所述多个应用进行分组之后得到的每种分组方式下各组的流量画像,每组包括至少一个应用;
    所述调整单元,用于根据所述各组的流量画像,调整所述多个应用和管道的映射关系,所述管道包括隧道和队列中的至少一个。
  23. 一种电子设备,其特征在于,所述电子设备包括处理器和存储器,所述存储器用于存储软件程序,所述处理器通过运行或执行存储在所述存储器内的软件程序,以使所述电子设备实现如权利要求1至13任一项所述的方法。
  24. 一种计算机可读存储介质,其特征在于,所述计算机可读存储介质用于存储处理器所执行的程序代码,所述程序代码包括用于实现如权利要求1至13任一项所述的方法的指令。
  25. 一种计算机程序产品,其特征在于,包括程序代码,当计算机运行所述计算机程序产品时,使得所述计算机执行如权利要求1至13任一项所述的方法。
PCT/CN2023/105128 2022-09-30 2023-06-30 调度策略的确定方法及相关装置 WO2024066623A1 (zh)

Applications Claiming Priority (2)

Application Number Priority Date Filing Date Title
CN202211215235.3 2022-09-30
CN202211215235.3A CN117812014A (zh) 2022-09-30 2022-09-30 调度策略的确定方法及相关装置

Publications (1)

Publication Number Publication Date
WO2024066623A1 true WO2024066623A1 (zh) 2024-04-04

Family

ID=90418969

Family Applications (1)

Application Number Title Priority Date Filing Date
PCT/CN2023/105128 WO2024066623A1 (zh) 2022-09-30 2023-06-30 调度策略的确定方法及相关装置

Country Status (2)

Country Link
CN (1) CN117812014A (zh)
WO (1) WO2024066623A1 (zh)

Citations (4)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN104052681A (zh) * 2014-06-25 2014-09-17 中国联合网络通信集团有限公司 流量控制方法和装置
US20160021013A1 (en) * 2014-07-21 2016-01-21 Cisco Technology, Inc. Network traffic control during limited power situations
CN111314236A (zh) * 2020-04-14 2020-06-19 杭州迪普科技股份有限公司 报文转发方法及装置
CN114051001A (zh) * 2021-11-10 2022-02-15 中国电信股份有限公司 流量数据处理方法及装置、存储介质及电子设备

Patent Citations (4)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN104052681A (zh) * 2014-06-25 2014-09-17 中国联合网络通信集团有限公司 流量控制方法和装置
US20160021013A1 (en) * 2014-07-21 2016-01-21 Cisco Technology, Inc. Network traffic control during limited power situations
CN111314236A (zh) * 2020-04-14 2020-06-19 杭州迪普科技股份有限公司 报文转发方法及装置
CN114051001A (zh) * 2021-11-10 2022-02-15 中国电信股份有限公司 流量数据处理方法及装置、存储介质及电子设备

Also Published As

Publication number Publication date
CN117812014A (zh) 2024-04-02

Similar Documents

Publication Publication Date Title
US9705783B2 (en) Techniques for end-to-end network bandwidth optimization using software defined networking
US7796507B2 (en) Method and apparatus for communication network congestion control
US7616572B2 (en) Call admission control/session management based on N source to destination severity levels for IP networks
US8782244B2 (en) Method and apparatus for a software programmable intelligent network
US20190014053A1 (en) Network Flow Control Method And Network Device
WO2020093780A1 (zh) 一种处理网络切片中用户接入的方法及装置
CN107135521B (zh) 一种流量控制方法、装置和系统
CN110177054B (zh) 一种端口队列调度方法、装置、网络控制器及存储介质
EP2469756A1 (en) Communications network management
CN104660507B (zh) 数据流转发路由的控制方法及装置
CN108476175B (zh) 使用对偶变量的传送sdn流量工程方法与系统
WO2021244247A1 (zh) 转发数据报文的方法、网络节点、系统及存储介质
CN102934403A (zh) 控制网络上的数据传输
US20220052949A1 (en) Dynamic deterministic adjustment of bandwidth across multiple hubs with Adaptive Per-Tunnel Quality of Service (QoS)
CN112152863B (zh) 分布式带宽分配方法及装置
CN111431811A (zh) 一种报文传输控制方法、装置和网络设备
KR20170033179A (ko) 소프트웨어 정의 네트워크 기반 가상 네트워크 사용 대역폭 관리 방법 및 가상 네트워크 관리장치
US20240007406A1 (en) Transmission control method and apparatus
CN109088822A (zh) 数据流量转发方法、装置、系统、计算机设备及存储介质
CN114466404A (zh) 业务流数据传输方法、装置、电子设备及存储介质
US20190306228A1 (en) Data prioritization and scheduling system
WO2024066623A1 (zh) 调度策略的确定方法及相关装置
US20240205155A1 (en) Protocol agnostic cognitive congestion control
EP3471351B1 (en) Method and device for acquiring path information about data packet
WO2022242243A1 (zh) 通信方法、设备及系统

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: 23869871

Country of ref document: EP

Kind code of ref document: A1