JP2004236262A - Packet communication quality measurement method and delay calculation apparatus - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To obtain the quality similar to the quality of an application performing retransmission control. <P>SOLUTION: In a delay calculation apparatus 121, the earliest passing time at the entrance of a path and the last passing time at the exit of the path are extracted for each packet to be measured, and the difference between the earliest passing time and the last passing time is calculated as the delay time of the packet to be measured which is used for packet delay distribution. Also, in the delay calculation apparatus, the earliest passing time and the last passing time at an observation point on the path are extracted for each packet, the correction time of the packet to be measured is calculated from the difference thereof, the packets to be measured are classified for each correction time and summed, and the number of packets is used as the weight of corresponding classification. The correction time of this classification is added to the delay time of the corresponding test packet to calculate a corrected delay time, and packet delay distribution is calculated by using the corrected delay time and a weight of each classification. <P>COPYRIGHT: (C)2004,JPO&NCIPI

Description

[0001]
TECHNICAL FIELD OF THE INVENTION
The present invention relates to a packet communication quality measurement method and a delay calculation device, and more particularly to a packet communication quality measurement method and a delay calculation device for obtaining a delay in an upper layer by retransmission control after retransmission control.
[0002]
[Prior art]
Conventionally, there has been a packet communication quality measuring method for measuring, for example, network delay distribution, packet delay distribution, loss rate, and the like as communication quality of a packet communication network. An active measurement method is known as a network delay distribution measurement method, and a passive measurement method and a measurement conversion type measurement method are known as packet delay distribution measurement methods.
In the network delay distribution measurement by the active measurement method, an agent that intermittently transmits test packets and an agent that receives the test packets are deployed in the network, and the network quality between agents is estimated from the communication quality of the test packets themselves. (See, for example, Patent Document 1).
[0003]
The delay distribution measurement of packets by the passive measurement method is a method of monitoring packets passing through a desired node or link in a packet communication network and calculating the quality of each packet from the monitoring result (for example, Patent Document 2 and the like). reference).
The packet delay distribution measurement by the rate conversion type quality measurement method is a method of estimating the delay distribution of user packets by weighting the test packet delay measurement result by active measurement with the number of user packets near the time of the test packet. (For example, see Non-Patent Document 1 and the like).
[0004]
The applicant has not found any prior art documents related to the present invention other than the prior art documents specified by the prior art document information described in this specification by the time of filing.
[0005]
[Patent Document 1]
JP-A-62-176239
[Patent Document 2]
JP 2001-53792 A
[Non-patent document 1]
Ishibashi, Kanazawa, Aida, "Proposal of Measure Conversion Quality Measurement Technology by Cooperation of Active / Passive Measurement", IEICE Technical Report, IN2001-72 (2001-09), IEICE
[0006]
[Problems to be solved by the invention]
In a packet communication network, generally, when a packet is exchanged between terminals, retransmission control for detecting an error of an unarrived packet and recovering the packet, that is, retransmitting, in order to ensure reliability of data communication from the start to the end of the packet flow. (Or flow control) is performed. For example, in TCP, when a receiving terminal receives a packet from a transmitting terminal, the receiving terminal notifies the receiving terminal of the reception by a confirmation notification. As a result, the transmission terminal confirms the completion of the delivery of the packet. If no confirmation notification is received within a predetermined time-out period from the packet transmission, the packet is determined to be undelivered and an error has occurred, and the packet is retransmitted. Therefore, when an error occurs, the same packet is repeatedly transferred on the communication path a plurality of times.
[0007]
In the above-described conventional packet delay measurement methods, in any method, a packet subjected to retransmission control during the measurement period is regarded as a user packet having no retransmission control, and is statistically processed. For this reason, the delay caused by the retransmission control is not taken into account, and there is a problem that the communication quality of the lower layer cannot be accurately measured as viewed from the upper layer that processes the packet after the retransmission control.
The present invention is intended to solve such a problem, and can measure communication quality under retransmission control by a lower layer, as viewed from an upper layer, and includes a packet delay due to retransmission control, which is close to application quality. It is an object of the present invention to provide a packet communication quality measurement method and a delay calculation device that can obtain the above.
[0008]
[Means for Solving the Problems]
In order to achieve such an object, a packet communication quality measurement method according to the present invention provides a packet flow of a packet exchanged using retransmission control between terminals connected via a measurement target path via a packet communication network. For the packet communication quality measurement method of calculating the delay time of each measurement target packet constituting the packet flow, and calculating the packet delay distribution of the packet flow based on the delay time of each measurement target packet exchanged within a predetermined period, Based on the passage time of each measurement target packet measured at the first observation point provided at the entrance of the measurement target path, the earliest passage time passing through the first observation point is extracted for each measurement target packet. , Based on the passing time of each measurement target packet measured at the second observation point provided at the exit of the measurement target path, Extracting, for each packet, the last passage time that passed through the second observation point, and for each measurement target packet, the difference between the earliest passage time of the measurement target packet and the last passage time. The method includes a step of calculating the delay time of the measurement target packet and a step of calculating a packet delay distribution of the packet flow based on the delay time of each measurement target packet.
[0009]
Further, another packet communication quality measurement method according to the present invention includes a packet flow of a packet exchanged using retransmission control between terminals connected through a measurement target path via a packet communication network, and a packet flow of the measurement target path. The test packet is transmitted from the entrance to the exit, and the delay time of the test packet is weighted based on the measurement target packet exchanged in the packet flow in the vicinity of the predetermined time of the test packet, so that the measurement conversion is performed on the test packet. In the packet communication quality measurement method of calculating the packet delay distribution of the packet flow, the observation point is set for each measurement target packet based on the passing time of each measurement target packet measured at the observation point provided on the measurement target path. Extracting the earliest passing time and the last passing time, and Calculating the correction time of the packet to be measured from the difference between the earliest passage time and the final passage time for each packet, and classifying and counting each measurement target packet for each correction time of the measurement target packet The number of packets is used as the weight of the classification, and the corrected delay time is calculated by adding the correction time of the classification to the delay time of the test packet obtained from the difference between the transmission time and the reception time of the test packet. And calculating the packet delay distribution of the packet flow using the corrected delay time and the weight of each classification.
[0010]
When extracting the earliest passing time and the last passing time for each measurement target packet, using order information indicating the order position in the packet flow given to each measurement target packet for reproduction control, Each measurement target packet may be identified.
[0011]
In addition, the delay calculation apparatus according to the present invention is capable of measuring each packet constituting a packet flow of a packet exchanged using retransmission control between terminals connected through a measurement target path via a packet communication network. In a delay calculating apparatus that calculates a delay time of a target packet and calculates a packet delay distribution of a packet flow based on a delay time of each measurement target packet exchanged within a predetermined period, a first observation provided at an entrance of a measurement target path. Based on the passing time of each measurement target packet measured at a point, the earliest passing time that has passed through the first observation point is extracted for each measurement target packet, and the second passing time provided at the exit of the measurement target path is extracted. Based on the passing time of each measurement target packet measured at the observation point, the final passing time of passing through the second observation point for each measurement target packet Passing time extracting means for extracting, and for each measurement target packet, delay time calculating means for calculating the difference between the earliest passing time and the final passing time of the measurement target packet as the delay time of the measurement target packet. And a delay distribution calculating means for calculating a packet delay distribution of a packet flow based on a delay time of each measurement target packet.
[0012]
Further, another delay calculating apparatus according to the present invention is configured such that, together with a packet flow of a packet exchanged using retransmission control between terminals connected via a measurement target path via a packet communication network, from the entrance of the measurement target path. A packet converted by the test packet by weighting the delay time of the test packet by transmitting the test packet to the egress and performing weighting on the delay time of the test packet based on the measurement target packet exchanged near a predetermined time of the test packet in the packet flow. In a delay calculating device that calculates a packet delay distribution of a flow, based on a passing time of each measurement target packet measured at an observation point provided on a measurement target path, for each measurement target packet, Passing time extracting means for extracting the passing time and the final passing time of the A correction time calculating means for calculating a correction time of the measurement target packet from a difference between the previous passage time and the final passage time, and classifying and counting each measurement target packet for each correction time of the measurement target packet; The number of packets is used as the weight of the classification, and the corrected delay time is calculated by adding the correction time of the classification to the delay time of the test packet obtained from the difference between the transmission time and the reception time of the test packet. A weight calculation unit; and a delay distribution calculation unit configured to calculate a packet delay distribution of a packet flow using the corrected delay time of each classification and the weight.
[0013]
When the passing time extracting means extracts the earliest passing time and the last passing time for each measurement target packet, it indicates the ordinal position in the packet flow assigned to each measurement target packet for reproduction control. The same measurement target packet may be identified using the order information.
[0014]
BEST MODE FOR CARRYING OUT THE INVENTION
Next, embodiments of the present invention will be described with reference to the drawings.
First, a packet communication quality measurement system according to a first embodiment of the present invention will be described with reference to FIG. FIG. 1 is a diagram showing the configuration of the packet communication quality measurement system according to the first embodiment of the present invention.
This packet communication quality measurement system is an example in which the present invention is applied to a passive quality measurement method. This system includes a plurality of passive measurement probes 111 and 112 and a delay calculation device 121. Hereinafter, a case where data communication with retransmission control is performed between the terminals 101 and 102 via the measurement target path 132 via the packet communication network 131 and the delay distribution at that time is obtained using the packet communication quality measurement system will be described. This will be described as an example.
[0015]
The passive measurement probes 111 and 112 are connected to links connecting the terminals 101 and 102 and the packet communication network 131, respectively, and obtain packets passing through these links, that is, packets passing through observation points, that is, packets entering and leaving the measurement target path 132. I do. This packet acquisition can be realized by, for example, a method in which a physical layer branching device is connected to a link, and all passing packets are drawn into a passive measurement probe.
The passive measurement probes 111 and 112 and the delay calculation device 121 are connected to each other via a predetermined communication network, and exchange various kinds of information necessary for measuring a packet flow delay.
[0016]
The passive measurement probes 111 and 112 respectively acquire packets passing through the observation points connected to the terminals 101 and 102, and among the packets, the packets constituting the measurement target flow, that is, the identifiers included in the measurement target packets and the identifiers of the packets It has a function of transmitting a pair with the acquisition time, that is, the passage time, to the delay calculation device 121 as packet observation information 141, 142.
The identifier may be newly added to the header information of the packet to be measured. However, by using the order information indicating the order position in the packet flow used for retransmission control of the packet flow, new information can be obtained. , The delay calculation device 121 can identify the same retransmitted measurement target packet.
[0017]
The delay calculation device 121 searches for an identifier of the measurement target packet included in the packet observation information 141 obtained from the passive measurement probe 111 to see if there is one that is already equal to the identifier notified from the passive measurement probe 111. If there is no equivalent, the pair of the identifier and the passing time is stored.
In addition, the delay calculation device 121 searches for an identifier of a measurement target packet included in the packet observation information 142 received from the passive measurement probe 112 for an identifier that is already equal to the identifier notified from the passive measurement probe 112. If there is a match, the stored passing time paired with the identifier is rewritten with the passing time paired with the received identifier. If there is no equivalent, the pair of the identifier and the passing time is stored.
[0018]
The entry time S based on the packet observation information 141 received from the passive measurement probe 111 and the packet observation information 142 received from the passive measurement probe 112 Then, a difference RS of the outgoing time R is determined, the value is set as the delay time D of the measurement target packet having the identifier, and the packet delay distribution of the packet flow is determined from the delay time of each measurement target packet.
[0019]
FIG. 2 is a block diagram showing a configuration of the delay calculation device 121. The delay calculation device 121 includes a computer having a communication function as a whole, and includes a communication interface unit (hereinafter, referred to as a communication I / F unit) 11, a screen display unit 12, an operation input unit 13, a storage unit 14, and A control unit 15 is provided.
The communication I / F unit 11 is a circuit unit that performs data communication with the passive measurement probes 111 and 112. The screen display unit 12 is a display device such as a CRT or an LCD for displaying various information such as obtained measurement results on a screen. The operation input unit 13 is an operation input device such as a keyboard and a mouse for inputting a user operation. The storage unit 14 is a storage device such as a hard disk or a memory that stores various information used by the control unit 15 and the program 14B.
[0020]
The control unit 15 is a functional unit that includes a microprocessor such as a CPU and its peripheral circuits, and executes the program 14B of the storage unit 14 to realize various functional units in cooperation with the hardware.
The control unit 15 includes a passing time extracting unit 15A, a delay time calculating unit 15B, and a delay distribution calculating unit 15C as various functional units.
The passing time extraction means 15A refers to the packet observation information 141, 142 obtained from the passive measurement probes 111, 112, and refers to the passing time of the measurement target packet at each observation point, that is, the entry time S to the measurement target path 132 and the output time. From the time R, the earliest entry time Ss (earliest passage time) and the last outgoing time Re (final passage time) are extracted for each measurement target packet.
[0021]
The delay time calculation unit 15B calculates, for each measurement target packet, the delay time D = Re−Ss of the measurement target packet from the difference between Ss and Re of each measurement target packet obtained by the passing time extraction unit 15A. Then, it is added to the packet delay information 14A of the storage unit 14.
The delay distribution calculating unit 15C calculates the packet delay distribution of the packet flow based on the delay time of each measurement target packet exchanged within a predetermined period with reference to the packet delay information 14A of the storage unit 14.
[0022]
Next, the operation of the packet communication quality measurement system according to the present embodiment will be described with reference to FIGS.
FIG. 3 is an example of the packet observation information 141 and 142 obtained by the passive measurement probes 111 and 112. FIG. 4 is an example of the packet delay information 14A obtained by the delay calculation device 121. FIG. 5 is a sequence diagram showing a packet flow at the time of packet flow delay measurement.
[0023]
Here, the packet communication network 131 is formed of an IP network, the flow is a TCP flow, the identifier used for retransmission control is a TCP sequence number, and the terminal 101 transmits the packet through the measurement target path 132 of the packet communication network 131 to the terminal. A case will be described as an example where a packet flow delay distribution of a packet to be measured included in a packet flow sent to 102 is obtained.
The passive measurement probes 111 and 112 are assumed to be time-synchronized, and can be time-synchronized using, for example, information from a GPS.
[0024]
First, the passive measurement probes 111 and 112 acquire all the packets passing through the observation point over a predetermined period, and select a packet of the measurement target flow from the acquired packets.
Here, there may be a plurality of measurement target flows. In the case of a plurality, each of these may be individually processed to determine the quality for each flow.
[0025]
From the selected packet to be measured, a set of an identifier such as a sequence number in the TCP header and the passage time of the packet is formed and transmitted to the delay calculation device 121 as packet observation information 141 and 142.
For example, the passive measurement probe 111 transmits the pair of the identifier of the first column and the entry time S of the second column in FIG. 3 to the delay calculation device 121 as the packet observation information 141. In addition, the passive measurement probe 112 transmits the pair of the identifier in the first column and the outgoing time T in the third column in FIG.
[0026]
Based on the packet observation information 141 and 142 received from the passive measurement probes 111 and 112, the delay calculation device 121 uses the passing time extraction means 15A of the control unit 15 to set, for each identifier, each entry time S and From the outgoing time R, the earliest value Ss of the incoming time and the final value Re of the outgoing time are obtained and stored in the storage unit 14 as packet delay information 14A.
In the present embodiment, since the passive measurement probes 111 and 112 send the packet observation information 141 and 142 to the delay calculation device 121 in the order of packet acquisition, the passing time extraction unit 15A of the delay calculation device 121 performs passive measurement for each identifier. The time based on the packet observation information 141 obtained first from the probe 111, that is, the first arrival time Ss, and the time based on the packet observation information 142 obtained last from the passive measurement probe 112, that is, the last exit time Re may be stored.
[0027]
In addition, the delay calculation device 121 calculates the delay time D = Re−Ss from the first-in first-out time Ss and the last outgoing time Re for each identifier by the delay time calculation means 15B of the control unit 15.
Thereby, as shown in FIG. 5, the same measurement target packet subjected to retransmission control can be regarded as one apparent measurement target packet, and the delay time of the measurement target packet is calculated.
[0028]
This process is repeated for each measurement target packet exchanged within a predetermined period, and a column of the identifier of the packet included in the flow and the delay time is obtained as the packet delay information 14A shown in FIG.
The delay distribution calculation means 15C of the control unit 15 calculates a packet delay distribution relating to a desired packet flow from the delay time and the frequency of each measurement target packet in a predetermined period based on the packet delay information 14A.
In the examples of FIGS. 3 and 4, P か ら delay = 10 パ ケ ッ ト = 3/6, P ｛delay as a delay distribution indicating the delay time D and the number of the packets, that is, the ratio of the weight, for the flow composed of six packets = 160 ° = ２, and P ｛delay = 180 ° = 1/6.
[0029]
As described above, for each packet constituting the test target flow, the earliest first arrival time Ss observed first at the entrance of the measurement target path and the last exit time last observed at the exit of the measurement target path before a predetermined time has elapsed from this Ss Since the delay time D of the packet is determined from the difference Re-Ss from Re, when the packet retransmission control is performed by the host application, Re is delayed by the time required for the retransmission control of the packet. That is, it is included in the packet delay time.
[0030]
By this means, in the passive quality measurement method, it is possible to measure the packet communication quality at the time of use by the upper layer, including the packet delay due to retransmission control, and to obtain a quality close to the application quality when information is passed to the application layer. it can.
Therefore, by using the packet flow quality obtained in this way for designing a network, it is not necessary to design the resources of the network excessively on the secure side, and the cost for constructing the network can be reduced.
[0031]
Next, a packet communication quality measurement system according to a second embodiment of the present invention will be described with reference to FIG. FIG. 6 is a diagram illustrating a configuration of the packet communication quality measurement system according to the embodiment of the present invention.
This packet communication quality measurement system is an example in which the present invention is applied to a measure conversion type quality measurement system. This system is provided with a plurality of quality measuring devices 241, 242, a passive measuring probe 211, and a delay calculating device 221. These devices are connected to each other via a predetermined communication network, and various types of devices necessary for the measurement are provided. Information is exchanged.
[0032]
The quality measurement devices 241 and 242 transmit a predetermined test packet to the other quality measurement device paired with the quality measurement device via the measurement target path 232 connecting the terminals 201 and 202 via the packet communication network 231. It has a function of transmitting, and a function of notifying the delay calculation device 221 of a reception status regarding a test packet received from the other quality measurement device, and performs quality measurement by switching these functions.
In this case, a packet transmitted by the quality measuring devices 241 and 242 is provided with a test packet identifier that indicates that the packet is a test packet and can be distinguished from other test packets. Detected separately from the packet. Examples of the test packet identifier include a method of writing a specific value in an empty field in a packet header, a method of writing a specific value in a specific position of a packet payload, and the like.
[0033]
The quality measuring devices 241, 242, the passive measuring probe 211, and the delay calculating device 221 can identify the test packet identifier.
The quality measuring devices 241 and 242 transmit the test packet at predetermined intervals to the other quality measuring device to be paired via the measurement target path 232. At this time, the transmission time is described in the test packet together with the test packet identifier. Further, when a test packet is received, a delay time D of the test packet is calculated from a difference between a transmission time described in the received test packet and a reception time in the own apparatus, and the delay time D is paired with the test packet identifier. It is transmitted to the calculation device 221 as test packet observation information 251 and 252.
[0034]
The passive measurement probe 211 acquires all the packets that pass through the link connecting the terminal 201 and the packet communication network 231, that is, the packets that pass through the observation point, and among the acquired packets, the packets that constitute the flow to be measured and satisfy predetermined conditions. The test packet is extracted as well as the packet to be measured.
Then, the identifier of the measurement target packet or the test packet identifier is transmitted to the delay calculation device 221 as packet observation information 261 in a pair with the packet passing time (time stamp) T.
[0035]
The predetermined condition may be that the passing time T is near the time of any of the test packet passing times. In the range near this time, the time width may be different before and after the test packet passing time, or these time widths may be equal. For example, the period may be from 50 ms before the acquisition of the test packet to 50 ms after the acquisition of the test packet.
The identifier may be newly added to the header information of the measurement target packet. However, by using the order information indicating the order position in the packet flow used for retransmission control of the packet flow, the identifier is newly added. The same packet to be measured retransmitted can be identified by the delay calculation device 221 without adding any unnecessary information.
[0036]
The delay calculation device 221 refers to the test packet observation information 251 and 252 and the packet observation information 261 received from the quality measurement devices 241 and 242 and the passive measurement probe 211 and, for each identifier of each packet, determines the last passage time Te and the test The packet delay information 24A composed of a set of the packet delay time D and the correction time d is created. The last passage time Te is a time when the measurement target packet having the same identifier is acquired a plurality of times and the last acquisition time is adopted. When only one measurement target packet having the same identifier is acquired, the passing time T of the packet becomes the final passing time Te.
[0037]
The test packet delay time D is information corresponding to only the test packet identifier, and is a delay time of the test packet. The correction time d indicates a time difference from the earliest acquisition to the last acquisition when a measurement target packet having the same identifier is acquired a plurality of times. Based on this correction time d, the measurement target packets corresponding to each test packet are classified.
For example, for each test packet, the number of packets in the vicinity of time without a correction time d is obtained, and the delay of the test packet is weighted by the number of packets. If the test packet has the correction time d, the corrected delay time D is obtained by adding the correction time d to the delay time D of the test packet. Weight. Then, a desired packet delay distribution is calculated using these weights and the corrected delay time Dd.
[0038]
FIG. 7 is a block diagram showing the configuration of the delay calculation device 221. The delay calculation device 221 includes a computer having a communication function as a whole, and includes a communication interface unit (hereinafter, referred to as a communication I / F unit) 21, a screen display unit 22, an operation input unit 23, a storage unit 24, and A control unit 25 is provided.
The communication I / F unit 21 is a circuit unit that performs data communication with the quality measurement devices 241, 242 and the passive measurement probe 211. The screen display unit 22 is a display device such as a CRT or an LCD that displays various information such as obtained measurement results on a screen. The operation input unit 23 is an operation input device such as a keyboard and a mouse for inputting a user operation. The storage unit 24 is a storage device such as a hard disk or a memory that stores various information used by the control unit 25 and the program 24C.
[0039]
The control unit 25 is a functional unit that includes a microprocessor such as a CPU and its peripheral circuits, and executes a program 24C in the storage unit 24 to realize various functional units in cooperation with the hardware.
The control unit 25 includes a passing time extracting unit 25A, a correction time calculating unit 25B, a weight calculating unit 25C, and a delay distribution calculating unit 25D as various functional units.
The passage time extracting unit 25A refers to the packet observation information 261 obtained from the passive measurement probe 211, and from the passage time T of the measurement target packet at the observation point, for each measurement target packet, the earliest passage time Ts and The last passing time Te is extracted.
[0040]
The correction time calculation unit 25B calculates the delay time Te-Ts of the measurement target packet from the difference between Ts and Te of each measurement target packet obtained by the passing time extraction unit 25A, and calculates this as the delay time of the measurement target packet. The correction time d is stored together with Te in the packet delay information 24A of the storage unit 24.
Further, the correction time calculating means 25B adds the delay time D of each test packet notified by the test packet observation information 251 and 252 from the quality measuring devices 241 and 242 to the packet delay information 24A.
[0041]
The weight calculation unit 25C classifies each measurement target packet for each combination of the identifier of the corresponding test packet and its correction time d with reference to the packet delay information 24A, and calculates the measurement target packet classified into each combination. The number of packets is set as a weight W, and this is stored in the storage unit 24 as packet weight information 24B. Further, for each combination, the delay time of each test packet, that is, the corrected delay time Dd is calculated by adding the correction time d of the measurement target packet to the delay time D of the test packet corresponding to the measurement target packet. Are added to the packet weight information 24B.
The delay distribution calculating unit 25D refers to the packet weight information 24B of the storage unit 24 and, based on the corrected delay time Dd and the weight W for each test packet, calculates the weight of the packet flow weighted by the measurement target packet near the time. Calculate the packet delay distribution.
[0042]
Next, the operation of the packet communication quality measurement system according to the present embodiment will be described with reference to FIGS.
FIG. 8 is an example of packet observation information obtained by the passive measurement probe 211. FIG. 9 is an example of the packet delay information 24A obtained by the delay calculation device 221. FIG. 10 is an example of the packet weight information 24B obtained by the delay calculation device 221. FIG. 11 is a sequence diagram showing a packet flow including retransmission control at the time of packet flow delay measurement.
Here, the packet communication network 231 is composed of an IP network, and the identifier used for retransmission control of the flow is an integer that increases by one for each packet, and the packet flow transmitted from the terminal 201 to the terminal 202 via the packet communication network 231. In the following, an example will be described in which a packet flow delay distribution of the packet included in the packet is obtained.
[0043]
First, the quality measurement device 241 intermittently transmits a test packet to the quality measurement device 242 over a predetermined measurement period. At this time, a test packet identifier and a transmission time are described in the test packet.
Upon receiving the test packet, the quality measuring device 242 calculates the delay time D of the test packet from the difference between the reception time of the test packet and the transmission time described in the test packet. To the delay calculation device 221 together with the test packet identifier.
[0044]
On the other hand, the passive measurement probe 211 acquires a packet passing through the observation point over the above measurement period, and, among the acquired test packet and the measurement target packets constituting the measurement target flow, a time near the time of the test packet passing time, for example, the test The packet having the absolute value of the difference from the passage time of the packet acquired within 50 ms is extracted, and the identifier or test packet identifier of the packet and the passage time are transmitted to the delay calculation device 221 as packet observation information 261. Here, it is assumed that the passive measurement probe 211 has obtained the packet observation information 261 as shown in FIG.
[0045]
The control unit 25 of the delay calculation device 221 creates packet delay information 24A shown in FIG. 9 based on the test packet observation information 252 from the quality measurement device 242 and the packet observation information 261 from the passive measurement probe 211.
In the control unit 25, the final value Te and the earliest value Ts of the passing time T are extracted from the packet observation information 261 for each measurement target packet other than the test packet by the passing time extracting unit 25A.
[0046]
Then, the correction time calculation means 25B of the control unit 25 calculates the correction time d for each measurement target packet using the final value Te and the earliest value Ts. Here, when the same identifier is obtained a plurality of times, the difference between the final value Te of the passage time T and the earliest value Ts is set as the correction time d, and when the same identifier is obtained only once, the correction time d = “none” ".
As a result, as shown in FIG. 11, the same measurement target packet subjected to retransmission control can be regarded as one apparent measurement target packet, and the delay time of the measurement target packet is calculated as the correction time d. Is done.
[0047]
In the example of FIG. 9, the identifiers # 6 to # 15 are measurement target packets obtained near the time of the passing time T of the test packet “Trial 1”. In this case, there is no retransmission control, and the correction time d is all set to "none".
Also, the identifiers # 96 to # 105 are measurement target packets obtained near the time of the passing time T of the test packet “Trial 2”. In this case, since the retransmission has been performed for the identifiers # 96 to # 101, the correction time d is set to "102", and the correction time d is set to "none" for the identifiers # 102 to # 105.
[0048]
Further, the correction time calculating means 25B acquires the delay time D of each test packet from the test packet observation information 252, and adds it to the packet delay information 24A.
Then, the weight calculation means 25C creates packet weight information 24B shown in FIG. 10 from the packet delay information 24A. Here, each measurement target packet is classified for each combination of the test packet corresponding to each measurement target packet and the correction time d, and the weight W and the corrected delay time Dd are calculated for each classification.
[0049]
As a specific example, for each test packet, the number of measurement target packets located near the time of the test packet and having a correction time d = “none” is counted, and the number of packets is set as the weight W of the classification.
In the example of FIG. 9, since there are 10 packets (# 6 to # 15) with the correction time d = “none” in the test packet “test 1”, the number of packets = 10 is set as the weight W of the classification. In addition, since there are four packets (# 102 to # 105) with the correction time d = “none” in the test packet “Trial 2”, the number of packets = 4 is set as the weight W of the classification.
[0050]
Next, for each test packet, the number of packets having a correction time d> 0 among the packets located near the time of the test packet is further counted for each correction time, and the number of packets is classified into the weight W of the classification. And In the example of FIG. 9, there are six packets (# 96 to # 101) with the correction time d = “102” in the test packet “Trial 2”, and the number of packets = 6 is set as the weight W of the classification.
Then, for each of these classifications, the sum D + d of the delay time D of the test packet and the correction time d of the packet is set as the corrected delay time Dd. When the correction time d = “none”, Dd = D
[0051]
Then, the delay distribution calculating means 25D of the control unit 25 calculates the delay distribution P from the corrected delay time Dd and the weight W for each classification of the packet weight information 24B. In this case, since the number of near-time packets corresponding to each test packet is “10”, the delay distribution P of the packet flow is P ｛delay = 50｝ = 4/20, P ｛delay = 100｝ = 10/20, P {delay = 152} = 6/20.
[0052]
In this way, the test packet is intermittently transmitted via the measurement target path, and the observation packet is measured for each measurement target packet based on the passing time of each measurement target packet measured at the observation point provided on the measurement target path. Extract the earliest passing time and the final passing time after passing the point, calculate the correction time of the measurement target packet from the difference between the earliest passing time and the final passing time, The target packet is classified and counted for each correction time of the measurement target packet, the number of packets is used as the weight of the classification, and the difference between the transmission time and the reception time of the test packet at the entrance and the exit of the measurement target path. The corrected delay time is calculated by adding the correction time of the class to the delay time of the test packet obtained from the test packet, and the packet is calculated using the corrected delay time of each class and the weight. When the packet retransmission control is performed by the upper-level application, the delay time of the test packet is equal to the time required for the retransmission control of the measurement target packet near that time, that is, the correction time. It will be corrected by d.
[0053]
By this means, in the measure conversion type quality measurement method, the packet flow delay of the target user, that is, the packet communication quality at the time of use by the upper layer after the retransmission control can be measured, which is close to the application quality when information is passed to the application layer. Quality can be obtained.
Therefore, by using the packet flow quality obtained in this way for designing a network, it is not necessary to design the resources of the network excessively on the secure side, and the cost for constructing the network can be reduced.
[0054]
【The invention's effect】
As described above, the present invention passes through the first observation point for each measurement target packet based on the passing time of each measurement target packet measured at the first observation point provided at the entrance of the measurement target path. The first observation point extracted at the second observation point provided at the exit of the measurement target path is extracted, and the second observation point is determined for each measurement target packet based on the passage time of each measurement target packet measured at the second observation point provided at the exit of the measurement target path. Extracting the last passing time that has passed, and for each packet to be measured, the difference between the earliest passing time and the last passing time of the packet to be measured is used to calculate the packet delay distribution. And the step of calculating the delay time of the packet, when the packet retransmission control is performed by the host application, the time required for the last retransmission It will be delayed, and those included in the delay time of a packet.
[0055]
In addition, based on the passage time of each measurement target packet measured at the observation point provided on the measurement target path, the earliest passage time and the last passage time that passed the observation point are extracted for each measurement target packet. Calculating the correction time of the measurement target packet from the difference between the earliest passage time and the final passage time for each measurement target packet, and correcting each measurement target packet to the measurement target packet. It classifies and counts by time, the number of packets is used as the weight of the classification, and the delay of the test packet obtained from the difference between the transmission time and the reception time of the test packet at the entrance and the exit of the measurement target path. Calculating a corrected delay time by adding the correction time of the classification to the time, and a packet using the corrected delay time and the weight of each classification. Calculating the packet delay distribution of the flow, if the packet retransmission control is performed by the upper application, the delay time of the test packet is the time required for the retransmission control of the measurement target packet near that time, The correction is performed in the correction time d.
[0056]
This makes it possible to measure the packet flow delay of the target user, that is, the packet communication quality at the time of use by the upper layer after the retransmission control, and obtain a quality close to the application quality when information is passed to the application layer.
Therefore, by using the packet flow quality obtained in this way for designing a network, it is not necessary to design the resources of the network excessively on the secure side, and the cost for constructing the network can be reduced.
[Brief description of the drawings]
FIG. 1 is a block diagram showing a configuration of a packet communication quality measurement system according to a first embodiment of the present invention.
FIG. 2 is a block diagram illustrating a configuration of a delay calculation device in FIG. 1;
FIG. 3 is an example of packet observation information obtained by the passive measurement probe of FIG.
FIG. 4 is an example of packet delay information obtained by the delay calculation device of FIG. 1;
FIG. 5 is a sequence diagram showing a packet flow when measuring the packet flow delay of FIG. 1;
FIG. 6 is a block diagram illustrating a configuration of a packet communication quality measurement system according to a second embodiment of the present invention.
FIG. 7 is a block diagram illustrating a configuration of the delay calculation device of FIG. 6;
FIG. 8 is an example of packet observation information obtained by the passive measurement probe of FIG. 6;
FIG. 9 is an example of packet delay information obtained by the delay calculation device of FIG. 6;
FIG. 10 is an example of packet weight information obtained by the delay calculation device of FIG. 6;
FIG. 11 is a sequence diagram showing a packet flow including retransmission control when measuring the packet flow delay in FIG. 6;
[Explanation of symbols]
101, 102 terminal, 111, 112 passive measurement probe, 121 delay calculation device, 131 packet communication network, 132 measurement path, 141 142 packet observation information, 11 communication I / F unit, 12 Screen display unit, 13 operation input unit, 14 storage unit, 14A packet delay information, 14B program, 15 control unit, 15A passage time extraction unit, 15B delay time calculation unit, 15C delay distribution calculation unit , 201, 202 ... terminal, 211 ... passive measurement probe, 221 ... delay calculation device, 231 ... packet communication network, 232 ... measurement target path, 241, 242 ... quality measurement device, 251, 252 ... test packet observation information, 261 ... Packet observation information, 21: communication I / F unit, 22: screen display unit, 23: operation input unit, 24: storage unit, 24A: packet DOO delay information, 24B ... packet weight information, 24C ... program, 25 ... control unit, 25A ... passage time extracting means, 25B ... correction time calculation means, 25C ... weight calculating means, 25D ... delay distribution calculating means.

Claims (6)

For a packet flow of a packet exchanged by using retransmission control between terminals connected via a measurement target path via a packet communication network, a delay time of each measurement target packet constituting the packet flow is calculated, and a predetermined period is calculated. A packet communication quality measurement method for calculating a packet delay distribution of the packet flow based on a delay time of each of the measurement target packets exchanged within,
Based on the passage time of each of the measurement target packets measured at a first observation point provided at the entrance of the measurement target path, the earliest passage time of passing through the first observation point is determined for each measurement target packet. Extraction and, based on the passing time of each of the measurement target packets measured at the second observation point provided at the exit of the measurement target path, for each measurement target packet, the final passing through the second observation point Extracting a transit time;
Calculating, for each of the measurement target packets, a difference between the earliest passage time and the final passage time of the measurement target packet as a delay time of the measurement target packet;
Calculating a packet delay distribution of the packet flow based on a delay time of each of the measurement target packets. Along with the packet flow of packets exchanged using retransmission control between terminals connected via the measurement target path via the packet communication network, a test packet is transmitted from the entrance to the exit of the measurement target path, and the The delay time of the test packet is weighted based on the measurement target packet exchanged in the vicinity of a predetermined time of the test packet to calculate a packet delay distribution of the packet flow which is measured and converted by the test packet. The packet communication quality measurement method
Based on the passage time of each of the measurement target packets measured at the observation point provided on the measurement target path, for each of the measurement target packets, the earliest passage time and the final passage time that passed the observation point Extracting
Calculating, for each of the measurement target packets, a correction time of the measurement target packet from a difference between the earliest passage time and the final passage time;
Each of the measurement target packets is classified and counted for each correction time of the measurement target packet, and the number of packets is set as the weight of the classification, and the number of packets is obtained from a difference between the transmission time and the reception time of the test packet. Calculating a corrected delay time by adding the correction time of the classification to the delay time of the test packet;
Calculating a packet delay distribution of the packet flow using the corrected delay time and the weight of each of the classifications. The method for measuring packet communication quality according to claim 1 or 2,
When extracting the earliest passing time and the last passing time for each of the measurement target packets, indicates an ordinal position in the packet flow given to each of the measurement target packets for the reproduction control. A packet communication quality measurement method characterized in that the same measurement target packet is identified using order information. For a packet flow of a packet exchanged by using retransmission control between terminals connected via a measurement target path via a packet communication network, a delay time of each measurement target packet constituting the packet flow is calculated, and a predetermined period is calculated. A delay calculation device that calculates a packet delay distribution of the packet flow based on a delay time of each measurement target packet exchanged within the
Based on the passage time of each of the measurement target packets measured at a first observation point provided at the entrance of the measurement target path, the earliest passage time of passing through the first observation point is determined for each measurement target packet. Extraction and, based on the passing time of each of the measurement target packets measured at the second observation point provided at the exit of the measurement target path, for each measurement target packet, the final passing through the second observation point Passing time extracting means for extracting a passing time;
For each of the measurement target packets, a delay time calculating unit that calculates a difference between an earliest passage time and a final passage time of the measurement target packet as a delay time of the measurement target packet,
A delay distribution calculating means for calculating a packet delay distribution of the packet flow based on a delay time of each of the measurement target packets. Along with the packet flow of packets exchanged using retransmission control between terminals connected via the measurement target path via the packet communication network, a test packet is transmitted from the entrance to the exit of the measurement target path, and the The delay time of the test packet is weighted based on the measurement target packet exchanged in the vicinity of a predetermined time of the test packet to calculate a packet delay distribution of the packet flow which is measured and converted by the test packet. In the delay calculation device that
Based on the passage time of each of the measurement target packets measured at the observation point provided on the measurement target path, for each of the measurement target packets, the earliest passage time and the final passage time that passed the observation point Passing time extracting means for extracting
For each of the measurement target packets, a correction time calculation unit that calculates a correction time of the measurement target packet from a difference between the earliest passage time and the final passage time,
Each of the measurement target packets is classified and counted for each correction time of the measurement target packet, and the number of packets is set as the weight of the classification, and the number of packets is obtained from a difference between the transmission time and the reception time of the test packet. Weight calculating means for calculating the corrected delay time by adding the correction time of the classification to the delay time of the test packet;
A delay calculation unit configured to calculate a packet delay distribution of the packet flow using the corrected delay time and the weight of each of the classifications. The delay calculation device according to claim 4 or 5,
The passing time extracting means,
A delay calculation device for identifying the same measurement target packet using order information indicating the order position in the packet flow assigned to each of the measurement target packets for the reproduction control.

Images