CN106330710B - Data flow scheduling method and device - Google Patents

Abstract

The present invention discloses a data flow scheduling method, the data flow scheduling method includes: according to the current scheduling parameters, determining the scheduling level that the preset port needs to enable; enabling the corresponding scheduling nodes of the determined scheduling level, and establishing different scheduling levels The link relationship between the scheduling nodes; the scheduling strategy is configured for each scheduling node enabled; when the preset port receives the data stream, according to the scheduling strategy configured by each scheduling node and the link relationship between each scheduling node, in The data streams are scheduled hierarchically among the enabled scheduling nodes. The invention also discloses a data flow scheduling device. The invention can improve the efficiency of scheduling data flow.

Description

Data flow scheduling method and device

technical field

The invention relates to the field of communication technology, in particular to a data flow scheduling method and device.

Background technique

With the rapid development of communication technology, services such as voice, video, and data that are sensitive to bandwidth, delay, and jitter are increasingly transmitted through the network. In order to provide different promises and guarantees for data transmission performance, QoS (Quality of Service, Quality of Service) technology is widely used to ensure network transmission quality. However, with the expansion of user scale and the increase of business types, network equipment is required not only to further refine and distinguish business traffic, but also to perform unified management and layering for multiple users, multiple services, and multiple traffic transmission objects. scheduling. This is difficult to realize for the traditional QoS technology. In order to achieve the purpose of hierarchical scheduling, HQoS (Hierarchical Quality of Service, hierarchical QoS) technology emerges as the times require. However, the existing HQoS technology has a fixed scheduling level and cannot perform flexible hierarchical scheduling according to user needs, and the efficiency of scheduling data streams is low.

Contents of the invention

The main purpose of the present invention is to provide a data flow scheduling method and device, aiming at improving the efficiency of scheduling data flow.

In order to achieve the above object, the present invention provides a data flow scheduling method, the data flow scheduling method includes:

Determine the scheduling level that the preset port needs to enable according to the current scheduling parameters;

Enable the corresponding scheduling nodes of the determined scheduling level, and establish the link relationship between scheduling nodes of different scheduling levels;

Configure scheduling policies for each enabled scheduling node;

When the preset port receives the data flow, the data flow is hierarchically scheduled among the enabled scheduling nodes according to the scheduling policy configured by each scheduling node and the link relationship between the scheduling nodes.

Preferably, the preset port includes a class I port and a class II port. When the preset port is a class I port, the preset port includes a 5-level scheduling hierarchy; when the preset port is a class II port When , the preset port includes three scheduling levels.

Preferably, the determining the scheduling level that needs to be enabled for the preset port according to the current scheduling parameters includes:

Always use the scheduling level with the lowest and highest scheduling priority of the preset port as the scheduling level that needs to be enabled, and determine that the preset port needs to be enabled according to the current scheduling parameters, except for the scheduling priority of the preset port Scheduling levels other than the lowest and highest scheduling levels.

Preferably, the configuring a scheduling policy for each enabled scheduling node includes:

Configure scheduling algorithms, traffic shaping policies, and/or congestion avoidance policies for each enabled scheduling node.

Preferably, before determining the scheduling level that the preset port needs to enable according to the current scheduling parameters, it also includes:

receiving scheduling parameters input by the user based on the scheduling parameter setting interface;

The received scheduling parameter is used as the current scheduling parameter.

In addition, in order to achieve the above object, the present invention also provides a data flow scheduling device, the data flow scheduling device includes:

A determining module, configured to determine the scheduling level that needs to be enabled for the preset port according to the current scheduling parameters;

An enabling module, configured to enable corresponding scheduling nodes of a determined scheduling level, and establish a link relationship between scheduling nodes of different scheduling levels;

A configuration module, configured to configure a scheduling policy for each enabled scheduling node;

A dispatching module, configured to, when the preset port receives the data stream, divide the data stream between the enabled scheduling nodes according to the scheduling strategy configured by each scheduling node and the link relationship between the scheduling nodes Layers are scheduled.

Preferably, the preset port includes a class I port and a class II port. When the preset port is a class I port, the preset port includes a 5-level scheduling hierarchy; when the preset port is a class II port When , the preset port includes three scheduling levels.

Preferably, the determining module is also configured to always use the scheduling level with the lowest and highest scheduling priority of the preset port as the scheduling level that needs to be enabled, and determine the level that needs to be enabled for the preset port according to the current scheduling parameters. , other scheduling levels except the lowest and highest scheduling levels of the preset port scheduling priority.

Preferably, the configuration module is further configured to configure scheduling algorithms, traffic shaping strategies and/or congestion avoidance strategies for each enabled scheduling node.

Preferably, the data flow scheduling device further includes a setting module, configured to receive a scheduling parameter input by a user based on a scheduling parameter setting interface; and use the received scheduling parameter as a current scheduling parameter.

Based on the existing HQoS technology, the present invention flexibly selects the scheduling level that needs to be enabled, and enables the corresponding scheduling nodes of the selected scheduling level, and each scheduling node cooperates with each other according to the configured scheduling strategy and the link relationship between scheduling nodes. Implement hierarchical scheduling of data streams. The invention solves the problem of fixed scheduling levels in the existing HQoS technology, better meets user needs, and avoids unnecessary waste of processing resources, so that the invention can improve the efficiency of scheduling data streams.

Description of drawings

FIG. 1 is a schematic flow diagram of a preferred embodiment of the data flow scheduling method of the present invention;

Fig. 2 is a schematic diagram of networking in a preferred embodiment of the data flow scheduling method of the present invention;

FIG. 3 is a schematic diagram of a scheduling hierarchy in a preferred embodiment of the data flow scheduling method of the present invention;

FIG. 4 is a schematic diagram of another scheduling level in a preferred embodiment of the data flow scheduling method of the present invention;

FIG. 5 is a schematic diagram of another scheduling level in a preferred embodiment of the data flow scheduling method of the present invention;

FIG. 6 is a schematic diagram of functional modules of a preferred embodiment of the data stream scheduling device of the present invention.

The realization of the purpose of the present invention, functional characteristics and advantages will be further described in conjunction with the embodiments and with reference to the accompanying drawings.

Detailed ways

It should be understood that the specific embodiments described here are only used to explain the present invention, not to limit the present invention.

The present invention provides a data flow scheduling method. Referring to FIG. 1, in a preferred embodiment of the data flow scheduling method of the present invention, the data flow scheduling method includes:

S10. Determine the scheduling level that needs to be enabled for the preset port according to the current scheduling parameter;

S20, enabling the corresponding scheduling nodes of the determined scheduling level, and establishing a link relationship between scheduling nodes of different scheduling levels;

The data stream scheduling method provided in this embodiment can be applied to an MPLS (Multi-Protocol Label Switching, Multi-Protocol Label Switching) network. For example, in an MPLS network, when a certain port of a certain network node receives a data stream, according to the pre-configured The scheduling level and the corresponding scheduling strategy perform hierarchical scheduling on the data flow, and forward the data flow efficiently. In the following, the method for scheduling data flow provided by the present invention is applied to an MPLS network for illustration.

It should be noted that MPLS is a new generation IP high-speed backbone network switching standard proposed by IETF (Internet Engineering Task Force, Internet Engineering Task Force). MPLS is independent of the second-layer and third-layer protocols. It uses Label (label) for data forwarding: when the data flow enters the network node, it is assigned a fixed-length label, and the label and the data flow are encapsulated together. During the forwarding process, the network nodes only forward according to the label. Wherein, data transmission occurs on LSP (Label Switching Path), and LSP is a label sequence of each network node along the path from the source end to the terminal end.

The data stream scheduling method proposed in this embodiment divides different scheduling levels in advance according to the characteristics of MPLS, and determines the scheduling levels that need to be enabled for the preset ports according to the current scheduling parameters; and enables the scheduling levels determined according to the current scheduling parameters. Correspondingly schedule nodes, and establish link relationships between scheduling nodes at different scheduling levels. Wherein, the scheduling parameters include, but are not limited to, scheduling level information that needs to be enabled, scheduling node information, scheduling policy information, and link relationship information between scheduling nodes at different scheduling levels.

S30, configuring a scheduling strategy for each enabled scheduling node;

After the scheduling nodes are enabled, you need to configure scheduling policies for each enabled scheduling node. Among them, the same scheduling policy can be configured between scheduling nodes of different scheduling levels or the same scheduling level, and different scheduling policies can also be configured, which is specifically configured according to the scheduling policy information in the scheduling parameters.

S40. When the preset port receives the data flow, perform hierarchical scheduling on the data flow among the enabled scheduling nodes according to the scheduling strategy configured by each scheduling node and the link relationship between the scheduling nodes .

It should be noted that, in this embodiment, each scheduling node of a different scheduling level is mounted with a scheduler of a corresponding scheduling level, and the schedulers of different scheduling levels implement hierarchical scheduling of data streams. Specifically, according to the scheduling hierarchy relationship between schedulers (the link relationship between scheduling nodes at different scheduling levels) and the corresponding scheduling strategy, the upper-level scheduler controls the total bandwidth of a group of lower-level schedulers, and according to the level of the lower-level scheduler The committed bandwidth and peak bandwidth of the lower-level scheduler are allocated with weights to realize hierarchical scheduling of data streams.

The data flow scheduling method proposed in this embodiment, on the basis of the existing HQoS technology, flexibly selects the scheduling level that needs to be enabled, and enables the corresponding scheduling nodes of the selected scheduling level, and each scheduling node follows the configured scheduling policy and scheduling The links between nodes cooperate with each other to realize the hierarchical scheduling of data flow. The invention solves the problem of fixed scheduling levels in the existing HQoS technology, better meets user needs, and avoids unnecessary waste of processing resources, so that the invention can improve the efficiency of scheduling data streams.

Further, in a preferred embodiment of the data flow scheduling method of the present invention, the preset ports include a first-class port and a second-class port, and when the preset port is a first-class port, the preset port includes 5 level scheduling hierarchy; when the preset port is a second-class port, the preset port includes a 3-level scheduling hierarchy.

It should be noted that in the MPLS network, there are two types of ports. With reference to FIG. 2, the ports connected to the network nodes in this network are called NNI (network node ports), and the ports connected to this network and other networks are called UNI ( user network port). According to the network structure of MPLS, this embodiment divides different levels of scheduling levels for different types of ports.

Specifically, when the preset port is a first-class port (NNI), the preset port includes 5 levels of scheduling levels, that is, there are 5 levels of scheduling levels for NNI, which are: port, VLAN subport, tunnel, For pseudowires and queues, each scheduling level includes one or more scheduling nodes. For example, referring to FIG. 3 , for NNI, scheduling nodes at all 5 scheduling levels are enabled; referring to FIG. 4 , for NNI, scheduling nodes at port, tunnel, pseudowire, and queue scheduling levels are enabled.

When the preset port is a second-class port (UNI), the preset port includes 3 levels of scheduling levels, that is, there are 3 levels of scheduling levels for UNI, which are: port, CIP (Client Computer's IP address, client computer's IP address, client address) and queues, and each scheduling level includes one or more scheduling nodes. For example, referring to FIG. 5 , for the UNI, scheduling nodes of all three scheduling levels are enabled.

In this embodiment, different scheduling levels are correspondingly divided for different types of ports to meet different network networking environments.

Further, in a preferred embodiment of the data flow scheduling method of the present invention, the determining the scheduling level that needs to be enabled for the preset port according to the current scheduling parameters includes:

Always use the scheduling level with the lowest and highest scheduling priority of the preset port as the scheduling level that needs to be enabled, and determine that the preset port needs to be enabled according to the current scheduling parameters, except for the scheduling priority of the preset port Scheduling levels other than the lowest and highest scheduling levels.

It should be noted that even if ordinary QoS scheduling is used, the scheduling at the port scheduling level and the queue scheduling level are always enabled. The scheduling level with the lowest preset port scheduling priority (port scheduling level) and the scheduling level with the highest scheduling priority (queue scheduling level) are used as the scheduling levels to be enabled. On the basis of always taking the port scheduling level and queue scheduling level as the scheduling levels that need to be enabled, according to the current scheduling parameters, determine the preset ports that need to be enabled, except that the preset port scheduling priority is the lowest and the highest Scheduling levels other than the scheduling level.

For example, for NNI, the port scheduling level and queue scheduling level are always used as the scheduling levels that need to be enabled, and the VLAN sub-interface scheduling level, tunnel scheduling level, and pseudowire scheduling level are dynamically selected as the scheduling levels that need to be enabled according to the current scheduling parameters Level: For UNI, always use the port scheduling level and queue scheduling level as the scheduling level that needs to be enabled, and dynamically select whether to use the CIP scheduling level as the scheduling level that needs to be enabled according to the current scheduling parameters.

Further, in a preferred embodiment of the data flow scheduling method of the present invention, the configuration of scheduling policies for each enabled scheduling node includes:

Configure scheduling algorithms, traffic shaping policies, and/or congestion avoidance policies for each enabled scheduling node.

In this embodiment, the scheduling strategy information in the scheduling parameters includes but is not limited to scheduling algorithms, traffic shaping strategies, and congestion avoidance strategies, and configuring scheduling strategies for each enabled scheduling node includes configuring a scheduling algorithm for each enabled scheduling node , traffic shaping strategy and/or congestion avoidance strategy.

Specifically, for scheduling nodes of different scheduling levels, the scheduling algorithms that can be configured include but are not limited to SP (Strict Priority, strict priority level), RR (Round Robin, round-robin scheduling), DRR (Deficit Round Robin, difference round-robin scheduling), WRR One of (Weighted Round Robin, weighted round robin), WDRR (WeightedDeficit Round Robin, weighted deficit round robin), and DWRR (Dynamic Weighted Round Robin, dynamic weighted round robin). For example, referring to FIG. 3 , an SP algorithm can be configured for the scheduling node null_vlan at the VLAN sub-interface scheduling level, an SP algorithm can be configured for the scheduling node null_lsp at the tunnel scheduling level, and a DWRR algorithm can be configured for the scheduling node Lsp_1 at the tunnel scheduling level.

In order to prevent the data flow forwarded by the preset port from exceeding the carrying capacity of the network, preferably, in this embodiment, a traffic shaping policy is configured on the scheduling nodes of each scheduling level to perform traffic shaping. The traffic shaping strategy is to use token buckets to evaluate data flows and shave peaks and fill valleys, buffer peak data flows in the shaping queue, and delay resending, so as to ensure that packets (data flows are composed of packets) are not lost as much as possible.

The congestion avoidance strategy can only be configured on the scheduling nodes at the queue scheduling level. Congestion avoidance is a flow control mechanism. It monitors the usage of the queue buffer, and when the congestion tends to intensify, it actively discards packets and adjusts the data flow. Remove network overload. The traditional congestion avoidance strategy adopts the tail drop (Tail-Drop) method. When the length of the queue reaches a certain maximum value, all new incoming packets will be discarded, and Tail-Drop will cause TCP global synchronization. Preferably, in order to avoid TCP global synchronization phenomenon, this embodiment adopts WRED (Weighted Random Early Detection, Weighted Random Early Detection) to avoid congestion. WRED sets an upper limit and a lower limit for each queue based on the packet discarding priority, and processes the packets in the queue as follows:

(1), when the length of the queue is less than the lower limit, the message is not discarded;

(2), when the length of the queue exceeds the upper limit, all incoming messages are discarded;

(3) When the length of the queue is between the upper limit and the lower limit, randomly discard incoming messages. The longer the queue, the higher the probability of discarding, but there is a maximum probability of discarding.

Further, in a preferred embodiment of the data flow scheduling method of the present invention, before determining the scheduling level that needs to be enabled for the preset port according to the current scheduling parameters, it also includes:

receiving scheduling parameters input by the user based on the scheduling parameter setting interface;

The received scheduling parameter is used as the current scheduling parameter.

Specifically, the user can set preset port scheduling parameters through the CLI (Command Line Interface, Command Line Interface) and/or GUI (Graphical User Interface, Graphical User Interface) scheduling parameter setting interface provided by the network node according to actual needs. For example, when it is judged that a greater scheduling effort is required to meet the data flow scheduling requirements, the user can enable more levels of scheduling levels by setting corresponding scheduling parameters to meet the data flow scheduling requirements and ensure that the network nodes forward the data flow normally.

Furthermore, the user can also update the set scheduling parameters according to actual needs, so that the network nodes can adapt to the ever-changing network environment.

The present invention also provides a data flow scheduling device. Referring to FIG. 6, in a preferred embodiment of the data flow scheduling device of the present invention, the data flow scheduling device includes:

A determining module 10, configured to determine the scheduling level that the preset port needs to enable according to the current scheduling parameter;

The enabling module 20 is configured to enable corresponding scheduling nodes of a determined scheduling level, and establish a link relationship between scheduling nodes of different scheduling levels;

The data stream scheduling device provided in this embodiment can be applied to an MPLS (Multi-Protocol Label Switching, Multi-Protocol Label Switching) network. For example, in an MPLS network, the data stream scheduling device is built into a network node to run. When a certain port of the network node When receiving the data flow, the data flow scheduling device performs hierarchical scheduling on the data flow according to the pre-configured scheduling level and corresponding scheduling strategy, and efficiently forwards the data flow. In the following, the data flow scheduling device provided by the present invention is applied to an MPLS network for description.

It should be noted that MPLS is a new generation IP high-speed backbone network switching standard proposed by IETF (Internet Engineering Task Force, Internet Engineering Task Force). MPLS is independent of the second-layer and third-layer protocols. It uses Label (label) for data forwarding: when the data flow enters the network node, it is assigned a fixed-length label, and the label and the data flow are encapsulated together. During the forwarding process, the network nodes only forward according to the label. Wherein, data transmission occurs on LSP (Label Switching Path), and LSP is a label sequence of each network node along the path from the source end to the terminal end.

The data flow scheduling device proposed in this embodiment divides different scheduling levels in advance according to the characteristics of MPLS, and the scheduling level that the preset port needs to be enabled is determined by the determination module 10 according to the current scheduling parameters; Scheduling parameters enable corresponding scheduling nodes of a determined scheduling level, and establish link relationships between scheduling nodes of different scheduling levels. Wherein, the scheduling parameters include, but are not limited to, scheduling level information that needs to be enabled, scheduling node information, scheduling policy information, and link relationship information between scheduling nodes at different scheduling levels.

The configuration module 30 is configured to configure a scheduling policy for each enabled scheduling node;

After the scheduling nodes are enabled, the configuration module 30 needs to configure scheduling policies for each enabled scheduling node. The same scheduling strategy or different scheduling strategies can be configured between scheduling nodes at different scheduling levels or at the same scheduling level, which is specifically configured by the configuration module 30 according to the scheduling policy information in the scheduling parameters.

The scheduling module 40 is configured to, when the preset port receives the data stream, according to the scheduling policy configured by each scheduling node and the link relationship between the scheduling nodes, to schedule the data stream between the enabled scheduling nodes Hierarchical scheduling.

It should be noted that, in this embodiment, the scheduling module 40 mounts schedulers of corresponding scheduling levels for each scheduling node of different scheduling levels, and the schedulers of different scheduling levels implement hierarchical scheduling of data streams. Specifically, according to the scheduling hierarchy relationship between schedulers (the link relationship between scheduling nodes at different scheduling levels) and the corresponding scheduling strategy, the upper-level scheduler controls the total bandwidth of a group of lower-level schedulers, and according to the level of the lower-level scheduler The committed bandwidth and peak bandwidth of the lower-level scheduler are allocated with weights to realize hierarchical scheduling of data streams.

The data flow scheduling device proposed in this embodiment, on the basis of the existing HQoS technology, flexibly selects the scheduling level that needs to be enabled, and enables the corresponding scheduling nodes of the selected scheduling level, and each scheduling node follows the configured scheduling policy and scheduling The links between nodes cooperate with each other to realize the hierarchical scheduling of data flow. The invention solves the problem of fixed scheduling levels in the existing HQoS technology, better meets user needs, and avoids unnecessary waste of processing resources, so that the invention can improve the efficiency of scheduling data streams.

Further, in a preferred embodiment of the data flow scheduling device of the present invention, the preset ports include a type I port and a type II port, and when the preset port is a type I port, the preset port includes 5 level scheduling hierarchy; when the preset port is a second-class port, the preset port includes a 3-level scheduling hierarchy.

It should be noted that in the MPLS network, there are two types of ports. With reference to FIG. 2, the ports connected to the network nodes in this network are called NNI (network node ports), and the ports connected to this network and other networks are called UNI ( user network port). In this embodiment, the data flow scheduling device further includes a division module, configured to divide different types of ports into different levels of scheduling levels according to the MPLS network structure.

Specifically, when the preset port is a first-class port (NNI), the preset port includes 5 levels of scheduling levels, that is, the division module divides 5 levels of scheduling levels for NNI, which are: port, VLAN subport, For tunnels, pseudowires and queues, each scheduling level includes one or more scheduling nodes. For example, referring to Fig. 3, for NNI, the enabling module 20 has enabled the scheduling nodes of all 5 levels of scheduling levels divided by the division module; referring to Fig. 4, for NNI, the enabling module 20 has enabled the ports and tunnels divided by the division module , Pseudowire, and scheduling nodes at the queue scheduling level.

When the preset port is a second-class port (UNI), the preset port includes a 3-level scheduling hierarchy, that is, the division module divides a 3-level scheduling hierarchy for the UNI, which are: port, CIP (Client Computer's IP address, client address) and queues, and each scheduling level includes one or more scheduling nodes. For example, referring to FIG. 5 , for UNI, the enabling module 20 enables scheduling nodes of all three scheduling levels divided by the dividing module.

In this embodiment, different scheduling levels are correspondingly divided for different types of ports to meet different network networking environments.

Further, in a preferred embodiment of the data flow scheduling device of the present invention, the determining module 10 is also configured to always use the scheduling level with the lowest and highest priority of the preset port scheduling as the scheduling level that needs to be enabled, and According to the current scheduling parameters, other scheduling levels other than the lowest scheduling level and the highest scheduling level with the highest scheduling priority of the preset port that need to be enabled for the preset port are determined.

It should be noted that even if ordinary QoS scheduling is adopted, the scheduling at the port scheduling level and the queue scheduling level are always enabled. The module 10 always takes the scheduling level with the lowest preset port scheduling priority (port scheduling level) and the scheduling level with the highest scheduling priority (queue scheduling level) as the scheduling levels to be enabled. On the basis of always taking the port scheduling level and the queue scheduling level as the scheduling levels that need to be enabled, the determination module 10 determines that the preset ports need to be enabled according to the current scheduling parameters, except for the preset port scheduling priority Scheduling levels other than the lowest and highest scheduling levels.

For example, for NNI, the determination module 10 always takes the port scheduling level and the queue scheduling level as the scheduling levels that need to be enabled, and dynamically selects the VLAN sub-interface scheduling level, the tunnel scheduling level, and the pseudowire scheduling level as the scheduling levels that need to be enabled according to the current scheduling parameters. For the UNI, the determination module 10 always uses the port scheduling level and the queue scheduling level as the scheduling level that needs to be enabled, and dynamically selects whether to use the CIP scheduling level as the scheduling level that needs to be enabled according to the current scheduling parameters.

Further, in a preferred embodiment of the data flow scheduling device of the present invention, the configuration module 30 is also configured to configure scheduling algorithms, traffic shaping strategies and/or congestion avoidance strategies for each enabled scheduling node.

In this embodiment, the scheduling strategy information in the scheduling parameters includes but not limited to scheduling algorithms, traffic shaping strategies and congestion avoidance strategies, and the configuration module 30 is also used to configure scheduling algorithms, traffic shaping strategies and congestion avoidance strategies for each enabled scheduling node. / or congestion avoidance strategy.

Specifically, for scheduling nodes of different scheduling levels, the scheduling algorithm that configuration module 30 can configure includes but is not limited to SP (Strict Priority, strict priority level), RR (Round Robin, round-robin scheduling), DRR (DeficitRound Robin, difference round-robin scheduling ), WRR (Weighted Round Robin, weighted round robin), WDRR (Weighted Deficit Round Robin, weighted deficit round robin), and one of DWRR (Dynamic Weighted Round Robin, dynamic weighted round robin). For example, referring to FIG. 3 , the configuration module 30 may configure the SP algorithm for the scheduling node null_vlan at the VLAN sub-interface scheduling level, configure the SP algorithm for the scheduling node null_lsp at the tunnel scheduling level, and configure the DWRR algorithm for the scheduling node Lsp_1 at the tunnel scheduling level.

In order to prevent the data flow forwarded by the preset port from exceeding the carrying capacity of the network, preferably, the configuration module 30 configures a traffic shaping policy on the scheduling nodes of each scheduling level to perform traffic shaping. The traffic shaping strategy is to use token buckets to evaluate data flows and shave peaks and fill valleys, buffer peak data flows in the shaping queue, and delay resending, so as to ensure that packets (data flows are composed of packets) are not lost as much as possible.

The congestion avoidance strategy can only be configured on the scheduling nodes at the queue scheduling level. Congestion avoidance is a flow control mechanism. It monitors the usage of the queue buffer, and when the congestion tends to intensify, it actively discards packets and adjusts the data flow. Remove network overload. The traditional congestion avoidance strategy adopts the tail drop (Tail-Drop) method. When the length of the queue reaches a certain maximum value, all new incoming packets will be discarded, and Tail-Drop will cause TCP global synchronization. Preferably, in order to avoid TCP global synchronization phenomenon, this embodiment adopts WRED (Weighted Random Early Detection, Weighted Random Early Detection) to avoid congestion. WRED sets an upper limit and a lower limit for each queue based on the packet discarding priority, and processes the packets in the queue as follows:

(1), when the length of the queue is less than the lower limit, the message is not discarded;

(2), when the length of the queue exceeds the upper limit, all incoming messages are discarded;

(3) When the length of the queue is between the upper limit and the lower limit, randomly discard incoming messages. The longer the queue, the higher the probability of discarding, but there is a maximum probability of discarding.

Further, in a preferred embodiment of the data flow scheduling device of the present invention, the data flow scheduling device further includes a setting module, configured to receive the scheduling parameters input by the user based on the scheduling parameter setting interface; and the received scheduling parameters parameter as the current scheduling parameter.

Specifically, the user can set the default port scheduling parameters through the setting module based on the CLI (CommandLine Interface, command line interface) and/or GUI (Graphical User Interface, Graphical User Interface) scheduling parameter setting interface provided by the network node according to actual needs . For example, when it is judged that a greater scheduling effort is required to meet the data flow scheduling requirements, the user can enable more levels of scheduling levels by setting corresponding scheduling parameters to meet the data flow scheduling requirements and ensure that the network nodes forward the data flow normally.

Furthermore, the user can also update the set scheduling parameters according to actual needs, so that the network nodes can adapt to the ever-changing network environment.

The above are only preferred embodiments of the present invention, and are not intended to limit the patent scope of the present invention. Any equivalent structure or equivalent process conversion made by using the description of the present invention and the contents of the accompanying drawings, or directly or indirectly used in other related technical fields , are all included in the scope of patent protection of the present invention in the same way.

Claims (10)

1. A data flow scheduling method, characterized in that, the data flow scheduling method comprises: Determine the scheduling level that the preset port needs to enable according to the current scheduling parameter, wherein the scheduling parameter includes at least information about the scheduling level that needs to be enabled, scheduling node information, scheduling policy information, and link relationship information between scheduling nodes at different scheduling levels ; Enable the corresponding scheduling nodes of the determined scheduling level, and establish the link relationship between scheduling nodes of different scheduling levels; Configure scheduling policies for each enabled scheduling node; When the preset port receives the data flow, the data flow is hierarchically scheduled among the enabled scheduling nodes according to the scheduling policy configured by each scheduling node and the link relationship between the scheduling nodes. 2. The data flow scheduling method according to claim 1, wherein the preset port includes a first-class port and a second-class port, and when the preset port is a first-class port, the preset port includes 5-level scheduling hierarchy; when the preset port is a Type 2 port, the preset port includes 3-level scheduling hierarchy. 3. The data flow scheduling method according to claim 2, wherein the determining the scheduling level that the preset port needs to enable according to the current scheduling parameters includes: Always use the scheduling level with the lowest and highest scheduling priority of the preset port as the scheduling level that needs to be enabled, and determine that the preset port needs to be enabled according to the current scheduling parameters, except for the scheduling priority of the preset port Scheduling levels other than the lowest and highest scheduling levels. 4. The data flow scheduling method according to any one of claims 1-3, wherein the configuring a scheduling strategy for each enabled scheduling node comprises: Configure scheduling algorithms, traffic shaping policies, and/or congestion avoidance policies for each enabled scheduling node. 5. The data flow scheduling method according to any one of claims 1-3, characterized in that before determining the scheduling level that needs to be enabled for the preset port according to the current scheduling parameters, it also includes: receiving scheduling parameters input by the user based on the scheduling parameter setting interface; The received scheduling parameter is used as the current scheduling parameter. 6. A data flow scheduling device, characterized in that, the data flow scheduling device comprises: A determining module, configured to determine the scheduling level that needs to be enabled for the preset port according to the current scheduling parameter, wherein the scheduling parameter includes at least information about the scheduling level that needs to be enabled, scheduling node information, scheduling policy information, and information about scheduling nodes at different scheduling levels. Link relationship information between; An enabling module, configured to enable corresponding scheduling nodes of a determined scheduling level, and establish a link relationship between scheduling nodes of different scheduling levels; A configuration module, configured to configure a scheduling policy for each enabled scheduling node; A scheduling module, configured to, when the preset port receives the data stream, divide the data stream between the enabled scheduling nodes according to the scheduling policy configured by each scheduling node and the link relationship between the scheduling nodes Layers are scheduled. 7. The data flow scheduling device according to claim 6, wherein the preset port includes a first-class port and a second-class port, and when the preset port is a first-class port, the preset port includes 5-level scheduling hierarchy; when the preset port is a Type 2 port, the preset port includes 3-level scheduling hierarchy. 8. The data flow scheduling device according to claim 7, wherein the determining module is further configured to always use the scheduling level with the lowest and highest priority of the preset port scheduling as the scheduling level that needs to be enabled, and According to the current scheduling parameters, other scheduling levels other than the lowest scheduling level and the highest scheduling level with the highest scheduling priority of the preset port that need to be enabled for the preset port are determined. 9. The data flow scheduling device according to any one of claims 6-8, wherein the configuration module is further configured to configure scheduling algorithms, traffic shaping strategies and/or congestion avoidance strategies for each enabled scheduling node . 10. The data flow scheduling device according to any one of claims 6-8, wherein the data flow scheduling device further comprises a setting module, configured to receive a scheduling parameter input by a user based on a scheduling parameter setting interface; and set The received scheduling parameters are used as current scheduling parameters.