WO2021227245A1 - 一种面向tsn和非tsn互联的工业异构网络调度方法 - Google Patents

一种面向tsn和非tsn互联的工业异构网络调度方法 Download PDF

Info

Publication number
WO2021227245A1
WO2021227245A1 PCT/CN2020/102199 CN2020102199W WO2021227245A1 WO 2021227245 A1 WO2021227245 A1 WO 2021227245A1 CN 2020102199 W CN2020102199 W CN 2020102199W WO 2021227245 A1 WO2021227245 A1 WO 2021227245A1
Authority
WO
WIPO (PCT)
Prior art keywords
tsn
data
path
network
delay
Prior art date
Application number
PCT/CN2020/102199
Other languages
English (en)
French (fr)
Inventor
魏旻
向雪琴
王平
晏先春
Original Assignee
重庆邮电大学
Priority date (The priority date is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the date listed.)
Filing date
Publication date
Application filed by 重庆邮电大学 filed Critical 重庆邮电大学
Priority to KR1020217040187A priority Critical patent/KR102629950B1/ko
Priority to US17/423,484 priority patent/US11736408B2/en
Publication of WO2021227245A1 publication Critical patent/WO2021227245A1/zh

Links

Images

Classifications

    • HELECTRICITY
    • H04ELECTRIC COMMUNICATION TECHNIQUE
    • H04LTRANSMISSION OF DIGITAL INFORMATION, e.g. TELEGRAPHIC COMMUNICATION
    • H04L41/00Arrangements for maintenance, administration or management of data switching networks, e.g. of packet switching networks
    • H04L41/08Configuration management of networks or network elements
    • H04L41/0803Configuration setting
    • HELECTRICITY
    • H04ELECTRIC COMMUNICATION TECHNIQUE
    • H04LTRANSMISSION OF DIGITAL INFORMATION, e.g. TELEGRAPHIC COMMUNICATION
    • H04L47/00Traffic control in data switching networks
    • H04L47/10Flow control; Congestion control
    • H04L47/24Traffic characterised by specific attributes, e.g. priority or QoS
    • H04L47/2458Modification of priorities while in transit
    • HELECTRICITY
    • H04ELECTRIC COMMUNICATION TECHNIQUE
    • H04LTRANSMISSION OF DIGITAL INFORMATION, e.g. TELEGRAPHIC COMMUNICATION
    • H04L41/00Arrangements for maintenance, administration or management of data switching networks, e.g. of packet switching networks
    • H04L41/12Discovery or management of network topologies
    • HELECTRICITY
    • H04ELECTRIC COMMUNICATION TECHNIQUE
    • H04LTRANSMISSION OF DIGITAL INFORMATION, e.g. TELEGRAPHIC COMMUNICATION
    • H04L41/00Arrangements for maintenance, administration or management of data switching networks, e.g. of packet switching networks
    • H04L41/12Discovery or management of network topologies
    • H04L41/122Discovery or management of network topologies of virtualised topologies, e.g. software-defined networks [SDN] or network function virtualisation [NFV]
    • HELECTRICITY
    • H04ELECTRIC COMMUNICATION TECHNIQUE
    • H04LTRANSMISSION OF DIGITAL INFORMATION, e.g. TELEGRAPHIC COMMUNICATION
    • H04L43/00Arrangements for monitoring or testing data switching networks
    • H04L43/08Monitoring or testing based on specific metrics, e.g. QoS, energy consumption or environmental parameters
    • H04L43/0852Delays
    • HELECTRICITY
    • H04ELECTRIC COMMUNICATION TECHNIQUE
    • H04LTRANSMISSION OF DIGITAL INFORMATION, e.g. TELEGRAPHIC COMMUNICATION
    • H04L43/00Arrangements for monitoring or testing data switching networks
    • H04L43/08Monitoring or testing based on specific metrics, e.g. QoS, energy consumption or environmental parameters
    • H04L43/0876Network utilisation, e.g. volume of load or congestion level
    • HELECTRICITY
    • H04ELECTRIC COMMUNICATION TECHNIQUE
    • H04LTRANSMISSION OF DIGITAL INFORMATION, e.g. TELEGRAPHIC COMMUNICATION
    • H04L45/00Routing or path finding of packets in data switching networks
    • H04L45/14Routing performance; Theoretical aspects
    • HELECTRICITY
    • H04ELECTRIC COMMUNICATION TECHNIQUE
    • H04LTRANSMISSION OF DIGITAL INFORMATION, e.g. TELEGRAPHIC COMMUNICATION
    • H04L47/00Traffic control in data switching networks
    • H04L47/50Queue scheduling
    • H04L47/52Queue scheduling by attributing bandwidth to queues
    • H04L47/522Dynamic queue service slot or variable bandwidth allocation
    • HELECTRICITY
    • H04ELECTRIC COMMUNICATION TECHNIQUE
    • H04LTRANSMISSION OF DIGITAL INFORMATION, e.g. TELEGRAPHIC COMMUNICATION
    • H04L47/00Traffic control in data switching networks
    • H04L47/50Queue scheduling
    • H04L47/56Queue scheduling implementing delay-aware scheduling
    • HELECTRICITY
    • H04ELECTRIC COMMUNICATION TECHNIQUE
    • H04LTRANSMISSION OF DIGITAL INFORMATION, e.g. TELEGRAPHIC COMMUNICATION
    • H04L41/00Arrangements for maintenance, administration or management of data switching networks, e.g. of packet switching networks
    • H04L41/08Configuration management of networks or network elements
    • H04L41/0803Configuration setting
    • H04L41/0813Configuration setting characterised by the conditions triggering a change of settings
    • H04L41/082Configuration setting characterised by the conditions triggering a change of settings the condition being updates or upgrades of network functionality
    • HELECTRICITY
    • H04ELECTRIC COMMUNICATION TECHNIQUE
    • H04LTRANSMISSION OF DIGITAL INFORMATION, e.g. TELEGRAPHIC COMMUNICATION
    • H04L41/00Arrangements for maintenance, administration or management of data switching networks, e.g. of packet switching networks
    • H04L41/08Configuration management of networks or network elements
    • H04L41/0895Configuration of virtualised networks or elements, e.g. virtualised network function or OpenFlow elements
    • HELECTRICITY
    • H04ELECTRIC COMMUNICATION TECHNIQUE
    • H04LTRANSMISSION OF DIGITAL INFORMATION, e.g. TELEGRAPHIC COMMUNICATION
    • H04L41/00Arrangements for maintenance, administration or management of data switching networks, e.g. of packet switching networks
    • H04L41/08Configuration management of networks or network elements
    • H04L41/0896Bandwidth or capacity management, i.e. automatically increasing or decreasing capacities
    • HELECTRICITY
    • H04ELECTRIC COMMUNICATION TECHNIQUE
    • H04LTRANSMISSION OF DIGITAL INFORMATION, e.g. TELEGRAPHIC COMMUNICATION
    • H04L43/00Arrangements for monitoring or testing data switching networks
    • H04L43/20Arrangements for monitoring or testing data switching networks the monitoring system or the monitored elements being virtualised, abstracted or software-defined entities, e.g. SDN or NFV

Definitions

  • the invention belongs to the field of industrial Ethernet, and relates to an industrial heterogeneous network scheduling method oriented to the interconnection of TSN and non-TSN.
  • TSN Time Sensitive Networking, Time Sensitive Networking
  • the key requirement of TSN lies in the transmission of the "same" network, that is, the transmission of heterogeneous network data.
  • the existing TSN scheduling mechanism does not provide a solution.
  • the purpose of the present invention is to provide an industrial heterogeneous network scheduling method oriented to the interconnection of TSN and non-TSN.
  • the present invention provides the following technical solutions:
  • TSSDN Time Sensitive Software Defined Networking
  • the application plane includes a user and a terminal equipped with an application program for controlling the TSSDN controller, and the application program in the terminal is used by the user to configure the TSSDN controller according to their own application requirements;
  • the control plane is composed of a TSSDN controller.
  • the TSSDN controller is used to discover network topology, formulate scheduling strategies, and path management.
  • the TSSDN controller schedules data streams in the industrial heterogeneous network according to application requests from the application layer;
  • the forwarding plane includes an industrial network and a TSN network, and the industrial network includes any network that supports industrial network protocols, such as EtherCAT, Profinet, and Powerlink. According to whether it is TSN data, the entire industrial heterogeneous network is divided into non-TSN industrial networks, TSN industrial networks, and TSN industrial networks across non-TSN networks; the forwarding plane specifically includes SDN switches, TSN switches, industrial switches, industrial field devices, and TSN Field devices; among them, SDN switches, TSN switches, and industrial switches are used to implement data forwarding.
  • Industrial field devices refer to the devices responsible for sending and receiving data in the industrial network.
  • TSN field devices include TSN Talker and TSN Listener, where TSN Talker sends data , TSN Listener receives data.
  • the TSSDN controller is the core of the industrial heterogeneous network architecture and has a centralized management and control function; the TSSDN controller includes a path management module, a topology management module, and a policy management module. The TSSDN controller is based on application requests from the application layer. , To schedule data in industrial heterogeneous networks;
  • the topology management module is used to discover and update the network topology, and feed it back to the terminal application; therefore, the user can grasp the global topology of the industrial heterogeneous network through the terminal application.
  • the path management module is a response to the application requirements from users, including a series of path management operations such as link delay and queue bandwidth settings;
  • Strategy management module includes TSN module and SDN module, through strategy management to meet the delay requirements of time-sensitive applications in industrial heterogeneous network;
  • TSN industrial network data is centralized and unified management by CUC/CNC in TSN module, CNC has centralized computing Capabilities include calculating transmission scheduling, determining data paths, scheduling TSN data streams through time-aware shapers and frame preemption scheduling algorithms;
  • non-TSN industrial network data and TSN industrial network data across non-TSN networks are centralized by the SDN controller of the SDN module Unified management, dispatching of non-TSN industrial network data and TSN industrial network data across non-TSN networks by issuing flow tables to SDN switches.
  • the TSSDN controller classifies data according to delay requirements
  • the TSSDN controller manages the scheduling priority of the data
  • TSSDN controller calculates the shortest path
  • S6 SDN switch flow table matching: There are one or more flow tables in the SDN switch. When the data matches the flow table, it always matches from the first flow table along the pipeline. When the data matches the flow table entry successfully , Then update the counter and execute the corresponding instructions in the flow table, including forwarding and discarding instructions; when the flow table matching fails, it will be sent to the TSSDN controller in the form of a PacketIn message, and the TSSDN controller will analyze and make a decision;
  • the present invention provides a feedback adjustment mechanism for real-time monitoring of the status of each queue of the port, and when network congestion occurs, adjustments are made according to the congestion status to prevent data retention caused by queue congestion.
  • different adjustment strategies need to be made for different degrees of network congestion.
  • the standard for judging the degree of congestion of a queue in the present invention is queue delay.
  • S8 Feedback adjustment of queue congestion: When the queue is congested, the queue bandwidth is adjusted to different degrees according to the congestion situation. The congestion degree is judged based on the queue delay. According to the traffic situation, the queue congestion degree is divided into three levels: normal state , General congestion state and severe congestion state, the queue delay is compared with the corresponding maximum bandwidth value, when the delay is less than 70% of the maximum bandwidth value, the current queue state is defined as the normal state; when the delay is greater than 90% of the maximum bandwidth value, then The current queue status is defined as severely congested; when the queue delay is between 70% and 90% of the maximum bandwidth value, the current queue status is defined as a general congestion state.
  • step S1 the user implements topology information management and flow information management through the application in the terminal according to their own application requirements.
  • the flow information management includes setting the link delay, setting the maximum and minimum queue bandwidth, and viewing the current network data forwarding.
  • the user sets the link delay and bandwidth as required.
  • step S2 let the path from node p to node q be Path(p, q), any link in the network Li ⁇ Path(p, q), d i , j i , and BD i respectively denote the path of link i Delay, jitter and bandwidth, dl is the delay level, the data priority of the TSN network is OP, and ⁇ and ⁇ respectively represent the proportions of dl and OP in the scheduling priority reference coefficient ⁇ i.
  • the TSSDN controller classifies links according to the distribution of TSN nodes and non-TSN nodes, and is divided into three categories:
  • the link between the TSN node and the TSN node is represented by the set S1 ⁇ TSN ⁇ TSN ⁇ ;
  • the link between the TSN node and the non-TSN node is represented by the set S2 ⁇ TSN ⁇ non-TSN ⁇ ;
  • the link between a non-TSN node and a non-TSN node is represented by the set S3 ⁇ TSN ⁇ non-TSN ⁇ .
  • the data stream can be divided into 8 categories, namely: background stream, best effort, best effort, key application, video, audio, Internet control, network control ;
  • the TSSDN controller uses formulas (1), (2), and (3) according to the delay requirements of the link to compare and calculate the industrial heterogeneity
  • the data flow in the network is classified.
  • the data stream whose delay requirement meets formula (1) is non-real-time (NRT)
  • the data stream whose delay requirement meets formula (2) is real-time (RT)
  • the data flow is time-critical data (Time-critical, TC), and those without delay requirements are best effort (BE);
  • S2 ⁇ TSN ⁇ non-TSN ⁇ is divided into 4 categories: time-critical data, real-time data, non-real-time data, best-effort; represented by TC span , RT span , NRT span , and BE span ;
  • S3 ⁇ Non-TSN ⁇ Non-TSN ⁇ is divided into 4 categories: time-critical data, real-time data, non-real-time data, best-effort, which are represented by TC non , RT non , NRT non , and BE non-respectively ;
  • each category can be further subdivided into different data stream types according to delay requirements.
  • the data stream classification is shown in Table 1;
  • step S3 specifically includes:
  • the S1 ⁇ TSN ⁇ TSN ⁇ data stream is scheduled and managed by the TSN module in the TSSDN controller, and its priority is specified by IEEE802.1Q. According to IEEE802.1Q, the priority of the TSN data stream is in the PCP field. There are 8 priorities, which are 0,1,2,3,4,5,6,7. The priority of the S1 ⁇ TSN ⁇ TSN ⁇ data stream is represented by OP.
  • the delay level dl is proposed, which is divided into four levels according to the delay requirements.
  • the delay level dl corresponds to the data type one to one, and TSN ⁇ non-TSN
  • the delay level of the non-TSN ⁇ non-TSN link is denoted by dl.
  • the best-effort delay level dl is 1, the non-real-time data delay level dl is 2, the real-time data delay level dl is 3, and the time-critical data delay level dl is 4.
  • Table 2 The original priority and delay level of TSSDN's industrial heterogeneous network
  • the TSSDN controller does not calculate the scheduling priority reference coefficient for the data stream of S1 ⁇ TSN ⁇ TSN ⁇ , and the data stream of S2 ⁇ TSN ⁇ non-TSN ⁇ uses formula (4), S3 ⁇ non-TSN ⁇ non-TSN ⁇ Use formula (5) for the data flow data.
  • the values of ⁇ and ⁇ can be adjusted according to the low latency requirements and real-time requirements of data in different application environments.
  • Table 3 The relationship between scheduling priority reference coefficient and scheduling priority
  • the priority of the data stream of S1 ⁇ TSN ⁇ TSN ⁇ is specified by IEEE802.1Q, and the scheduling priority value and scheduling priority of the data stream of S2 ⁇ TSN ⁇ non-TSN ⁇ and S3 ⁇ non-TSN ⁇ non-TSN ⁇ Reference coefficient correlation. Then, the classification and scheduling priority of data flow in the industrial heterogeneous network of TSSDN are shown in Table 4.
  • step S4 it specifically includes:
  • the TSSDN controller After the TSSDN controller calculates the scheduling priority SP of a certain data flow, it is necessary to find a suitable path for the flow. If you only consider the number of hops of the route when selecting a route in the network, you only need to use the Dijkstra algorithm to select the route. However, with the continuous growth of network scale, network routing needs to consider more measurement factors, such as bandwidth, delay, jitter, packet loss rate and other constraints.
  • the CSPF algorithm is Dijkstra's algorithm with constraints.
  • the invention adopts an improved CSPF algorithm, that is, a multi-constrained multi-objective routing algorithm to select the best path of the network, so as to ensure the efficient use of network resources under a series of constraint conditions.
  • the multi-constrained multi-objective routing algorithm combines network status and business application requirements to provide network traffic with an optimal path that satisfies multiple constraints.
  • the multiple constraint conditions of the present invention include bandwidth, delay, and jitter, and the goal is that the selected path has the smallest delay and the largest available bandwidth.
  • the path from node p to node q is Path(p,q), any link in the network Li ⁇ Path(p,q), d i , j i , and BD i represent the delay, jitter, and bandwidth of link i, respectively.
  • the path is the shortest path; when there are multiple paths from node p to node q, the shortest path is calculated using the method of the present invention. Then each constraint attribute value of the path is expressed as:
  • d(Path(p,q)) represents the sum of delays on the path Path(p,q)
  • j(Path(p,q)) represents the sum of jitters on the path Path(p,q)
  • BD(Path( p,q)) represents the minimum bandwidth on the path Path(p,q).
  • the multi-constraint condition of the present invention is that the multi-target routing algorithm takes delay, jitter and bandwidth as the constraint conditions, and takes the minimum delay and the maximum available bandwidth as the target.
  • the specific formula is:
  • the objective function F(x) can be further expressed as:
  • dc, jc, and BDc represent the path delay constraint value, path jitter constraint value, and path bandwidth constraint value specified by the user.
  • min ⁇ D(Path(p,q)) ⁇ represents the minimum delay from node p to node q
  • max ⁇ AvailableBD(Path(p,q)) ⁇ represents the maximum available bandwidth from node p to node q.
  • the multi-constrained multi-objective routing algorithm solves the problem of target routing selection through the dictionary sorting method.
  • the dictionary sorting method is an analysis method for solving multi-objective optimization problems.
  • the core idea is to express the importance of the objective function by the order of the objective function .
  • the algorithm proposed in the present invention ranks the objective function f1(x) before f2(x), that is, in the process of calculating the path, the minimum delay is the primary goal, and the maximum available bandwidth is selected when there are multiple paths with the minimum delay. path of.
  • step S5 it specifically includes:
  • the present invention uses the TSSDN controller to divide the scheduling priorities of data streams from different field devices.
  • the 8 scheduling priorities correspond to the 8 queues one-to-one, as shown in Table 5. As shown; then according to the scheduling priority label of the data packet for scheduling in and out of the queue, and then arrive at the SDN switch in turn.
  • the characteristics of network topology information can be obtained by using the TSSDN controller, and the industrial data flow forwarding behavior of the SDN switch can be centrally controlled
  • the characteristic of the network is to perform bandwidth allocation and congestion control for large flows in the form of flow table distribution, so as to achieve the effect of improving the quality of network service.
  • the TSSDN controller passes this The method proposed by the invention calculates the scheduling priority SP of the flow, and maps the value of the scheduling priority SP to the corresponding priority code point (PCP) value.
  • PCP priority code point
  • the TSSDN controller issues a flow entry of "Mark scheduling priority and forward to destination port" to the switch, and then when all data flows of this type arrive at the switch, their PCP field is marked with the scheduling priority and forwarded to the corresponding outgoing port .
  • step S6 it specifically includes:
  • the present invention modifies the traditional flow table, mainly expanding the matching field, so that the SDN switch can match the TSN data packet more accurately.
  • Protocol type the network protocol used to mark the TSN network, such as 802.1AS, 802.1Qbu, 802.1Qbv, etc.
  • Source MAC address Mark the MAC address of the source of the TSN data.
  • Destination MAC address Mark the MAC address of the destination of the TSN data.
  • the present invention modifies the traditional flow table, mainly expanding the matching field, so that the SDN switch can match the TSN data packet more accurately.
  • Protocol type the network protocol used to mark the TSN network, such as 802.1AS, 802.1Qbu, 802.1Qbv, etc.
  • Source MAC address Mark the MAC address of the source of the TSN data.
  • Destination MAC address Mark the MAC address of the destination of the TSN data.
  • the SDN switch When the data flow passes through the SDN switch, the SDN switch first encapsulates the data into a PacketIn message and sends it to the TSSDN controller. According to the VLAN, source MAC address and destination MAC address in the matching field, the TSSDN controller allocates the TSN data flow to the TSN module. Dispatched and managed by the TSN module; the non-TSN data stream and the data stream of the TSN network across the non-TSN network are allocated to the SDN module, and the TSN module is dispatched and managed.
  • the TSSDN controller After the TSSDN controller finishes executing the scheduling algorithm according to the characteristics of the network traffic, it decides whether the data stream should be forwarded or discarded according to the scheduling priority and bandwidth reaching the SDN switch, and sends it to the SDN switch in the form of a flow table.
  • the SDN switch executes the command to decide the controller according to the matching situation of the flow table.
  • step S7 in order to more accurately reflect the current network congestion, the random early detection RED algorithm is used to calculate the average length of the queue, avgQ_k represents the average length of queue k, and q_k is the real-time length of queue k.
  • the enqueue rate and dequeue rate are calculated.
  • w is the weight, which is equivalent to the time constant, then the average length of the queue is:
  • a suitable w can balance the average queue length of the queue and avoid the problem of the average queue length jitter caused by data surge or short-term congestion of the queue. If w is too large, it will not be possible Filter out short-term congestion; if w is too small, it cannot reasonably reflect the congestion status.
  • the queue delay calculation formula is:
  • que_dealy_k represents the delay of queue k
  • now_rate_k represents the current rate of queue k, that is, the current bandwidth of the queue.
  • step S8 when the queue is congested, different degrees of queue bandwidth adjustment is performed according to the congestion situation. Judgment and classification of congestion is a key to data forwarding optimization. The congestion degree is judged based on the queue delay. According to the traffic situation, the congestion degree of the queue is divided into three levels: normal state, general congestion state and severe congestion state, as shown in Table 6.
  • the queue delay is compared with the corresponding maximum bandwidth value.
  • the current queue status is defined as normal; when the queue delay is greater than 90% of the maximum bandwidth value, the current queue status is defined as serious Congestion:
  • the current queue state is defined as a general congestion state.
  • the feedback adjustment mechanism will adjust the bandwidth according to the current queue priority and queue congestion.
  • the higher the priority of the eight queues the stronger the ability to seize the bandwidth of others when adjusting the bandwidth, and the lower the priority, the more likely it is that the bandwidth will be preempted. Larger, the worse the ability to guarantee network QoS;
  • the queues with higher priority only need to ensure that there is no serious congestion when preempting the bandwidth of the queues with lower priority, while the low-priority queues must ensure that the high-priority queues are always in a normal state when preempting bandwidth.
  • the beneficial effects of the present invention are that the present invention performs high-efficiency and high-quality scheduling in industrial heterogeneous networks where TSN and non-TSN are interconnected, and ensures low-delay and deterministic transmission of different types of data in the network.
  • Figure 1 is a flow chart of the industrial heterogeneous network scheduling method for the interconnection of TSN and non-TSN according to the present invention
  • FIG. 2 is a schematic diagram of an industrial heterogeneous network architecture based on TSSDN
  • FIG. 3 is a schematic diagram of the composition of the TSSDN controller
  • Figure 4 is a schematic diagram of an Ethernet frame with a VLAN tag
  • Figure 5 is an expanded structure diagram of the matching field
  • Figure 6 is a schematic diagram of the queue delay calculation process
  • Figure 7 is a schematic diagram of queue preemption of bandwidth
  • FIG. 8 is a diagram of an industrial heterogeneous network architecture according to a specific embodiment of the present invention.
  • the present invention provides an industrial heterogeneous network scheduling method oriented to the interconnection of TSN and non-TSN, which specifically includes the following steps:
  • Step 1 In the network deployment phase, users configure and manage the TSSDN controller according to application requirements through terminal applications.
  • Flow information management includes setting link delay, setting the maximum and minimum queue bandwidth, viewing the current network data forwarding situation, and setting according to needs.
  • Link delay and bandwidth After the user configuration is successful, there is a set of links, the delay demand of each link, the jitter demand of each link, the bandwidth of each link, the set of path constraints specified by the user, and each queue Six sets of bandwidth:
  • Step 2 The TSSDN controller classifies data according to the delay requirements.
  • d i , j i , and BD i represent the delay, jitter, and bandwidth of link i, respectively
  • dl is the delay level
  • the data priority of the TSN network is OP
  • ⁇ and ⁇ respectively represent the proportions of dl and OP in the scheduling priority reference coefficient ⁇ i.
  • the TSSDN controller classifies links according to the distribution of TSN nodes and non-TSN nodes, and is divided into 3 categories:
  • the link between the TSN node and the TSN node is represented by the set S1 ⁇ TSN ⁇ TSN ⁇ ;
  • the link between the TSN node and the non-TSN node is represented by the set S2 ⁇ TSN ⁇ non-TSN ⁇ ;
  • the link between a non-TSN node and a non-TSN node is represented by the set S3 ⁇ TSN ⁇ non-TSN ⁇ .
  • the data stream can be divided into 8 categories, namely: background stream, best effort, best effort, key application, video, audio, Internet control, network control ;
  • the TSSDN controller uses formulas (1), (2), and (3) according to the delay requirements of the link to compare and calculate the industrial heterogeneity
  • the data flow in the network is classified.
  • the data stream whose delay requirement meets formula (1) is non-real-time (NRT)
  • the data stream whose delay requirement meets formula (2) is real-time (RT)
  • the data flow is time-critical data (Time-critical, TC), and those without delay requirements are best effort (BE);
  • S2 ⁇ TSN ⁇ non-TSN ⁇ is divided into 4 categories: time-critical data, real-time data, non-real-time data, best-effort; represented by TC span , RT span , NRT span , and BE span ;
  • S3 ⁇ Non-TSN ⁇ Non-TSN ⁇ is divided into 4 categories: time-critical data, real-time data, non-real-time data, best-effort, which are represented by TC non , RT non , NRT non , and BE non-respectively ;
  • each category can be further subdivided into different data stream types according to delay requirements.
  • the data stream classification is shown in Table 1;
  • the third step The TSSDN controller manages the scheduling priority of the data.
  • the S1 ⁇ TSN ⁇ TSN ⁇ data stream is scheduled and managed by the TSN module in the TSSDN controller, and its priority is specified by IEEE802.1Q. According to IEEE802.1Q, the priority of the TSN data stream is in the PCP field. There are 8 priorities, which are 0,1,2,3,4,5,6,7. The priority of the S1 ⁇ TSN ⁇ TSN ⁇ data stream is represented by OP.
  • the delay level dl is proposed, which is divided into four levels according to the delay requirements.
  • the delay level dl corresponds to the data type one to one, and TSN ⁇ non-TSN
  • the delay level of the non-TSN ⁇ non-TSN link is denoted by dl.
  • the best-effort delay level dl is 1, the non-real-time data delay level dl is 2, the real-time data delay level dl is 3, and the time-critical data delay level dl is 4.
  • Table 2 The original priority and delay level of TSSDN's industrial heterogeneous network
  • the TSSDN controller does not calculate the scheduling priority reference coefficient for the data stream of S1 ⁇ TSN ⁇ TSN ⁇ , and the data stream of S2 ⁇ TSN ⁇ non-TSN ⁇ uses formula (4), S3 ⁇ non-TSN ⁇ non-TSN ⁇ Use formula (5) for the data flow data.
  • the values of ⁇ and ⁇ can be adjusted according to the low latency requirements and real-time requirements of data in different application environments.
  • Table 3 The relationship between scheduling priority reference coefficient and scheduling priority
  • the priority of the data stream of S1 ⁇ TSN ⁇ TSN ⁇ is specified by IEEE802.1Q, and the scheduling priority value and scheduling priority of the data stream of S2 ⁇ TSN ⁇ non-TSN ⁇ and S3 ⁇ non-TSN ⁇ non-TSN ⁇ Reference coefficient correlation. Then, the classification and scheduling priority of data flow in the industrial heterogeneous network of TSSDN are shown in Table 4.
  • Step 4 The TSSDN controller calculates the shortest path.
  • the TSSDN controller After the TSSDN controller calculates the scheduling priority SP of a certain data flow, it is necessary to find a suitable path for the flow. If you only consider the number of hops of the route when selecting a route in the network, you only need to use the Dijkstra algorithm to select the route. However, with the continuous growth of network scale, network routing needs to consider more measurement factors, such as bandwidth, delay, jitter, packet loss rate and other constraints.
  • the CSPF algorithm is Dijkstra's algorithm with constraints.
  • the invention adopts an improved CSPF algorithm, that is, a multi-constrained multi-objective routing algorithm to select the best path of the network, so as to ensure the efficient use of network resources under a series of constraint conditions.
  • the multi-constrained multi-objective routing algorithm combines network status and business application requirements to provide network traffic with an optimal path that satisfies multiple constraints.
  • the multiple constraint conditions of the present invention include bandwidth, delay, and jitter, and the goal is that the selected path has the smallest delay and the largest available bandwidth.
  • the path from node p to node q is Path(p,q), any link in the network Li ⁇ Path(p,q), d i , j i , and BD i represent the delay, jitter, and bandwidth of link i, respectively.
  • the path is the shortest path; when there are multiple paths from node p to node q, the shortest path is calculated using the method of the present invention. Then each constraint attribute value of the path is expressed as:
  • d(Path(p,q)) represents the sum of delays on the path Path(p,q)
  • j(Path(p,q)) represents the sum of jitters on the path Path(p,q)
  • BD(Path( p,q)) represents the minimum bandwidth on the path Path(p,q).
  • the multi-constraint condition of the present invention is that the multi-target routing algorithm takes delay, jitter and bandwidth as the constraint conditions, and takes the minimum delay and the maximum available bandwidth as the target.
  • the specific formula is:
  • the objective function F(x) can be further expressed as:
  • dc, jc, and BDc represent the path delay constraint value, path jitter constraint value, and path bandwidth constraint value specified by the user.
  • min ⁇ D(Path(p,q)) ⁇ represents the minimum delay from node p to node q
  • max ⁇ AvailableBD(Path(p,q)) ⁇ represents the maximum available bandwidth from node p to node q.
  • the multi-constrained multi-objective routing algorithm solves the problem of target routing selection through the dictionary sorting method.
  • the dictionary sorting method is an analysis method for solving multi-objective optimization problems.
  • the core idea is to express the importance of the objective function by the order of the objective function .
  • the algorithm proposed in the present invention ranks the objective function f1(x) before f2(x), that is, in the process of calculating the path, the minimum delay is the primary goal, and the maximum available bandwidth is selected when there are multiple paths with the minimum delay. path of.
  • Step 5 Marking of scheduling priority
  • the present invention uses the TSSDN controller to divide the scheduling priorities of the data streams from different field devices.
  • the 8 scheduling priorities correspond to the 8 queues one-to-one, as shown in Table 6.
  • the characteristics of network topology information can be obtained by using the TSSDN controller, and the industrial data flow forwarding behavior of the SDN switch can be centrally controlled
  • the characteristics of the flow table are used to distribute bandwidth and congestion control for large flows to achieve the effect of improving network service quality, as shown in Table 5.
  • the TSSDN controller passes this The method proposed by the invention calculates the scheduling priority SP of the flow, and maps the value of the scheduling priority SP to the corresponding priority code point (PCP) value.
  • PCP priority code point
  • the TSSDN controller issues a flow entry of "Mark scheduling priority and forward to destination port" to the switch, and then when all data flows of this type arrive at the switch, their PCP field is marked with the scheduling priority and forwarded to the corresponding outgoing port .
  • Step 6 SDN switch flow table matching
  • the present invention modifies the traditional flow table, mainly expanding the matching field, so that the SDN switch can match the TSN data packet more accurately.
  • Protocol type the network protocol used to mark the TSN network, such as 802.1AS, 802.1Qbu, 802.1Qbv, etc.
  • Source MAC address Mark the MAC address of the source of the TSN data.
  • Destination MAC address Mark the MAC address of the destination of the TSN data.
  • the present invention modifies the traditional flow table, mainly expanding the matching field, so that the SDN switch can match the TSN data packet more accurately.
  • Protocol type the network protocol used to mark the TSN network, such as 802.1AS, 802.1Qbu, 802.1Qbv, etc.
  • Source MAC address Mark the MAC address of the source of the TSN data.
  • Destination MAC address Mark the MAC address of the destination of the TSN data.
  • the SDN switch When the data flow passes through the SDN switch, the SDN switch first encapsulates the data into a PacketIn message and sends it to the TSSDN controller. According to the VLAN, source MAC address and destination MAC address in the matching field, the TSSDN controller allocates the TSN data flow to the TSN module. Dispatched and managed by the TSN module; the non-TSN data stream and the data stream of the TSN network across the non-TSN network are allocated to the SDN module, and the TSN module is dispatched and managed.
  • the TSSDN controller After the TSSDN controller finishes executing the scheduling algorithm according to the characteristics of the network traffic, it decides whether the data stream should be forwarded or discarded according to the scheduling priority and bandwidth reaching the SDN switch, and sends it to the SDN switch in the form of a flow table.
  • the SDN switch executes the command to decide the controller according to the matching situation of the flow table.
  • Step 7 Judgment of queue congestion
  • the present invention designs a feedback adjustment mechanism to monitor the status of each queue of the port in real time. When network congestion occurs, adjustment is made according to the congestion status to prevent data retention caused by queue congestion. In order to improve the possible malicious packet loss during the data forwarding process, different adjustment strategies need to be made for different degrees of network congestion.
  • the standard for judging the degree of congestion of a queue in the present invention is queue delay.
  • the random early detection RED algorithm is used to calculate the average length of the queue.
  • AvgQ_k represents the average length of queue k
  • q_k is the real-time length of queue k.
  • the team velocity is calculated.
  • w is the weight, which is equivalent to the time constant, then the average length of the queue is:
  • a suitable w can balance the average queue length of the queue and avoid the problem of the average queue length jitter caused by data surge or short-term congestion of the queue. If w is too large, it will not be possible Filter out short-term congestion; if w is too small, it cannot reasonably reflect the congestion status.
  • the queue delay calculation formula is:
  • que_dealy_k represents the delay of queue k
  • now_rate_k represents the current rate of queue k, that is, the current bandwidth of the queue.
  • Step 8 Feedback adjustment of queue congestion
  • the data forwarding optimization of the present invention is mainly concentrated on the queue of the SDN switch port.
  • the queue bandwidth is adjusted to different degrees according to the congestion situation. Judgment and classification of congestion is a key to data forwarding optimization.
  • the congestion degree is judged based on the queue delay. According to the traffic situation, the congestion degree of the queue is divided into three levels: normal state, general congestion state and severe congestion state.
  • the queue delay is compared with the corresponding maximum bandwidth value.
  • the current queue status is defined as normal; when the delay is greater than 90% of the maximum bandwidth value, the current queue status is defined as severely congested;
  • the current queue state is defined as a general congestion state.
  • the feedback adjustment mechanism will adjust the bandwidth according to the current queue priority and queue congestion. On the whole, the higher the priority of the eight queues, the stronger the ability to preempt others' bandwidth when adjusting the bandwidth, the lower the priority, the greater the possibility of bandwidth being preempted, and the worse the ability to ensure network QoS.
  • the queues with higher priority only need to ensure that there is no serious congestion when preempting the bandwidth of the queues with lower priority, while the low-priority queues must ensure that the high-priority queues are always in a normal state when preempting bandwidth. .
  • the three links are denoted as L1, L2, and L3 in sequence.
  • L1 belongs to S1 ⁇ TSN ⁇ TSN ⁇
  • L2 belongs to S2 ⁇ TSN ⁇ non-TSN ⁇
  • L3 belongs to S3 ⁇ non-TSN ⁇ non-TSN ⁇ , each chain
  • the data types of roads are shown in the table below.
  • the level is divided.
  • TSN link does not need to be divided into delay levels
  • Priority is 4, scheduled and managed by the TSN module of the TSSDN controller;
  • L1 has only one path that does not need to be calculated, and the existing path is the shortest path
  • L2 has only one path that does not need to be calculated, and the existing path is the shortest path
  • L3 has multiple paths and needs to calculate the traffic generator 2 The shortest path to the traffic receiver 2.
  • delay is an important reference for constraint conditions, so the TSSDN controller selects Path1 as the transmission path of the traffic generator 2.
  • the scheduling priority corresponds to the queue one by one.
  • the TSSDN controller When the data flow of the subnet in the heterogeneous network of TSSDN arrives at the switch for the first time, because there is no flow entry in the switch that matches the data flow, the data flow will be packaged into a PacketIn message and sent to the TSSDN controller; the TSSDN controller will schedule The value of the priority SP is mapped to the corresponding priority code point (PCP) value.
  • PCP priority code point
  • the TSSDN controller issues a flow entry of "Mark scheduling priority and forward to destination port" to the switch, and then when all data flows of this type arrive at the switch, their PCP field is marked with the scheduling priority and forwarded to the corresponding outgoing port .
  • the TSSDN controller generates a flow table and sends it to the SDN switch.
  • the flow table is matched.

Landscapes

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

Abstract

一种面向TSN和非TSN互联的工业异构网络调度方法,属于工业以太网领域,包括以下步骤:网络部署阶段用户通过终端应用程序,根据应用需求对TSSDN控制器进行配置管理;TSSDN控制器按照延迟需求进行数据分类;TSSDN控制器对数据进行调度优先级的管理;TSSDN控制器计算最短路径;调度优先级的标记;SDN交换机流表匹配;判断队列拥塞情况;对队列拥塞反馈调整。其在TSN和非TSN互联的工业异构网络中进行高效率、高质量的调度,保障了网络中不同类型数据的低延迟、确定性传输。

Description

一种面向TSN和非TSN互联的工业异构网络调度方法 技术领域
本发明属于工业以太网领域,涉及一种面向TSN和非TSN互联的工业异构网络调度方法。
背景技术
工业以太网在任务关键型工业应用和汽车控制等应用中获得越来越多的认可,但是随着工厂业务的需要,特别是当前大数据和云计算等进入工业控制领域、要求IT和OT融合的背景下,不仅要保证大数据传输,而且要保证传输的实时性和确定性,这时现有的工业以太网协议就显得更力不从心。并且当网络中存在大量的不同类型的数据时,如何进行合理的调度,保证数据传输的QoS是一个亟需解决的问题。
正是工业以太网存在的这些瓶颈,才催生了TSN网络的产生。TSN(Time Sensitive Networking,时间敏感网络)是在IEEE802.1标准框架下,基于特定应用需求制定的一组“子标准”,旨在为以太网协议建立“通用”的时间敏感机制,以确保网络数据传输的实时性、确定性和低延迟。TSN的关键诉求在于“同一”网络的传输,即异构网络数据的传输。
然而在异构网络中如何执行高效率、高质量的调度,保障网络中不同类型数据的低延迟、确定性传输,现有的TSN调度机制没有提供解决方案。
发明内容
有鉴于此,本发明的目的在于提供一种面向TSN和非TSN互联的工业异构网络调度方法。
为达到上述目的,本发明提供如下技术方案:
一种面向TSN和非TSN互联的工业异构网络调度方法,基于时间敏感的软件定义网络(Time Sensitive Software Defined Networking,TSSDN)的工业异构网络架构,包括依次连接的应用平面、控制平面和转发平面;
所述应用平面包括用户和装有控制TSSDN控制器应用程序的终端,终端中的应用程序用于用户根据自己的应用需求对TSSDN控制器进行相应的配置;
所述控制平面由TSSDN控制器构成,TSSDN控制器用于发现网络拓扑、制定调度策略、路径管理,TSSDN控制器根据来自应用层的应用请求,对工业异构网络中的数据流进行调度;
所述转发平面包括工业网络和TSN网络,所述工业网络包含任意个支持工业网络协议的 网络,如EtherCAT、Profinet和Power link等。按照是否为TSN数据,整个工业异构网络分为非TSN工业网络、TSN工业网络以及TSN工业网络跨非TSN网络;所述转发平面具体包括SDN交换机、TSN交换机、工业交换机、工业现场设备、TSN现场设备;其中SDN交换机、TSN交换机、工业交换机用于实现数据转发,工业现场设备是指在工业网络中负责发送和接收数据的设备,TSN现场设备包括TSN Talker和TSN Listener,其中TSN Talker发送数据,TSN Listener接收数据。
进一步,所述TSSDN控制器是工业异构网络架构的核心,具有集中式管理控制功能;所述TSSDN控制器包括路径管理模块、拓扑管理模块和策略管理模块TSSDN控制器根据来自应用层的应用请求,对工业异构网络中的数据进行调度;
拓扑管理模块用于发现网络拓扑和更新网络拓扑,并反馈给终端应用程序;因此用户通过终端应用程序,可以掌握工业异构网络的全局拓扑。
路径管理模块是对来自用户的应用需求的响应,包括链路延迟、队列带宽设置的一系列路径管理操作;
策略管理模块包括TSN模块和SDN模块,通过策略管理来满足工业异构网络中的时间敏感应用的延迟要求;TSN工业网络数据由TSN模块中的CUC/CNC进行集中统一管理,CNC具有集中计算的能力,包括计算传输调度,确定数据路径,通过时间感知整形器、帧抢占调度算法对TSN数据流调度;非TSN工业网络数据以及TSN工业网络跨非TSN网络数据由SDN模块的SDN控制器进行集中统一管理,通过下发流表到SDN交换机对非TSN工业网络数据以及TSN工业网络跨非TSN网络数据调度。
进一步,具体包括以下步骤:
S1:网络部署阶段用户通过终端应用程序,根据应用需求对TSSDN控制器进行配置管理;
S2:TSSDN控制器按照延迟需求进行数据分类;
S3:TSSDN控制器对数据进行调度优先级的管理;
S4:TSSDN控制器计算最短路径;
S5:调度优先级的标记;
S6:SDN交换机流表匹配:SDN交换机中有一个或多个流表,数据与流表进行匹配时,总是沿着流水线从第一个流表开始匹配,当数据与流表项匹配成功时,则更新计数器并执行流表中的相应指令,包括转发、丢弃指令;当流表匹配失败时,则以PacketIn消息的形式发送至TSSDN控制器,并由TSSDN控制器分析并做出决定;
S7:判断队列拥塞情况:在数据转发的过程中可能出现数据量增加导致当前网络环境变差的情况,如果不能及时的调整转发的流量,那么将会出现比较严重的丢包现象,其中一些重要的、时间敏感的数据有可能会在这个过程中丢失,无法到达指定的接收端。因此,本发明提供一种反馈调节机制实时监控端口各队列的状态,当网络出现拥塞时,根据拥堵状况做出调整,防止队列拥塞导致的数据滞留。为了改善数据转发过程可能出现的恶意丢包,需要针对不同程度的网络拥塞情况做不同的调整策略。本发明判断一个队列的拥塞程度的标准是队列延迟。
S8:对队列拥塞反馈调整:当队列出现拥塞,则根据拥塞情况进行不同程度的队列带宽调整,拥塞程度的评判依据是队列延迟,根据流量情况将队列的拥塞程度分成了三个等级:正常状态、一般拥塞状态与严重拥塞状态,队列延迟与对应最大带宽值进行比较,当延迟小于最大带宽值的70%,则定义当前队列的状态为正常状态;当延迟大于最大带宽值的90%,则定义当前队列的状态为严重拥塞;当队列延迟处于最大带宽值的70%与90%之间,则定义当前的队列状态为一般拥塞状态。
进一步,步骤S1中,用户根据自己的应用需求通过终端中的应用程序实现拓扑信息管理、流量信息管理,其中流量信息管理包括设置链路延迟、设置队列带宽最大和最小值、查看当前网络数据转发情况,用户按需设置链路延迟和带宽,用户配置成功后,存在链路集合、每条链路的延迟需求、每条链路的抖动需求、每条链路的带宽、用户指定的路径约束条件集合、每个队列的带宽六个集合:
链路集合{L 1…..L i…};
每条链路的延迟需求{d 1…d i.…};
每条链路的抖动需求{j 1…j i.…};
每条链路的带宽{BD 1…BD i.…};
用户指定的路径约束条件集合{d c、j c、BD c};
队列k的最大和最小带宽{max_bw_k、min_bw_k}。
进一步,步骤S2中,设节点p到节点q的路径为Path(p,q),网络中任意链路Li∈Path(p,q),d i、j i、BD i分别表示链路i的延迟、抖动和带宽,dl为延迟等级,TSN网络的数据优先级为OP,α和β分别表示调度优先级参考系数Θ i中dl、OP所占的比重。TSSDN控制器根据TSN节点与非TSN节点的分布情况对链路进行分类,共分成3类:
TSN节点与TSN节点之间的链路,用集合S1{TSN←→TSN}表示;
TSN节点与非TSN节点之间的链路,用集合S2{TSN←→非TSN}表示;
非TSN节点与非TSN节点之间的链路,用集合S3{TSN←→非TSN}表示。
针对S1{TSN←→TSN},可根据IEEE 802.1Q的规定,将数据流分为8类,分别是:背景流、尽力而为、最大努力、关键应用、视频、音频、互联网控制、网络控制;
针对S2{TSN←→非TSN}、S3{非TSN←→非TSN},TSSDN控制器根据链路的延迟需求使用公式(1)、(2)、(3)通过比较计算,将工业异构网络中的数据流进行分类。延迟需求满足公式(1)的数据流为非实时数据(non-real-time,NRT),延迟需求满足公式(2)的数据流为实时数据(real-time,RT),延迟需求满足公式(3)的数据流为时间关键数据(Time-critical,TC),没有延迟要求的属于尽力而为(best effort,BE);
d i≥100ms       (1)
1ms<d i<100ms      (2)
d i<1ms       (3)
因此S2{TSN←→非TSN}分为4类:时间关键数据、实时数据、非实时数据、尽力而为;用TC 、RT 、NRT 、BE 表示;
S3{非TSN←→非TSN}分为4类:时间关键数据、实时数据、非实时数据、尽力而为,分别用TC 、RT 、NRT 、BE 表示;
TSSDN的工业异构网络中共有若干条链路,根据TSN节点与非TSN节点的分布情况将链路分成集合S1{TSN←→TSN}、S2{TSN←→非TSN}、S3{非TSN←→非TSN}三类,每一类按照延迟需求可继续细分为不同的数据流类型,数据流分类如表1所示;
表1工业异构网络数据流分类
Figure PCTCN2020102199-appb-000001
Figure PCTCN2020102199-appb-000002
进一步,步骤S3具体包括:
S1{TSN←→TSN}数据流由TSSDN控制器中的TSN模块进行调度管理,其优先级由IEEE802.1Q规定。根据IEEE802.1Q规定,TSN数据流的优先级在PCP字段,共有8个优先级,分别是0,1,2,3,4,5,6,7。S1{TSN←→TSN}数据流的优先级用OP表示。
针对S2{TSN←→非TSN}、S3{非TSN←→非TSN},提出延迟等级dl,既按照延迟需求划分为四个等级,延迟等级dl与数据类型一一对应,TSN←→非TSN类、非TSN←→非TSN类链路的延迟等级用dl表示。尽力而为的延迟等级dl为1,非实时数据延迟等级dl为2,实时数据延迟等级dl为3,时间关键数据延迟等级dl为4。
表2 TSSDN的工业异构网络的原有优先级与延迟等级
Figure PCTCN2020102199-appb-000003
Figure PCTCN2020102199-appb-000004
设链路i的调度优先级参考系数Θ i为:
Figure PCTCN2020102199-appb-000005
其中,α和β分别表示调度优先级参考系数Θ i中dl、OP所占的比重,且α+β=1。TSSDN控制器对S1{TSN←→TSN}的数据流不计算其调度优先级参考系数,S2{TSN←→非TSN}的数据流则使用公式(4),S3{非TSN←→非TSN}的数流据则使用公式(5)。
α和β的值可以根据不同应用环境下对数据低延迟要求和实时性的需求程度调整。本发明综合考虑数据低延迟要求与确定性,令α=0.4,β=0.6,则异构网络的调度优先级如表3所示。最低优先级为1,最高优先级为6。
表3调度优先级参考系数与调度优先级的关系
Θ i 调度优先级SP
0<Θ i<=0.5 1
0.5<Θ i<=1.0 2
1<Θ i<=1.3 3
1.3<Θ i<=1.6 4
1.6<Θ i<=1.9 5
Θ i>1.9 6
S1{TSN←→TSN}的数据流的优先级由IEEE802.1Q规定,S2{TSN←→非TSN}、S3{非TSN←→非TSN}的数据流的调度优先级取值与调度优先级参考系数相关。则TSSDN的工业异构网络中数据流的分类与调度优先级如表4所示。
表4调度优先级
Figure PCTCN2020102199-appb-000006
进一步,步骤S4中,具体包括:
TSSDN控制器计算出某一数据流的调度优先级SP后,有必要为该流寻找合适的路径。在网络中进行路由选择的时候如果只考虑到路由的跳数,那只需要使用Dijkstra算法来选择路由。但是,随着网络规模的不断增长,网络的路由需要考虑更多的度量因素,如带宽、延迟、抖动、丢包率等的约束。CSPF算法是带有约束条件的Dijkstra算法。本发明采用改进的CSPF算法即多约束多目标路由选择算法来选择网络的最佳路径,以此来保障在一系列的约束条件下,网络资源的高效利用。
多约束多目标路由选择算法结合网络状态和业务应用需求为网络流量提供提一条满足多约束条件的最优路径。本发明的多约束条件包括带宽、延迟、抖动,目标是选出的路径延迟最小、可用带宽最大。
设节点p到节点q的路径为Path(p,q),网络中任意链路Li∈Path(p,q),d i、j i、BD i分别表示链路i的延迟、抖动和带宽,当节点p到节点q有且仅有一条路径时,该路径即为最短路径;当节点p到节点q有存在多条路径时,使用本发明所提方法计算最短路径。则路径的各个约束属性值表示为:
d(Path(p,q))=∑ Li∈Path(p,q)d i     (6)
j(Path(p,q))=∑ Li∈Path(p,q)j i     (7)
BD(Path(p,q))=min{BD i,Li∈Path(p,q)}   (8)
d(Path(p,q))代表路径Path(p,q)上的延迟之和,j(Path(p,q))代表路径Path(p,q)上的抖动之和、BD(Path(p,q))代表路径Path(p,q)上的最小带宽。
本发明的多约束条件为多目标路由算法以延迟、抖动和带宽为约束条件,以最小延迟和最大可用带宽为目标,具体公式为:
Figure PCTCN2020102199-appb-000007
其中目标函数F(x)可进一步表示为:
Figure PCTCN2020102199-appb-000008
dc、jc、BDc表示的是用户指定的路径延迟约束值、路径抖动约束值和路径带宽约束值。min{D(Path(p,q))}表示节点p到节点q的最小延迟,max{AvailableBD(Path(p,q))}表示节点p到节点q的最大可用带宽。
多约束多目标路由选择算法通过字典排序方法来解决目标路由的选择问题,字典排序方法是一种求解多目标优化问题的分析方法,核心思想是用目标函数的先后顺序来表达目标函数的重要程度。本发明所提算法是将目标函数f1(x)排在f2(x)的前面,即在计算路径的过程中,把最小延迟作为首要目标,当出现多条最小延迟的路径时选择最大可用带宽的路径。
进一步,步骤S5中,具体包括:
为满足不同调度优先级的数据流的服务质量,本发明利用TSSDN控制器,对来自不同现场设备的数据流进行调度优先级划分,8个调度优先级与8个队列一一对应,如表5所示;然后根据数据包的调度优先级标签进行出入队列的调度,依次到达SDN交换机。当某一调度优先级的数据流突然增加而导致TSSDN的工业异构网络出现拥塞时,通过利用TSSDN控制器可以获取网络拓扑信息的特性,以及能够对SDN交换机的工业数据流转发行为进行集中控制的特点,以流表下发的形式来对大流进行带宽分配和拥塞控制,达到提升网络服务质量的 效果。
表5调度优先级与队列的映射
调度优先级SP 对应队列
0 0
1 1
2 2
3 3
4 4
5 5
6 6
7 7
当TSSDN的异构网络内子网的数据流首次到达交换机时,由于交换机内没有与该数据流相匹配的流表项,数据流将打包成PacketIn消息被发送至TSSDN控制器;TSSDN控制器通过本发明所提方法计算出该流的调度优先级SP,并将调度优先级SP的数值映射成相应的优先级代码点(PCP)值。由TSSDN控制器向交换机下发“标记调度优先级并转发至目的端口”流表项,然后所有该类数据流到达交换机时,其PCP字段均被标记调度优先级,并转发至相应的出端口。
进一步,步骤S6中,具体包括:
为了提高TSSDN控制器对TSN数据流的识别的准确性,本发明对传统的流表进行了修改,主要是对匹配字段进行了拓展,使SDN交换机能够更精确的匹配TSN的数据包。
协议类型:用于标记TSN网络的网络协议,如802.1AS、802.1Qbu、802.1Qbv等。
源MAC地址:标记TSN数据来源的MAC地址。
目的MAC地址:标记TSN数据目的地的MAC地址。
为了提高TSSDN控制器对TSN数据流的识别的准确性,本发明对传统的流表进行了修改,主要是对匹配字段进行了拓展,使SDN交换机能够更精确的匹配TSN的数据包。
协议类型:用于标记TSN网络的网络协议,如802.1AS、802.1Qbu、802.1Qbv等。
源MAC地址:标记TSN数据来源的MAC地址。
目的MAC地址:标记TSN数据目的地的MAC地址。
当数据流经过SDN交换机时,SDN交换机先把数据封装成PacketIn消息,发送至TSSDN控制器,根据匹配字段中VLAN、源MAC地址和目的MAC地址,TSSDN控制器把TSN数据流分配至TSN模块,由TSN模块调度管理;把非TSN数据流和TSN网络跨非TSN网络 的数据流分配至SDN模块,由TSN模块调度管理。待TSSDN控制器根据网络流量特征执行完毕调度算法,根据到达SDN交换机的调度优先级以及带宽大小来决策该数据流是否转发或者丢弃,以流表的形式下发至SDN交换机。SDN交换机根据流表匹配情况执行决策该控制器的命令。
进一步,步骤S7中,为更准确的反映当前网络拥塞情况,采用随机早期检测RED算法来计算队列的平均长度,avgQ_k代表的是队列k的平均长度,q_k为队列k的实时长度,由队列k的入队速率与出队速率计算得到。w为权值,相当于时间常数,则队列的平均长度为:
avgQk=(1-w)*avgQk+q*w  (7)
q_k=in_rate-out_rate  (8)
在随机早期检测算法中,设置一个合理的权值很重要,合适的w能够平衡队列的平均队列长度,规避因为数据激增或队列短暂拥塞导致的平均队列长度抖动的问题,如果w太大将会无法过滤掉短暂的拥塞;如果w太小则不能合理地反映拥塞状况。
为了计算队列的延迟,需要获取队列当前的速率,用户通过终端应用程序为每个队列设置最大最小带宽,当最大值与最小值不相同时,数据会在规定的速率范围内进行传输;而当最大值设定成等于最小值时,数据在队列中进行传输时会尽量保持在所设定的速率附近,以此获取队列的当前速率,则队列延迟计算公式为:
que_dealy_k=avgQk/now_rate_k      (9)
que_dealy_k表示队列k的延迟,now_rate_k表示队列k的当前速率,即队列的当前带宽。
进一步,步骤S8中,当队列出现拥塞,则根据拥塞情况进行不同程度的队列带宽调整。拥塞情况的判断与分级是数据转发优化的一个关键。拥塞程度的评判依据是队列延迟,根据流量情况将队列的拥塞程度分成了三个等级:正常状态、一般拥塞状态与严重拥塞状态,如表6所示。
表6拥塞等级
队列延迟与最大带宽的关系 拥塞等级
que_dealy<70%max_bd 正常状态
70%<que_dealy<90%max_bd 一般拥塞状态
que_dealy>90%max_bd 严重拥塞状态
队列延迟与对应最大带宽值进行比较,当队列延迟小于最大带宽值的70%,则定义当前队列的状态为正常状态;当队列延迟大于最大带宽值的90%,则定义当前队列的状态为严重 拥塞;当队列延迟处于最大带宽值的70%与90%之间,则定义当前的队列状态为一般拥塞状态。
反馈调整机制会根据当前的队列优先级和队列拥塞程度进行带宽的调整,八个队列中优先级越高,调整带宽时抢占别人带宽的能力越强,优先级越低带宽被抢占的可能性越大,保证网络QoS的能力也越差;
队列抢占带宽的原则:
(1)优先级较高的队列在抢占优先级较低的队列带宽时只要保证其不出现严重拥塞情况即可,而低优先级队列在抢占带宽时则必须保证高优先级队列一直处于正常状态;
(2)当出现不同的优先级队列要抢占队列带宽时,遵循先高后低原则,即当高优先级队列先抢占完成后,再开始低优先级队列的抢占。这种做法可以确保高优先级的数据的优先传输,以保障其时间敏感性。
本发明的有益效果在于:本发明在TSN和非TSN互联的工业异构网络中进行高效率、高质量的调度,保障了网络中不同类型数据的低延迟、确定性传输。
本发明的其他优点、目标和特征在某种程度上将在随后的说明书中进行阐述,并且在某种程度上,基于对下文的考察研究对本领域技术人员而言将是显而易见的,或者可以从本发明的实践中得到教导。本发明的目标和其他优点可以通过下面的说明书来实现和获得。
附图说明
为了使本发明的目的、技术方案和优点更加清楚,下面将结合附图对本发明作优选的详细描述,其中:
图1为本发明所述面向TSN和非TSN互联的工业异构网络调度方法流程图;
图2为基于TSSDN的工业异构网络架构示意图;
图3为TSSDN控制器的组成示意图;
图4为具有VLAN标签的以太网帧示意图;
图5为匹配字段的拓展结构图;
图6为队列延迟计算流程示意图;
图7为队列抢占带宽示意图;
图8为本发明具体实施例的工业异构网络架构图。
具体实施方式
以下通过特定的具体实例说明本发明的实施方式,本领域技术人员可由本说明书所揭露 的内容轻易地了解本发明的其他优点与功效。本发明还可以通过另外不同的具体实施方式加以实施或应用,本说明书中的各项细节也可以基于不同观点与应用,在没有背离本发明的精神下进行各种修饰或改变。需要说明的是,以下实施例中所提供的图示仅以示意方式说明本发明的基本构想,在不冲突的情况下,以下实施例及实施例中的特征可以相互组合。
其中,附图仅用于示例性说明,表示的仅是示意图,而非实物图,不能理解为对本发明的限制;为了更好地说明本发明的实施例,附图某些部件会有省略、放大或缩小,并不代表实际产品的尺寸;对本领域技术人员来说,附图中某些公知结构及其说明可能省略是可以理解的。
本发明实施例的附图中相同或相似的标号对应相同或相似的部件;在本发明的描述中,需要理解的是,若有术语“上”、“下”、“左”、“右”、“前”、“后”等指示的方位或位置关系为基于附图所示的方位或位置关系,仅是为了便于描述本发明和简化描述,而不是指示或暗示所指的装置或元件必须具有特定的方位、以特定的方位构造和操作,因此附图中描述位置关系的用语仅用于示例性说明,不能理解为对本发明的限制,对于本领域的普通技术人员而言,可以根据具体情况理解上述术语的具体含义。
如图1-图7所示,本发明提供一种面向TSN和非TSN互联的工业异构网络调度方法,具体包括以下步骤:
第一步:网络部署阶段用户通过终端应用程序,根据应用需求对TSSDN控制器进行配置管理。
用户根据自己的应用需求通过终端中的应用程序实现拓扑信息管理、流量信息管理,其中流量信息管理包括设置链路延迟、设置队列带宽最大和最小值、查看当前网络数据转发情况,用户按需设置链路延迟和带宽,用户配置成功后,存在链路集合、每条链路的延迟需求、每条链路的抖动需求、每条链路的带宽、用户指定的路径约束条件集合、每个队列的带宽六个集合:
链路集合{L 1…..L i…};
每条链路的延迟需求{d 1…d i.…};
每条链路的抖动需求{j 1…j i.…};
每条链路的带宽{BD 1…BD i.…};
用户指定的路径约束条件集合{d c、j c、BD c};
队列k的最大和最小带宽{max_bw_k、min_bw_k}。
第二步:TSSDN控制器按照延迟需求进行数据分类。
设节点p到节点q的路径为Path(p,q),网络中任意链路Li∈Path(p,q),d i、j i、BD i分别表示链路i的延迟、抖动和带宽,dl为延迟等级,TSN网络的数据优先级为OP,α和β分别表示调度优先级参考系数Θ i中dl、OP所占的比重。TSSDN控制器根据TSN节点与非TSN节点的分布情况对链路进行分类,共分成3类:
TSN节点与TSN节点之间的链路,用集合S1{TSN←→TSN}表示;
TSN节点与非TSN节点之间的链路,用集合S2{TSN←→非TSN}表示;
非TSN节点与非TSN节点之间的链路,用集合S3{TSN←→非TSN}表示。
针对S1{TSN←→TSN},可根据IEEE 802.1Q的规定,将数据流分为8类,分别是:背景流、尽力而为、最大努力、关键应用、视频、音频、互联网控制、网络控制;
针对S2{TSN←→非TSN}、S3{非TSN←→非TSN},TSSDN控制器根据链路的延迟需求使用公式(1)、(2)、(3)通过比较计算,将工业异构网络中的数据流进行分类。延迟需求满足公式(1)的数据流为非实时数据(non-real-time,NRT),延迟需求满足公式(2)的数据流为实时数据(real-time,RT),延迟需求满足公式(3)的数据流为时间关键数据(Time-critical,TC),没有延迟要求的属于尽力而为(best effort,BE);
d i≥100ms       (1)
1ms<d i<100ms       (2)
d i<1ms         (3)
因此S2{TSN←→非TSN}分为4类:时间关键数据、实时数据、非实时数据、尽力而为;用TC 、RT 、NRT 、BE 表示;
S3{非TSN←→非TSN}分为4类:时间关键数据、实时数据、非实时数据、尽力而为,分别用TC 、RT 、NRT 、BE 表示;
TSSDN的工业异构网络中共有若干条链路,根据TSN节点与非TSN节点的分布情况将链路分成集合S1{TSN←→TSN}、S2{TSN←→非TSN}、S3{非TSN←→非TSN}三类,每一类按照延迟需求可继续细分为不同的数据流类型,数据流分类如表1所示;
表1工业异构网络数据流分类
Figure PCTCN2020102199-appb-000009
Figure PCTCN2020102199-appb-000010
第三步:TSSDN控制器对数据进行调度优先级的管理。
S1{TSN←→TSN}数据流由TSSDN控制器中的TSN模块进行调度管理,其优先级由IEEE802.1Q规定。根据IEEE802.1Q规定,TSN数据流的优先级在PCP字段,共有8个优先级,分别是0,1,2,3,4,5,6,7。S1{TSN←→TSN}数据流的优先级用OP表示。
针对S2{TSN←→非TSN}、S3{非TSN←→非TSN},提出延迟等级dl,既按照延迟需求划分为四个等级,延迟等级dl与数据类型一一对应,TSN←→非TSN类、非TSN←→非TSN类链路的延迟等级用dl表示。尽力而为的延迟等级dl为1,非实时数据延迟等级dl为2,实时数据延迟等级dl为3,时间关键数据延迟等级dl为4。
表2 TSSDN的工业异构网络的原有优先级与延迟等级
Figure PCTCN2020102199-appb-000011
Figure PCTCN2020102199-appb-000012
设链路i的调度优先级参考系数Θ i为:
Figure PCTCN2020102199-appb-000013
其中,α和β分别表示调度优先级参考系数Θ i中dl、OP所占的比重,且α+β=1。TSSDN控制器对S1{TSN←→TSN}的数据流不计算其调度优先级参考系数,S2{TSN←→非TSN}的数据流则使用公式(4),S3{非TSN←→非TSN}的数流据则使用公式(5)。
α和β的值可以根据不同应用环境下对数据低延迟要求和实时性的需求程度调整。本发明综合考虑数据低延迟要求与确定性,令α=0.4,β=0.6,则异构网络的调度优先级如表3所示。最低优先级为1,最高优先级为6。
表3调度优先级参考系数与调度优先级的关系
Θ i 调度优先级SP
0<Θ i<=0.5 1
0.5<Θ i<=1.0 2
1<Θ i<=1.3 3
1.3<Θ i<=1.6 4
1.6<Θ i<=1.9 5
Θ i>1.9 6
S1{TSN←→TSN}的数据流的优先级由IEEE802.1Q规定,S2{TSN←→非TSN}、S3{非TSN←→非TSN}的数据流的调度优先级取值与调度优先级参考系数相关。则TSSDN的工业异构网络中数据流的分类与调度优先级如表4所示。
表4调度优先级
Figure PCTCN2020102199-appb-000014
第四步:TSSDN控制器计算最短路径。
TSSDN控制器计算出某一数据流的调度优先级SP后,有必要为该流寻找合适的路径。在网络中进行路由选择的时候如果只考虑到路由的跳数,那只需要使用Dijkstra算法来选择路由。但是,随着网络规模的不断增长,网络的路由需要考虑更多的度量因素,如带宽、延迟、抖动、丢包率等的约束。CSPF算法是带有约束条件的Dijkstra算法。本发明采用改进的CSPF算法即多约束多目标路由选择算法来选择网络的最佳路径,以此来保障在一系列的约束条件下,网络资源的高效利用。
多约束多目标路由选择算法结合网络状态和业务应用需求为网络流量提供提一条满足多约束条件的最优路径。本发明的多约束条件包括带宽、延迟、抖动,目标是选出的路径延迟最小、可用带宽最大。
设节点p到节点q的路径为Path(p,q),网络中任意链路Li∈Path(p,q),d i、j i、BD i分别表示链路i的延迟、抖动和带宽,当节点p到节点q有且仅有一条路径时,该路径即为最短路径;当节点p到节点q有存在多条路径时,使用本发明所提方法计算最短路径。则路径的各个约束属性值表示为:
d(Path(p,q))=∑ Li∈Path(p,q)d i      (6)
j(Path(p,q))=∑ Li∈Path(p,q)j i    (7)
BD(Path(p,q))=min{BD i,Li∈Path(p,q)}    (8)
d(Path(p,q))代表路径Path(p,q)上的延迟之和,j(Path(p,q))代表路径Path(p,q)上的抖动之和、BD(Path(p,q))代表路径Path(p,q)上的最小带宽。
本发明的多约束条件为多目标路由算法以延迟、抖动和带宽为约束条件,以最小延迟和最大可用带宽为目标,具体公式为:
Figure PCTCN2020102199-appb-000015
其中目标函数F(x)可进一步表示为:
Figure PCTCN2020102199-appb-000016
dc、jc、BDc表示的是用户指定的路径延迟约束值、路径抖动约束值和路径带宽约束值。min{D(Path(p,q))}表示节点p到节点q的最小延迟,max{AvailableBD(Path(p,q))}表示节点p到节点q的最大可用带宽。
多约束多目标路由选择算法通过字典排序方法来解决目标路由的选择问题,字典排序方法是一种求解多目标优化问题的分析方法,核心思想是用目标函数的先后顺序来表达目标函数的重要程度。本发明所提算法是将目标函数f1(x)排在f2(x)的前面,即在计算路径的过程中,把最小延迟作为首要目标,当出现多条最小延迟的路径时选择最大可用带宽的路径。
第五步:调度优先级的标记
为满足不同调度优先级的数据流的服务质量,本发明利用TSSDN控制器,对来自不同现场设备的数据流进行调度优先级划分,8个调度优先级与8个队列一一对应,如表6所示;然后根据数据包的调度优先级标签进行出入队列的调度,依次到达SDN交换机。当某一调度 优先级的数据流突然增加而导致TSSDN的工业异构网络出现拥塞时,通过利用TSSDN控制器可以获取网络拓扑信息的特性,以及能够对SDN交换机的工业数据流转发行为进行集中控制的特点,以流表下发的形式来对大流进行带宽分配和拥塞控制,达到提升网络服务质量的效果,如表5所示。
表5调度优先级与队列的映射
调度优先级SP 对应队列
0 0
1 1
2 2
3 3
4 4
5 5
6 6
7 7
当TSSDN的异构网络内子网的数据流首次到达交换机时,由于交换机内没有与该数据流相匹配的流表项,数据流将打包成PacketIn消息被发送至TSSDN控制器;TSSDN控制器通过本发明所提方法计算出该流的调度优先级SP,并将调度优先级SP的数值映射成相应的优先级代码点(PCP)值。由TSSDN控制器向交换机下发“标记调度优先级并转发至目的端口”流表项,然后所有该类数据流到达交换机时,其PCP字段均被标记调度优先级,并转发至相应的出端口。
第六步:SDN交换机流表匹配
为了提高TSSDN控制器对TSN数据流的识别的准确性,本发明对传统的流表进行了修改,主要是对匹配字段进行了拓展,使SDN交换机能够更精确的匹配TSN的数据包。
协议类型:用于标记TSN网络的网络协议,如802.1AS、802.1Qbu、802.1Qbv等。
源MAC地址:标记TSN数据来源的MAC地址。
目的MAC地址:标记TSN数据目的地的MAC地址。
为了提高TSSDN控制器对TSN数据流的识别的准确性,本发明对传统的流表进行了修改,主要是对匹配字段进行了拓展,使SDN交换机能够更精确的匹配TSN的数据包。
协议类型:用于标记TSN网络的网络协议,如802.1AS、802.1Qbu、802.1Qbv等。
源MAC地址:标记TSN数据来源的MAC地址。
目的MAC地址:标记TSN数据目的地的MAC地址。
当数据流经过SDN交换机时,SDN交换机先把数据封装成PacketIn消息,发送至TSSDN控制器,根据匹配字段中VLAN、源MAC地址和目的MAC地址,TSSDN控制器把TSN数据流分配至TSN模块,由TSN模块调度管理;把非TSN数据流和TSN网络跨非TSN网络的数据流分配至SDN模块,由TSN模块调度管理。待TSSDN控制器根据网络流量特征执行完毕调度算法,根据到达SDN交换机的调度优先级以及带宽大小来决策该数据流是否转发或者丢弃,以流表的形式下发至SDN交换机。SDN交换机根据流表匹配情况执行决策该控制器的命令。
第七步:队列拥塞情况的判断
在数据转发的过程中可能出现数据量增加导致当前网络环境变差的情况,如果不能及时的调整转发的流量,那么将会出现比较严重的丢包现象,其中一些重要的、时间敏感的数据有可能会在这个过程中丢失,无法到达指定的接收端。因此,本发明设计了反馈调节机制实时监控端口各队列的状态,当网络出现拥塞时,根据拥堵状况做出调整,防止队列拥塞导致的数据滞留。为了改善数据转发过程可能出现的恶意丢包,需要针对不同程度的网络拥塞情况做不同的调整策略。本发明判断一个队列的拥塞程度的标准是队列延迟。
为更准确的反映当前网络拥塞情况,采用随机早期检测RED算法来计算队列的平均长度,avgQ_k代表的是队列k的平均长度,q_k为队列k的实时长度,由队列k的入队速率与出队速率计算得到。w为权值,相当于时间常数,则队列的平均长度为:
avgQk=(1-w)*avgQk+q*w   (7)
q_k=in_rate-out_rate    (8)
在随机早期检测算法中,设置一个合理的权值很重要,合适的w能够平衡队列的平均队列长度,规避因为数据激增或队列短暂拥塞导致的平均队列长度抖动的问题,如果w太大将会无法过滤掉短暂的拥塞;如果w太小则不能合理地反映拥塞状况。
为了计算队列的延迟,需要获取队列当前的速率,用户通过终端应用程序为每个队列设置最大最小带宽,当最大值与最小值不相同时,数据会在规定的速率范围内进行传输;而当最大值设定成等于最小值时,数据在队列中进行传输时会尽量保持在所设定的速率附近,以此获取队列的当前速率,则队列延迟计算公式为:
que_dealy_k=avgQk/now_rate_k      (9)
que_dealy_k表示队列k的延迟,now_rate_k表示队列k的当前速率,即队列的当前带宽。
第八步:队列拥塞的反馈调整
本发明的数据转发优化主要集中在SDN交换机端口的队列上。当队列出现拥塞,则根据 拥塞情况进行不同程度的队列带宽调整。拥塞情况的判断与分级是数据转发优化的一个关键。拥塞程度的评判依据是队列延迟,根据流量情况将队列的拥塞程度分成了三个等级:正常状态、一般拥塞状态与严重拥塞状态。
队列延迟与对应最大带宽值进行比较,当延迟小于最大带宽值的70%,则定义当前队列的状态为正常状态;当延迟大于最大带宽值的90%,则定义当前队列的状态为严重拥塞;当队列延迟处于最大带宽值的70%与90%之间,则定义当前的队列状态为一般拥塞状态。
反馈调整机制会根据当前的队列优先级和队列拥塞程度进行带宽的调整。整体来看,八个队列中优先级越高,调整带宽时抢占别人带宽的能力越强,优先级越低带宽被抢占的可能性越大,保证网络QoS的能力也越差。
如图7所示,队列抢占带宽的原则:
(1)优先级较高的队列在抢占优先级较低的队列带宽时只要保证其不出现严重拥塞情况即可,而低优先级队列在抢占带宽时则必须保证高优先级队列一直处于正常状态。
(2)当出现不同的优先级队列要抢占队列带宽时,遵循先高后低原则。即当高优先级队列先抢占完成后,在开始低优先级队列的抢占。这种做法可以确保高优先级的数据的优先传输,以保障其时间敏感性。
具体实施例:假设一个工业异构网络如图8所示,现使用本方案所提出的异构网络路由和调度相结合的方法对网络中的流量进行调度。流量发生器1的目的端是流量发生器1,摄像头的目的端是显示器,流量发生器2的目的端是流量发生器2。
(1)用户使用终端应用程序,设置各链路的延迟依次为d1=125us且优先级为4、d2=58ms且优先级为3、d3=9ms。为了方便后续的叙述,把三条链路依次记为L1、L2、L3。
(2)数据分类
(3)TSSDN控制器根据所提算法中第二步,L1属于S1{TSN←→TSN},L2属于S2{TSN←→非TSN},L3属于S3{非TSN←→非TSN},各链路的数据类型如下表所示。
链路 数据类型
L1 视频
L2 实时数据
L3 实时数据
(4)调度优先级的计算
根据所提算法中第三步进行等级划分。
L1:TSN链路不需延迟等级划分,;
L2:实时数据延迟等级为3
L3:实时数据延迟等级为3
链路 数据类型 原有优先级 延迟等级
L1 视频 4  
L2 实时数据   3
L3 实时数据   3
根据所提算法中第三步对其进行优先级管理。
L1:优先级为4,由TSSDN控制器的TSN模块进行调度管理;
L2:根据公式(4)Θ 2=0.4*3+0.6=1.8,则调度优先级为5;
L3:根据公式(5)Θ 4=0.4*3=1.2,则调度优先级为3。
链路 数据类型 原有优先级 延迟等级 调度优先级
L1 视频 4    
L2 实时数据   3 5
L3 实时数据   3 3
(4)生成流表
L1只有一条路径不需要计算,所存在的现有路径即为最短路径;L2只有一条路径不需要计算,所存在的现有路径即为最短路径;L3存在多条路径,需要计算流量发生器2到流量接收器2最短路径。
Figure PCTCN2020102199-appb-000017
根据本方案所提最短路径算法,延迟作为约束条件的重要参考,因此TSSDN控制器选择Path1作为流量发生器2的传输路径。
(5)调度优先级的标记
调度优先级与队列一一对应。
调度优先级SP 对应队列
3 3
4 4
5 5
当TSSDN的异构网络内子网的数据流首次到达交换机时,由于交换机内没有与该数据流相匹配的流表项,数据流将打包成PacketIn消息被发送至TSSDN控制器;TSSDN控制器将调度优先级SP的数值映射成相应的优先级代码点(PCP)值。由TSSDN控制器向交换机下发“标记调度优先级并转发至目的端口”流表项,然后所有该类数据流到达交换机时,其PCP字段均被标记调度优先级,并转发至相应的出端口。
(6)生成流表
TSSDN控制器生成流表并下发至SDN交换机,数据经过SDN交换机时,进行流表匹配。
最后说明的是,以上实施例仅用以说明本发明的技术方案而非限制,尽管参照较佳实施例对本发明进行了详细说明,本领域的普通技术人员应当理解,可以对本发明的技术方案进行修改或者等同替换,而不脱离本技术方案的宗旨和范围,其均应涵盖在本发明的权利要求范围当中。

Claims (10)

  1. 一种面向TSN和非TSN互联的工业异构网络调度方法,其特征在于:时间敏感的软件定义网络TSSDN的工业异构网络架构包括应用平面、控制平面和转发平面;
    所述应用平面包括用户和装有控制TSSDN控制器应用程序的终端,终端中的应用程序用于用户根据自己的应用需求对TSSDN控制器进行相应的配置;
    所述控制平面由TSSDN控制器构成,TSSDN控制器用于发现网络拓扑、制定调度策略、路径管理,TSSDN控制器根据来自应用层的应用请求,对工业异构网络中的数据流进行调度;
    所述转发平面包括工业网络和TSN网络,所述工业网络是任意个支持不同工业网络协议的网络,包括EtherCAT、Profinet和powerlink,按照是否为TSN数据流,整个工业异构网络分为非TSN工业网络、TSN工业网络以及TSN工业网络跨非TSN网络;所述转发平面具体包括SDN交换机、TSN交换机、工业交换机、工业现场设备、TSN现场设备;其中SDN交换机、TSN交换机、工业交换机用于实现数据转发,工业现场设备是指在工业网络中负责发送和接收数据的设备,TSN现场设备包括TSN Talker和TSN Listener,其中TSN Talker发送数据,TSN Listener接收数据。
  2. 根据权利要求1所述的面向TSN和非TSN互联的工业异构网络调度方法,其特征在于:所述TSSDN控制器是工业异构网络架构的核心,具有集中式管理控制功能;所述TSSDN控制器包括路径管理模块、拓扑管理模块和策略管理模块;TSSDN控制器根据来自应用层的应用请求,对工业异构网络中的数据流进行调度;
    拓扑管理模块用于发现网络拓扑和更新网络拓扑,并反馈给终端应用程序;
    路径管理模块是对来自用户的应用需求的响应,包括链路延迟、队列带宽设置的一系列路径管理操作;
    策略管理模块包括TSN模块和SDN模块,通过策略管理来满足工业异构网络中的时间敏感应用的延迟要求;TSN工业网络由TSN模块中的CUC/CNC进行集中统一管理,CNC具有集中计算的能力,包括计算传输调度,确定数据路径,通过时间感知整形器、帧抢占调度算法对TSN数据流调度;非TSN工业网络以及TSN工业网络跨非TSN网络由SDN模块的SDN控制器进行集中统一管理,通过下发流表到SDN交换机对非TSN工业网络以及TSN工业网络跨非TSN网络进行调度。
  3. 根据权利要求2所述的面向TSN和非TSN互联的工业异构网络调度方法,其特征在于:具体包括以下步骤:
    S1:网络部署阶段用户通过终端应用程序,根据应用需求对TSSDN控制器进行配置管理;
    S2:TSSDN控制器按照延迟需求进行数据分类;
    S3:TSSDN控制器对数据进行调度优先级的管理;
    S4:TSSDN控制器计算最短路径;
    S5:调度优先级的标记;
    S6:SDN交换机流表匹配:SDN交换机中有一个或多个流表,数据与流表进行匹配时,总是沿着流水线从第一个流表开始匹配,当数据与流表项匹配成功时,则更新计数器并执行流表中的相应指令,包括转发、丢弃指令;当流表匹配失败时,则以PacketIn消息的形式发送至TSSDN控制器,并由TSSDN控制器分析并做出决定;
    S7:判断队列拥塞情况:通过反馈调节机制实时监控端口各队列的状态,当网络出现拥塞时,根据拥堵状况做出调整,防止队列拥塞导致的数据滞留,针对不同程度的网络拥塞情况做不同的调整策略,判断一个队列的拥塞程度的标准是队列延迟;
    S8:对队列拥塞反馈调整:当队列出现拥塞,则根据拥塞情况进行不同程度的队列带宽调整,拥塞程度的评判依据是队列延迟,根据流量情况将队列的拥塞程度分成了三个等级:正常状态、一般拥塞状态与严重拥塞状态,队列延迟与对应最大带宽值进行比较,当延迟小于最大带宽值的70%,则定义当前队列的状态为正常状态;当延迟大于最大带宽值的90%,则定义当前队列的状态为严重拥塞;当队列延迟处于最大带宽值的70%与90%之间,则定义当前的队列状态为一般拥塞状态。
  4. 根据权利要求3所述的面向TSN和非TSN互联的工业异构网络调度方法,其特征在于:步骤S1中,用户根据自己的应用需求通过终端中的应用程序实现拓扑信息管理、流量信息管理,其中流量信息管理包括设置链路延迟、设置队列带宽最大和最小值、查看当前网络数据转发情况,用户按需设置链路延迟和带宽,用户配置成功后,存在链路集合、每条链路的延迟需求、每条链路的抖动需求、每条链路的带宽、用户指定的路径约束条件集合、每个队列的带宽六个集合:
    链路集合{L 1…..L i…};
    每条链路的延迟需求{d 1…d i.…};
    每条链路的抖动需求{j 1…j i.…};
    每条链路的带宽{BD 1…BD i.…};
    用户指定的路径约束条件集合{d c、j c、BD c};
    队列k的最大和最小带宽{max_bw_k、min_bw_k}。
  5. 根据权利要求4所述的面向TSN和非TSN互联的工业异构网络调度方法,其特征在 于:步骤S2中,设节点p到节点q的路径为Path(p,q),网络中任意链路Li∈Path(p,q),d i、j i、BD i分别表示链路i的延迟、抖动和带宽,dl为延迟等级,TSN网络的数据优先级为OP,α和β分别表示调度优先级参考系数Θ i中dl、OP所占的比重;TSSDN控制器根据TSN节点与非TSN节点的分布情况对链路进行分类,共分成3类:
    TSN节点与TSN节点之间的链路,用集合S1{TSN←→TSN}表示;
    TSN节点与非TSN节点之间的链路,用集合S2{TSN←→非TSN}表示;
    非TSN节点与非TSN节点之间的链路,用集合S3{TSN←→非TSN}表示;
    针对S1{TSN←→TSN},根据IEEE 802.1Q的规定,将数据流分为8类,分别是:背景流、尽力而为、最大努力、关键应用、视频、音频、互联网控制、网络控制;
    针对S2{TSN←→非TSN}、S3{非TSN←→非TSN},TSSDN控制器根据链路的延迟需求使用公式(1)、(2)、(3)通过比较计算,将工业异构网络中的数据流进行分类;延迟需求满足公式(1)的数据流为非实时数据NRT,延迟需求满足公式(2)的数据流为实时数据RT,延迟需求满足公式(3)的数据流为时间关键数据TC,没有延迟要求的属于尽力而为BE;
    d i≥100ms    (1)
    1ms<d i<100ms     (2)
    d i<1ms     (3)
    因此S2{TSN←→非TSN}分为4类:时间关键数据、实时数据、非实时数据、尽力而为;用TC 、RT 、NRT 、BE 表示;
    S3{非TSN←→非TSN}分为4类:时间关键数据、实时数据、非实时数据、尽力而为,分别用TC 、RT 、NRT 、BE 表示;
    TSSDN的工业异构网络中共有若干条链路,根据TSN节点与非TSN节点的分布情况将链路分成集合S1{TSN←→TSN}、S2{TSN←→非TSN}、S3{非TSN←→非TSN}三类,每一类按照延迟需求可继续细分为不同的数据流类型,数据流分类如下:
    Figure PCTCN2020102199-appb-100001
    Figure PCTCN2020102199-appb-100002
  6. 根据权利要求5所述的面向TSN和非TSN互联的工业异构网络调度方法,其特征在于:步骤S3具体包括:
    S1{TSN←→TSN}数据流由TSSDN控制器中的TSN模块进行调度管理,其优先级为:TSN数据流的优先级在PCP字段,共有8个优先级,分别是0,1,2,3,4,5,6,7;S1{TSN←→TSN}数据流的优先级用OP表示;
    针对S2{TSN←→非TSN}、S3{非TSN←→非TSN},提出延迟等级dl,既按照延迟需求划分为四个等级,延迟等级dl与数据类型一一对应,TSN←→非TSN类、非TSN←→非TSN类链路的延迟等级用dl表示;尽力而为的延迟等级dl为1,非实时数据延迟等级dl为2,实时数据延迟等级dl为3,时间关键数据延迟等级dl为4:
    Figure PCTCN2020102199-appb-100003
    Figure PCTCN2020102199-appb-100004
    设链路i的调度优先级参考系数Θ i为:
    Figure PCTCN2020102199-appb-100005
    其中,α和β分别表示调度优先级参考系数Θ i中dl、OP所占的比重,且α+β=1;TSSDN控制器对S1{TSN←→TSN}的数据流不计算其调度优先级参考系数,S2{TSN←→非TSN}的数据流则使用公式(4),S3{非TSN←→非TSN}的数流据则使用公式(5);
    α和β的值根据不同应用环境下对数据低延迟要求和实时性的需求程度调整,令α=0.4,β=0.6,则异构网络的调度优先级如下,最低优先级为1,最高优先级为6:
    Θ i 调度优先级SP 0<Θ i<=0.5 1 0.5<Θ i<=1.0 2 1<Θ i<=1.3 3 1.3<Θ i<=1.6 4 1.6<Θ i<=1.9 5 Θ i>1.9 6
    S1{TSN←→TSN}的数据流的优先级由IEEE802.1Q规定,S2{TSN←→非TSN}、S3{非TSN←→非TSN}的数据流的调度优先级取值与调度优先级参考系数相关;则TSSDN的工业异构网络中数据流的分类与调度优先级如下:
    Figure PCTCN2020102199-appb-100006
  7. 根据权利要求6所述的面向TSN和非TSN互联的工业异构网络调度方法,其特征在于:步骤S4中,具体包括:
    TSSDN控制器计算出某一数据流的调度优先级SP后,为该流寻找合适的路径,采用改进的CSPF算法即多约束多目标路由选择算法来选择网络的最佳路径,多约束多目标路由选择算法结合网络状态和业务应用需求为网络流量提供提一条满足多约束条件的最优路径;
    设节点p到节点q的路径为Path(p,q),网络中任意链路Li∈Path(p,q),d i、j i、BD i分别表示链路i的延迟、抖动和带宽,当节点p到节点q有且仅有一条路径时,该路径即为最短路径;当节点p到节点q有存在多条路径时,计算最短路径;则路径的各个约束属性值表示为:
    d(Path(p,q))=∑ Li∈Path(p,q)d i   (6)
    j(Path(p,q))=∑ Li∈Path(p,q)j i   (7)
    BD(Path(p,q))=min{BD i,Li∈Path(p,q)}    (8)
    d(Path(p,q))代表路径Path(p,q)上的延迟之和,j(Path(p,q))代表路径Path(p,q)上的抖动之和、BD(Path(p,q))代表路径Path(p,q)上的最小带宽;
    多约束条件为多目标路由算法以延迟、抖动和带宽为约束条件,以最小延迟和最大可用带宽为目标,具体公式为:
    Figure PCTCN2020102199-appb-100007
    其中目标函数F(x)进一步表示为:
    Figure PCTCN2020102199-appb-100008
    dc、jc、BDc表示的是用户指定的路径延迟约束值、路径抖动约束值和路径带宽约束值;min{D(Path(p,q))}表示节点p到节点q的最小延迟,max{AvailableBD(Path(p,q))}表示节点p到节点q的最大可用带宽;
    多约束多目标路由选择算法通过字典排序方法来解决目标路由的选择问题,字典排序方法是一种求解多目标优化问题的分析方法,核心思想是用目标函数的先后顺序来表达目标函数的重要程度;将目标函数f1(x)排在f2(x)的前面,即在计算路径的过程中,把最小延迟作为首要目标,当出现多条最小延迟的路径时选择最大可用带宽的路径。
  8. 根据权利要求7所述的面向TSN和非TSN互联的工业异构网络调度方法,其特征在于:步骤S5中,具体包括:
    利用TSSDN控制器,对来自不同现场设备的数据流进行调度优先级划分,8个调度优先级与8个队列一一对应,如下所示;然后根据数据包的调度优先级标签进行出入队列的调度,依次到达SDN交换机;当某一调度优先级的数据流突然增加而导致TSSDN的工业异构网络出现拥塞时,通过利用TSSDN控制器获取网络拓扑信息的特性,以及能够对SDN交换机的工业数据流转发行为进行集中控制的特点,以流表下发的形式来对大流进行带宽分配和拥塞控制,达到提升网络服务质量的效果;
    调度优先级SP 对应队列 0 0 1 1 2 2 3 3
    4 4 5 5 6 6 7 7
    当TSSDN的异构网络内子网的数据流首次到达交换机时,由于交换机内没有与该数据流相匹配的流表项,数据流将打包成PacketIn消息被发送至TSSDN控制器;TSSDN控制器计算出该流的调度优先级SP,并将调度优先级SP的数值映射成相应的优先级代码点PCP值;由TSSDN控制器向交换机下发“标记调度优先级并转发至目的端口”流表项,然后所有该类数据流到达交换机时,其PCP字段均被标记调度优先级,并转发至相应的出端口。
  9. 根据权利要求8所述的面向TSN和非TSN互联的工业异构网络调度方法,其特征在于:步骤S6中,具体包括:
    为了提高TSSDN控制器对TSN数据流的识别的准确性,对匹配字段进行拓展,使SDN交换机能够更精确的匹配TSN的数据包;
    协议类型:用于标记TSN网络的网络协议,包括802.1AS、802.1Qbu、802.1Qbv;
    源MAC地址:标记TSN数据来源的MAC地址;
    目的MAC地址:标记TSN数据目的地的MAC地址;
    当数据流经过SDN交换机时,SDN交换机先把数据封装成PacketIn消息,发送至TSSDN控制器,根据匹配字段中VLAN、源MAC地址和目的MAC地址,TSSDN控制器把TSN数据流分配至TSN模块,由TSN模块调度管理;把非TSN数据流和TSN网络跨非TSN网络的数据流分配至SDN模块,由TSN模块调度管理;待TSSDN控制器根据网络流量特征执行完毕调度算法,根据到达SDN交换机的调度优先级以及带宽大小来决策该数据流是否转发或者丢弃,以流表的形式下发至SDN交换机;SDN交换机根据流表匹配情况执行决策该控制器的命令。
  10. 根据权利要求9所述的面向TSN和非TSN互联的工业异构网络调度方法,其特征在于:步骤S8中,当队列出现拥塞,则根据拥塞情况进行不同程度的队列带宽调;拥塞程度的评判依据是队列延迟,根据流量情况将队列的拥塞程度分成了三个等级:正常状态、一般拥塞状态与严重拥塞状态:
    队列延迟与最大带宽的关系 拥塞等级 que_dealy<70%max_bd 正常状态 70%<que_dealy<90%max_bd 一般拥塞状态
    que_dealy>90%max_bd 严重拥塞状态
    队列延迟与对应最大带宽值进行比较,当队列延迟小于最大带宽值的70%,则定义当前队列的状态为正常状态;当队列延迟大于最大带宽值的90%,则定义当前队列的状态为严重拥塞;当队列延迟处于最大带宽值的70%与90%之间,则定义当前的队列状态为一般拥塞状态;
    反馈调整机制会根据当前的队列优先级和队列拥塞程度进行带宽的调整,六个队列中优先级越高,调整带宽时抢占别人带宽的能力越强,优先级越低带宽被抢占的可能性越大,保证网络QoS的能力也越差;
    队列抢占带宽的原则:
    (1)优先级较高的队列在抢占优先级较低的队列带宽时只要保证其不出现严重拥塞情况即可,而低优先级队列在抢占带宽时则必须保证高优先级队列一直处于正常状态;
    (2)当出现不同的优先级队列要抢占队列带宽时,遵循先高后低原则,即当高优先级队列先抢占完成后,再开始低优先级队列的抢占。
PCT/CN2020/102199 2020-05-11 2020-07-15 一种面向tsn和非tsn互联的工业异构网络调度方法 WO2021227245A1 (zh)

Priority Applications (2)

Application Number Priority Date Filing Date Title
KR1020217040187A KR102629950B1 (ko) 2020-05-11 2020-07-15 Tsn과 비tsn의 상호연결에 사용되는 산업 이기종 네트워크의 스케줄링 방법
US17/423,484 US11736408B2 (en) 2020-05-11 2020-07-15 Scheduling method applied in industrial heterogeneous network in which TSN and non-TSN are interconnected

Applications Claiming Priority (2)

Application Number Priority Date Filing Date Title
CN202010400917.6A CN111600754B (zh) 2020-05-11 2020-05-11 一种面向tsn和非tsn互联的工业异构网络调度方法
CN202010400917.6 2020-05-11

Publications (1)

Publication Number Publication Date
WO2021227245A1 true WO2021227245A1 (zh) 2021-11-18

Family

ID=72183712

Family Applications (1)

Application Number Title Priority Date Filing Date
PCT/CN2020/102199 WO2021227245A1 (zh) 2020-05-11 2020-07-15 一种面向tsn和非tsn互联的工业异构网络调度方法

Country Status (4)

Country Link
US (1) US11736408B2 (zh)
KR (1) KR102629950B1 (zh)
CN (1) CN111600754B (zh)
WO (1) WO2021227245A1 (zh)

Cited By (20)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN113923172A (zh) * 2021-07-28 2022-01-11 北京大学 具有时间确定性的ff现场总线交换机实现方法
CN114024916A (zh) * 2021-12-17 2022-02-08 湖北亿咖通科技有限公司 数据传输方法、装置、计算机可读存储介质及处理器
CN114172867A (zh) * 2021-12-02 2022-03-11 国网山东省电力公司信息通信公司 一种基于ims的海量信令诊断系统及方法
CN114257554A (zh) * 2021-12-23 2022-03-29 同济大学 一种提高tsn网络be流带宽利用率的调度方法
CN114362872A (zh) * 2021-12-30 2022-04-15 上海孟伯智能物联网科技有限公司 一种基于tsn网络的多网协同边缘智能设备
CN114422437A (zh) * 2021-12-24 2022-04-29 网络通信与安全紫金山实验室 一种异构报文的转发方法及装置
CN114513467A (zh) * 2022-04-18 2022-05-17 苏州浪潮智能科技有限公司 一种数据中心的网络流量负载均衡方法及装置
CN114615204A (zh) * 2022-02-17 2022-06-10 三一汽车起重机械有限公司 一种空中下载升级方法、系统及起重机
CN114785738A (zh) * 2022-06-16 2022-07-22 北京邮电大学 时间敏感流的调度方法、装置和电子设备
CN114802058A (zh) * 2022-06-23 2022-07-29 北京理工大学 智能电动汽车区域架构车载网络化控制系统及调度方法
CN114866426A (zh) * 2022-05-23 2022-08-05 北京计算机技术及应用研究所 一种基于时间敏感网络的嵌入式实时网络检测系统
CN114884899A (zh) * 2022-07-12 2022-08-09 之江实验室 一种多模态核心网转发和调度方法及装置
CN115051938A (zh) * 2022-05-31 2022-09-13 中国电子技术标准化研究院 Opc ua-tsn传输时延测试系统及方法
CN115189983A (zh) * 2022-07-07 2022-10-14 北京东土科技股份有限公司 一种访问autbus网络的tsn装置
CN115604170A (zh) * 2022-09-22 2023-01-13 长沙理工大学(Cn) 用于分布式设备聚合的边缘网络容器数据路由方法及系统
CN115951989A (zh) * 2023-03-15 2023-04-11 之江实验室 一种基于严格优先级的协同流量调度数值模拟方法与系统
CN116095055A (zh) * 2023-01-31 2023-05-09 安徽工业大学 一种支持事件触发的异步tsn流调度方法
CN116192651A (zh) * 2023-03-06 2023-05-30 重庆邮电大学 一种基于带宽预留的时间敏感网络调度方法
WO2023134091A1 (zh) * 2022-01-17 2023-07-20 重庆邮电大学 一种时间敏感网络中路径选择和门控调度的联合优化方法
CN117750405A (zh) * 2023-12-21 2024-03-22 山东未来网络研究院(紫金山实验室工业互联网创新应用基地) 一种跨地域5g专网管理方法

Families Citing this family (36)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US11121889B1 (en) 2020-06-15 2021-09-14 Moxa Inc. Apparatuses and methods for routing packets between a time-sensitive networking (TSN) network and a non-TSN network by virtual local area network (VLAN) tag manipulation
US11133959B1 (en) 2020-06-15 2021-09-28 Moxa Inc. Apparatuses and methods for routing packets for a time-sensitive networking (TSN) network by virtual local area network (VLAN) tag replacement
EP4211872A1 (de) * 2020-09-07 2023-07-19 Hirschmann Automation and Control GmbH Verfahren zum betreiben eines netzwerks
CN114363182A (zh) * 2020-09-29 2022-04-15 中国移动通信有限公司研究院 确定性网络系统及确定性业务的传输方法
CN112260957B (zh) * 2020-10-19 2022-05-13 重庆大学 一种时间感知整形器分级交叉流量调度系统
CN112311623B (zh) * 2020-10-22 2022-10-25 中车大连电力牵引研发中心有限公司 应用于列车的时间敏感网络测试平台及测试方法
CN112565068B (zh) * 2020-11-20 2021-12-21 华南理工大学 一种应用于tsn网络的冗余流调度方法
CN114553350B (zh) * 2020-11-24 2023-09-05 中国科学院沈阳自动化研究所 一种确定性低延时报文处理方法
CN112511462B (zh) * 2020-12-17 2022-07-26 上海交通大学 一种软件定义工业异构时间敏感网络系统及资源调度方法
CN112671673A (zh) * 2020-12-28 2021-04-16 广州西麦科技股份有限公司 一种基于sdn的流量控制系统和方法
CN112738843B (zh) * 2021-01-18 2022-05-27 哈尔滨工业大学 确定通信延迟的高速率5g上行拥塞控制方法
CN112887226B (zh) * 2021-01-26 2022-07-22 重庆邮电大学 有线无线融合的卫星时间敏感网络队列管理调度方法
CN112929117B (zh) * 2021-01-28 2022-03-15 杭州瓦良格智造有限公司 一种兼容可定义确定性通信以太网
CN113219909A (zh) * 2021-03-15 2021-08-06 之江实验室 基于工业级确定性网络的数据传送方法以及融合控制系统
CN115379525A (zh) * 2021-05-21 2022-11-22 华为技术有限公司 通信处理方法及相关设备
CN115484161A (zh) * 2021-05-28 2022-12-16 南宁富联富桂精密工业有限公司 Tsn网络配置方法、装置及计算机可读存储介质
CN113472570B (zh) * 2021-06-26 2024-02-02 华北电力大学 一种面向电力通信的时延敏感前传网流量调度方法
CN113726656B (zh) * 2021-08-09 2023-04-07 北京中电飞华通信有限公司 时延敏感流转发的方法及装置
CN114039936B (zh) * 2021-10-22 2023-12-26 北京邮电大学 基于时间敏感的虚拟电厂通信业务流量调度方法及装置
CN114390009B (zh) * 2021-12-20 2024-04-26 中国电子科技集团公司第五十四研究所 一种时间敏感网络中的数据处理方法及装置
CN114598413B (zh) * 2022-01-25 2024-04-02 浙江大学 一种支持时间敏感网络功能的安全分布式控制系统
CN114448894B (zh) * 2022-02-10 2024-01-05 上海交通大学 一种面向时间敏感网络的多等级业务调度引擎及实现方法
CN114553697B (zh) * 2022-02-14 2023-09-15 重庆邮电大学 一种面向工业无线与tsn融合的网络调度方法
KR102620137B1 (ko) * 2022-02-22 2024-01-02 네스트필드(주) Tsn 트래픽 스케줄링에 대한 구현 및 평가 방법
CN114567602B (zh) * 2022-03-08 2023-09-05 中国人民解放军国防科技大学 一种延迟保障传输系统及方法
CN114666280B (zh) * 2022-03-10 2023-09-01 山东大学 基于时间敏感软件定义网络的工业互联网时延优化方法
CN115037616B (zh) * 2022-04-22 2023-09-29 东北大学 一种基于tssdn网络的确定性流调度及其更新方法
CN115086239B (zh) * 2022-08-23 2022-11-04 中国人民解放军国防科技大学 一种共享式tsn整形调度装置
CN115604127A (zh) * 2022-09-30 2023-01-13 重庆邮电大学(Cn) 一种Modbus TCP跨TSN联合调度方法
CN116015711A (zh) * 2022-11-29 2023-04-25 国网山东省电力公司信息通信公司 基于电力交换网跨区安全交互方法及系统
CN115865810B (zh) * 2023-02-22 2023-06-09 浙江中控研究院有限公司 一种时间敏感网络中信用值流量调度系统及方法
CN116319598B (zh) * 2023-05-19 2023-11-07 工业富联(佛山)产业示范基地有限公司 时间敏感机械手臂控制方法、交换机及存储介质
CN116566903B (zh) * 2023-07-05 2023-09-26 南京信息工程大学 一种指控网络异构链路汇聚流量的端到端时延分析方法
CN116708280B (zh) * 2023-08-08 2023-10-24 合肥综合性国家科学中心人工智能研究院(安徽省人工智能实验室) 一种基于乱序度容忍的数据中心网络多路径传输方法
CN116743653B (zh) * 2023-08-16 2023-11-10 南京赛宁信息技术有限公司 一种网络靶场中流量发生器的自动选择方法与系统
CN117596200B (zh) * 2024-01-15 2024-05-07 之江实验室 一种时间敏感网络路由调度方法、电子设备、介质

Citations (8)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
WO2009026012A1 (en) * 2007-08-22 2009-02-26 Intel Corporation Methods, apparatus and systems configured for heterogeneous secure associations in wi-fi pan
CN105897493A (zh) * 2016-06-28 2016-08-24 电子科技大学 一种sdn规则冲突的检测方法
CN107294746A (zh) * 2016-03-30 2017-10-24 华为技术有限公司 一种部署业务的方法及设备
CN108289104A (zh) * 2018-02-05 2018-07-17 重庆邮电大学 一种工业SDN网络DDoS攻击检测与缓解方法
CN108809707A (zh) * 2018-05-30 2018-11-13 浙江理工大学 一种面向实时应用需求的tsn调度方法
CN108965171A (zh) * 2018-07-19 2018-12-07 重庆邮电大学 工业无线wia-pa网络与时间敏感网络转换方法及装置
CN109450943A (zh) * 2018-12-25 2019-03-08 重庆邮电大学 一种工业时间敏感网络多级安全数据调度方法
CN110996368A (zh) * 2019-11-22 2020-04-10 中国科学院计算机网络信息中心 一种应用于智能工厂的异构融合网络架构及路由配制方法

Family Cites Families (10)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN104994033A (zh) * 2015-05-13 2015-10-21 南京航空航天大学 一种资源动态管理的SDN网络QoS保障方法
KR102482102B1 (ko) * 2016-02-26 2022-12-27 현대자동차주식회사 네트워크에서 자원 예약의 해지 방법
CN105847151B (zh) * 2016-05-25 2019-03-29 安徽大学 一种面向软件定义网络的多约束QoS路由策略设计方法
CN109691038B (zh) * 2016-07-19 2021-12-17 施耐德电器工业公司 时间敏感的软件定义的网络
WO2018024809A1 (en) * 2016-08-03 2018-02-08 Schneider Electric Industries Sas Industrial software defined networking architecture for deployment in a software defined automation system
CN106341346B (zh) * 2016-09-08 2019-07-19 重庆邮电大学 基于SDN的数据中心网络中一种保障QoS的路由算法
TWI643478B (zh) * 2016-12-15 2018-12-01 新加坡商雲網科技新加坡有限公司 Sdn控制器及網路服務動態部署系統及方法
CN114008987A (zh) * 2019-05-02 2022-02-01 诺基亚技术有限公司 时间敏感联网系统中通信网络的集成
CN114826891A (zh) * 2019-05-20 2022-07-29 华为技术有限公司 Tsn中控制器间通信的方法、装置及系统
CN111010350B (zh) * 2019-12-09 2023-04-07 交控科技股份有限公司 基于tsn的列车多专业以太网综合承载方法及装置

Patent Citations (8)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
WO2009026012A1 (en) * 2007-08-22 2009-02-26 Intel Corporation Methods, apparatus and systems configured for heterogeneous secure associations in wi-fi pan
CN107294746A (zh) * 2016-03-30 2017-10-24 华为技术有限公司 一种部署业务的方法及设备
CN105897493A (zh) * 2016-06-28 2016-08-24 电子科技大学 一种sdn规则冲突的检测方法
CN108289104A (zh) * 2018-02-05 2018-07-17 重庆邮电大学 一种工业SDN网络DDoS攻击检测与缓解方法
CN108809707A (zh) * 2018-05-30 2018-11-13 浙江理工大学 一种面向实时应用需求的tsn调度方法
CN108965171A (zh) * 2018-07-19 2018-12-07 重庆邮电大学 工业无线wia-pa网络与时间敏感网络转换方法及装置
CN109450943A (zh) * 2018-12-25 2019-03-08 重庆邮电大学 一种工业时间敏感网络多级安全数据调度方法
CN110996368A (zh) * 2019-11-22 2020-04-10 中国科学院计算机网络信息中心 一种应用于智能工厂的异构融合网络架构及路由配制方法

Cited By (30)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN113923172B (zh) * 2021-07-28 2022-10-21 北京大学 一种具有时间确定性的ff现场总线交换机
CN113923172A (zh) * 2021-07-28 2022-01-11 北京大学 具有时间确定性的ff现场总线交换机实现方法
CN114172867A (zh) * 2021-12-02 2022-03-11 国网山东省电力公司信息通信公司 一种基于ims的海量信令诊断系统及方法
CN114172867B (zh) * 2021-12-02 2023-09-29 国网山东省电力公司信息通信公司 一种基于ims的海量信令诊断系统及方法
CN114024916A (zh) * 2021-12-17 2022-02-08 湖北亿咖通科技有限公司 数据传输方法、装置、计算机可读存储介质及处理器
CN114257554A (zh) * 2021-12-23 2022-03-29 同济大学 一种提高tsn网络be流带宽利用率的调度方法
CN114257554B (zh) * 2021-12-23 2023-05-12 同济大学 一种提高tsn网络be流带宽利用率的调度方法
CN114422437B (zh) * 2021-12-24 2024-05-10 网络通信与安全紫金山实验室 一种异构报文的转发方法及装置
CN114422437A (zh) * 2021-12-24 2022-04-29 网络通信与安全紫金山实验室 一种异构报文的转发方法及装置
CN114362872A (zh) * 2021-12-30 2022-04-15 上海孟伯智能物联网科技有限公司 一种基于tsn网络的多网协同边缘智能设备
WO2023134091A1 (zh) * 2022-01-17 2023-07-20 重庆邮电大学 一种时间敏感网络中路径选择和门控调度的联合优化方法
CN114615204A (zh) * 2022-02-17 2022-06-10 三一汽车起重机械有限公司 一种空中下载升级方法、系统及起重机
CN114513467B (zh) * 2022-04-18 2022-07-15 苏州浪潮智能科技有限公司 一种数据中心的网络流量负载均衡方法及装置
CN114513467A (zh) * 2022-04-18 2022-05-17 苏州浪潮智能科技有限公司 一种数据中心的网络流量负载均衡方法及装置
CN114866426A (zh) * 2022-05-23 2022-08-05 北京计算机技术及应用研究所 一种基于时间敏感网络的嵌入式实时网络检测系统
CN115051938B (zh) * 2022-05-31 2024-02-13 中国电子技术标准化研究院 Opc ua-tsn传输时延测试系统及方法
CN115051938A (zh) * 2022-05-31 2022-09-13 中国电子技术标准化研究院 Opc ua-tsn传输时延测试系统及方法
CN114785738B (zh) * 2022-06-16 2022-09-20 北京邮电大学 时间敏感流的调度方法、装置和电子设备
CN114785738A (zh) * 2022-06-16 2022-07-22 北京邮电大学 时间敏感流的调度方法、装置和电子设备
CN114802058A (zh) * 2022-06-23 2022-07-29 北京理工大学 智能电动汽车区域架构车载网络化控制系统及调度方法
CN115189983A (zh) * 2022-07-07 2022-10-14 北京东土科技股份有限公司 一种访问autbus网络的tsn装置
CN115189983B (zh) * 2022-07-07 2024-02-13 北京东土科技股份有限公司 一种访问autbus网络的tsn装置
CN114884899A (zh) * 2022-07-12 2022-08-09 之江实验室 一种多模态核心网转发和调度方法及装置
CN115604170A (zh) * 2022-09-22 2023-01-13 长沙理工大学(Cn) 用于分布式设备聚合的边缘网络容器数据路由方法及系统
CN116095055A (zh) * 2023-01-31 2023-05-09 安徽工业大学 一种支持事件触发的异步tsn流调度方法
CN116095055B (zh) * 2023-01-31 2024-04-12 安徽工业大学 一种支持事件触发的异步tsn流调度方法
CN116192651A (zh) * 2023-03-06 2023-05-30 重庆邮电大学 一种基于带宽预留的时间敏感网络调度方法
CN115951989A (zh) * 2023-03-15 2023-04-11 之江实验室 一种基于严格优先级的协同流量调度数值模拟方法与系统
CN115951989B (zh) * 2023-03-15 2023-06-20 之江实验室 一种基于严格优先级的协同流量调度数值模拟方法与系统
CN117750405A (zh) * 2023-12-21 2024-03-22 山东未来网络研究院(紫金山实验室工业互联网创新应用基地) 一种跨地域5g专网管理方法

Also Published As

Publication number Publication date
KR20220006588A (ko) 2022-01-17
CN111600754B (zh) 2022-02-25
KR102629950B1 (ko) 2024-01-29
CN111600754A (zh) 2020-08-28
US20220353195A1 (en) 2022-11-03
US11736408B2 (en) 2023-08-22

Similar Documents

Publication Publication Date Title
WO2021227245A1 (zh) 一种面向tsn和非tsn互联的工业异构网络调度方法
US6104700A (en) Policy based quality of service
CN104378309B (zh) OpenFlow网络中实现QoS的方法、系统和相关设备
US7835279B1 (en) Method and apparatus for shared shaping
US20060101144A1 (en) Bandwidth management system
US20040081093A1 (en) Policy based quality of service
EP1708441B1 (en) A method, network element and communication network for fairly adjusting bandwidth among distributed network elements
CN107454015B (zh) 一种基于OF-DiffServ模型的QoS控制方法及系统
CN104994033A (zh) 一种资源动态管理的SDN网络QoS保障方法
CN113746748B (zh) 一种命名数据网络中显式拥塞控制方法
CN101964758A (zh) 一种基于区分服务的队列调度方法
JP2004201304A (ja) 高速パケット網のためのパケットスケジューリングシステム及び方法
US20210250301A1 (en) Scalable deterministic services in packet networks
CN110247859B (zh) 一种基于airt-ros节点的数据发送方法及装置
Baklizi Stabilizing average queue length in active queue management method
CN115643220A (zh) 基于抖动时延的确定性业务传输方法和装置
CN106685849B (zh) 基于sdn的wred报文丢弃方法及系统
CN115883490A (zh) 基于sdn的分布式计算通信一体化调度方法及相关组件
Cisco Configuring Quality of Service
Cisco Configuring Quality of Service
Tokmakov et al. A case for data centre traffic management on software programmable ethernet switches
Asaduzzaman et al. The Eight Class of Service Model-An Improvement over the Five Classes of Service
Elbediwy et al. DR-PIFO: A Dynamic Ranking Packet Scheduler Using a Push-In-First-Out Queue
Xue et al. Fall: A fair and low latency queuing scheme for data center networks
KR20090062878A (ko) 패킷 스케줄링 시스템 및 방법

Legal Events

Date Code Title Description
ENP Entry into the national phase

Ref document number: 20217040187

Country of ref document: KR

Kind code of ref document: A

121 Ep: the epo has been informed by wipo that ep was designated in this application

Ref document number: 20935611

Country of ref document: EP

Kind code of ref document: A1

NENP Non-entry into the national phase

Ref country code: DE

122 Ep: pct application non-entry in european phase

Ref document number: 20935611

Country of ref document: EP

Kind code of ref document: A1