RU2350027C2 - Method of multilevel planning supporting plural ports and set of services - Google Patents

Abstract

FIELD: information technologies.

SUBSTANCE: method of multilevel planning of the traffic, supporting set of ports and set of services, contains stages on which provide turns of the second level for classes A, B and C of traffic in each port of the leg of the user as a part of the separate device of an adaptive ring of a packet transmission (RPR), and the data frames accepted in port of the leg of the user, are subject to a premise in respective turns of the second level according to their identifiers of classes of the traffic; carry out the two-level planning consisting of primary planning according to a class of the traffic and the subsequent planning according to weight of port, on turns of the second level for port of the leg of the user; enter shots of data after two-level planning into turns of the first level for classes A, B and C of traffic accordingly; also carry out introduction planning in RPR for three classes of the traffic in turns of the first level.

EFFECT: improvement of planning of traffic transmission.

11 cl, 6 dwg

Description

FIELD OF THE INVENTION

The present invention generally relates to data frame transmission technology, and more particularly to adaptive packet transmission ring (RPR) technology in a city-wide network, and more particularly, to a scheduling method for an RPR device using multiple user side ports, each user side port supporting multiple services.

State of the art

Adaptive packet ring technology is a newly emerging technology whose purpose is to create a data frame network where the bandwidth can be multiplexed and each node contains a fair algorithm for guaranteeing bandwidth occupancy and has ring protection and quality of service (QoS) capabilities and basically It is oriented on the main ring networks of the city scale and the ring access networks of the city scale. In December 2000, the Institute of Electrical and Electronics Engineers (IEEE) specifically established the IEEE 802.17 standards group to formulate the IEEE 802.17 standard of the MAC layer (media access control), which is based on RPR technology and has the ability to create an adaptive packet transmission ring.

The main feature of RPR technology is the classification of all traffic into three classes: class A, which is real-time traffic; Class B, divided into two parts, namely, B-CIR (coordinated information transfer rate), corresponding to the guaranteed transmission rate level traffic, and B-EIR (exceeded information transmission rate), corresponding to traffic that exceeds the guaranteed transmission rate level; and class C, which is the best that’s possible traffic. These three traffic classes differ according to the class of service (ServiceClass) specified in the frame structure. In order to enter the RPR ring from a separate RPR node, traffic is formed and planned according to their respective classes A, B or C and then enters the RPR ring.

As shown in FIG. 1, an RPR device typically contains two ring side RPR ports and a set of user side ports, and frames entering or leaving the ring side RPR ports are MAC layer RPR frames containing traffic class identifiers (IDs) corresponding to A, B or C while the user side ports are ordinary Ethernet LAN ports, and the frames entering or leaving these ports are Ethernet LAN frames that do not have traffic class identifiers A, B, or C inside the frame structure. Many RPR devices are organized in RPR through their ring side RPR ports. The user side ports of the prior art RPR device do not have queues, and when a data frame enters the user side port, after some forwarding operation, it enters the A / B / C queue of the RPR ring and then goes through the usual shaping and scheduling operations, enters the parcel waiting queue and finally enters the RPR ring, as shown in FIG. 2. There are several difficulties in the aforementioned method of scheduling the processing of frames received at user side ports: first, now that there are usually many user side ports in an RPR device, it may happen that some port has a large traffic load and some other a port has a small traffic load, and in the prior art, scheduling is usually done on a first-come-first-served basis, which can lead to a situation where a port with a small traffic load is running gives with underload. Secondly, traffic coming from some port on the user side can be class A, or B, or C traffic, and if a large amount of class C traffic comes in first and class A traffic comes in next, in accordance with the “first come first Served ”a large amount of Class C traffic with a lower priority will be processed first, and Class A traffic with a higher priority will be processed only after Class C traffic is processed, thus the problem of delay and delay time spread for class traffic It can not be solved.

SUMMARY OF THE INVENTION

The technical problem to be solved by the present invention is the provision of a multi-level scheduling method supporting multiple ports and multiple services, which is aimed at eliminating the disadvantage corresponding to the prior art of inappropriate traffic planning between ports in the case where the RPR device contains many user side ports, in order to better meet the requirements for RPR.

A multi-level scheduling method supporting multiple ports and multiple services proposed in the present invention comprises the following steps:

provide second-level queues for traffic classes A, B and C in each port of the user side of a separate RPR device, and data frames received at this port of the user side must be placed in the corresponding second level queues according to their traffic class identifiers;

perform two-level planning, which consists of initial planning according to the class of traffic and subsequent planning according to the weight of the port in the queues of the second level of the port of the user side;

data frames after two-level planning are included in the first level queues for traffic classes A, B and C, respectively; and

Perform implementation planning in RPR on three classes of traffic in first-level queues.

The minimum bandwidths for the three classes of traffic A, B, and C allowed to enter the RPR ring from each port on the user side are user-configurable.

Planning according to the class of traffic is performed in priority order, such that class A traffic takes precedence over class B traffic, and class B traffic takes precedence over class C traffic; scheduling according to the weight of the port is performed in accordance with the weights of the respective traffic classes for the respective ports of the user side.

The weight value is obtained based on the ratio between the minimum bandwidth of traffic of class A, B or C in the port of the user side and the bandwidth of traffic of the same class in all ports of the user side in the RPR device.

The weights for different classes of traffic may be different.

The stage of performing two-level planning on the second-level queues specifically contains:

step 1, in which second-level queues for class A traffic for ports of the user side are polled one by one for whether they are empty, and if so, perform step 2, otherwise go to step 3;

step 2, in which it is determined whether the given port of the user side is the last port of the user side, and if so, go to step 5, otherwise they return to stage 1, in which the second level queue for class A traffic for the next port of the user side is polled;

step 3, in which, in accordance with the weight value for the class A traffic of the port, the user parties prioritize class A traffic in the second level queue into the first level queue for the class A traffic;

step 4, in which it is determined whether the queue of the first level of traffic class A is full, and if it is not, go to step 2, and if it is, then go to step 5;

step 5, in which second-level queues for class B traffic for user side ports are polled one by one for whether they are empty, and if so, perform step 6, otherwise go to step 7;

step 6, which determines whether the user side port is the last user side port, and if so, go to step 9, otherwise return to step 5, in which the second level queue for traffic class B for the next user side port is polled;

step 7, in which, in accordance with the weight value for class B traffic of the port, the user parties prioritize class B traffic in the second level queue into the first level queue used for class B traffic;

step 8, in which it is determined whether the first level queue for traffic of class B is full, and if it is not, go to step 6, and if so, go to step 9;

step 9, in which a survey is carried out to determine whether the queues of the second level for traffic class C for ports of the user side are empty, and if so, perform step 10, otherwise go to step 11;

step 10, which determines whether the user side port is the last user side port, and if so, go to step 1, otherwise return to step 9, in which the second level queue for traffic class C is polled for the next user side port;

step 11, in which, in accordance with the weight value for class C traffic of the port, the user parties prioritize class C traffic in the second level queue into the first level queue for class C traffic;

step 12, in which it is determined whether the first level queue for traffic class C is full, and if it is not, go to step 10, and if it is, then go to step 1.

The step of implementing implementation planning on three traffic classes in the first level queues further comprises the formation of three traffic classes, respectively, and then planning data frames for the respective traffic to be output to the corresponding RPR ring.

Formation and scheduling is performed in such a way that class A traffic takes precedence over class B traffic, and class B traffic takes precedence over class C traffic.

At the formation stage, traffic of the corresponding class should be generated in the first level queues in accordance with the value of the loan for each traffic class.

The credit value for each traffic class is determined in accordance with the full bandwidth for traffic of class A, B or C, valid for entering the RPR ring from the RPR device.

The full bandwidth for Class A, B, or C traffic allowed to enter the RPR ring from a separate RPR device is user-configurable.

The implementation planning steps specifically include:

stage 1, which generates class A traffic in the first level queue in accordance with the credit value for class A traffic, consuming the corresponding credit value;

stage 2, in which the generated frames of traffic data of class A are to be output to the RPR ring;

step 3, in which it is determined whether the credit value for class A traffic has been used up, and if this is not the case, they return to step 1, and if so, they suspend the formation and planning of class A traffic and perform step 4;

step 4, which generates class B traffic in a first level queue in accordance with the credit value for class B traffic, consuming the corresponding credit value;

step 5, in which the generated data frames are planned for class B traffic to be output to the RPR ring;

step 6, in which it is determined whether the credit value for class B traffic has been used up, and if it is not, they return to step 4, and if so, the generation and planning of class B traffic is stopped and step 7 is performed;

step 7, which generates class C traffic in the first level queue in accordance with the credit value for class C traffic, consuming the corresponding credit value;

step 8, where the generated data frames are planned for class C traffic to be output to the RPR ring;

step 9, in which it is determined whether the credit value of class C traffic has been used up, and if it is not, they return to step 7, and if so, the generation and planning of class C traffic is stopped and step 1 is performed.

The present invention, by providing a second level queue for three traffic classes A, B and C, and providing a two-level scheduling mechanism in which scheduling according to the traffic class and then scheduling according to the weight of the port on the respective second level queues is performed, is able to satisfy the priority order so that Class A traffic entering ports is scheduled first, Class B traffic is planned second, and Class C traffic is scheduled last, thus satisfying requirements for delay and delay time spread for Class A traffic and Class B traffic while eliminating the phenomenon of underloading ports if they have the same traffic class. The present invention effectively solves the problem with respect to the planning strategy in the event that traffic of multiple ports of the user side must be included in the RPR ring, and each port of the user side at the same time receives traffic of three classes A, B and C.

List of drawings

FIG. 1 is a schematic representation of an RPR device;

FIG. 2 is a schematic diagram of a traffic planning infrastructure according to the prior art;

FIG. 3 is a schematic representation of multi-level planning according to the present invention;

FIG. 4 is a flowchart of receiving traffic data frames in accordance with the present invention;

FIG. 5 is a flowchart of performing two-time scheduling in second-level queues for corresponding traffic according to the present invention; and

FIG. 6 is a flowchart of implementing implementation planning according to the present invention.

The implementation of the invention

The technical solution according to the present invention will be described below with reference to the accompanying drawings and embodiments.

In FIG. 1 and FIG. 2 shows an introduction to the prior art, which has been described in detail above in the section "Prior art" and will not be repeated in this.

As shown in FIG. 3, in each port of the user side, three queues are provided as part of the RPR device in the form of second-level queues for three traffic classes A, B and C. The data frames received by the user side ports first enter the second level queues for the corresponding traffic classes, and two-level planning is performed, consisting of initial planning according to the traffic class and subsequent planning according to the port weight, on the data frames of various traffic in the second level queues, and the data frames after two-level planning enter the first level queues for traffic of classes A, B and C, and after the formation of the data block, the data frames enter the queue parcel expectations and finally enter the RPR ring.

The full bandwidth for class A traffic, the full bandwidth for class B traffic and the full bandwidth for class C traffic of a single RPR device that are allowed to enter the RPR ring, as well as the minimum bandwidth for class A traffic, the minimum bandwidth for class traffic B, the minimum bandwidth for class C traffic at each port of the user side, valid for entering the RPR ring, can all be configured by users, with the sum of the minimum bandwidth for class A traffic allowed at each port on the user side is equal to the full bandwidth for class A traffic of the RPR device, and the same ratio is true for class B traffic and class C traffic.

The sequence of operations for user side ports receiving data frames is as shown in FIG. 4, where the corresponding user-side port receives a data frame, then classifies the data frame into traffic classes A, B or C in accordance with its class identifier (ID) and places the data frame in the second-level queue for the corresponding traffic class.

The flowchart for two-level scheduling is as shown in FIG. 5. If a separate RPR network node contains N ports of the user side, the planning is carried out first according to the traffic class relative to the first level, that is, the planning is performed in the following order: first, traffic of class A, then traffic of class B and finally traffic of class C one by one . First, one after another, second-level queues are polled for class A traffic for user-side ports 1-N, and if the polled second-level queue for the current user side port is empty, then second-level queues for class A traffic for the next user side port are polled, otherwise, the priority is set for traffic data frames that are in the second level queue for class A traffic of this port of the user side, in the first level queue for class A traffic in tvetstvii with the weight value for the traffic class of the user side port. The weight value for class A traffic is obtained from the relationship between the bandwidth for class A traffic of the port of the user side configured by the user and the bandwidth for class A traffic of the RPR device as a whole. If all the second level queues for Class A traffic have been planned, or the first level queue for Class A traffic is full, then the second level queues are scheduled for Class B traffic. The second level queue planning for Class B traffic is the same as the above method scheduling second-level queues for Class A traffic. If all second-level queues for Class B traffic have been planned, or if the first level queue for Class B traffic is full, schedule s second tier queue for traffic class C. The process of the second level scheduling queues for traffic of class C is the same as the above method the second level scheduling queues for traffic class A. And the above-mentioned steps are repeated in a cycle repeatedly. It should be noted that port weight values for the three traffic classes A, B, and C may be different.

In FIG. 6 is a flowchart of implementation planning, in which conventional RPR implementation planning is performed on first-level queue traffic, including the generation operation and planning process on traffic data, in the following priority order: first, traffic of class A, then traffic of class B and finally class C traffic. As defined in the mentioned standard, class A traffic is further divided into class A0 and class A1, with class A0 being a higher priority, therefore, if it is planned, I’m a class A traffic, planning is performed in the following order: first, class A0 traffic and then class A1 traffic. If traffic of class A0 is planned, class A traffic is first generated in the first level queue in accordance with the credit value for A0, consuming the corresponding credit value A0, and the generated data frames are sent to the send waiting queue to enter the RPR ring. If the credit value A0 is used up, class A1 traffic is planned, and class A traffic in the first level queue is formed in accordance with the credit value A1, consuming the corresponding credit value A1, and then the generated data frames go through the generating unit D, consuming the corresponding credit value D , and the generated data is sent to the send waiting queue to enter the RPR ring. If credit A1 has been used up, class A traffic planning is stopped, and class B traffic is planned, and then class C traffic is planned further, the process for planning class B traffic and class C traffic is the same as that for class A traffic. The credit values for the various traffic are determined according to the corresponding full bandwidth of the RPR device for traffic of classes A, B or C, valid for entering the RPR ring.

Although a method for classifying traffic classes according to the present invention has been illustrated and described, it is obvious that the present invention is not limited thereto. And numerous modifications, substitutions, changes and alternative and equivalent elements will be readily considered by experts in the field of technology without going beyond the essence and scope of the present invention defined in accordance with the attached claims.

Claims (11)

1. The method of multi-level traffic planning, supporting many ports and many services, characterized in that it contains the stages at which
provide second level queues for traffic classes A, B and C at each port of the user side as part of a separate adaptive packet transmission ring (RPR) device, and the data frames received at the user side port must be placed in the corresponding second level queues according to their class identifiers traffic
perform two-level planning, consisting of initial planning according to the class of traffic and subsequent planning according to the weight of the port, in the second level queues for the port of the user side;
enter data frames after two-level planning in the first level queues for traffic classes A, B and C, respectively; and they plan implementation in RPR for the three classes of traffic in the first level queues. 2. The method according to claim 1, characterized in that the planning according to the traffic class is performed in priority order, so that class A traffic takes precedence over class B traffic, and class B traffic gets priority over class C traffic; scheduling according to the weight of the port is performed according to the weights of the respective traffic classes for the respective ports of the user side. 3. The method according to claim 2, characterized in that the weight value is obtained based on the relationship between the minimum bandwidth for traffic of classes A, B or C in the port of the user side and the bandwidth of traffic of the same class in all ports of the user side as part of the RPR device . 4. The method according to claim 3, characterized in that the minimum bandwidths for the three classes A, B and C of traffic that are valid for input into the RPR ring from each port of the user side are user-configurable. 5. The method according to claim 4, characterized in that the values of the weights for different classes of traffic may be different. 6. The method according to claim 1, characterized in that the step of performing two-level planning in the queues of the second level specifically comprises
stage 1, in which second-level queues for class A traffic for user side ports are polled one by one for whether they are empty, and if so, perform stage 2, otherwise go to stage 3;
step 2, in which it is determined whether the given port of the user side is the last port of the user side, and if so, go to step 5, otherwise return to step 1, in which the second level queue for traffic class A for the next user side port is polled;
step 3, in which, in accordance with the weight value for the class A traffic of the port of the user side, the priority is set for class A traffic in the second level queue in the first level queue for class A traffic;
step 4, in which it is determined whether the first level queue for traffic class A is full, and if it is not, then go to step 2, and if so, then go to step 5;
step 5, in which second-level queues for traffic class B for ports of the user side are polled one by one for whether they are empty, and if so, then perform step 6, otherwise go to step 7;
step 6, in which it is determined whether the given user side port is the last user side port, and if so, go to step 9, otherwise return to step 5, in which the second level queue for traffic class B is polled for the next user side port ;
step 7, in which, in accordance with the weight value for the class B traffic of the port of the user side, the priority is set for class B traffic in the second level queue in the first level queue for the class B traffic;
step 8, in which it is determined whether the first level queue for traffic class B is full, and if it is not, then go to step 6, and if it is, then go to step 9;
step 9, in which a survey is performed as to whether the second level queues for traffic class C for ports of the user side are empty, and if so, then step 10 is performed, otherwise go to step 11;
step 10, in which it is determined whether the given port of the user side is the last port of the user side, and if so, go to step 1, otherwise return to step 9, in which the second level queue for traffic class C of the next port of the user side is polled;
step 11, in which, in accordance with the weight value for the class C traffic of the port of the user side, the order is established for the traffic of class C in the second level queue in the first level queue for the class C traffic; and
step 12, in which it is determined whether the first level queue for traffic class C is full, and if it is not, then go to step 10, and if it is, then go to step 1. 7. The method according to claim 1, characterized in that the stage of implementation planning for three classes of traffic in the first level queues further comprises the steps of forming three classes of traffic, respectively, and then planning data frames for the respective traffic to be output to the corresponding ring RPR 8. The method according to claim 7, characterized in that the formation and planning are performed in such a way that class A traffic takes precedence over class B traffic, and class B traffic takes precedence over class C traffic. 9. The method according to claim 8, characterized in that the implementation planning step specifically comprises
stage 1, which generates class A traffic in the first level queue in accordance with the credit value for class A traffic, consuming the corresponding credit value;
step 2, in which the generated data frames are planned for class A traffic to be output to the RPR ring;
stage 3, in which it is determined whether the credit value for class A traffic has been used up, and if this is not so, then they return to stage 1, and if so, then the formation and planning of class A traffic is stopped and stage 4 is performed;
step 4, in which class B traffic is generated in a first level queue in accordance with the credit value for class B traffic, consuming the corresponding credit value;
step 5, in which the generated data frames are planned for class B traffic to be output to the RPR ring;
step 6, in which it is determined whether the credit value for class B traffic has been used up, and if it is not, then they return to step 4, and if so, then the generation and planning of class B traffic is stopped and step 7 is performed;
step 7, which generates class C traffic in the first level queue in accordance with the credit value for class C traffic, consuming the corresponding credit value;
step 8, in which the generated data frames for class C traffic are scheduled to be output to the RPR ring;
step 9, in which it is determined whether the credit value for class C traffic has been used up, and if it is not, then return to step 7, and if so, then the generation and planning of class C traffic is stopped and step 1 is performed. 10. The method according to claim 9, characterized in that the credit value for each class of traffic is determined in accordance with the full bandwidth of class A, B or C traffic allowed to enter the RPR ring from the RPR device. 11. The method according to claim 10, characterized in that the full bandwidth for traffic of class A, B or C, valid for entering the RPR ring from a separate RPR device, is user-configurable.

Images