JP2003244236A - Method and device for deciding packet switching network quality by test packet - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To provide a method and device for deciding the quality of a packet switching network in a short period by reducing the number of test packets according to the fact of a packet loss rate and network quality to be decided. <P>SOLUTION: A test packet transmission part 2 transmits N pieces of test packets to a network to be measured in a batch, and a packet receiving part 3 receives the test packets returned through the network to be measured, and a quality deterioration decision part 4 calculates the number n (i) of lost packets from the number of the received packets and N, and decides whether the quality is satisfactory/defective/unclear by a statistical detecting method by using the calculated number of lost packets, the threshold of the packet loss rate for deciding that the quality is satisfactory, X1 and the threshold of the packet loss rate for deciding that the quality is defective, X2 (X1<X2), the I type error probability α, the II type error probability β. When the quality is unclear, the N pieces of test packets are transmitted again, and the above mentioned decision is repeated. <P>COPYRIGHT: (C)2003,JPO

Description

Detailed Description of the Invention

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a technique for managing a packet switching network such as the Internet,
More specifically, the present invention relates to a technique for measuring the state of the network and a technique for managing the quality of service provided on the network.

[0002]

2. Description of the Related Art A conventional method for judging the quality of a packet-switched network is to transmit a test packet from one end of a measurement target network at a fixed interval in advance and to receive it at the other end. Measure the number of packets and the number of received packets, estimate the packet loss rate from them (statistically),
Methods have been used to determine quality by comparison with pre-given values.

[0003]

However, in this method, it is necessary to estimate the index value (packet loss rate) of the network quality, and the number of test packets required for the estimation is generally large, and the load on the network is heavy. There was a problem.

It is an object of the present invention to provide a method and apparatus for reducing the number of test packets according to the packet loss rate to be judged and the actual condition of the network quality and capable of judging the quality of a packet switching network in a short period of time. .

[0005]

The packet switching network quality judgment method of the present invention comprises the steps of transmitting an arbitrary number of test packets to the measurement target network, receiving the test packets from the measurement target network, and transmitting the test packets. Calculating a packet loss rate, which is the ratio of the number of packets lost in the network, from the number of test packets and the number of received test packets, and determining the quality state of the packet switching network to be good, degraded, or judgment pending. When the determination of the quality state is suspended, the step of transmitting and receiving the test packet and the step of determining the quality state of the packet switching network is repeated with an arbitrary number of times specified in advance as an upper limit.

Further, the packet switching network quality judgment apparatus of the present invention is a means for transmitting an arbitrary number of test packets to the measurement object network, a means for receiving the test packet from the measurement object network, the number of transmitted test packets and The packet loss rate, which is the ratio of the number of packets lost in the network, is calculated from the number of received test packets, and the quality of the packet-switched network is judged to be good, degraded, or judgment pending. And a means for repeating the transmission and reception of the test packet and the determination of the quality of the packet switching network with an arbitrary number of times specified in advance as an upper limit.

[0007]

DESCRIPTION OF THE PREFERRED EMBODIMENTS Next, embodiments of the present invention will be described with reference to the drawings.

FIG. 1 shows an example of an embodiment of a packet switching network quality judging device of the present invention. In the figure, 1 is an input unit, 2 is a test packet transmitting unit, 3 is a test packet receiving unit, and 4 is a test packet receiving unit. The quality deterioration determination unit 5 is an output unit.

The input unit 1 has parameters input from the outside, here, "the number N of test packets to be transmitted at one time (N is an integer of 1 or more)", "threshold value X1 of packet loss rate judged to be good quality", "Threshold X2 of packet loss rate judged to be quality deterioration" (where X1 <X2), "Type I error probability α", "Type II error probability β", "Upper limit number of repetitions of judgment indicating termination of measurement" M (M is an integer greater than or equal to 1) ", these are sent to the quality deterioration determination unit 4, and" the number N of test packets to be transmitted at one time (N is an integer greater than or equal to 1) "is sent to the test packet transmission unit 2. Also send.

The test packet transmission unit 2 transmits N test packets to the measurement object network at one time in accordance with the instruction from the quality deterioration determination unit 4. The test packet receiving unit 3 receives the test packet returned through the measurement target network,
The number of received packets is stored.

The quality deterioration determination unit 4 calculates the number of lost packets n (i) from the number of received packets and the number of transmitted packets N stored in the test packet receiving unit 3 (where i = 1,
2, ...: Number of times N test packets have been transmitted), the calculated number of lost packets n (i), the threshold X1, and the threshold X
2. Using the type I error probability α and the type II error probability β,
As will be described later, the quality is judged by a statistical test method, and here, it is judged whether the quality is good / quality deteriorated / unknown.

When the judgment result is good quality or quality deterioration, the quality deterioration judgment unit 4 immediately outputs these results through the output unit 5, but when it is unknown, the transmission of N test packets up to that time is transmitted. If the number of times has not reached the upper limit number M, the test packet transmitting unit 2 transmits N test packets again, and the above determination is repeated. If it has reached M, an unknown result is output.

FIG. 2 is a flow chart showing the flow of the packet switching network quality judgment method of the present invention, and the operation will be described in detail below. It is assumed that the various parameters described above have been previously input and stored.

First, the initial value of the counter value m indicating the number of test packets to be sent is set to 1, the initial value of the number of lost packets n (i) is set to N, and N test packets are measured. Send to the target network. Also, the initial value of the counter value j representing the number of received packets is set to 0 (step s
1-s4).

The transmitted test packet passes through the measurement object network and is received. When received, the packet reception number j is incremented by 1, and the lost packet number n (i) is decremented by 1. The reception ends when all the transmitted test packets are received (j is N or more) or no packets arrive for a long time (steps s5 to s7).

Upon completion of reception, a statistical test is performed. Here, an example using a sequential test for the test is shown.

First, two statistical hypotheses about the quality state based on the above parameters, that is, the null hypothesis that the quality is good, H1: packet loss rate is less than X1, and the alternative hypothesis that the quality is deteriorated, H2: packet. Set a loss rate of X2 or more (however, X1 <X2).

The error probabilities of the respective hypotheses, that is, the type I error probabilities α, which are the probabilities that the packet loss rate is less than X1 and other judgments are made when it should be judged that the quality is good, and the packet loss. The type II error probability β 1 which is the probability that the rate is X2 or more and a determination other than the above is made when it should be determined that the quality is deteriorated is obtained as the parameter.

Next, the cumulative lost packet number k is

[0020]

[Equation 1] Calculated as a probability event that packet loss occurs,
Assuming a binomial distribution, the probability distribution under hypothesis H1 and hypothesis H
The probability distribution ratio λm under 2 is λm = X2 ^ k × (1-X2) ^ (mN-k) / X1 ^
It is calculated from k * (1-X1) ^ (mN-k) (step s8).

At this time, if λm ≦ β / (1-α) is established, it can be determined that it is H1, so that “good quality” is output as a result, and if λm ≧ (1-β) / α is established, it is determined as H2. Since it is possible, "quality deterioration" is output as a result (step s9).

If neither is satisfied (when the packet loss rate is X1 or more and less than X2), the number m of transmissions of N test packets is compared with the upper limit number M of repetitions of the determination indicating the termination of measurement, and M is determined. If not satisfied, N test packets are transmitted again, and the processing from step s2 is repeated.
When m is equal to M or exceeds M, "unknown" is output as a result (steps s10 and s1).
1).

In the above-described embodiment, the quality judgment is repeated by comparing the number m of transmissions of N test packets with a pre-input "upper limit number M of judgment repetitions indicating termination of measurement". However, the judgment is made by comparing the total number (m × N) of the transmitted test packets with the preliminarily inputted “upper limit test packet number Z (Z is an integer equal to or greater than N) indicating the termination of measurement”. May be.

[0024]

As described above, in the conventional method by simple sampling, it is necessary to estimate the packet loss rate, and in general, a large number of test packets are required, and there is a problem that the load on the network is heavy. However, according to the present invention, since attention can be paid only to detection of goodness or deterioration, the transmission of the test packet can be stopped when the determination becomes possible,
The advantage is that the number of required test packets can be reduced. In particular, in the quality control of service networks where the quality deterioration phenomenon easily spreads, the difference between the “good quality state” and the “quality degradation state” is significant, so the number of test packets can be significantly reduced compared to the conventional method. The problem that the network load becomes heavy due to the test packet can be solved.

[Brief description of drawings]

FIG. 1 is a block diagram showing an example of an embodiment of a packet switching network quality determination device of the present invention.

FIG. 2 is a flowchart showing a flow of a packet switching network quality judgment method of the present invention.

[Explanation of symbols]

1: input unit, 2: test packet transmission unit, 3: test packet reception unit, 4: quality deterioration determination unit, 5: output unit.

   ─────────────────────────────────────────────────── ─── Continued front page    (72) Inventor Mitsufumi Shinomiya             2-3-1, Otemachi, Chiyoda-ku, Tokyo             Inside Telegraph and Telephone Corporation F term (reference) 5K030 GA14 HA08 HB06 JA10 LE16                       MA01 MB05 MB13 MC02 MC03                 5K035 AA02 DD02 EE22 GG02 JJ04

Claims (6)

[Claims] 1. A method for determining the quality of a packet-switched network using test packets, the method comprising: transmitting an arbitrary number of test packets to a measurement target network; receiving a test packet from the measurement target network; The packet loss rate, which is the ratio of the number of packets lost in the network, is calculated from the number of packets and the number of test packets received, and the quality state of the packet switched network is good, deteriorated,
If the judgment of any of the judgments is suspended, and if the judgment of the quality is suspended, the steps of transmitting and receiving the test packet and judging the quality of the packet switching network are repeated up to an arbitrary number of times specified in advance. And a packet-switched network quality judging method comprising: 2. The packet-switched network quality determination method according to claim 1, further comprising the step of transmitting a pre-input number N of test packets to the measurement target network, and receiving the test packet from the measurement target network and receiving the received test packet. A packet switching network quality judgment method comprising: recording the number of packets; and recording the number of times of repeating transmission of N test packets by suspending the quality judgment. 3. The packet switching network quality determination method according to claim 1, wherein a packet loss rate threshold value X1 for determining good quality, a packet loss rate threshold value X2 for determining quality deterioration, and a type I sequential determination method. Steps of receiving error probability α and type II error probability β as inputs, and two statistical hypotheses regarding the quality state: null hypothesis that quality is good H1: packet loss rate is less than X1, quality is degraded Alternative hypothesis H2: packet switching characterized by including a step of setting a packet loss rate of X2 or more (however, X1 <X2), and a step of performing successive verification to determine quality good / quality deterioration / decision pending. Network quality judgment method. 4. An apparatus for determining the quality of a packet-switched network based on test packets, means for transmitting an arbitrary number of test packets to a measurement target network, means for receiving test packets from the measurement target network, and the transmitted test. The packet loss rate, which is the ratio of the number of packets lost in the network, is calculated from the number of packets and the number of test packets received, and the quality state of the packet switched network is good, deteriorated,
And a means for repeating the transmission and reception of the test packet and the determination of the quality of the packet switching network up to an arbitrary number of times specified in advance when the determination is suspended. Packet switching network quality determination device. 5. The packet-switched network quality determination device according to claim 4, further comprising means for transmitting a pre-input number N of test packets to the measurement target network, and a test packet received from the measurement target network and the received test. A packet switching network quality judging device comprising: a unit for recording the number of packets; and a unit for holding the quality judgment and recording the number of times of repeating N test packet transmissions. 6. The packet switching network quality determination device according to claim 4, wherein a packet loss rate threshold value X1 for determining good quality, a packet loss rate threshold value X2 for determining quality degradation, and a type I sequential determination method. The error hypothesis α and the type II error probability β are received as inputs, and two statistical hypotheses about the quality state are null null hypothesis H1: the packet loss rate is less than X1, and the quality is degraded. Hypothesis H2: A packet switching network quality determination device characterized by having a means for determining a good quality / quality deterioration / decision pending by setting a packet loss rate of X2 or more (where X1 <X2) and performing sequential verification. .