阅读说明：本技术 一种时间敏感网络尽力而为流调度方法 (Best effort flow scheduling method for time sensitive network ) 是由 刘勤 张楠浠 黄鹏宇 于 2021-01-29 设计创作，主要内容包括：本发明属于网络通信技术领域,公开了一种时间敏感网络尽力而为流调度方法,该方法包括：获取当前时间敏感网络数据；从缓存中选择使尽力而为流带宽利用率最大化的一组帧；获取尽力而为流调度结果；对尽力而为流进行调度。本发明针对Qbv协议中在每个带宽周期都使尽力而为流时隙的带宽利用率达到最大化,提高了尽力而为流的带宽资源利用率。随着带宽周期的累积,在整体服务上很大程度的提高了带宽资源的利用率。(The invention belongs to the technical field of network communication, and discloses a best-effort flow scheduling method for a time-sensitive network, which comprises the following steps: acquiring current time sensitive network data; selecting a set of frames from the cache that maximizes best effort stream bandwidth utilization; obtaining a best effort flow scheduling result; best effort flows are scheduled. The invention maximizes the bandwidth utilization rate of the best-effort flow time slot in each bandwidth period in the Qbv protocol and improves the bandwidth resource utilization rate of the best-effort flow. With the accumulation of bandwidth periods, the utilization rate of bandwidth resources is greatly improved on the whole service.)

1. A method for scheduling a best effort flow of a time sensitive network is characterized by specifically comprising the following steps:

acquiring current time sensitive network data;

selecting a set of frames from the cache that maximizes best effort stream bandwidth utilization;

obtaining a best effort flow scheduling result;

best effort flows are scheduled.

2. The method of claim 1, wherein the data set F ═ F is set to₁,f₂,…,f_i,…,f_nI 1 ≦ i ≦ n } represents a best effort flow through the time sensitive network switch cache with a corresponding frame length C ≦ C₁,c₂,…,c_i,…,c_nI is more than or equal to 1 and less than or equal to n, wherein f_iI-th best effort flow representing incoming switch, c_iRepresents the ith best effort flow f_iA corresponding frame length;

the method comprises the steps that a time perception shaper divides communication on an Ethernet into fixed-length repeated time periods, scheduling is mainly carried out on high-priority planned flows and best-effort flows, high-priority planned flows are scheduled by high-priority planned flow time slots in each period, the best-effort flows are scheduled by the best-effort flow time slots, and the current time sensitive network bandwidth period is CT, the best-effort flows are scheduled by the best-effort flow time slots, and the transmission speed of a link in the time sensitive network is u.

3. The method for best effort flow scheduling for a time sensitive network of claim 2 wherein the high priority scheduled flow slots at each cycle schedule the high priority scheduling mechanism according to a time aware shaper in the time sensitive network; at the best-effort stream time slot of each cycle, a set of frames from the buffer is taken to be selected that maximizes the best-effort stream bandwidth utilization:

(1) establishing a mathematical model eta which maximizes the utilization rate of the best effort flow time slot bandwidth:

wherein x_iAs decision variables, x_i1 denotes the ith best effort flow f_iIs selected, x_i0 denotes the ith best effort flow f_iIs not selected;

(2) solving a mathematical model eta for maximizing the best effort flow time slot bandwidth utilization rate by adopting a solution algorithm of a 0-1 knapsack problem to obtain a group of frames for maximizing the best effort flow time slot bandwidth utilization rate

Obtaining a best-effort flow scheduling result and scheduling the best-effort flow: the set of frames that maximizes best effort stream slot bandwidth utilizationAnd sending the sequence number i from small to large to finish the scheduling of the best-effort flow.

Technical Field

The invention belongs to the technical field of network communication, and particularly relates to a best-effort flow scheduling method for a time-sensitive network.

Background

At present: with the continuous development of network communication technology, new information technology is emerging continuously, for example, services such as audio and video technology which have requirements on real-time performance are increased continuously, and standard ethernet only provides best-effort services, which cannot meet the real-time services with high requirements on time delay and jitter. Therefore, a Time Sensitive Network (TSN) is produced, which is a group of revisions to IEEE802.1 and aims to introduce a real-Time function into an ethernet-based network, and has the advantages of high-precision Time synchronization, bounded Time delay and zero jitter, coexistence of multiple services, high reliability and compatibility, and the like, and is widely applied to the fields of industrial automation, vehicle-mounted networks, remote control, and the like. So far, the standard of the TSN is still under development, each sub-protocol in the TSN is researched and the cooperative work among the protocols is carried out, so that the essence of the technology guarantee real-time communication of each protocol of the TSN is researched, and the TSN has profound practical significance for more widely and better applying the TSN to each field in the future.

The technical core of the time-sensitive network is the traffic scheduling and shaping algorithm, which is a hot spot of deep research in academic circles. Most studies of scheduling algorithms focus on how to guarantee deterministic delay for Time-sensitive flows, such as Time-aware shapers (TAS), which aim to divide the traffic over ethernet networks into fixed-length, repeating Time periods. During these periods, different Time slices can be configured according to the concept of Time Division Multiple Access (TDMA), and these Time slices can be allocated to one or several of the eight ethernet priorities. By doing so, time-critical traffic can be separated from noncritical traffic and transmitted strictly according to cycles and time slots, ensuring limited delay and jitter of time-sensitive traffic. The TAS considers two main stream types, namely a high priority Scheduled Stream (ST) and a low priority Best Effort stream (BE), with the ST buffered in the ST queue and the BE buffered in the BE queue. The TAS implements frame priority isolation by traffic class. The TAS ensures that the ST delay is bounded and protects the ST from interference by any cross traffic. In each time period, the ST stream and the BE stream are transmitted within the corresponding time slice.

When the BE stream fails to complete transmission in the corresponding bandwidth, the bandwidth of the ST stream in the next period is occupied, and the delay of the ST stream cannot BE guaranteed, which affects the service quality of the high-priority stream. The TSN task group thus proposed a guard band mechanism in the enhanced Qbv protocol in 2015, entitled "IEEE std802.1 qbvt-2015. IEEE Standard for Local and metric area Networks-Bridges and Bridge Networks evaluation 25: Enhancement for Scheduled Traffic", the principle of which is to place a guard band before the end of the best effort stream time slice, i.e. no new frame is allowed to be transmitted within the guard band, but the frame being transmitted can continue to be transmitted, so the size of the guard band should be set to the time of the size of the largest frame in the transport ethernet. Although the introduction of the guard band ensures the deterministic and reliable transmission of the high-priority service, it also causes a serious problem of bandwidth resource waste. The TSN task group proposes a length-aware scheduling mechanism to partially mitigate the bandwidth waste due to the guard band. The principle of this length-aware scheduling mechanism is that when using store-and-forward transmission, when an ethernet frame needs to be transmitted within a guard band time slot, the scheduler will check the length of the frame to be transmitted, and if the frame can be completely transmitted within the guard band without encroaching on the next cycle time sensitive stream time slice, the scheduler can transmit the frame despite the guard band time slot, to reduce the waste of bandwidth resources by the guard band. Although the length-aware scheduling mechanism alleviates the waste of bandwidth resources, the problem of more serious waste of bandwidth resources still exists.

Through the above analysis, the problems and defects of the prior art are as follows: the guard band mechanism and the length perception scheduling mechanism in the existing Qbv protocol have the problem of bandwidth resource waste, the guard band in each period can cause the waste of bandwidth resources, and the problem of bandwidth resource waste caused by the accumulation of the period time is not a little worth. When there are many BE frames in the buffer, a large number of BE frames may BE left in the buffer, and in a serious case, the buffer may overflow, and then data may BE lost.

The significance of solving the problems and the defects is as follows: the present best-effort flow scheduling method is carried out according to FIFO principle, most researches only focus on the transmission scheduling of high priority flow, ensure the reliable transmission of the high priority flow, and neglect the scheduling of the best-effort flow, therefore, the invention mainly analyzes the problem of bandwidth waste caused by a guard band mechanism, improves the utilization rate of the best-effort flow time slot bandwidth resource, and improves the integral service quality of the best-effort flow.

Disclosure of Invention

Aiming at the problems in the prior art, the invention provides a best effort flow scheduling method for a time sensitive network.

The invention is realized in this way, a time-sensitive network best-effort flow scheduling method, the time-sensitive network best-effort flow scheduling method includes:

acquiring current time sensitive network data;

selecting a set of frames from the cache that maximizes best effort stream bandwidth utilization;

obtaining a best effort flow scheduling result;

best effort flows are scheduled.

Further, let F be { F ═ F₁,f₂,…,f_i,…,f_nI 1 ≦ i ≦ n } represents a best effort flow through the time sensitive network switch cache with a corresponding frame length C ≦ C₁,c₂,…,c_i,…,c_nI is more than or equal to 1 and less than or equal to n, wherein f_iI-th best effort flow representing incoming switch, c_iRepresents the ith best effort flow f_iA corresponding frame length;

further, the communication on the Ethernet is divided into fixed length and repeated time periods by the time perception shaper, the scheduling is mainly carried out aiming at the high-priority planned flow and the best-effort flow, the high-priority planned flow is scheduled by the high-priority planned flow time slot in each period, the best-effort flow time slot is scheduled by the best-effort flow time slot, and the current time sensitive network bandwidth period is CT, the best-effort flow time slot is D, and the transmission speed of a link in the time sensitive network is u.

Further, scheduling the high-priority scheduling mechanism in the high-priority scheduled stream time slot of each period according to a time perception shaper in the time sensitive network; at the best-effort stream time slot of each cycle, a set of frames from the buffer is taken to be selected that maximizes the best-effort stream bandwidth utilization:

(1) establishing a mathematical model eta which maximizes the utilization rate of the best effort flow time slot bandwidth:

wherein x_iAs decision variables, x_i1 denotes the ith best effort flow f_iIs selected, x_i0 denotes the ith best effort flow f_iIs not selected;

(2) solving a mathematical model eta for maximizing the best effort flow time slot bandwidth utilization rate by adopting a solution algorithm of a 0-1 knapsack problem to obtain a group of frames for maximizing the best effort flow time slot bandwidth utilization rate

Further, a best effort flow scheduling result is obtained, and best effort flow scheduling is completed: the set of frames that maximizes best effort stream slot bandwidth utilizationAnd sending the sequence number i from small to large to finish the scheduling of the best-effort flow.

By combining all the technical schemes, the invention has the advantages and positive effects that: the invention substantially carries out time sequence rearrangement on the best-effort flow, selects a group of frames which enables the bandwidth utilization rate to be maximum from the best-effort flow in the cache, and then sends the selected frame combination according to the sequence. When the number of BE frames in the cache is large, namely when the best effort traffic is large, the bandwidth utilization rate of the BE time slot is improved in each bandwidth period, the utilization rate of bandwidth resources is integrally improved along with the accumulation of the bandwidth periods, and the retention and overflow loss of cache data frames are avoided.

Drawings

In order to more clearly illustrate the technical solutions of the embodiments of the present application, the drawings needed to be used in the embodiments of the present application will be briefly described below, and it is obvious that the drawings described below are only some embodiments of the present application, and it is obvious for those skilled in the art that other drawings can be obtained from the drawings without creative efforts.

Fig. 1 is a flowchart of a method for scheduling best effort flows in a time-sensitive network according to an embodiment of the present invention.

Fig. 2 is a schematic diagram of the Qbv protocol provided by the embodiment of the present invention.

Fig. 3 is a diagram of simulation results provided by the embodiment of the present invention.

Detailed Description

In order to make the objects, technical solutions and advantages of the present invention more apparent, the present invention is further described in detail with reference to the following embodiments. It should be understood that the specific embodiments described herein are merely illustrative of the invention and are not intended to limit the invention.

In view of the problems in the prior art, the present invention provides a best effort flow scheduling method for a time sensitive network, which is described in detail below with reference to the accompanying drawings.

As shown in fig. 1, the best effort flow scheduling method for a time-sensitive network provided by the present invention includes the following steps:

s101: acquiring current time sensitive network data;

s102: selecting a set of frames from the cache that maximizes best effort stream bandwidth utilization;

s103: obtaining a best effort flow scheduling result;

s104: best effort flows are scheduled.

Obtaining best effort flow F ═ F in current time sensitive network switch cache₁,f₂,…,f_i,…,f_nI is not less than 1 and not more than n and the corresponding frame length is C ═ C₁,c₂,…,c_i,…,c_nI is more than or equal to 1 and less than or equal to n, wherein f_iI-th best effort flow representing incoming switch, c_iRepresents the ith best effort flow f_iA corresponding frame length;

as shown in fig. 2, the time-aware shaper divides the communication on the ethernet network into fixed-length and repeated time periods, the time-sensitive network mainly schedules high-priority scheduled flows and best-effort flows, and in each period, time slice 1 transmits the high-priority scheduled flows, and time slice 2 transmits the best-effort flows, that is, in each period, time slice 1 schedules the high-priority scheduled flows, and time slice 2 schedules the best-effort flow slots. The bandwidth cycle of the current time sensitive network is CT, the bandwidth of the time slice 2 is D, and the transmission speed of the link in the time sensitive network is u.

The time perception shaper relies on global time synchronization, and after the bandwidth period of the current time sensitive network and the bandwidth sizes of the time slice 1 and the time slice 2 are obtained at the beginning, the time slice 1 in each period carries out corresponding scheduling on a high-priority scheduling mechanism according to the time perception shaper in the time sensitive network; and in time slice 2 of each period, scheduling the best-effort flow by adopting the best-effort flow scheduling method provided by the invention.

The steps of obtaining the best-effort flow scheduling result and scheduling the best-effort flow comprise the following specific steps:

step 1) aiming at the bandwidth D of the time slice 2 and the best-effort flow in the switch cache in the current period, establishing a mathematical model eta which maximizes the bandwidth utilization rate of the time slice 2:

wherein x_iAs decision variables, x_i1 denotes the ith best effort flow f_iIs selected, x_i0 denotes the ith best effort flow f_iIs not selected.

Step 2) solving a mathematical model eta for maximizing the bandwidth utilization rate of the time slice 2by adopting a solution algorithm of a 0-1 knapsack problem to obtain a group of frames for maximizing the bandwidth utilization rate of the time slice 2

Step 3) the group of framesAccording to the size of the serial number i, from small to largeAnd transmitting the data on the bandwidth of the time slice 2 to finish scheduling the best-effort flow.

Those skilled in the art can also implement the best-effort flow scheduling method for the time-sensitive network provided by the present invention by using other steps, and fig. 1 shows that the best-effort flow scheduling method for the time-sensitive network provided by the present invention is only one specific embodiment.

The technical solution of the present invention is further described below with reference to the accompanying drawings.

As shown in fig. 2, the Qbv protocol divides the ethernet-transmitted traffic into fixed-length, repeating time periods, each of which time slice 1 transmits a high-priority scheduled flow and time slice 2 transmits a best-effort flow. The scheduling method of the best-effort flow in the Qbv protocol is to transmit according to the FIFO principle, the size of the last transmission frame in each period is uncertain randomly, thus causing the situation that the bandwidth is wasted in each period, therefore, the invention breaks through the mode of transmitting in sequence in the original protocol, and the invention rearranges the transmission sequence of the best-effort flow, and aims to improve the utilization rate of bandwidth resources. The concrete implementation is as follows: and selecting a group of frames which can occupy the largest time slot of the best effort stream from the buffer, and then sending the selected group of frames from small to large according to the sequence number, wherein the sending is carried out according to the sequence number from small to large so as to ensure the fairness of data transmission.

The technical effects of the present invention will be described in detail with reference to simulation experiments.

1. Simulation experiment conditions are as follows:

the hardware platform of the simulation experiment of the invention is as follows: the processor is an Intel Pentium CPU, the main frequency is 3.30GHz, and the memory is 8 GB.

The software platform of the simulation experiment of the invention is as follows: windows7 operating system and MATLAB.

The best-effort flow scheduling is verified by the experiment, the best-effort flow time slot, namely the bandwidth of the time slice 2 in fig. 2, is set to BE D50 us, the transmission speed of the link is u 1Gbps, different numbers of best-effort frames are randomly generated in the cache, and the frame length of all BE frames is in the range of 84byte to 1542byte, so that the frame length range of the Ethernet is met.

2. And (3) simulation results: under the experimental conditions, the results of comparing the bandwidth utilization rate of the best-effort stream time slot are mainly carried out aiming at three scheduling mechanisms, one is FIFO principle scheduling with a protection band mechanism with the maximum frame transmission bandwidth of Ethernet in a time sensitive network Qbv protocol, one is a length perception scheduling mechanism provided by a TSN task group for relieving bandwidth waste produced by a protection band machine, and the other is a best-effort stream time sequence rearrangement scheduling method provided by the invention. Fig. 3 is a simulation diagram of calculating the bandwidth utilization rate of the current period by scheduling best-effort flows by using the three scheduling methods when different numbers of BE frames are stored in the cache, and performing an experiment 1000 times to obtain an average value. It can BE found that when the number of BE frames in the buffer is small, the bandwidth utilization rates of the three scheduling methods are the same, because all services can BE sent in one cycle. With the increase of the traffic, that is, the number of BE frames in the buffer, the utilization rate of the bandwidth is also increased. For the scheduling method of time sequence rearrangement, when the traffic is greater than 11, the utilization rate of the bandwidth almost reaches 100%, and the method enters a stable stage. For length-aware scheduling, the bandwidth utilization fluctuates around 92% after traffic is greater than 11. For the transmission method using the maximum frame length as the guard band FIFO principle, after the traffic is greater than 11, the utilization rate of the bandwidth fluctuates around 84%. It can be seen from the figure that, when the traffic volume is large, the scheduling method for time sequence rearrangement can greatly improve the bandwidth utilization ratio compared with the other two methods, so that the bandwidth utilization ratio is maximized.

The above description is only for the purpose of illustrating the present invention and the appended claims are not to be construed as limiting the scope of the invention, and any modifications, equivalents and improvements made by those skilled in the art within the spirit and principle of the present invention are intended to be covered by the present invention.