CN1878167A - Method for ensuring service quality in multi protocol label switching protocol - Google Patents

Abstract

The invention discloses a method for guaranteeing service quality in a multi-protocol label switching protocol. Including: (1) Create three queues for EF, AF, and BF LSP messages, each queue contains several queues; (2) Create three schedulers, and hang the three queues of EF, AF, and BF to the three schedulers (3) allocate scheduling resources for the port; (4) judge whether each queue under the EF scheduler has messages to send, if not, go to step (5); if so, go to step (7); (5) judge Whether each queue under the AF scheduler has a message to send, if not, go to step (6); if so, go to step (7); (6) judge whether each queue under the BF scheduler has a message to send, if not, End; if so, go to step (7); (7) determine the queue scheduling that needs to be scheduled until all queues that have messages are scheduled, and then execute the judgment of the queue in the next priority scheduler.

Description

A Method of Guaranteeing Quality of Service in Multiprotocol Label Switching Protocol

technical field

The present invention relates to the field of multiprotocol label switching protocol MPLS (Multiprotocol Lable Switch), in particular to a method for guaranteeing service quality in the multiprotocol label switching protocol.

Background technique

With the development of the Internet, the amount of data communication is growing explosively. To accommodate this growth, communication networks are undergoing various technological innovations. Due to the popularity of the Internet, IP applications are increasingly widespread, and IP networks have penetrated into various traditional communication areas. It is possible to build a multi-service network based on IP. However, different services have different requirements on the network. For example, services such as voice and video have very high requirements on QoS (Quality of Service). How to implement multiple real-time and non-real-time services in a packetized IP network has become an important topic. People have proposed the concept of QoS, which provides bandwidth guarantee and low delay. Due to the best-effort mechanism of the IP network itself, MPLS technology is introduced to ensure QOS.

At present, to ensure the quality of service in the multi-protocol label switching protocol, a relatively mature solution is to use DiffServ (differentiated service), that is, to monitor and classify the service flow at the edge of the network, and mark the message according to the classification result. That is, different label switching paths (LSPs) are selected, and queue scheduling is performed on the intermediate LSR device according to the priority indicated by the experimental domain (EXP domain) in the MPLS label. For the business of a virtual private network (VPN), the data transmitted between different sites passes through the backbone network in the form of a label switching path (LSP). Therefore, the QoS control between the edge equipment (PE equipment) and the backbone network can be realized based on the label switching path (LSP). However, the specific definition of the EXP value is currently only a draft, and different manufacturers may have different understandings of the EXP value. Therefore, using the EXP value for priority scheduling has certain risks in connection.

Contents of the invention

The technical problem to be solved by the present invention is to provide a method for guaranteeing service quality in the multi-protocol label switching protocol. The method of the invention can provide bandwidth guarantee for different data communication services and reduce time delay.

In order to solve the problems of the technologies described above, the present invention provides the following solutions:

A method for ensuring quality of service in a multi-protocol label switching protocol, comprising the steps of:

Step 1: Create three queues for fast forwarding, guaranteed forwarding, and best-effort forwarding three types of label-switched path packets, and each queue contains one or more queues;

Step 2: Create three schedulers with different priorities, and hang the three queues respectively corresponding to the fast forwarding, guaranteed forwarding, and best-effort forwarding three types of label switching path messages under the three schedulers;

Step 3: Allocate scheduling resources for the port;

Step 4: Determine whether there is a message to be sent in each queue under the scheduler corresponding to the fast forwarding label switching channel message, if not, perform step 5; if exist, perform step 7;

Step 5: Determine whether there is a message to be sent in each queue under the scheduler corresponding to the guaranteed forwarding label switching channel message. If there is no message, perform step 6; if it exists, perform step 7;

Step 6: Determine whether there is a message to be sent in each queue under the scheduler corresponding to the best-effort forwarding label switching channel message. If there is no message, do no processing and end; The judgment of the queue in the priority scheduler;

Step 7: Find the queues of all messages, determine the queues that need to be scheduled for scheduling, and send them out from the port until all the queues of existing messages are scheduled.

In the method for ensuring the quality of service in the multi-protocol label switching protocol of the present invention, a step is added between the first step and the second step: a leaky bucket is created for the label switching path queue, and when the message passes through the label switching path First, the rate limit is performed through the leaky bucket; if the node where the label switching path is located is an edge node, the leaky bucket takes effect and the rate limit is performed; if the node where the label switching path is located is a core node, the leaky bucket does not take effect, No rate limiting is performed.

In the method for guaranteeing the quality of service in the multi-protocol label switching protocol of the present invention, the maximum bandwidth limit parameter and the burst rate parameter of the label switching path message are added into the leaky bucket.

In the method for guaranteeing the quality of service in the multi-protocol label switching protocol of the present invention, the three schedulers in the second step are performed according to the absolute priority of fast forwarding scheduler>guaranteed forwarding scheduler>best effort forwarding scheduler scheduling.

In the method for guaranteeing the quality of service in the multi-protocol label switching protocol of the present invention, the allocation of scheduling resources for the port in the step 3 is to ensure that the port scheduling bandwidth is greater than or equal to the packet bandwidth of all fast forwarding label switching paths and Guaranteed Forwarding The sum of the packet guaranteed bandwidth of the label-switched path.

According to the method for guaranteeing service quality in the multi-protocol label switching protocol of the present invention, in the step seven, if the current scheduler is the fast forwarding scheduler, the next priority scheduler is the guaranteed forwarding scheduler.

In the method for guaranteeing service quality in the multi-protocol label switching protocol of the present invention, in step seven, if the current scheduler is a guaranteed forwarding scheduler, the next priority scheduler is a best-effort forwarding scheduler.

In the method for guaranteeing the quality of service in the multi-protocol label switching protocol of the present invention, in the step 7, if the current scheduler is a best-effort forwarding scheduler, the judgment operation of the queue in the next priority scheduler is not performed, Finish.

In the method for guaranteeing the quality of service in the multi-protocol label switching protocol of the present invention, in the step seven, the current need for scheduling is determined according to the weighted round-robin scheduling algorithm among multiple queues in the same scheduler queue for scheduling.

In the method for guaranteeing the quality of service in the multi-protocol label switching protocol of the present invention, the scheduling weight of the scheduler corresponding to the guaranteed forwarding label switching path message is set according to the guaranteed bandwidth value of the label switching path message.

Compared with the prior art, the present invention has the advantages of:

The method for ensuring quality of service in the multi-protocol label switching protocol of the present invention does not need to judge the EXP value, and uses the same mechanism at the edge node (PE node) and the core node (P node) to directly set different QOS for LSP Different scheduling is performed to provide bandwidth guarantees and reduce delays for different services.

The technical problems to be solved by the present invention, the key points and advantages of the technical solutions will be further described in conjunction with the embodiments and with reference to the accompanying drawings.

Description of drawings

Figure 1 is a flow chart of the method of the present invention.

Fig. 2 is a schematic diagram of the scheduling of ports, schedulers, and queues when sending packets according to the present invention.

Figure 3 shows the processing procedure when creating an LSP.

Detailed ways

With reference to Fig. 1, the flowchart of the method of the present invention.

A method for ensuring quality of service in a multi-protocol label switching protocol, comprising the steps of:

Step 1. Create three queues for fast forwarding, guaranteed forwarding, and best-effort forwarding three types of label-switched path messages, and each queue includes one or more queues;

Step 2. Create a leaky bucket for the label switching channel queue. When the packet passes through the label switching channel, it first passes through the leaky bucket to limit the rate;

Here, we add parameters such as the maximum bandwidth limit parameter and the burst rate of the label switching path message into the leaky bucket.

If the node where the label switching path is located is an edge node, the leaky bucket takes effect, and rate limitation is performed; if the node where the label switching path is located is a core node, the leaky bucket is not effective, and rate limitation is not performed.

Step 3, create three schedulers with different priorities, hang fast forwarding, guaranteed forwarding, and best-effort forwarding three kinds of queues respectively corresponding to three types of label switching path messages under the three schedulers respectively;

Here, the three schedulers perform scheduling according to the absolute priority of fast forwarding scheduler>guaranteed forwarding scheduler>best effort forwarding scheduler.

Step 4, allocate scheduling resources for the port;

Here, the allocation of scheduling resources for the port is to ensure that the port scheduling bandwidth is greater than or equal to the sum of the packet bandwidths of all fast-forwarding label-switching paths and guaranteed forwarding-type label-switching path packets. For example, the port scheduling capability is 100M.

Step 5, judging whether there is a message to be sent in each queue under the scheduler corresponding to the fast forwarding label switching path message, if not, then perform step 6; if exist, then perform step 8;

Step 6, judging whether there is a message to be sent in each queue under the scheduler corresponding to the guaranteed forwarding label switching path message, if there is no message, then perform step 7; if it exists, then perform step 8;

Step 7, judging whether there is a message to be sent in each queue under the scheduler corresponding to the best-effort forwarding label switching path message, if not, then do not do any processing, end; if exist, then perform step 8;

Step 8. Find the queues of all messages, determine the queues that currently need to be scheduled for scheduling, and send them out from the port until all queues with messages are dispatched, and then execute the judgment on the queues in the next priority scheduler.

Here, if the current scheduler is the fast forwarding scheduler, the next priority scheduler is the guaranteed forwarding scheduler; if the current scheduler is the guaranteed forwarding scheduler, the next priority scheduler is the best effort forwarding scheduler ; If the current scheduler is a best-effort forwarding scheduler, then do not perform the judgment operation on the queue in the next priority scheduler, and end.

In the above step 8, the queue to be scheduled is determined by the weighted round-robin scheduling algorithm among multiple queues in the same scheduler.

For example, the scheduling weight of the scheduler corresponding to the fast-forwarding label-switched channel packet is set to 1. The scheduling weight of the scheduler corresponding to the guaranteed-forwarding label-switched path message is set according to the guaranteed bandwidth value of the label-switched path message. The scheduling weight of the scheduler corresponding to the best-effort forwarding label-switched path packets is set according to the limit bandwidth value of the label-switched path packets.

We know that according to different QOS settings of LSP: it can be divided into three forwarding behavior sets: fast forwarding (EF), used for real-time multimedia services, such as voice, video, etc.; guaranteed forwarding (AF), used for high reliability Businesses such as e-commerce and VPN user data; best effort forwarding (BF) for general Internet services.

Among them, EF-type LSPs implement bandwidth limitation and prioritize forwarding to ensure low latency; AF-type LSPs have a restricted bandwidth and guaranteed bandwidth; BF-type LSPs have only a restricted bandwidth and have the lowest scheduling priority and cannot perform bandwidth guarantees.

Referring to Figure 2, three-level scheduling is performed according to ports, schedulers, and queues. The schedulers are scheduled according to the absolute priority relationship, that is, if there is a message to be sent in the EF-type LSP transmission queue, the EF-type LSP message is sent first. The second is the message of AF type, and the third is the message of BF type LSP. For PE nodes, there is a leaky bucket in front of the queue, which limits the rate of each LSP, and only when it meets the bandwidth setting can the message be stored in the queue; but for P nodes, this leaky bucket does not work.

Referring to Figure 3, when creating an LSP, first create a queue for this LSP, forward the message to this LSP into this queue, and hang this queue under the corresponding scheduler according to the QOS attribute of the LSP, such as LSP is EF class, the queue is hung under the scheduler A corresponding to the EF class LSP; when the LSP is AF class, the queue is hung under the scheduler B corresponding to the AF class LSP; when the LSP is BF class, the queue is hung under the BF class LSP Under the corresponding scheduler C.

Create a leaky bucket for this LSP, and fill it into the leaky bucket according to the maximum bandwidth limit parameter and burst rate parameter of the LSP. When the message enters the LSP, it first passes through the leaky bucket to limit the rate. If the LSP is at this node If it is a P node, then the leaky bucket does not take effect, and no rate limit is performed.

Since there are multiple queues under the scheduler, the scheduler schedules the linked queues according to the weighted round-robin (WRR) algorithm, so a weight must be specified for this queue. If it is an EF-type LSP, the weight is 1; if it is an AF-type LSP For an LSP, the weight is the value of the guaranteed bandwidth. If it is a BF type LSP, the weight is the value of the limited bandwidth.

When the port obtains scheduling resources, it first schedules the queues under the scheduler A corresponding to the EF LSP, finds all the queues under the scheduler A that have packets, and finds out a queue to be scheduled according to the weighted round-robin scheduling WRR algorithm. queue, send the packets in this queue, and this scheduling ends. If there are no packets in all the queues under the scheduler A corresponding to the EF LSP, then it is necessary to schedule the queues under the scheduler B corresponding to the AF type LSP, and find out all the queues under the scheduler B that have packets. Similarly, according to The WRR algorithm finds a queue to be scheduled, sends the packets in this queue, and the scheduling ends. If there are no packets in all the queues under scheduler A corresponding to EF LSP and scheduler B corresponding to AF LSP, then it is necessary to schedule the queues under scheduler C corresponding to BF LSP to find out all queues under this scheduler C. For queues with messages, similarly, according to the WRR algorithm, a queue to be scheduled is found, and the messages in this queue are sent out, and the scheduling ends.

A method for ensuring quality of service in the multi-protocol label switching protocol described in the present invention is not limited to the use listed in the specification and implementation, it can be applied to various fields suitable for the present invention, and those who are familiar with the present invention Additional advantages and modifications will readily occur to those skilled in the art, so the invention is not to be limited to the specific details, representative examples, and the like without departing from the spirit and scope of the general concept defined by the claims and equivalents thereof. equipment and illustrated examples shown and described herein.

Claims (10)

1. A method for ensuring quality of service in a multi-protocol label switching protocol, characterized in that it comprises the following steps: (1), create three kinds of queues for fast forwarding, guaranteed forwarding, and best effort forwarding three types of label switching path messages, and each queue includes one or more queues; (2), create three schedulers with different priorities, hang fast forwarding, guaranteed forwarding, and best-effort forwarding three kinds of queues corresponding to three types of label switching path messages respectively under the three schedulers; (3) Allocate scheduling resources for the port; (4), judge whether there is a message to be sent in each queue under the dispatcher corresponding to the fast forwarding class label switching path message, if not exist, then perform step (5); if exist, then perform step (7); (5), judging whether there is a message to be sent in each queue under the dispatcher corresponding to the forwarding class label switching path message, if not, then perform step (6); if exist, then perform step (7); (6), judge whether there is a message to be sent in each queue under the dispatcher corresponding to the best-effort forwarding class label switching path message, if not exist, then do not do any processing, end; If exist, then perform step (7); (7), find the queues of all messages, determine the queues that currently need to be scheduled for scheduling, and send them out from the port until all queues that have messages have been scheduled, and then execute the judgment of the queues in the next priority scheduler. 2. The method for ensuring quality of service in an MPLS protocol according to claim 1, characterized in that: a further step is added between said steps 1 and 2: creating a leaky bucket for the label switching path queue, When the message passes through the label switching path, it first passes through the leaky bucket for rate limiting; if the node where the label switching path is located is an edge node, the leaky bucket takes effect and performs rate limitation; if the node where the label switching path is located is a core node, then The leaky bucket does not take effect, and no rate limiting is performed. 3. The method for guaranteeing service quality in an MPLS protocol according to claim 2, characterized in that: the maximum bandwidth limit parameter and the burst rate parameter of the Label Switching path message are added to the leaky bucket. 4. The method for guaranteeing quality of service in an MPLS protocol according to claim 1, characterized in that: said three schedulers in said step (2) are fast forwarding scheduler>guaranteed forwarding scheduler> The absolute priority of the best-effort forwarding scheduler is used for scheduling. 5. The method for guaranteeing quality of service in an MPLS protocol according to claim 1, characterized in that: in said step (3), allocating scheduling resources for a port is to ensure that the port scheduling bandwidth is greater than or equal to all fast The sum of the packet bandwidth of the forwarding label-switched path and the guaranteed bandwidth of the packet of the guaranteed forwarding label-switched path. 6. The method for guaranteeing quality of service in an MPLS protocol according to claim 1, characterized in that: in said step (7), if the current scheduler is a fast forwarding scheduler, the next priority scheduling The scheduler is the guaranteed forwarding scheduler. 7. The method for guaranteeing quality of service in an MPLS protocol according to claim 1, characterized in that: in the step (7), if the current scheduler is a guaranteed forwarding scheduler, the next priority scheduling The scheduler is a best-effort forwarding scheduler. 8. The method for guaranteeing quality of service in an MPLS protocol according to claim 1, characterized in that: in said step (7), if the current scheduler is a best-effort forwarding scheduler, the following The judgment operation of the queue in the priority scheduler ends. 9. The method for ensuring quality of service in an MPLS protocol according to claim 1, characterized in that: in the step (7), multiple queues in the same scheduler are arranged according to weighted The round-robin scheduling algorithm is used to determine the queues that currently need to be scheduled for scheduling. 10. The method for guaranteeing quality of service in a multi-protocol label switching protocol according to claim 10, characterized in that: the scheduling weight of the scheduler corresponding to the guaranteed forwarding label switching path message is based on the guaranteed bandwidth of the label switching path message value to set.

Images