JP2006186889A - End-to-end quality estimation device and method - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To attain end-to-end quality estimation for saving a memory source in a communication quality estimation technology, and for preventing the estimated quality from being affected by the transfer speed of a network. <P>SOLUTION: This end-to-end quality estimation device is configured to extract a flow since the start to end of communication between transmitting/receiving terminals being the object of measurement from inputted packet header information, and to decide whether or not the packet is re-transmitted each time the packet arrives, and to operate a re-transmission number counter and an arrival number counter in a block to which the packet is belonging, and to calculate end-to-end quality by using the values of the re-transmission number counter and an arrival number counter by using the temporal transition of the sequence number of the arriving packet as a trigger. <P>COPYRIGHT: (C)2006,JPO&NCIPI

Description

  The present invention relates to an end-to-end quality estimation apparatus and method, and more particularly to an end-to-end quality estimation apparatus for measuring communication quality in a performance management technique in a communication service using a packet switching network such as the Internet, and the like. Regarding the method.

  Examples of general performance measurement modes include the following.

(1) MIB (Management Information Base) information (number of passing packets, amount of information, number of lost packets in a buffer, etc.) implemented in a device or the like is acquired by SNMP (Simple Network Management Protocol) (for example, non-patent literature) 1):
(2) A measurement device is installed end-to-end of the network, and trial communication is performed between them to measure performance (end-to-end packet loss rate, delay time, etc.) (for example, see Non-Patent Document 2). :
(3) A test packet is transmitted from a certain point such as Ping to the opposite point, and a packet loss rate and a round-trip delay time in a round-trip route are measured (for example, see Non-Patent Document 3):
(4) On the premise of a specific application (VoIP: packets are sent from the transmission side at regular intervals), on the assumption that the transmission side is sending packets with a constant sense, from the information of the sequence number and its arrival time, Calculate the IP packet loss rate from only one point observation information (at the receiving side):
There is.

  The method (1) has a problem that it cannot measure the end-to-end quality of the network (for example, the end-to-end packet loss rate) although it has a small influence on the actual communication traffic because it is passive measurement. .

  Further, the method (2) has a point that an end-to-end measuring device has to be arranged, and a test packet is applied to the network for the measurement, so that the actual communication traffic is affected. There are problems such as.

  In the method (3), it is not necessary to arrange a measuring device end-to-end in the network, but it is impossible to measure a one-way delay or a one-way packet loss rate (that is, only a quality in a round-trip route can be measured). There's a problem.

  Further, the method (4) has a problem that applications are limited.

In order to solve the above problems, there is a method that focuses on estimating end-to-end quality only from information obtained by passively measuring communication packets at one point (on the receiving side) of the network. The method can be used to minimize the impact on the quality of higher layers than IP when there are multiple packet transfer paths from the source to the destination, and when packets are continuously lost in the network. In the measurement information at a single point, the sequence number of the arriving (or passing) packet is not always in the order because the packet order is not sent out in sequence number order (interleave function). In such a situation, as a function in the TCP layer or higher layer, it is necessary to retransmit the packet in the IP layer (that is, a loss occurs in the network). That is, the packet loss rate cannot be calculated simply by counting the sequence numbers that have not arrived (or have not passed). As a solution to these problems, the measurement period is divided into fixed unit times, packet capture information (sequence number, arrival time) collected in the unit time is held, and each packet is ordered in arrival packet order or sequence number order. An algorithm for determining whether or not to retransmit has been proposed (see, for example, Non-Patent Document 4).
Konel Terplan "Lan network management technique" published by Soft Research Center, 1994, pp. 108-110 Takadoi et al., “Evaluation of the effect of test packet transmission pattern on quality degradation detection of video streaming, 2002 IEICE General Conference B-11-7 Translated by Jun Murai, “Internet System Handbook”, Ipress, 1996, p. 555 Kawamura, Terada "Network quality estimation method by one-point observation on the receiving side", IEICE General Conference, B-7-107, 2004, 3

  However, a method that focuses on estimating end-to-end quality using only information obtained by passively measuring the communication packet at one point (on the receiving side) of the network is packet capture information (packet header) at one point on the receiving side. Can be used to estimate (measure) end-to-end packet loss rate, etc. using only the sequence number, packet arrival time, etc.), but all of the packet capture information must be retained for a certain period of time. There are problems such as the need for a large capacity data storage memory and the CPU processing time required to process a large amount of information.

  The present invention has been made in view of the above points. By improving the method for calculating end-to-end performance (IP packet loss rate, etc.), the information that must be retained for a certain period is greatly reduced. An object of the present invention is to provide an end-to-end quality estimation method and apparatus capable of measuring end-to-end performance in real time only by observation information at one point on the receiving side.

  FIG. 1 is a principle configuration diagram of the present invention.

According to the present invention (Claim 1), there is a data loss in the middle in order to ensure data transfer by a function of a protocol including TCP / IP or RTP or an application function positioned above a communication protocol such as a video distribution application. An end-to-end quality estimation device for measuring communication quality by monitoring at a single point between the transmission side and the reception side in communication for performing data transfer again,
A flow extraction unit 120 that extracts a flow from the start to the end of communication between transmission and reception terminals, which is a measurement target, from the input packet header information;
A retransmission number counter 151 provided on the memory for counting the number of retransmissions for each sequence section obtained by dividing the space composed of sequence numbers;
An arrival number counter 152 provided on the memory for counting the number of arrivals of packets for each sequence section obtained by dividing the space composed of sequence numbers;
Retransmission packet determination means 131 for determining whether or not the packet is retransmitted every time a packet arrives, and operating a retransmission number counter and an arrival number counter in a section to which the packet belongs;
And a quality information calculation unit 132 for calculating end-to-end quality using the values of the retransmission number counter 151 and the arrival number counter 152, triggered by the temporal transition of the sequence number of the arrived packet.

Further, the present invention (Claim 2) is a retransmission packet determination means,
When determining whether or not a packet is a retransmission, storage means for storing information on the maximum value of the sequence number, the arrival time of the packet, and information on the average increase rate of the sequence number among the information on the packets that have arrived in the past;
Update means for sequentially updating information stored in the storage means;
And means for determining whether or not the packet is a retransmission packet based on only the sequence number and arrival time of the packet and information stored in the storage means when the packet to be determined arrives.

Further, according to the present invention (Claim 3), in the quality information calculation means 132,
When calculating end-to-end quality,
Based on the packet header information triggered by packet arrival, it has means for updating the counter value of the arrival number counter and retransmission number counter for each sequence section obtained by dividing the space consisting of the sequence number,
Arrival number counter and retransmission number counter
As the sequence number of the arriving packet progresses, it has an algorithm mechanism in which the sequence number can be used as a counter of the further sequence section.

  FIG. 2 is a diagram illustrating the principle of the present invention.

According to the present invention (Claim 4), when there is data loss in the middle in order to perform data transfer reliably by a function of a protocol including TCP / IP or an application function positioned above a communication protocol such as a video distribution application. In addition, in communication for performing data transfer again, in an end-to-end quality estimation method for measuring communication quality by monitoring at one point between the transmission side and the reception side,
A flow extraction step (step 1) for extracting a flow from the start to the end of communication between transmitting and receiving terminals, which is a measurement target, from the input packet header information;
Each time a packet arrives, it is determined whether or not the packet is a retransmission, and the number of retransmissions is counted for each sequence interval to which the packet belongs. A retransmission packet determination step (step 2) for operating an arrival number counter provided on the memory for counting the arrival number;
A quality information calculation step (step 3) for calculating end-to-end quality using the values of the number of retransmission counters and the number of arrivals counter using the time transition of the sequence number of the arrived packet as a trigger is performed.

  As described above, according to the present invention, when measuring end-to-end performance with a measuring device at one point on the receiving side, the large amount of data required to hold packet capture information over the entire period of the analysis unit time. A large-capacity memory and large-capacity data processing are not required, and the memory of the measuring apparatus can be reduced and the result can be output in real time.

  Therefore, a light measurement device can be applied to quality measurement of services provided via a large-scale network, and quality degradation can be detected in real time to provide services with good quality.

  Hereinafter, embodiments of the present invention will be described with reference to the drawings.

  FIG. 3 shows a system configuration according to an embodiment of the present invention.

  The system shown in the figure is connected to a network 30, a plurality of terminals 40A and 40B connected to the network 30, a plurality of routers 10A, 10B and 10C in the network 30, a server 20 connected to the network 30, and a router 10A. The single-point observation type measuring device 100 is configured.

  In the figure, it is assumed that the single point observation type measuring device 100 is connected to the terminal 40A and receives communication packets from the server 20 and the terminal 40B via the routers 10B and 10C.

  The system configuration and the configuration of the single point observation type measuring instrument in the first embodiment and the second embodiment shown below are the same, but the first embodiment shows the case of TCP / IP, In the second embodiment, the case of RTP is shown.

[First Embodiment]
In this embodiment, the case of TCP / IP will be described.

  FIG. 4 shows the configuration of a single point observation type measuring instrument according to the first embodiment of the present invention.

  The single-point observation type measuring instrument 100 shown in the figure includes a measurement target setting IF (interface) 101, an estimation parameter setting IF 102, an output parameter setting IF 103, a measurement unit 110, a flow extraction unit 120, a quality estimation unit 130, an output unit 140, And a storage unit 150.

  The measurement target setting IF 101 extracts information (communication source address, communication source port number, communication destination address, communication destination port number, communication direction) in the packet header (IP header and TCP header) of the communication packet, and the flow The data is input to a memory (not shown) of the extraction unit 120.

  The estimation parameter setting IF 102 includes a retransmission determination parameter (T) (W) for estimating the network quality, a preset arrival packet number (L), and the transmission side overlaps before the retransmission packet reaches the transmission side. A parameter (α) indicating how many times it has been sent, the number of section counters to be prepared (N1), a counter switching parameter (N2), and the like are input to a memory (not shown) of the quality estimation unit 130.

  The output parameter setting IF 103 inputs a constant (information unit to be transferred from the transmission side toward the reception side) C to a memory (not shown) of the quality estimation unit 130.

  The setting target setting IF 101, the estimated parameter setting IF 102, and the output parameter setting IF 103 are connected to the terminal 40A and accept input from the operator.

  When the measurement time and the measurement target are input from an arbitrary IF, the measurement unit 110 measures the measurement time and the communication packet of the measurement target, and inputs the measurement information to the flow extraction unit 120.

  From the packet header information of the measurement information input from the measurement unit 110, the flow extraction unit 120 is configured to measure a flow to be measured (communication between the transmission and reception terminals based on the packet header information set by the measurement target setting IF 101). Extract from start to end).

  The quality estimation unit 130 uses each parameter input from the estimation parameter setting IF 102 and uses a retransmission number counter and arrival number on a memory (not shown) for each sequence section obtained by dividing a space (sequence number space) composed of sequence numbers. A number counter is installed, and each time a packet arrives, it is determined whether or not the packet is retransmitted, and the counter (retransmission number counter, arrival number counter) of the section to which the packet belongs is operated. In this embodiment, since TCP / IP is adopted, the sequence number is information on the bytes to be transmitted. That is, in the case of TCP, since information to be transferred is divided and transferred as IP packets, it indicates from what byte to what byte in each IP packet (S1) to (S2-1). A plurality of sequence numbers are assigned.

  In addition, the quality estimation unit 130 uses the counter information of the retransmission number counter and the arrival number counter as described above, triggered by the time transition of the sequence number of the arrival packet as a trigger, and the end-to-end quality (IP packet loss rate, loss Byte rate, retransmission rate, etc.).

  The output unit 140 outputs the end-to-end quality calculated by the quality estimation unit 130 as network quality information.

  The storage unit 150 stores the maximum value of the sequence number, the arrival time of the packet, and the average increase rate of the sequence number among the information on the packets that have arrived in the past.

  FIG. 4 is a flowchart of the operation in the first embodiment of the present invention.

  Step 110) Measurement data including measurement start time, arrival time, and packet header information is input to the measurement unit 110. The packet measurement can be realized using, for example, Tcpdump, which is one of network packet capturing tools. In the present embodiment, TCP / IP communication will be described as an example. FIG. 6 shows information that can be acquired in that case. The information output by Tcpdump includes items such as time, transmission source address, arrow (communication direction), flag, sequence number, etc., shown in bold in FIG. Assume that the information shown in FIG. 6 (hereinafter referred to as packet information) is obtained every time a packet arrives. The sequence number is set or processed so as to be assigned from 1.

  Step 120) The flow extraction unit 120 sets the target based on the measurement target parameters (communication source address, communication source port number, communication destination address, communication destination port number, communication direction) set by the measurement target setting IF 101. The arrival packet information is extracted from the measurement result output from the measurement unit 110. Here, the TCP protocol is identified by looking at “protocol” in the IP header.

  Next, referring to the TCP header information of the arriving packet in time series order shown in FIG. 7, the flag is set to S (next connection establishment time) from the time of the packet whose “flag” is “S” (connection establishment time). ), Or F (connection release time), R (connection forced disconnection time) is one flow (that is, communication from establishment of TCP session to release).

  (Step 130) The quality estimation unit 160 performs a process of determining whether or not the packet is a retransmission packet (Step 131), a process of counting arrival packets for each sequence space (Step 132), and a process of calculating network quality information (Step 133). The network quality (such as packet loss rate at the IP layer) is output to the output unit 140. Each process will be described later.

  Below, each process in the quality estimation part 160 is demonstrated. In the following, the calculation of output items is specified in units of sequence numbers (in the case of TCP, the number of bytes). Here, the unit constant is C in advance. That is, it means that the network quality is calculated every time the sequence number of the packet (s) sent by the transmission side increases by C.

  Further, since the circulation problem due to the upper limit of the sequence number counter can be easily avoided by using the residue class, the sequence number of the packet (data) transmitted first from the transmission side is set to “1” here.

Step 131) Processing for determining whether or not the packet is a retransmission packet:
First, the following symbols are defined in advance in the estimation parameter setting IF 102.

to: Measurement start time;
Pj: the jth packet arrived is called packet j;
Sj: Sequence number of Pj (in the case of TCP, it is expressed using S1 and S2, but here it is S1 value);
m (t): average increase rate of sequence number from measurement start to time t (defined by increase of sequence number per unit time) (stored in storage unit 150);
Smax (t): the maximum value (stored in the storage unit 150) of the entire sequence numbers of packets that arrived up to a certain time t;
Tmax (t): time when the above packet arrives (stored in the storage unit 150);
T: retransmission determination parameter (time unit threshold);
W: Retransmission determination parameter (threshold with sequence number as unit);
Processing for determining whether or not the packet is a retransmission packet is as follows.

  (1) The quality estimation unit 130 updates Smax (t), Tmax (t), and m (t) stored in the storage unit 150 every time a packet arrives (time t).

  The values of Smax (t) and Tmax (t) in the storage unit 150 are updated if the sequence number of the current arrival packet is larger than the maximum sequence number at the time of arrival of the previous packet.

  For example, the following method can be used to update m (t) in the storage unit 150.

  (i) Let m (t) be the slope of a straight line connecting the coordinates (T1, S1) consisting of the first packet arrival time and the sequence number and (Smax (t), Tmax (t)).

  (ii) The increase in sequence number per unit time is calculated and set to m (t).

  (iii) In the above (ii), it is assumed that the sequence number is increased only in the latest (for example, within ΔT) range.

  (iv) m (t) is obtained by multiplying m (t) in (ii) above by a safety factor.

(2) Processing for determining whether or not the packet is a retransmission packet:
First, the quality estimation unit 130 does not determine retransmission when it is the first predetermined arrival packet within the Lth, and regards it as not all retransmission packets.

Next, for the Lth and subsequent arrival packets (referred to as the jth arrival packet),
If Sj <Smax (Tj) + m (Tj) × (Tj− (T + Tmax (Tj))), it is determined as a retransmission packet.

  As described above, it can be determined whether the packet is a retransmission packet.

  An example of determining whether or not the packet is the retransmission packet is shown in FIG. In the figure, the sequence number is shown for the sequence number S1, which is the parameter of FIG.

  Note that the method (2) for determining whether or not the packet is a retransmission packet can also be implemented as follows using the retransmission determination parameter W.

  When it is the first preset arrival packet within the Lth, retransmission determination is not performed and all packets are regarded as not being retransmission packets.

Next, for the Lth and subsequent arrival packets (referred to as the jth arrival packet),
If Sj <(Smax (Tj) −W) + m (Tj) × (Tj−Tmax (Tj)), it is determined as a retransmission packet.

  As described above, it can be determined whether the packet is a retransmission packet.

  An example of determining whether or not the packet is a retransmission packet using the retransmission determination parameter W is shown in FIG. In the figure, the sequence number is shown for the sequence number S1, which is the parameter of FIG.

Step 132) Counting of arrival packets for each sequence space:
The quality estimation unit 130 divides the space of the sequence number into designated constant C units and defines sections A1, A2,... (Hereinafter referred to as sequence space).

For example, as shown in FIG. 10, when C = 10 ⁴ (bytes), the network quality is calculated for each section An = [(n−1) · Cn · C]. This means that the quality index is calculated for each information unit (C) that the transmission side intends to transfer toward the reception side. The processing here is to increase the values of the retransmission number counter 151 and the arrival number counter 152 at the time of packet arrival.

  The counter includes a retransmission number counter Rn that is a counter of the number of bytes (packets) of retransmission packets and an arrival number counter Kn that is a counter of the number of arrival bytes (packets) of the interval An.

  For an arrival packet determined to be a retransmission packet, its sequence number (note that it is the width [S1, S2] in the case of TCP; see FIG. 6) is read, and the part (bytes) belonging to the corresponding section A1 Unit) is added to the retransmission counter Rn. However, when there are a plurality of sections, the following is performed on the retransmission number counter Rn for the plurality of sections.

Rn = Rn + # {[s1, S2) ∩ [(n-1) · C, n · C)]
However, # {·} is the number of sequence numbers included in the section {·}, and A∩B is a section common to the sections A and B.

For example, if section A = [5, 15) and section B = [10, 25),
# A = 10, # B = 15, A∩B = [10,15), # {A∩B} = 5
For an arrival packet that has not been determined as a retransmission packet, its sequence number (note that it is the width [S1, S2] in the case of TCP; see FIG. 6) is read, and the part belonging to the corresponding section A1 ( (In bytes) is added to the arrival number counter Kn. However, when there are a plurality of sections, the following is performed on the arrival number counter Kn for the plurality of sections.

Kn = Kn + # {[s1, S2) ∩ [(n−1) · C, n · C)]
However, # {·} is the number of sequence numbers included in the section {·}, and A∩B is a section common to the sections A and B.

Step 133) Network quality information calculation process:
The quality estimation unit 130 uses the processing for determining whether or not it is a retransmitted packet in steps 131 and 132 and the processing for counting the arrival packets for each sequence space as follows to apply the network quality (hereinafter applied to TCP communication). The method of calculating the loss byte rate in the IP layer).

  FIG. 11 is a flowchart of network quality information calculation processing according to the first embodiment of this invention.

  First, the quality estimation unit 130 uses counters Kn, Rn (n = 1, n) for the arrival packet counters of the sections A1, A2,... AN1 created by dividing the sequence number space {1, ∞) for each width C. 2,..., N1) are initialized and prepared (step 1331). n2 is set as a counter switching parameter N2 (step 1332). j = 1 is initialized (step 1333).

  Next, it is assumed that “the arrival of a packet belonging to the section An, n ≧ n2” is an event serving as an aggregation trigger of the counters K1, R1. That is, it is confirmed for each arrival packet whether the section to which the arrival packet belongs is after An2 (step 1334). If it belongs to the section after section An2, the values of the counters K1, R1 are summed up, and the section As for the quality index in “”, “Implementation of data in section An” described later is performed. After the processing, the counter itself is released, and new counters Kn + 1 and Rn + 1 are newly provided (step 1339). This process realizes a mechanism for saving the memory source by dynamically adapting the counter. This mechanism is not affected by the transfer rate of the network. In other words, the timing of aggregation processing for a certain section is set when a packet belonging to a section ahead of N2 (counter switching parameter) arrives for the first time. On the other hand, there is an advantage that the setting parameter value need not be changed according to the number of arrival packets per unit time.

  Further, in order to set an event that triggers the counting of the counters K2 and R2 to be processed next, n2 = n2 + 1, that is, n2 = N2 + 1 is set (step 1340).

  In step 1334, when n ≧ n2 is not satisfied (when it belongs to a young section), the processing of step 131 described above is executed as the determination processing for Pj (step 1335).

  Next, the processing of step 132 described above is executed as a counter processing for Pj (step 1336).

  The above processing is repeated until the last arrival packet (j = jmax) (step 1337). Finally, the sections for which data has not been aggregated are aggregated (step 1338).

  In addition, in the method in which the counting triggers (dynamic allocation) of the counters Kn and Rn are timed and the counting is sequentially performed according to the progress of time, an appropriate counting trigger time depends on the magnitude of the packet transfer rate. However, in the above method, since the relative position of the section (triggered if there is an arrival packet in the section ahead N2) is used as a trigger, it does not depend on the packet transfer rate.

  Next, a method of performing data aggregation for the section An performed in the above step 1339 will be described.

The loss byte rate in the section _{An2-N2 + 1} is calculated as follows using the counter values Kn, Rn at the time of step 1334.

但し、αは再送パケットが送信側に届くまでに、送信側が重複して何回送ったかを示すパラメータであり、例えば、高々１回の再送で届くならば、α＝１、通常はα＝１とする。

However, α is a parameter indicating how many times the transmission side has repeatedly sent before the retransmission packet reaches the transmission side. For example, α = 1, usually α = 1 if the retransmission packet arrives at most once. And

  Next, a method for performing data aggregation in step 1338 will be described.

The section to which P _jmax belongs is _assumed to be Am = [((m−1) · C, m · C).

At this time, it is assumed that (m−1) · C <S _max (P _jmax ) ≦ m · C. In addition, regarding the unaggregated section of An (n <m), the above-described method (1) is followed. Am is corrected as follows so that the denominator is not excessive.

但し、Ｃ’＝Ｓ_ｍａｘ（Ｐ_ｊｍａｘ）−（ｍ−１）・Ｃ
上記式（１）、（２）の損失バイト率の右辺は、正確には、「送信側における再送バイト数の割合（率）」であり、ここでは、ＩＰ損失バイト数を推定している。

However, C ′ = S _max (P _jmax ) − (m−1) · C
The right side of the loss byte rate in the above formulas (1) and (2) is precisely the “ratio (rate) of the number of retransmission bytes on the transmission side”, and here, the number of IP loss bytes is estimated.

[Second Embodiment]
In this embodiment, a case of RTP (overUDP) will be described.

  The configuration of the one-point observation type measuring instrument is the same as the configuration shown in FIG. 3 in the first embodiment described above.

  The measurement unit 110 acquires information (sequence number and the like) of the RTP header by Tcpdump as in the case of TCP in the first embodiment. However, the sequence number of RTP is a serial number that increases by +1, such as 1, 2, 3, 4,. Of the information shown in FIG. 5 used in the case of TCP, the sequence number portion is changed to that shown in FIG. 12, but is basically the same as the case of TCP in the first embodiment.

  In the case of UDP, the flow extraction unit 120 is the information in the packet header (IP header and UDP header), the communication source address, the communication source port number, the communication destination address, the communication destination port number, and the communication direction. Communication uniquely determined by the combination is defined as a flow. Therefore, in the flow extraction, header information is extracted for each communication source address, communication source port number, communication destination address, communication destination port number, and communication direction.

The quality estimation unit 130 is basically the same as in the case of TCP, but when the sequence number assignment system is TCP, there is a width such as “from S1 byte to S2 byte for each packet”. In the case of RTP, it is different in that it is assigned with a serial number for each packet. Therefore, the sequence number (S1, S1) described in the case of TCP is set to S1 = S2-1 in the case of RTP, and the sequence number of RTP is interpreted as the S1 value.
If steps 131 to 133 in the flowchart shown in FIG. 4 are executed, a packet loss rate in the IP layer can be obtained. The packet loss rate is a byte loss rate in the case of TCP, but differs in that a serial number is assigned to each packet in the case of RTP.

  In addition, the operation of the single point observation type measuring device described in the first embodiment and the second embodiment is constructed as a program, and is installed and executed on a computer used as the single point observation type measuring device. Alternatively, it can be distributed via a network.

  The present invention is not limited to the above-described embodiment, and various modifications and applications can be made within the scope of the claims.

  The present invention is applicable to a technique for measuring communication quality in a communication service using a packet switching network.

It is a principle block diagram of this invention. It is a principle explanatory view of the present invention. 1 is a system configuration diagram according to an embodiment of the present invention. It is a block diagram of the single point observation type measuring device in the 1st Embodiment of this invention. It is a flowchart of the operation | movement in the 1st Embodiment of this invention. It is an output example of Tcpdump in the first embodiment of the present invention. It is a figure which shows the IPv4 header structure in the 1st Embodiment of this invention. It is a determination example (an example using a retransmission determination parameter T) as to whether or not it is a retransmission packet in the first embodiment of the present invention. It is a determination example (example using the retransmission determination parameter W) whether it is a retransmission packet in the first embodiment of the present invention. It is an example of the sequence area in the 1st Embodiment of this invention. It is a flowchart of the calculation process of the network quality information in the 1st Embodiment of this invention. It is an output example of Tcpdump in the second embodiment of the present invention.

Explanation of symbols

10 router 20 server 30 network 40 terminal 100 communication quality measuring device, one-point observation type measuring device 101 measurement object setting IF (interface)
102 Estimated parameter setting IF (interface)
103 Output parameter setting IF (interface)
110 Measurement unit 120 Flow extraction unit, flow extraction unit 130 Quality estimation unit 131 Retransmission packet determination unit 132 Quality information calculation unit 140 Output unit 150 Storage unit, storage unit 151 Retransmission number counter 152 Arrival number counter

Claims (4)

In order to ensure data transfer by the functions of protocols including TCP / IP (Transmission Control Protocol / Internet protocol) or RTP (Real-time Transport Protocol), or application functions positioned above communication protocols such as video distribution applications An end-to-end quality estimation device for measuring communication quality by monitoring at a single point between a transmission side and a reception side in communication in which data transfer is performed again when there is data loss in the middle. ,
A flow extraction means for extracting a flow from the start to the end of communication between transmission / reception terminals to be measured from the input packet header information;
A retransmission number counter provided on the memory for counting the number of retransmissions for each sequence section obtained by dividing the space composed of sequence numbers;
An arrival number counter provided on the memory for counting the number of arrivals of packets for each sequence section obtained by dividing the space consisting of sequence numbers;
Retransmission packet determination means for determining whether or not the packet is retransmitted every time a packet arrives, and operating the retransmission number counter and the arrival number counter in a section to which the packet belongs;
Quality information calculation means for calculating end-to-end quality using the value of the retransmission number counter and the arrival number counter, triggered by the temporal transition of the sequence number of the packet that has arrived,
An end-to-end quality measuring apparatus comprising: The retransmission packet determination means includes
When determining whether or not the packet is a retransmission, storage means for storing information on the maximum value of the sequence number, the arrival time of the packet, and information on the average increase rate of the sequence number among the information on the packets that have arrived in the past;
Updating means for sequentially updating the information stored in the storage means;
The end-to-end quality measurement apparatus according to claim 1, further comprising means for determining whether or not the packet is a retransmission packet based on only the sequence number and arrival time of the packet and information stored in the storage means when the packet to be determined arrives. . The quality information calculation means includes
When calculating the end-to-end quality,
Based on the packet header information triggered by packet arrival, the means for updating the counter value of the arrival number counter and the retransmission number counter for each sequence section obtained by dividing the space consisting of the sequence number,
The arrival number counter and the retransmission number counter are:
2. The end-to-end quality measurement apparatus according to claim 1, further comprising an algorithm mechanism in which the sequence number can be used as a counter of a further sequence section as the sequence number of the arrived packet progresses. In communication that performs data transfer again when there is data loss in the middle to ensure data transfer by the function of the protocol including TCP / IP or the application function positioned above the communication protocol such as a video distribution application In the end-to-end quality estimation method for measuring the communication quality by monitoring at one point between the transmission side and the reception side,
A flow extraction step for extracting a flow from the start to the end of communication between transmission and reception terminals, which is a measurement target, from the input packet header information;
Each time a packet arrives, it is determined whether or not the packet is a retransmission, and the number of retransmissions is counted for each sequence interval to which the packet belongs. A retransmission packet determination step of operating an arrival number counter provided on the memory for counting the arrival number;
A quality information calculation step of calculating end-to-end quality using a value of the retransmission number counter and the arrival number counter, triggered by a temporal transition of the sequence number of the packet that has arrived,
An end-to-end quality measurement method characterized by:

Images