CN115314444B

CN115314444B - SDN-based time-sensitive network gating decision method and device

Info

Publication number: CN115314444B
Application number: CN202211237243.8A
Authority: CN
Inventors: 孙雷; 胡文学; 王健全; 朱渊; 孙志权; 毕紫航
Original assignee: University of Science and Technology Beijing USTB
Current assignee: University of Science and Technology Beijing USTB
Priority date: 2022-10-11
Filing date: 2022-10-11
Publication date: 2023-01-17
Anticipated expiration: 2042-10-11
Also published as: CN115314444A

Abstract

The invention provides a time-sensitive network gating decision method and device based on an SDN (software defined network), and relates to the technical field of industrial communication. The method comprises the following steps: establishing a TSN control architecture based on a software definition concept; a device for managing and controlling the TSN network switching equipment is constructed based on a software definition concept by adopting a fully centralized control architecture model; and uniformly scheduling and arranging a TSN end-to-end gating list based on the gating decision mechanism of the SDN to finish the time-sensitive network gating decision based on the SDN. The method comprises the steps that an end-to-end delay upper bound of a service under the scene of multiple groups of network nodes is quantitatively analyzed aiming at a TAS mechanism widely adopted by a TSN in an industrial scene; and (3) a TSN centralized control architecture based on an SDN is set up, and GCL configuration control decision is carried out on the switch through cooperative cooperation among the functional modules. Under the condition of guaranteeing the time delay performance QoS requirement of the service based on a centralized network controller architecture and simplifying the gating setting complexity, a feasible solution is provided for the service scheduling based on the queue.

Description

SDN-based time-sensitive network gating decision method and device

Technical Field

The invention relates to the technical field of industrial communication, in particular to a time-sensitive network gating decision method and device based on an SDN.

Background

A time-sensitive network (TSN) has a deterministic delay guarantee and a multi-service carrying capability, solves the problem of data transmission in the same network in the industrial internet, and becomes a research hotspot of the existing industrial field network. However, the standard protocol of the TSN only defines a method for forwarding and processing data, and does not specify the networking of the TSN in an industrial environment. Therefore, in a complex industrial network environment, an end-to-end delay analysis mode in a time sensitive network multi-node networking environment is established for the quality-of-service (QoS) requirement index constraints such as delay, jitter, packet loss rate and rate, and the delay performance of the network under extreme conditions is quantitatively analyzed, which is a difficult problem for the current time sensitive network application.

The TSN is a new type of network with deterministic latency data transmission capability built on the basis of the conventional ethernet. The main reason for the lack of certainty in conventional ethernet is that it is essentially a shared transmission medium, congestion occurs when the traffic in the network is too large, the queuing time cannot be predicted, and certainty is difficult to guarantee. For the TSN, the main task is to guarantee the end-to-end delay and jitter requirements of the time sensitive traffic flow, and therefore, the boundedness of the end-to-end delay is a leading feature of the transmission determinacy of the TSN. Network calculation is a deterministic queuing theory based on minimum additive numbers, and in recent years, the network performance boundary is widely drawn, including the boundary of an individual network node or the end-to-end time delay of a network. The document 1[ Axer P, thiele D, ernst R, et al, application relating to application context to implementation requirements of ethernet AVB networks [ C ]// IEEE Design Automation Conference (DAC) ] San Francisco: IEEE, 2014: 1-6] utilizes the inherent characteristics of the Ethernet AVB service to analyze the worst end-to-end time delay of the service under the non-frame preemption strategy; document 2 [ Mohammadpour E, stai E, mohimuddin M, et al, latex and backlog bases in time-sensitive network with credit based shapers and asynchronous traffic shapers (ITC 30) ] IEEE, 2018: 1-6 analyzes the formation factors of network congestion under credit-based shapers (CBS) and Asynchronous Traffic Shapers (ATS) and evaluates the delay performance of network nodes at different rates; document 3 [ Zhao L, paul P, cracinuas S. Work-case latency analysis for IEEE802.1Qbv time sensitive networks using network clock [ j ]. IEEE Access, 2018, 6: 41803-41815.] for the influence of gate-control list (GCL) time distribution in IEEE802.1Qbv on end-to-end delay, a GCL scheduling model based on flexible windows is proposed by using network calculus, and the upper bound of end-to-end delay of the model under different configurations is analyzed.

In the existing time delay upper bound analysis based on network calculation in the TSN, on one hand, the arriving model of the service is characterized by very simple drawing, and most of the time delay upper bound analysis adopts the quantity of burst length to describe; on the other hand, the existing research is mainly based on the CBS mechanism defined in IEEE802.1 Qav to perform end-to-end delay analysis of services. How to model the service through network calculation and guide the design of the traffic shaper are currently challenging and of research value.

The TAS is defined by IEEE802.1Qbv standard, and specifies whether or not message transmission is possible for a priority service at every moment by using the GCL, and in the case of full network clock synchronization, the GCL in the TAS periodically controls the opening and closing of each queue exit gate and performs transmission in accordance with a strict priority scheme. Under the multi-node networking scene, to ensure the end-to-end delay certainty of time-sensitive services, the gating lists of multiple switching nodes need to be cooperatively set, which is a difficult problem faced by IEEE802.1Qbv in real networking. In addition, when a multi-node cooperative networking is performed, in order to prevent interference of low-priority traffic on high-priority traffic, the gating setting of the high-priority traffic is generally not "overlapped" with the gating setting of the low-priority traffic, which not only increases the difficulty of queue gating cooperative arrangement, but also sacrifices the network bandwidth in exchange for the "certainty" of transmission of the high-priority traffic.

Disclosure of Invention

The invention provides a time-sensitive network gating decision method and device based on an SDN (software defined network), aiming at the problems that in the prior art, queue gating cooperative arrangement is difficult, and network bandwidth is sacrificed to replace certainty of high-priority service transmission.

In order to solve the technical problems, the invention provides the following technical scheme:

in one aspect, a time-sensitive network gating decision method based on an SDN is provided, and the method is applied to an electronic device and includes the following steps:

s1: a fully centralized control architecture model is adopted, and a device framework for managing and controlling the time sensitive network TSN switching equipment is constructed based on a software definition concept;

s2: based on the device framework for managing and controlling the TSN switching device, a Software Defined Network (SDN) technology is adopted, a TSN gating decision mechanism based on the SDN is established, and a centralized network controller is obtained;

s3: and based on the centralized network controller, performing TSN end-to-end gating unified scheduling and arrangement to complete SDN-based time-sensitive network gating decision.

Optionally, in S2, a Software Defined Network (SDN) technology is adopted to establish a TSN gating decision mechanism based on SDN, including:

and acquiring a global network topology view of the TSN by adopting a Software Defined Network (SDN) technology, performing centralized control on network switching equipment, and establishing a TSN gating decision mechanism based on the SDN.

Optionally, in step S3, based on the centralized network controller, performing unified scheduling and scheduling of TSN end-to-end gating, including:

s31: presetting time nodes, and acquiring the time delay performance QoS requirement of a terminal service through centralized user configuration on the basis of time synchronization of each node;

s32: and converting the acquired time delay performance QoS requirement to a centralized network controller, and performing TSN end-to-end gating unified scheduling and arranging through the centralized network controller.

Optionally, in step S31, collecting the delay performance QoS requirement of the terminal service through centralized user configuration includes:

through centralized user configuration, collecting service information, the capability of a TSN switch and queue gating state parameters, wherein the service information comprises: the length of the data frame, the generation time of the service information at the sending end and the service period.

Optionally, in step S32, performing end-to-end unified scheduling and scheduling of TSNs by the centralized network controller includes:

analyzing the upper bound of the end-to-end time delay of the TSN through the centralized network controller, evaluating whether the gating setting of each switch on the end-to-end path meets the QoS requirement of the terminal service according to the upper bound of the time delay, outputting gating meeting the QoS requirement of the time delay performance of the user, and configuring the gating result to the corresponding switch.

Optionally, step S32 further includes:

and for the service which does not meet the QoS requirement, adjusting switch gating on the transmission path, returning to the switch database, and re-analyzing the time delay upper bound of the service after the gating information is updated until the QoS requirement of the terminal service is met.

Optionally, analyzing, by the centralized network controller, an upper bound of the end-to-end delay of the TSN includes:

the network calculation-based time delay upper bound analysis model is used for carrying out quantitative analysis on an end-to-end time delay upper bound under the scene of multiple groups of network nodes, wherein the network calculation-based time delay upper bound analysis model comprises the following steps: a traffic arrival model and a switch service model.

Optionally, the traffic arrival model comprises:

for periodic traffic, according to a GCRA model of the generic cell rate algorithm

For a target flow, the arrival model is expressed as

For a network topology, if traffic is present

From the source side, a cycle of

Then the arrival model is as shown in equation (1),

wherein，

Which represents the length of one data frame,

is as follows

And a switching node.

In one aspect, a time-sensitive network gating decision making apparatus based on SDN is provided, where the apparatus is applied to an electronic device, and the apparatus includes:

the centralized network control module is used for constructing a device framework for managing and controlling the TSN switching equipment based on a software definition concept by adopting a fully centralized control architecture model;

a decision mechanism establishing module, configured to establish a TSN gating decision mechanism based on an SDN by using a Software Defined Network (SDN) technology based on the device framework for management and control of the TSN switching device, and obtain a centralized network controller;

and the gating decision module is used for performing end-to-end gating unified scheduling and arrangement on the basis of the centralized network controller to finish the time sensitive network gating decision based on the SDN.

Optionally, the decision mechanism establishing module is configured to adopt a software defined network SDN, obtain a global view of the network through the SDN, perform centralized management and control on the network, and establish a gating decision mechanism based on the SDN.

In one aspect, a computer-readable storage medium is provided, where at least one instruction is stored, and the at least one instruction is loaded and executed by a processor to implement the SDN-based time-sensitive network gating decision method.

The technical scheme of the embodiment of the invention at least has the following beneficial effects:

in the above scheme, 1) the delay upper bound quantization analysis model based on network calculation: aiming at a multi-node networking application scene, transmission interference of low-priority service to high-priority service is introduced based on an overlapping mechanism among different queue gating windows, the model solves the upper bound of end-to-end time delay of a TSN (time delay network) under a TAS (time delay simulator) mechanism aiming at the worst condition of gating, and the upper bound of the time delay is expressed by the maximum horizontal deviation between two curves in a service arrival-exchange service model calculated by a network at an exchanger. Based on the establishment of a network calculus time delay quantitative analysis and evaluation mechanism, an important judgment basis is provided for the evaluation of the feasibility of gating setting.

2) TSN centralized control architecture and decision function module based on SDN: the method comprises the steps of building a TSN control architecture based on an SDN to conduct centralized control on a network, deploying a delay upper bound analysis model, and judging whether gating configuration of a TSN switch meets requirements or not according to collected service characteristics and user delay performance QoS requirements, so that the method is used for decision making of multi-switch gating list configuration in the TSN.

Drawings

In order to more clearly illustrate the technical solutions in the embodiments of the present invention, the drawings required to be used in the description of the embodiments are briefly introduced below, and it is obvious that the drawings in the description below are only some embodiments of the present invention, and it is obvious for those skilled in the art that other drawings can be obtained according to the drawings without creative efforts.

Fig. 1 is a flowchart of a time-sensitive network gating decision method based on SDN according to an embodiment of the present invention;

fig. 2 is a flowchart of a time-sensitive network gating decision method based on SDN according to an embodiment of the present invention;

fig. 3 is a diagram of a centralized controller and a functional module of an SDN-based TSN of a SDN-based time-sensitive network gating decision method according to an embodiment of the present invention;

fig. 4 is a time delay upper bound analysis model diagram of a time-sensitive network gating decision method based on SDN according to an embodiment of the present invention;

fig. 5 is a service arrival model diagram of a time-sensitive network gating decision method based on SDN according to an embodiment of the present invention;

fig. 6 is a model diagram of "service arrival-switch service" of a SDN-based time-sensitive network gating decision method according to an embodiment of the present invention;

fig. 7 is a network topology structure diagram of a time-sensitive network gating decision method based on SDN according to an embodiment of the present invention;

fig. 8 is a schematic diagram of an arrival process of a same queue service in a SDN-based time-sensitive network gating decision method according to an embodiment of the present invention;

fig. 9 is a graph of arrival of switch 1 high priority traffic for a SDN-based time-sensitive network gating decision method according to an embodiment of the present invention;

fig. 10 is a block diagram of a time-sensitive network gating decision device based on SDN according to an embodiment of the present invention;

fig. 11 is a schematic structural diagram of an electronic device according to an embodiment of the present invention.

Detailed Description

In order to make the technical problems, technical solutions and advantages of the present invention more apparent, the following detailed description is given with reference to the accompanying drawings and specific embodiments.

The embodiment of the invention provides a time-sensitive network gating decision method based on an SDN (software defined network). As shown in fig. 1, a flowchart of a time-sensitive network gating decision method based on SDN, a processing flow of the method may include the following steps:

s101: a device for managing and controlling the TSN network switching equipment is constructed based on a software definition concept by adopting a fully centralized control architecture model;

s102: establishing a TSN gating decision mechanism based on SDN by adopting a Software Defined Network (SDN) technology;

s103: and performing TSN end-to-end gating unified scheduling and arrangement based on the gating decision mechanism of the SDN to finish the time sensitive network gating decision based on the SDN.

Optionally, in step S102, a software defined network SDN is adopted, and a gating decision mechanism based on the SDN is established, including:

Optionally, in step S103, performing end-to-end unified scheduling and orchestration of TSNs based on the gating decision mechanism of the SDN includes:

s131: presetting time nodes, and collecting the time delay performance QoS requirement of the terminal service through the centralized user configuration on the basis of time synchronization of each node;

s132: and converting the acquired delay performance QoS requirement to a centralized network controller, and performing end-to-end unified scheduling and arranging of TSN through the centralized network controller.

Optionally, in step S131, acquiring a delay performance QoS requirement of a terminal service through the centralized user configuration includes:

through the centralized user configuration, collecting service information, the capacity of a TSN switch and queue gating state parameters, wherein the service information comprises: the length of the data frame, the generation time and the generation period of the data frame at the transmitting end.

Optionally, in step S132, performing end-to-end unified scheduling and scheduling of TSNs by the centralized network controller includes:

analyzing the upper bound of the end-to-end time delay of the TSN through the centralized network controller, evaluating whether the gating setting of each switch on the end-to-end path meets the QoS requirement of the terminal service according to the upper bound of the time delay, outputting gating meeting the QoS requirement of the time delay performance of a user, and configuring the gating to the corresponding switch.

Optionally, step S132 further includes:

and for the service which does not meet the QoS requirement, adjusting switch gating on a transmission path of the service, returning to a switch database, and re-analyzing the time delay upper bound of the service after the gating information is updated until the QoS requirement of the terminal service is met.

the method comprises the steps of quantitatively analyzing an end-to-end delay upper bound under the scene of multiple groups of network nodes based on a network calculation-based delay upper bound analysis model, wherein the network calculation-based delay upper bound analysis model comprises the following steps: a traffic arrival model and a switch service model.

Optionally, the traffic arrival model comprises:

for periodic services, according to a GCRA model of the general cell rate algorithm

For the target flow, the arrival model is expressed as

For a network topology, if traffic is present

From the source side, a cycle of

Then the arrival model is as shown in equation (1),

wherein, the first and the second end of the pipe are connected with each other,

which represents the length of one data frame,

is as follows

And a switching node.

In the embodiment of the invention, aiming at a TAS mechanism widely adopted by TSN in an industrial scene, a time delay upper bound analysis and evaluation mechanism based on network calculation is provided, and the end-to-end time delay upper bound of a service under the scene of a plurality of groups of network nodes is quantitatively analyzed; and (3) a TSN centralized control architecture based on an SDN is set up, and GCL configuration control decision is carried out on the switch through cooperative cooperation among the functional modules. Under the condition of ensuring the time delay performance QoS requirement of the service and simplifying the complexity of gating setting based on the IEEE802.1Qbv and the centralized network controller architecture, a feasible solution is provided for the service scheduling based on the queue.

The embodiment of the invention provides a time-sensitive network gating decision method based on an SDN (software defined network). Fig. 2 is a flowchart of a time-sensitive network gating decision method based on SDN, and fig. 3 is a network topology diagram of energy efficiency optimization of a communication-sensitive integrated network. The processing flow of the method can comprise the following steps:

s201: a device framework for managing and controlling the time-sensitive network TSN switching equipment is constructed by adopting a fully centralized control architecture model and based on a software definition concept.

In the embodiment of the invention, a full-centralized control architecture model is adopted, a TSN (software defined network) control architecture is realized based on a software definition concept, and an SDN (software-defined network) performs centralized control on a network by acquiring a global view of the network, so that high-efficiency resource scheduling capability is provided.

S202: a TSN switching device management and control-based device framework adopts Software Defined Network (SDN) technology to establish a TSN gating decision mechanism based on SDN and obtain a centralized network controller.

In the embodiment of the invention, on the premise that the whole network is always synchronous, a fully centralized control architecture model transfers user requirements to centralized user configuration for centralized acquisition and unified conversion, so that the guarantee capability of QoS (quality of service) is increased, the QoS (quality of service) guarantee capability is transmitted to a centralized network controller through a northbound interface after conversion, and the centralized network controller has network functions of calculating a topological path and the like, and after optimal calculation of a global network view is realized, updated routing path, GCL and other configuration information are transmitted to a TSN (switched traffic network) switching network through a southbound interface.

S203: and presetting time nodes, and acquiring the time delay performance QoS requirement of the terminal service through centralized user configuration on the basis of time synchronization of each node.

In a possible embodiment, parameters such as traffic information, TSN switch capability, queue gating status, etc. are collected through the centralized user configuration, where the traffic information includes: the length of the data frame, the generation time and the generation period of the data frame at the transmitting end.

S204: and converting the acquired time delay performance QoS requirement to a centralized network controller, and performing TSN end-to-end gating unified scheduling and arrangement through the centralized network controller to complete the SDN-based time sensitive network gating decision.

In a feasible implementation manner, when TSN end-to-end unified scheduling and arrangement are performed in combination with a gating decision mechanism based on an SDN, firstly, a centralized user configures, on the basis of time synchronization of each node, the delay performance QoS requirements of terminal services to be acquired, and the acquired requirements are converted and sent to a centralized network controller, as shown in fig. 3, the centralized network controller serves as a controller constructed based on the SDN, firstly, parameters such as service information, the capacity of a TSN switch, a queue gating state and the like are collected so as to perform centralized decision and evaluation, and finally, gating meeting the user QoS requirements is output by evaluating an upper bound of end-to-end transmission delay of a data frame, and the gating is configured to a corresponding switch.

As shown in fig. 3, the SDN-based TSN centralized controller and function module includes:

a switch database: storing parameters (including bandwidth of a switch port, GCL and the like) of all switches in the current network;

a topology discovery module: counting service information (including data frame length, generation time and generation period of the data frame at a transmitting end, and the like) in the network, and forwarding the service information downwards;

the time delay upper bound analysis and evaluation module comprises: analyzing the upper bound of the end-to-end time delay of the TSN through the information, and evaluating whether the gating setting of each switch on the end-to-end path meets the requirement of the service QoS or not according to the upper bound of the time delay;

a GCL calculation module: for the service which does not meet the QoS requirement, the switch gating on the transmission path is adjusted, the switch gating is returned to the switch database, and the delay upper bound of the service after the gating information is updated is reanalyzed;

a network configuration module: and configuring the GCL value meeting the requirements of all service time delay performance QoS to the switch, and finally finishing the service scheduling decision based on the queue.

In a possible implementation manner, the centralized network controller analyzes an upper bound of end-to-end delay of the TSN, evaluates whether gating setting of each switch on an end-to-end path meets a QoS requirement of a terminal service according to the upper bound of the delay, outputs gating meeting the QoS requirement of the delay performance of a user, and configures the gating to a corresponding switch.

In a feasible implementation manner, for a service which does not meet the QoS requirement, the switch gating on the transmission path is adjusted, and the service returns to the switch database, and the delay upper bound of the service after the gating information is updated is reanalyzed until the QoS requirement of the terminal service is met.

In a possible implementation manner, the network-calculus-based upper delay bound analysis model performs quantitative analysis on an upper end-to-end delay bound under a plurality of groups of network node scenarios, where the network-calculus-based upper delay bound analysis model includes: a traffic arrival model and a switch service model.

In a feasible implementation manner, in the delay analysis and evaluation module, the analysis of the upper delay bound is based on a delay upper bound analysis model of network calculation, and the quantitative analysis is performed on the upper end-to-end delay bound under the scene of multiple groups of network nodes. The service arrival model and the switch service model are two basic tools of network calculus, and are important basis for evaluating the upper bound of the delay by the network calculus.

When the invention analyzes the end-to-end delay upper bound of the TSN under the TAS mechanism by utilizing network calculation, the switch service model is only related to the GCL of the TSN switch. The upper bound of the end-to-end delay is the maximum level difference formed by overlapping the two models under the worst condition. Fig. 4 shows a solution for the upper time delay bound.

In a possible embodiment, the GCRA model is based on the general cell rate algorithm for periodic services

For the target flow, the arrival model is expressed as

Because on the egress port side of the switch, the same priority traffic arriving at the switch will enter the same queue. For a network topology, if traffic

From the source side, a cycle of

Then the arrival model is as shown in equation (1),

wherein the content of the first and second substances,

which represents the length of one data frame,

is as follows

And a switching node.

The traffic arrival model is represented as a graph according to equation (1), as shown in fig. 5.

In addition, the same priority service data frame of the multiple input ports of the switch has self-queuing competition, and the priority is in the moment

Arrival model for business

The sum of the arrival model functions for the traffic of the same priority at each input port is expressed as the formula,

wherein

Indicating the kind of the same priority service;

the number of the same kind of service on the same queue.

Traffic arrival model for path successor switching nodes

With the current node

The delay bound is determined by the maximum horizontal distance between two curves in the 'traffic arrival-switch service' model, so that the switch service model is required

And (4) performing representation. Thus giving it again in the subsequent analysis

。

In one possible implementation, the worst case switch service model:

switch service model

The service capability of a switching node in a network is related to factors such as the opening and closing condition of an internal queue door structure of a switch, the opening time of a window and the like. For the service model, when the traffic arrives at the corresponding queue, because there is an overlap between the windows, its transmission on the queue will be interfered by other priority queues, and the patent claims an upper bound of delay, so the service model with the worst interference situation is considered.

Like the traffic arrival model, the switch service model is also a cumulative function. For a TSN switch, there are 8 different priority queues per egress port for traffic

Is an object, in a queue

Service it, wherein

To do so

Representing a set of queues whose service model is determined by the service model function of each gated window transmission slot, and is obtained by summing them. When traffic is subjected to non-frame-preemption transmission on the corresponding queues, the traffic is interfered by high-priority and low-priority queues and is influenced by guard bands. Under the influence of the three interference factors, the guaranteed service time slot is passed

And offset between time slots

Maximum waiting time

Equation of isoparametric and willWhich is substituted into the service model function to represent the switch service model,

wherein

，

The dequeue forwarding rate for the switch.

While the service model is intended to represent any one transmission slot

Then it is required to

As a starting point, for a supercycle

Of all time slots in

The summation is performed.

So that the target flow rate

In the first place

A queue of switching nodes

Business arrival model over

And switch service model

Respectively, indicating completion. The two models are shown in the same figure, which is the "traffic arrival-switch service" model applied in this patent, as shown in fig. 6.

Maximum horizontal distance between two curves in fig. 6

I.e. the upper bound of the delay sought,

is obtained in the case where the traffic arrival model curve at the location is above the switch service model curve. Representing the upper delay bound of the current node

Then, the arrival model described above is combined

Then the arrival model of the service at the next node can be obtained

Expressed as formula (4),

in one possible embodiment, the gating decision mechanism:

by the technical scheme, the establishment of a time delay quantitative analysis model based on network calculation is completed aiming at a TAS mechanism and aiming at a multi-node networking application scene. And evaluating the upper bound of the end-to-end delay under the given gating condition by comparing the requirement with the QoS (quality of service) delay performance so as to judge the rationality of gating setting. If the delay bound is smaller than the QoS requirement during comparison, GCL configuration can be directly carried out on the exchanger; otherwise, the gating setting does not meet the requirement of time delay, a new GCL value needs to be calculated, time delay upper bound analysis is carried out for multiple times, and finally a feasible GCL value meeting the requirement of service QoS is configured to the switch. Therefore, the gating decision of multiple switches is completed, and the requirement of the centralized network controller on the dynamic configuration of the switches is realized, namely the implementation process of the mechanism in fig. 2.

In a feasible implementation manner, according to the above scheme, a delay upper bound analysis model based on network calculation is modularized to form a centralized network controller delay upper bound analysis and evaluation module, so as to realize calculation of an upper bound of network end-to-end delay under a given topology. In practical application, because the end-to-end delay of the high-priority service is preferentially guaranteed according to the order of the service priorities, the upper bound of the end-to-end delay of the high-priority service is considered in a centralized manner in the embodiment.

The network topology structure is shown in fig. 7, in which there are 5 service sending terminals, 1 service receiving terminal and two TSN switches.

The service arrival model adopted in this embodiment is a step function, that is, the service is periodically generated by the sending end and sent to the switch, so the time for the service to arrive at the switch is also periodic, and for a data stream

Its arrival period is recorded as

，

The period is constant, because the propagation rate of the data frame on the link is the same, the period is the same as the generation period of the traffic data frame at the transmitting end.

The service priority is first defined. After being input to the switch, the services with different priorities enter different queues for buffering, the service with the high priority has strong sensitivity to time and needs to be transmitted within a certain time, otherwise, the service with the low priority has poor sensitivity to time. Artificially set flow

Priority of (2)

And arrival period

In inverse proportion to each other, i.e.

The smaller the size of the tube is,

the higher. Therefore, if the generation cycles of the data streams are the same, the priorities of the data streams are also the same, and after the data streams arrive at the switch, the data streams with the same priority enter the same queue for buffering, and are transmitted after the gating window is opened. Due to the fact that the time of the traffic arriving at the switch is different, the arrival model of the data flow on the same queue is an aperiodic function, and the graph is represented to be an irregular graph. Two services with the same priority on one queue are taken as an example for explanation, as shown in fig. 8. In the figure, the position of the first and second end faces,

representing a frame of data.

Four kinds of services are set and sent to a TSN switch from five service sending ends respectively, the four kinds of services are divided into three priorities according to an arrival period, and after the four kinds of services arrive at the switch, the four kinds of services are transmitted on three queues inside the switch, so that the queues are defined as high-priority queues, medium-priority queues and low-priority queues respectively. According to the foregoing, the present embodiment only discusses the high-priority service, and therefore, only the high-priority service information needs to be given and the arrival model function is shown, and other priority service information does not need to be defined. Of these four services, there are two high priority services, symbolized

Indicating that they are generated by sender 1, sender 2 and sender 5, respectively, with an arrival period of

. According to the topology structure diagram, relative to the time 0, in the initial period, the time of two services generated by the transmitting end 1 and the transmitting end 2 reaching the switch 1 is respectively

And

and the traffic generated by the sender 5 arrives at the exchange 2 at the time of

. As with the solution objective, only the arrival model of the high priority traffic is fitted, taking switch 1 as an example, and the arrival model of the high priority traffic is shown in fig. 9.

Combining with a step function model, expressing the arrival model of the high-priority service of the two switches as a formula,

arrival curve of exchange 1

Since the traffic received by switch 2 comes from two parts, its arrival model function is represented in the form of an addition,

in the above two formulas, the first and second groups,

represents the time offset of arrival of high priority traffic at the switch relative to time 0, where

、

；

Representing the upper bound of the delay of traffic flowing through switch 1.

For a switch service model, its ability to express network service data, in relation to network state, involves the gating setup problem of the switch. According to the topology structure diagram, two TSN switches are shared in the network, and gating setting is needed. It is known that the queue gating window cycle is opened and closed, and thus the parameter settings include the opening and closing time of the gating window in the initial cycle, and the queue gating cycle. It is known that "1" represents the gating on state, and "0" represents the gating off state, and when the gating state is "1", data is transmitted, otherwise, data is not transmitted. Because the end-to-end delay of the high-priority service is guaranteed first, queue gating is set according to the sequence of the queue priority from high to low, and a high-priority queue gating window needs to be guaranteed to have a certain width. Defining gating periods of different queues as

The opening time of each queue gating window in the initial period of the switch 1 is

Corresponding off-time of

Then the opening and closing time of other periodic gating windows can pass

、

And with

In the form of a sum:

，

wherein

And the sequence number of the gating window is represented, so that after the gating period is determined, only the opening and closing time of the gating window in the initial period needs to be defined. Firstly, setting equal gating periods of all queues

Based on the queue gating period, the GCL hypercycle of the switch is also determined, also

Therefore, each priority queue has a gating window in a super-cycle. Then according to the sequence of priority from high to low, combining the time of data frame arriving at the exchanger, setting the opening and closing time of every queue gate control window in the initial period in turn, setting the opening and closing time of high priority queue gate control window as

And

when the gating windows of the other two priority queues are set, in order to explain the problem, the gating windows of the other two priority queues are overlapped with the gating window of the high priority queue to a certain degree, and the opening and closing time of the gating window of the medium priority queue is set as

And

low priority is

And

. At this point, a set of GCLs for switch 1 is set to completion.

Similarly, the same is true for the gating definition of switch 2, which has the same gating period as switch 1, but the opening and closing times of the queue gating window are different. After the gating parameter configuration is determined, the upper bound of the end-to-end delay of the high-priority service under the corresponding GCL is solved according to the formulas (5) and (6).

In the formula (I), the compound is shown in the specification,

as a priority queue

At one is

The number of inner door control windows is equal to

；

Is the upper bound on the latency inside both switches.

And quantitatively analyzing the upper bound of the end-to-end time delay of the TSN under the scene of multiple groups of network nodes through the established end-to-end time delay upper bound analysis model so as to judge whether the gating setting can meet the time delay performance QoS requirement of the service, thereby completing the construction of a gating decision model and deploying the gating decision model in a centralized network controller module to realize the GCL configuration of multiple switches.

In the embodiment of the invention, 1) a time delay upper bound quantitative analysis model based on network calculation: aiming at a multi-node networking application scene, transmission interference of low-priority service to high-priority service is introduced based on an overlapping mechanism among different queue gating windows, the model solves the upper bound of end-to-end time delay of a TSN (time delay network) under a TAS (time delay simulator) mechanism aiming at the worst condition of gating, and the upper bound of the time delay is expressed by the maximum horizontal deviation between two curves in a service arrival-exchange service model calculated by a network at an exchanger. Based on the establishment of a network calculus time delay quantitative analysis and evaluation mechanism, an important judgment basis is provided for the evaluation of the feasibility of gating setting.

Figure 10 is a block diagram illustrating an SDN based time sensitive network gating decision device, according to an example embodiment. Referring to fig. 10, the apparatus 300 includes:

the centralized network controller module 310 is configured to construct a device framework for managing and controlling the TSN switch device in the time sensitive network based on a software definition concept by using a fully centralized control architecture model;

a decision mechanism establishing module 320, configured to establish a TSN gating decision mechanism based on an SDN by using a Software Defined Network (SDN) technology based on the device framework for managing and controlling the TSN switching device, so as to obtain a centralized network controller;

and the gating decision module 330 is configured to perform unified scheduling and scheduling of end-to-end gating of the TSN based on the centralized network controller, and complete a time-sensitive network gating decision based on the SDN.

Optionally, the decision mechanism establishing module is configured to obtain a global network topology view of the TSN by using a Software Defined Network (SDN) technology, perform centralized control on the network switching device, and establish a TSN gating decision mechanism based on the SDN.

Optionally, the gating decision module 330 is configured to preset time nodes, and collect the delay performance QoS requirement of the terminal service through the centralized user configuration on the basis of time synchronization of each node;

and converting the acquired time delay performance QoS requirement to a centralized network controller, and performing TSN end-to-end unified scheduling and arrangement through the centralized network controller to complete the SDN-based time-sensitive network gating decision.

Optionally, the gating decision module 330 is configured to collect, through the centralized user configuration, parameters such as traffic information, a capability of a TSN switch, a queue gating state, and the like, where the traffic information includes: the length of the data frame, the generation time and the generation period of the data frame at the transmitting end.

Optionally, the gating decision module 330 is configured to analyze an upper bound of end-to-end delay of the TSN through the centralized network controller, evaluate whether a gating setting of each switch on an end-to-end path meets a QoS requirement of a terminal service according to the upper bound of delay, output a gating meeting the QoS requirement of user delay performance, and configure the gating to a corresponding switch.

Optionally, the gating decision module 330 is further configured to:

Optionally, the gating decision module 330 is configured to perform quantitative analysis on an end-to-end delay upper bound under a plurality of sets of network nodes based on a network calculus delay upper bound analysis model, where the network calculus delay upper bound analysis model includes: a traffic arrival model and a switch service model.

Optionally, the traffic arrival model comprises:

For the target flow, the arrival model is expressed as

For a network topology, if traffic is present

From the source side, a cycle of

Then the arrival model can be expressed as,

wherein the content of the first and second substances,

which represents the length of one data frame,

is a first

And a switching node.

Fig. 11 is a schematic structural diagram of an electronic device 400 according to an embodiment of the present invention, where the electronic device 400 may generate relatively large differences due to different configurations or performances, and may include one or more processors (CPUs) 401 and one or more memories 402, where the memory 402 stores at least one instruction, and the at least one instruction is loaded and executed by the processor 401 to implement the following steps of the SDN-based time-sensitive network gating decision method:

s1: a fully centralized control architecture model is adopted, and a device framework for managing and controlling the time-sensitive network TSN switching equipment is constructed based on a software definition concept;

In an exemplary embodiment, a computer-readable storage medium, such as a memory, is also provided that includes instructions executable by a processor in a terminal to perform the SDN based time sensitive network gating decision method described above. For example, the computer readable storage medium may be a ROM, a Random Access Memory (RAM), a CD-ROM, a magnetic tape, a floppy disk, an optical data storage device, and the like.

It will be understood by those skilled in the art that all or part of the steps for implementing the above embodiments may be implemented by hardware, or may be implemented by a program instructing relevant hardware, where the program may be stored in a computer-readable storage medium, and the above-mentioned storage medium may be a read-only memory, a magnetic disk or an optical disk, etc.

The above description is only for the purpose of illustrating the preferred embodiments of the present invention and should not be taken as limiting the scope of the present invention, which is intended to cover any modifications, equivalents, improvements, etc. within the spirit and scope of the present invention.

Claims

1. A time-sensitive network gating decision method based on an SDN is characterized by comprising the following steps:

s2: based on the device framework for managing and controlling the TSN switching equipment, a Software Defined Network (SDN) technology is adopted, a TSN gating decision mechanism based on the SDN is established, and a centralized network controller is obtained;

s3: based on the centralized network controller, performing TSN end-to-end gating unified scheduling and arrangement to complete SDN-based time-sensitive network gating decision;

in step S2, a Software Defined Network (SDN) technology is adopted to establish a TSN gating decision mechanism based on SDN, including:

adopting Software Defined Network (SDN) technology to obtain a global network topology view of the TSN, performing centralized control on network switching equipment, and establishing a TSN gating decision mechanism based on the SDN;

in step S3, based on the centralized network controller, performing TSN end-to-end gating unified scheduling and scheduling, including:

s31: presetting time nodes, and acquiring the time delay performance QoS requirement of the terminal service through centralized user configuration on the basis of time synchronization of each node;

s32: converting the acquired time delay performance QoS requirement to a centralized network controller, and performing TSN end-to-end gating unified scheduling and arrangement through the centralized network controller to complete the SDN-based time sensitive network gating decision;

in step S31, acquiring the delay performance QoS requirement of the terminal service through centralized user configuration includes:

through centralized user configuration, collecting service information, the capability of a TSN switch and queue gating state parameters, wherein the service information comprises: the length of a data frame, the generation time of service information at a sending end and a service period;

in step S32, performing end-to-end unified scheduling and scheduling of TSNs by the centralized network controller includes:

analyzing the upper bound of the end-to-end time delay of the TSN through the centralized network controller, evaluating whether the gating setting of each switch on the end-to-end path meets the QoS requirement of the terminal service according to the upper bound of the time delay, outputting gating meeting the QoS requirement of the time delay performance of a user, and configuring the gating result to the corresponding switch;

the step S32 further includes:

for the service which does not meet the QoS requirement, the switch gating on the transmission path is adjusted, the switch gating is returned to the switch database, and the delay upper bound of the service after the gating information is updated is re-analyzed until the QoS requirement of the terminal service is met;

analyzing, by the centralized network controller, an upper bound of a TSN end-to-end delay includes:

the network calculation-based time delay upper bound analysis model is used for carrying out quantitative analysis on an end-to-end time delay upper bound under the scene of multiple groups of network nodes, wherein the network calculation-based time delay upper bound analysis model comprises the following steps: a business arrival model and a switch service model;

the business arrival model comprises:

For the target flow, the arrival model is expressed as

For a network topology, if traffic is present

From the source side, periodically to the input port of the first switch, and with a period of

Then the arrival model is as shown in equation (1),

(1)

which represents the length of one data frame,

is a first

A switching node;

the switch service model is an accumulation function, and guarantees service time slots when the server is influenced by interference factors

Offset between time slots

Maximum waiting time

The parameters are formulated and substituted into the switch service model function, namely, the switch service model is represented:

(3)

wherein

，

Dequeue forwarding rate for the switch;

the service model for representing any transmission time slot needs to be expressed by

As a starting point, for a supercycle

Of all time slots in

The summation is carried out, and in the analysis of the upper bound of the time delay, the switches are taken as one after the summation

Minimum of amount of service data, hence in queues

Service model of a switch, for example

Can be represented as, for example,

as a priority queuemAt one is

The number of the internal door control windows.

2. An SDN-based time-sensitive network gating decision-making apparatus, adapted to the method of claim 1, the apparatus comprising:

a decision mechanism establishing module, configured to establish a TSN gating decision mechanism based on an SDN by using a Software Defined Network (SDN) technology based on the TSN switch device management and control device framework, and obtain a centralized network controller;

the gate control decision module is used for performing TSN end-to-end gate control unified scheduling and arrangement based on the centralized network controller to complete the SDN-based time sensitive network gate control decision;

the decision mechanism establishing module is used for acquiring a global network topology view of the TSN by adopting a Software Defined Network (SDN) technology, performing centralized control on network switching equipment and establishing a TSN gating decision mechanism based on the SDN;

the gating decision module is used for presetting time nodes and acquiring the time delay performance QoS requirement of the terminal service through the centralized user configuration on the basis of time synchronization of all the nodes;

converting the acquired delay performance QoS requirement to a centralized network controller, and performing end-to-end unified scheduling and arranging on TSN through the centralized network controller to complete the time sensitive network gating decision based on SDN;

a gating decision module, configured to collect, through the centralized user configuration, service information, a capability of a TSN switch, and a queue gating state parameter, where the service information includes: the length of a data frame, the generation time and the generation period of the data frame at a transmitting end;

the gate control decision module is used for analyzing the upper bound of the end-to-end time delay of the TSN through the centralized network controller, evaluating whether the gate control setting of each switch on the end-to-end path meets the QoS requirement of the terminal service according to the upper bound of the time delay, outputting the gate control meeting the QoS requirement of the time delay performance of the user, and configuring the gate control to the corresponding switch;

a gating decision module further configured to:

the gate control decision module is used for carrying out quantitative analysis on an end-to-end delay upper bound under the scene of a plurality of groups of network nodes based on a network calculation-based delay upper bound analysis model, wherein the network calculation-based delay upper bound analysis model comprises: a business arrival model and a switch service model;

the business arrival model includes: