JP5597094B2 - Quality class control method - Google Patents

Description

  The present invention relates to a quality class control method, and more particularly, to a quality class control method that can reduce processing load when a node on a network schedules packet transfer for each quality class.

  In Non-Patent Documents 1, 2, and 3, target quality is set for each of a plurality of quality classes, and each node is provided with a plurality of different packet queues for each quality class. A technique for satisfying the target quality in each quality class by calculating the number of packets that should be continuously transmitted from each queue to the outgoing link based on the total packet length and the like staying in the queue is disclosed.

AK Parekh and RG Gallager, "A generalized processor sharing approach to flow control in integrated services networks: The single-node case," IEEE / ACM Trans. On Networking, Vol. 1, No. 3, pp. 344-357, June 1993. M. Shreedhar and G. Varghese, "Efficient fair queuing using deficit round-robin," IEEE / ACM Trans. On Networking, Vol. 4, No. 3, pp. 375-385, June 1996. S. S. Kanhere, H. Sethu, and A. B. Parekh, "Fair and efficient packet scheduling using elastic round robin," IEEE Trans. On Parallel and Distributed Systems, Vol. 13, No. 3, pp. 324-336, March 2002.

  When each node has packet queues for the number of quality classes and schedules the reception of input packets in each queue and the reading of output packets from each queue, for example, the nth quality class Cn and the (n + 1) th If there is no substantial difference in the quality achieved with the quality class Cn + 1, the scheduling processing load can be reduced by integrating these two quality classes into one.

  However, in the conventional quality class control method, since the number of quality classes and the number of corresponding packet queues in each node are fixed, they can be integrated even if the realization qualities of two adjacent quality classes are equal. There is a technical problem that an unnecessary scheduling processing load occurs.

  An object of the present invention is to solve the above-described problems of the prior art and to provide a quality class control method that can reduce the processing load when a node on a network schedules packet transfer for each quality class.

  In order to achieve the above object, the present invention is characterized in that the following measures are taken in a quality class control method for scheduling packet transfer based on the quality required for each packet.

  (1) A plurality of packet queues provided for each quality class having different target qualities, a packet distribution unit that accommodates input packets in a packet queue corresponding to the required quality, and a packet accommodated in each packet queue, Implemented for packet transfer units that transfer with priority according to the quality class of each packet queue, means for detecting the actual quality of each quality class, and two adjacent quality classes whose target quality is relatively high and low Quality class integration means for integrating a lower class whose quality satisfies the target quality of the higher class into the higher class, and the quality class integration means classifies the packets accommodated in the lower class packet queue to the higher class. Was accommodated in the packet queue.

  (2) Provided with a quality class separation means for separating a quality class having a relatively low target quality from a quality class in which the realized quality cannot satisfy the target quality, and the quality class separation means satisfies the target quality. Packets that have been accommodated in the quality class packet queue that can no longer be accommodated in the separated quality class packet queue according to the required quality.

  (3) Means for increasing and correcting the target quality of the separated quality class within a range lower than the target quality of the quality class that cannot satisfy the target quality is provided.

  According to the present invention, the following effects are achieved.

  (1) If there is no substantial difference in the realization quality of packets classified into two adjacent quality classes whose target quality is relatively high and low, these two quality classes are integrated into one, It becomes possible to reduce the processing load when a node on the network schedules packet transfer by quality class.

  (2) After integrating two adjacent quality classes whose target quality is relatively high and low, if the realized quality of the integrated quality class cannot satisfy the target quality of the higher quality class before integration, Since relatively low quality classes are separated and packets with low required quality are accommodated in the packet queue of the low quality class, the realization quality of the high quality class can be improved.

  (3) After separating the lower quality class (child) from the integrated quality class, the target quality of the lower quality class (child) is increased and corrected if the realized quality of the higher quality class (parent) is not improved. Since more packets are distributed to the lower quality class (child), the realization quality of the higher quality class (parent) can be improved.

It is the block diagram which showed the structure of the network node to which the quality class control system of this invention is applied. FIG. 5 is a diagram showing a relationship among quality class C, target quality TQ, required quality RQ, and realized quality SQ in the present embodiment. It is the time chart which showed the operation | movement of one Embodiment of this invention.

  Hereinafter, embodiments of the present invention will be described in detail with reference to the drawings. FIG. 1 is a block diagram showing the configuration of a network node to which the quality class control method of the present invention is applied. In this embodiment, packets are divided into n stages of quality classes C1, C2,. Queue PCs (PC1, PC2,... PCn) are provided.

  The packet distributor 1 accommodates the input packet in the packet queue PC corresponding to the required quality RQ. The packet transfer unit 2 reads out the packets accommodated in each packet queue PC with a priority according to the quality class C of each packet queue PC and transfers it to the next node. The quality class review unit 3 integrates or separates the packet queue PCs based on the realized quality SQ of each quality class C packet.

  FIG. 2 is a diagram showing the relationship among “quality class C”, quality class “target quality TQ”, packet “required quality RQ”, and “realized quality SQ” in the present embodiment.

  The required quality RQ of the packet is the quality required by the node or network for each packet. In this embodiment, the required quality RQ is appropriate for each packet according to the application and user attributes (contract conditions, service contents, etc.). The required quality RQ is preset.

  When the required quality RQ is set according to the application, a high required quality is set for a packet with a high real-time property such as VoIP or streaming playback, and a packet with a low real-time property such as an e-mail is set. A low required quality is set, and a medium required quality is set in an intermediate application packet like a Web browser. Further, when the required quality RQ is set according to the user attribute, a higher required quality is set for a packet transmitted and received by a user who has contracted for a service with a higher fee.

  The target quality TQ is a target value of the quality promised for the packets classified in each quality class C, and the quality of the target quality TQ1 or higher is promised for the packets classified in the highest quality class C1. Similarly, packets classified as quality class C2 are promised quality of target quality TQ2 or higher, and packets classified as quality class C3 are promised quality of target quality TQ3 or higher and are classified as quality class C4 Packets are guaranteed to have a target quality of TQ4 or better.

  The realization quality SQ is the actual quality of each packet on the network. If the network or node is free, it can exceed the target quality TQ as shown by the solid line in FIG. If the node is congested, the target quality TQ may be lower than the target quality TQ, as indicated by a broken line.

  In the example of FIG. 2, the packet of the required quality RQ1 is classified into the quality class C1, and the packet transfer is scheduled so that the realized quality SQ exceeds the target quality TQ1. Similarly, packets of required quality RQ2 and RQ3 are classified into quality class C2, and packet transfer is scheduled so that the realized quality SQ exceeds the target quality TQ2. The packets of required quality RQ4 and RQ5 are classified into quality class C3, and packet transfer is scheduled so that the realized quality SQ exceeds the target quality TQ3. The packet of the required quality RQ6 is classified into the quality class C4, and the packet transfer is scheduled so that the realized quality SQ exceeds the target quality TQ4.

  Returning to FIG. 1, in the quality class review unit 3, the realized quality detection unit 30 detects the realized quality SQ of the transferred packet. The quality class integration unit 31 selects the lower class Ci + 1 that satisfies the target quality of the higher class Ci with the realized quality SQ for two adjacent quality classes Ci and Ci + 1 whose target quality TQ is relatively higher and lower. Integrate into the higher class Ci. At this time, the packet connected to the packet queue of the lower class Ci + 1 is accommodated in the packet queue of the higher class Ci.

  The quality class separation unit 32 separates a quality class (child) having a relatively low target quality TQ from a quality class (parent) whose realized quality SQ cannot satisfy the target quality TQ. At this time, the packet connected to the packet queue of the quality class (parent) that cannot satisfy the target quality TQ is distributed to the packet queue of the separated quality class (child) according to the required quality. The target quality TQ of the quality class (child) is set higher than the target quality of the adjacent quality class whose target quality is lower by one level than the quality class (parent) that cannot satisfy the target quality before the separation.

  If the realized quality of the quality class (parent) that can no longer satisfy the target quality cannot satisfy the target quality before separation, the target quality update unit 33 sets the target quality of the separated quality class (child) to Increase the correction within a range lower than the target quality of the quality class (parent) that cannot satisfy the target quality. Thereby, a part of the packets classified into the quality class (parent) is classified into the quality class (child), so that the realized quality RQ of the quality class (parent) can be improved.

  FIG. 3 is a time chart showing the operation of one embodiment of the present invention. The quality class review unit 3 is configured so that the realization quality SQi + 1 of the lower class Ci + 1 is the higher class for two adjacent quality classes C (Ci, Ci + 1) having the target quality TQ relatively higher and lower. Monitor the presence or absence of Ci's target quality TQi. The realization quality SQi of each quality class Ci can be represented by the average value of the realization quality SQ of the packets classified into the quality class Ci.

  At time t0, the realization quality SQi of all quality classes Ci satisfies the target quality TQi, and the realization quality SQi + 1 of the lower class Ci + 1 satisfies the target quality TQi of the higher class Ci It has not occurred.

  When the realized quality SQ3 of the quality class C3 also satisfies the target quality TQ2 of the quality class C2 at the time t1, the quality class integration unit 31 changes the lower quality class C3 to the higher quality class C2. The integrated (absorbed) quality class C3 packet queue PC3 is abolished, and all the packets accommodated in the packet queue PC3 are accommodated in the quality class C2 packet queue PC2. Further, after this time t1, not only packets whose required quality RQ belongs to the quality class C2 but also packets belonging to the quality class C3 before integration are accommodated in the packet queue PC2 of the quality class C2.

  As described above, according to the present embodiment, when there is no substantial difference in the realization quality of the packets classified into the two adjacent quality classes Ci and Ci + 1 whose target quality is relatively high and low, Since the two quality classes are integrated into one, it is possible to reduce the processing load when a node on the network schedules packet transfer for each quality class.

  Thereafter, the quality class review unit 3 monitors whether or not the realized quality SQ2 of the integrated quality class C2 satisfies the target quality TQ2, and at time t2, the realized quality SQ2 of the quality class C2 is the target. When the quality is lower than TQ2, the quality class separation unit 32 separates a lower quality class C3 ′ (child) from the integrated quality class C2 (parent), and the packet of quality class C2 (parent) Packets whose required quality RQ belongs to the quality class C3 ′ are separated from the queue and accommodated in the packet queue of the quality class C3 ′ (child). Further, after this time t2, packets whose required quality RQ belongs to the quality class C3 ′ are accommodated in the packet queue of the quality class C3 ′.

  As described above, according to the present embodiment, after integrating two adjacent quality classes whose target quality is relatively high and low, the realized quality of the integrated quality class is the quality class of the higher level before integration. When the target quality cannot be satisfied, the lower quality class is separated and packets with lower required quality are accommodated in the packet queue of the lower quality class, so that the realization quality of the higher quality class can be improved. Become.

  Thereafter, the quality class review unit 3 monitors whether the realized quality SQ2 of the quality class C2 (parent) satisfies the target quality, and at the time t3 immediately after the time t2, the quality class C2 (parent) When it is detected that the realized quality SQ2 is still not satisfying the target quality, the target quality update unit 33 performs an increase correction on the target quality TQ3 ′ of the quality class C3 ′ (child). Such increase correction of the target quality TQ3 ′ is repeated until the realization quality SQ2 of the quality class C2 reaches the target quality TQ2 at time t4.

  As a result, packets satisfying the required quality even in the quality class C3 ′ are moved from the quality class C2 to the C3 ′, and the number of packets classified into the quality class C2 is reduced, so that the quality SQ2 achieved in the quality class C2 can be improved. It becomes like this.

  DESCRIPTION OF SYMBOLS 1 ... Packet distribution part, 2 ... Packet transfer part, 3 ... Quality class review part, 30 ... Realization quality detection part, 31 ... Quality class integration part, 32 ... Quality class separation part, 33 ... Target quality update part

Claims (4)

In a quality class control scheme that schedules packet transfers based on the quality required for each packet,
Multiple packet queues set up for different quality classes with different target qualities,
A packet distributor that accommodates input packets in a packet queue according to the required quality;
A packet transfer unit that transfers packets connected to each packet queue with a priority according to the quality class of each packet queue;
Means to detect the realization quality of each quality class;
A quality class integrating means for integrating a lower class satisfying the target quality of the high-order class into the high-order class for two adjacent quality classes whose target quality is relatively high and low;
The quality class control unit is characterized in that all the packets corresponding to the lower class are accommodated in the higher class packet queue and scheduling to the lower class packet queue is not performed . A quality class separation means for separating a quality class whose target quality is relatively low from a quality class whose realized quality cannot satisfy the target quality;
The quality class separation means accommodates a packet stored in a packet queue of a quality class that can no longer satisfy the target quality in the packet queue of the separated quality class according to the required quality. The quality class control method according to claim 1.   The target quality of the separated quality class is set to be higher than the target quality of the quality class adjacent to the lower side with respect to the target quality in the quality class that cannot satisfy the target quality before the separation. The quality class control method according to claim 2.   The means for increasing and correcting the target quality of the separated quality class within a range lower than the target quality of the quality class that can no longer satisfy the target quality is provided. Quality class control method.

Images