KR100720917B1 - Differential Active Queue Management Method for Guaranteeing Multimedia Data Usage - Google Patents

Abstract

 The present invention relates to a queue management method that can guarantee the quality of service of multimedia data by managing a queue in consideration of traffic characteristics.

The present invention provides a method for managing a queue of multimedia data, wherein a differential active queue that preferentially guarantees bandwidth for first traffic that should not be interrupted in data flow as compared to second traffic requiring an excess band. Provide management methods.

RED, traffic, active queue, multimedia

Description

Differential Active Queue Management Method for Guaranteeing Multimedia Data QoS {Method of adaptive multi-queue management to guarantee QoS}

1 is a diagram illustrating a RED algorithm,

2 is a conceptual diagram of a strict priority method;

3 is a diagram illustrating a differential active queue management method according to an embodiment of the present invention;

4 is a diagram illustrating a network topology used for an experiment;

5A and 5B are graphs showing the results of comparative experiments.

The present invention relates to a queue management method for multimedia data, and to a queue management method capable of guaranteeing a quality of service of multimedia data by managing a queue in consideration of traffic characteristics.

Another object of the present invention is to provide a differential active queue management method for minimizing packet loss by discarding packets in consideration of the characteristics of traffic in each queue in a multi-queue.

As applications diversify and the use of the Internet increases, the demand for quality of service guarantees for users is increasing.

In order to guarantee the quality of service in IP network, network structures such as IntServ (Integrated Services), DiffServ (Differentiated Service), MPLS (Multi Protocol Label Switching), IEEE 802.1p / Q have been proposed, and RSVP (Resource Reservation Protocol), RTP Protocols such as Real Time Protocol have been proposed.

In addition, researches on quality assurance of multimedia data are being conducted in various fields.

However, the current Internet provides the best effort service that delivers all packets equally, and thus does not guarantee the requirements according to the characteristics of the service.

In particular, network application service types such as large-capacity high-speed multimedia data transmission service, video on demand (VoD), and VoIP require high bandwidth, thereby increasing the possibility of congestion in the network. Needs to be increased.

In order to avoid such congestion of network resources, a method of efficiently managing big pipes and network resources is proposed.

In this case, the big pipe method provides a sufficient bandwidth to the network, thereby avoiding congestion and ensuring quality of service. However, this is a simple method, but the large capacity of the system is expensive and time-consuming.

Therefore, there is a need for a differentiated queue management method that can guarantee the quality of service to users by classifying packets according to the characteristics of multimedia data services and minimizing packet loss, rather than the big pipe method.

For example, when congestion occurs in a network router, an overflow occurs in a queue, and a method of how to discard a packet is required. Queue management provides a way to discard packets for each queue.

The interest in this queue management method was due to the tail-drop algorithm and the issue of fairness for each link.

The tail-drop algorithm discards all incoming packets when a queue overflow occurs, thereby causing lock-out and global synchronization. In this case, this phenomenon causes the performance and the network equipment to become unstable due to the sudden fluctuation of the traffic volume.

In order to solve this problem, a BLUE algorithm using a RED (Random Early Detection) algorithm and link utilization, which can be solved by selecting a TCP flow to reduce the transmission speed, has emerged.

The RED is an algorithm for keeping the average queue size low, and is a method of reducing delay of a queue rather than throughput.

1 is a diagram for explaining such a RED algorithm.

As shown in Figure 1, the RED algorithm has an average queue size of TH _min If smaller, queue the packet.

In addition, when the average queue size is between TH _min and TH _max , the packet is discarded with a specific probability value according to the queue size.

In addition, when the average queue size is larger than TH _max , it is determined that congestion is severe, and all received packets are discarded before entering the queue. That is, as the degree of congestion increases, the algorithm attempts to reduce the amount of incoming traffic by discarding many packets.

 Therefore, if congestion is likely to occur while maintaining the average queue size, the TCP congestion avoidance function operates by discarding the packet with a discarding probability.

On the other hand, a priority method called strict priority queuing (Strict Priority Queueing), instead of using a single queue, it uses multiple queues.

2 is a conceptual diagram of a strict priority scheme.

As shown in FIG. 2, the strict priority scheme uses multiple first-in, first-out queues, so that each queue is mapped to a different traffic class.

Using this priority method, the scheduling method is very simple. That is, packets stored in the low priority queue become service after all the packets stored in the high priority queue are serviced.

However, the above-described RED and priority schemes in order to guarantee a quality of service (QoS) have certain limitations in the present situation in which network applications are diversified and differentiated quality of service is required for each traffic.

Disclosure of Invention The present invention has been made to solve the above-mentioned problems, and manages a queue in a manner of applying differential time weights in consideration of the characteristics of traffic, thereby enabling differentiated active queue management to guarantee the quality of service of multimedia data. It is an object to provide a method.

Another object of the present invention is to provide a differential active queue management method for minimizing packet loss by discarding packets in consideration of the characteristics of traffic in each queue in a multi-queue.

In order to achieve the above object, the present invention, in the queue management method of multimedia data, after scanning the queue of the multimedia data, obtain the state and length of the queue, the maximum of the voice traffic or video traffic in the queue Applying a maximum threshold to be less than a queue length, and applying a tail-drop algorithm if the queue exceeds the maximum threshold; It is characterized in that the bandwidth for the voice traffic or video traffic in the queue, preferentially compared to the web data or file transfer data.

delete

delete

In another preferred embodiment, in a class using multi-queues for multimedia data, the packet interval time is preferentially guaranteed for a first queue in which data flow should not be interrupted in the multi-queues, followed by packet intervals for the remaining queues. A differential active queue management method using differential time weights that allocates time is provided.

In addition, for the queue that requires the quality of the surplus band of the remaining queue, the packet that does not receive the service for a predetermined time is characterized in that discarded from the long wait in the queue.

The first cue is audio data, and the remaining cues are video data and web data.

In addition, a packet interval time of 5 ms is applied to the audio data, α ms is applied to the video data, and (5-α) ms is applied to the web data.

Where 0 ≦ α <5.

Hereinafter, preferred embodiments of the present invention will be described in detail with reference to the accompanying drawings.

3 is a diagram illustrating a differential active queue management method according to an embodiment of the present invention.

Referring to FIG. 3, the differential active queue management method according to an embodiment of the present invention shows that scheduling is performed by applying differential time weights to a class using a multi-queue for multimedia data.

First, a large amount of multimedia data is classified into data that data flow should not be interrupted and data that is not.

That is, audio data Q1 and image data Q2 may correspond to data that data flow should not be interrupted, and data such as web data or file transfer may correspond to data Q3 that is not.

After that, since voice data is not interrupted when the packet interval (inter-arrival time) is 5ms, the voice data queue (Q1) is guaranteed at least 5ms as the packet interval time.

The queue Q2, which is video data, is also video data whose bandwidth is to be guaranteed, thereby ensuring αms as the packet interval time.

For the data queue Q3 where the data flow may be interrupted, a packet interval time of (5-?) Ms is guaranteed. That is, since the web data Q3 requires the quality of the redundant band, the packet interval time is given low. Thus, starvation may occur because web data has a low time weight.

The solution is to discard packets that have not been serviced for a certain period of time starting from a long queue.

If the queue is empty, the time weight is passed to another queue. If the queue is not empty, the packet interval time of 5ms, Q2 is αms, and Q3 is (5-α) ms is applied. That is, after sending the audio data, α ms is applied to the video data and (5-α) ms time weight is applied to the web data. The range of values here is 0 ≦ α <5.

As such, the differential active queue management method according to an embodiment of the present invention allocates time to the video data and the web data after guaranteeing the priority of the audio data, which ensures a certain bandwidth or more, which is a characteristic of the video data. Allocates time considering the characteristics of web data.

The following is a queue scheduling algorithm as an example of the differential active queue management method of FIG. 3.

...

if (subq_empty (0) == FALSE) packet_send (0);

time (current_time);

current_time_first = current_time + α;

while (current_time <current_time_first)

{

 if (subq_empty (1) == TRUE) {

    time (current_time); break;

} else {

  packet_send (1); time (current_time);

  }

}

...

Referring to this, after scanning the multimedia data queue, the queue state and length are obtained, and the maximum threshold TH _max is applied to be smaller than the maximum queue length of Q1.

If we exceed the maximum threshold of the queue, we apply a tail-drop algorithm that discards all packets, and Q2 guarantees maximum throughput.

In addition, Q3 gives a low time weight because the end-to-end transmission time of web data is guaranteed to be 500 to 800 ms.

As described above, the differential active queue management method according to an embodiment of the present invention provides a surplus band such as web data or file transmission for traffic of a special type such that voice and video data should not be interrupted. This algorithm assigns and controls network traffic by class so that it has priority over traffic of quality that requires.

Hereinafter, a result of comparing and comparing the case where RED is applied to each queue of multimedia data and the case of applying the differential active queue management method according to an embodiment of the present invention will be described with reference to FIGS. 4 to 5. .

Figure 4 shows the network topology used for the experiment, Figures 5a and 5b is a graph showing the results of the comparative experiment, Figure 5a shows the bandwidth before and after the bottleneck, Figure 5b is a differential active queue algorithm Application results are shown.

Here, the voice traffic is guaranteed at least 5ms, the video traffic is αms in consideration of the bandwidth, the web data is applied to the value (5-α) ms.

In addition, the link corresponding to the bottleneck section is set to 2Mbps, the remaining link is set to 3.2Mbps. As the queue policy of the bottleneck section, RED and differential active queue management method according to an embodiment of the present invention are applied.

As shown in FIG. 5B, it can be seen that the packet loss is less when the differential active queue algorithm considering the traffic of the network is applied, rather than the same result of applying the RED algorithm to the multimedia data queue.

Therefore, packet loss can be reduced by applying a differential active queue algorithm considering the characteristics of traffic for each queue.

While the differential active queue management method according to the present invention has been described above, if the differential active queue algorithm is applied to a QoS device in the future, it is expected that the differential active queue management method can be efficiently applied to network traffic management and operation.

In addition, it will be possible to extend the output flow control method as well as the queue policy in the future.

Therefore, the present invention is not limited to the above-described embodiments, and a person having ordinary skill in the art may change the design or avoid the design without departing from the scope of the technical idea of the present invention. Will be in range.

As described above, the present invention provides a simple and efficient differential queue scheduling suitable for application in a multi-queue by applying a differential active queue algorithm in consideration of traffic characteristics of each queue.

This ensures multimedia data QoS.

Claims (7)

In the queue management method of multimedia data, After scanning the queue of multimedia data, obtaining a state and a length of the queue, applying a maximum threshold value smaller than the maximum queue length of voice traffic or video traffic in the queue; Applying a tail-drop algorithm if the queue exceeds the maximum threshold;  Differential active queue management method, characterized in that for the voice traffic or video traffic in the queue, bandwidth is preferentially guaranteed compared to web data or file transfer data. In a class that uses multi-queues for multimedia data, Differential active, characterized in that for the first queue that the flow of data in the multi-queue should not be interrupted preferentially guarantee the packet interval time, and then apply a differential time weight to allocate the packet interval time to the remaining queues How to manage queues. The method of claim 2, And among the remaining queues, packets that do not receive service for a predetermined time are discarded from the long queued queues for the queues requiring the quality of the surplus band. The method of claim 2, And the first queue is voice data, and the remaining queues are video data and web data. The method of claim 4, wherein And a packet interval time of 5 ms for the audio data, α ms for the video data, and (5-α) ms for the web data. Where 0 ≦ α <5. delete delete

Images