JP3392343B2 - Traffic data correction method, traffic measurement system, and computer network - Google Patents

Description

Detailed Description of the Invention

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention is used in a computer network such as the Internet for traffic analysis such as trend analysis for future capital investment, fault management, billing management, and the like. The present invention relates to a traffic data correction method for measuring numbers and the like and a traffic measurement system for realizing such a method.

[0002]

2. Description of the Related Art In a network node (router, hub, switch, etc.), MIB is used as device management information.
(Management Information Base) and SNMP (Si
There is also one that implements the mple Network Management Protocol) to obtain and set such management information remotely. For example, most current routers implement these.

Generally, for measuring traffic from a network node, a method of using SNMP to acquire items relating to traffic in the MIB of the network node is used. The mechanism of traffic measurement by SNMP is such that an SNMP request for acquiring a specific MIB value is sent to an SNMP agent of a network node, and the SNMP agent returns an SNMP response containing the specific MIB value. Hereinafter, from the network node to M using SNMP
Obtaining the IB value is referred to as “measurement by SNMP”. Since the traffic is concentrated on the network node, the traffic measurement using the management information of the network node is more effective in terms of resources than the method of newly connecting a dedicated network measurement device to the network for measurement. Then, there are the following methods for measuring traffic.

[Method Using MIB2] MIB2 has become a standard for network management. Among them, the number of transmission and reception octets for each interface of a network node (ifInOctets, i in the MIB2 Interface Group).
fOutOctets) exists. These values are collectively called an "octet counter". According to the specifications of MIB2, the octet counter is a counter that increases monotonically (maximum value 2 ³² −
1) is to be used. In traffic measurement using this octet counter, a method is used in which the octet counter value is periodically acquired at each time interval to be statistically calculated and the difference from the previous measured value is taken. For example, 1
If you want hourly statistical data, you only need to measure once an hour. Hereinafter, a set of octet counters measured periodically will be referred to as “measurement data”.

[Method Using Private MIB] Some vendors use their own private MIBs to calculate the average used bandwidth in the past fixed time, and hereinafter, these are collectively referred to as “throughput variable”. For example, in a router of Cisco, Inc., there is an item that represents the moving average of the used bandwidth for the past 5 minutes, so the traffic can be measured by periodically acquiring this value by SNMP. However, this method has a problem that traffic cannot be acquired at all when the measurement by SNMP fails. As the cause of the measurement failure by SNMP, since SNMP is UDP (User Datagram Protocol), packet loss occurs during transmission,
The network node has a low priority for reply processing to the SNMP request and therefore does not reply.

[0006]

[Problems to be Solved by the Invention] MI using SNMP
The following problems occur when acquiring the value of the octet counter of B2 and calculating the accurate statistical value of the traffic of the network node using this value. (1) When the octet counter reaches the maximum value, it returns to "0" (hereinafter referred to as "rounding of the octet counter").
However, this phenomenon occurs in a short time in a wide band transmission line. (2) Since the octet counter returns to "0" even when the network node is reset, it is indistinguishable from the circulation of the octet counter. (3) Measurement failure or measurement delay occurs due to SNMP.

Regarding the problem (1), for example, 100M
There is a risk that the octet counter will make a full turn in about 5.7 minutes in the (mega) band and in about 3.7 minutes in the 155M (mega) band (theoretical value). If the number of revolutions of the octet counter between SNMP measurements is one or less, the flow rate can be corrected by a simple correction of adding the maximum value of the octet counter (2 ³² -1) to the value of the octet counter measured this time. Although required, it is impossible to accurately calculate the flow rate when there is a possibility that the number of rotations is two or more. In order to prevent this, it is necessary to perform the measurement at a time interval (that is, the maximum value of the octet counter / network band) that guarantees that the octet counter does not make two or more rounds. However, if the measurement is performed at short time intervals, there is a problem in that the network or network node is overloaded.

On the other hand, the cause of the problem (3) is, as described above, that the packet is lost during the transmission because the SNMP is UDP, and the priority of the reply process to the SNMP request is high in the network node. Because it is low, it may be that the response is delayed or the response is delayed when the load is high. As described above, there are various problems in acquiring management information of a network node using SNMP and using this value to calculate an accurate measurement value for a certain period of time.

The present invention has been made in view of the above points, and it is an object of the present invention to cause a problem when calculating measured values at constant time intervals using management information of network nodes acquired by SNMP. Even under various conditions,
It is an object of the present invention to provide a traffic data correction method, a traffic measurement system, and a computer network equipped with such a system, which can accurately calculate statistical values of traffic.

[0010]

In order to solve the above problems, the invention according to claim 1 is a traffic data correction method for measuring traffic of a network, which network node provided in the network has. When the management information including the counter type data is periodically measured from the management information and the difference between the counter type data at the time of the previous measurement and the current measurement time is obtained to obtain the traffic statistical value at regular time intervals, the management information is used. Based on the above, the flow rate that has flowed between the previous measurement time point and the current measurement time point is estimated, and from the counter type data and the flow rate at the previous measurement time point and
It is characterized in that the number of rounds of the counter type data between the previous measurement time point and the current measurement time point is estimated, and the counter type data at the current measurement time point is corrected with a count value according to the number of rounds.

A second aspect of the present invention is a traffic data correction method for measuring traffic of a network, wherein management information including counter type data is included in the management information held by network nodes provided in the network. Is measured periodically, and when the difference between the counter type data at the time of the previous measurement and this time is calculated and the statistical value of the traffic is calculated at regular intervals, based on the management information, the measurement time of the previous measurement and the measurement of this time are measured. It is determined whether or not the network node is reset between the time points, and when the reset is detected, the counter type data at the previous measurement time point is added to the counter type data at the current measurement time point, and the “ The counter type data returned to 0 "is corrected. The invention according to claim 3 is characterized in that, in the invention according to claim 1 or 2, the counter type data at a specific time is obtained by linearly interpolating the counter type data before and after the time.

[0012] According to a fourth aspect of the invention, in a traffic measuring system for measuring the traffic of a network, a measuring means for measuring management information including counter type data from a network node provided in the network at regular intervals. Based on the management information, the flow rate estimating means for estimating the flow rate that has flowed between the previous measurement time point and the current measurement time point, and the previous measurement time point from the counter type data at the previous and current measurement time points and the flow rate. And the number of laps of the counter type data between the measurement time of this time, the correction means for correcting the counter type data at the time of this measurement with a count value according to the number of laps, and The difference between the counter-type data and the corrected counter-type data at the current measurement time point is obtained to obtain the traffic at the fixed time It is characterized by comprising a calculation means for calculating a statistical value.

Further, the invention according to claim 5 is a traffic measuring system for measuring the traffic of a network, and a measuring means for measuring management information including counter type data from a network node provided in the network at regular intervals. , Detecting means for detecting a reset of the network node between a previous measurement time point and a current measurement time point based on the management information, and measuring the counter type data at the previous measurement time point by the reset detection. A correction unit that adds to the counter type data at the time point and corrects the counter type data that has returned to “0” by the reset, and the previous counter type data and the corrected counter type data at the present measurement time point. And a calculating means for obtaining a statistical value of the traffic for each fixed time by obtaining a difference. It is characterized. Further, the invention according to claim 6 is the invention according to claim 4 or 5, further comprising an interpolating unit for linearly interpolating the counter type data before and after the specific time. It has a feature.

The invention according to claim 7 is the invention according to claims 4 to 4.
7. A computer network having the traffic measurement system according to any one of 6 above and a network node, wherein a plurality of the traffic measurement systems are provided, and each traffic measurement system has a plurality of ports that the network node has. The feature is that it is measured by a traffic measurement system of one unit. In the invention according to claim 8, in the invention according to claim 7, while monitoring that each of the traffic measurement systems is operating normally, the fact that a failure has occurred in a specific measurement system is detected. A management system for switching to another measurement system other than the specific measurement system is provided. The invention according to claim 9 is the computer network according to the invention according to claim 8, which is divided into a general network and a management network by the network node, wherein the traffic measurement system and the management system are the management networks. It is characterized in that it is provided in the communication network.

[0015]

DETAILED DESCRIPTION OF THE INVENTION An embodiment of the present invention will be described below with reference to the drawings. In the following, first, an outline of a method for calculating an accurate traffic statistical value will be described, and then the traffic measurement system according to the present embodiment will be described in detail. First, the octet counter, throughput variable, and sysUpTime (MIB2 system group) are measured at regular time intervals using SNMP. Here, sysUpTime is a counter value that counts the number of seconds since the network node started up in milliseconds. Then, for the measured octet counter value,
Correct and interpolate by the following procedure to calculate accurate traffic statistics.

(1) Counter Correction Reset Correction The sysUpTime measured this time and the sysUpTime measured last time are compared to determine whether or not there was a reset between the last measurement and the present measurement. If it is determined that there is a reset, the correction is performed by adding the value of the previous octet counter to the octet counter measured this time (see FIG. 1; details will be described later). In contrast,
If it is determined that there is no reset, do nothing.

The number of octets that has flowed between the previous measurement and the current measurement is calculated from the number of laps correction throughput variable, and the number of laps is calculated from the value of the calculated octet number. If the number of laps is “1” or more, the number of laps × (2 ³² −1) is added to the current throughput variable for correction (see FIG. 2; details will be described later). On the other hand, if there is no lap, nothing is done.

(2) Interpolation of measurement data If a leak occurs in the measurement data necessary for statisticalization, the values of the octet counters before and after are used to linearly interpolate them, and the value of the leaked octet counter is lost. (See FIG. 3A; details will be described later). Further, when the value of the octet counter at a specific time required for performing the statistic cannot be obtained due to the measurement delay, the value is obtained by linear interpolation between the measured values before and after (see FIG. 3).
See (b); details are given below). (3) Calculation of statistic value The statistic is calculated by obtaining the difference between the data necessary for statisticization (see FIG. 4; details will be described later). That is, the statistical value can be obtained by obtaining the difference between the values before and after the octet counter of the time required for the statistic obtained in “(2) Interpolation of measurement data”.

Next, a traffic measuring system according to an embodiment of the present invention will be described with reference to FIG. FIG. 1 is a block diagram showing the configuration of a computer network equipped with the traffic measurement system according to this embodiment. In the following, the configuration of the traffic measurement system will be described first, and then the operation inside the traffic measurement system will be described.

[Description of Configuration] FIG. 5 exemplifies a traffic measurement system that calculates the traffic of each port of the router as a statistical value. For that, two Cisc
o Measuring systems 3 and 4 are installed for each of the routers 1 and 2 manufactured by the company, and each measuring system measures the traffic entering and leaving the port of each router. In addition, router 1
Connects to customers C1 to C3, and router 2 connects customers C4 to C5
Connected with. Since the routers 1 and 2 have the same configuration, only the router 1 is shown in detail in FIG. This also applies to the measurement systems 3 and 4, and in FIG.
Only the detailed configuration is shown.

(1) Mutual backup of system This measuring system can measure a plurality of ports of a plurality of routers with one measuring system. With such a configuration, the load is distributed by a plurality of measurement systems, and when a failure occurs in one measurement system, the other measurement system covers it. Further, the measurement data are backed up between the measurement systems.

(2) Introduction of management system Management system 5 for managing the measurement systems 3 and 4
To introduce. The functions of the management system 5 include monitoring that the measurement system is operating normally, notification by a trap when an error occurs, and GUI (Graphical User Interfac).
This includes support for setting the measurement system using e), automatic switching of measurement to another measurement system when a specific measurement system fails.

(3) Construction of Management Network A management network 7 is constructed independently of the general network 6 normally used. By separating the traffic for management in this way, it is configured to support free communication between the measurement systems 3 and 4 and monitoring of the measurement system by the management system 5.

As described above, the router 1 has the MIB 11 and the SNMP agent 12 which are the management information of the equipment.
have. When the SNMP request is sent from the measurement system 1 side, the SNMP agent 12 receives the MIB
11, an SNMP response (octet counter, throughput variable, sysUpTime in the figure) including the specific MIB value is returned to the measurement system 1 side. The same applies to the router 2 as well. Further, each measurement system has a measurement process 31, a statistic process 32, and a data file 33 for storing measurement data, which performs the operations described in detail below.

[Explanation of Inside of Measurement System] In the measurement system 3 or the measurement system 4, the measurement process 31 and the statistic process 32 perform the measurement process and the statistic process, respectively. Therefore, each of these processes will be sequentially described below. ■ Measurement process The measurement process 31 consists of five MIBs listed below.
The value is measured every 5 minutes and is output to the data file 33 together with the measurement date and time and the measurement target router name (or IP (Internet Protocol) address).・ IfInOctets (MIB2 interface Group) ・ ifOutOctets (MIB2 interface Group) ・ sysUpTime (MIB2 system Group) ・ locIfInBitsSec (Cisco private MIB linterface
group) locIfOutBitsSec (Cisco private MIB linterface
group)

(3) Statistical Process The statistical process 32 calculates a statistical value for each hour based on the data file 33 created by the measurement process 31. Therefore, specific correction methods and statistical methods will be described below.

(1) Reset Correction The following reset detection / correction algorithm is used to determine the presence or absence of reset and perform correction. <Reset detection / correction algorithm> Calculation of the theoretical value of sysUpTime in this measurement The time interval (measurement interval T) between the time of the previous measurement and the measurement time of this time is calculated, and the sysUpTime previously measured
Add to. Here, sysUpTime (theoretical value) is the theoretical value that sysUpTime is supposed to reach at the time of this measurement,
The value of sysUpTime measured in the previous measurement is set to sysUpTi
Let me (previous measurement value). Then, sysUpTime (theoretical value) is calculated by the following equation. sysUpTime (theoretical value) = sysUpTime (previous measurement value) +
T

Reset decision Now, the theoretical value of sysUpTime calculated in the procedure is set to sysUpTi
me (theoretical value), sysUpTim measured in this measurement
The value of e is defined as sysUpTime (measured value), and the presence or absence of reset is determined as follows. sysUpTime (measured value) ≥ sysUpTime (theoretical value) → no reset sysUpTime (measured value) <sysUpTime (theoretical value) sysUpTime (theoretical value) ≥ 2 ³² -1 → no reset sysUpTime (theoretical value) < In the case of 2 ³² -1 → Reset is performed. And, if reset is detected, the value of the octet counter measured last time is added to the value of the octet counter measured this time to perform correction. On the other hand, if there is no reset, nothing is done and the process proceeds to “(2) Correction of number of laps”.

Here, FIG. 1 is a diagram showing how the reset correction is performed when the octet counter makes a revolution between the measurements. In the figure, the horizontal axis represents time, and the vertical axis represents the octet counter value (Octe
t), which is the same in FIGS. 2 to 4 described later. In FIG. 1, time t ₁₁ is the last measurement time, time t
₁₂ is the time when the network node is down, time t ₁₃ is the time when the network node is restarted due to the reset performed after the down, time t ₁₄ is the current measurement time, T
Is the measurement interval. Then, as shown in the figure, since the octet counter value at the previous measurement time point (time t ₁₁ ) is “r”, the measured octet counter value becomes “r” at the time of the current measurement (time t ₁₄ ). "I am making a correction.

(2) Correction of number of revolutions From loclfInBitsSec of Cisco, an approximate value of the flow rate for the past 5 minutes is calculated according to the following formula. Approximate flow rate for the last 5 minutes = locIfInBitsSec x 300
/ 8 where "300" is 5 minutes in seconds,
Further, "8" is for converting bit into byte. This is because the unit of locIfInBitsSec is [bit / sec] and the unit is adjusted to the unit of [byte] of ifInOctets. It should be noted that this formula is for obtaining an input value, and when calculating an output value, locIfOutBitsSec is used instead of loclfInBitsSec.

By using the approximate value calculated in, the number of laps is estimated according to the following method of estimating the number of laps. <Round number estimation method> The number of rounds n is calculated by the following formula. Number of revolutions n = int [(p + a-b) / max] where "a" is the value of the octet counter measured last time, "b" is the value of the octet counter measured this time,
“Max” is a constant (2 ³² −1), and “p” is the number of octets estimated from (an approximate value of the throughput variable). Further, "int []" represents an operation for cutting off the fractional part of the value in parentheses.

The octet counter is corrected using the number of revolutions n calculated in. Then, the number of octets flowing from the previous measurement time to the current measurement time is calculated by the following “circulation number correction method”. <Method of correcting the number of laps> octets = b−a + n × max

Here, FIG. 2 shows the reset detection and
FIG. 9 is a diagram showing a state of lap number correction performed when it is determined that there is a reset as a result of determining the presence or absence of a reset according to a correction algorithm. In the figure, times t ₂₁ and t ₂₂
Are the previous measurement time and the current measurement time, respectively. In addition, in FIG. 2, the result of the calculation based on the “circulation number estimation method” is 2
A case is shown in which it is calculated that one round (ie, n = 2) has occurred. Therefore, in FIG. 2, “2 × max” is added to the value “b” of the octet counter measured this time at time t ₂₂ .

(3) Interpolation of data required for statisticization When data required for statisticization (here, measured values at each time) are leaked, or data at that time is delayed and measured due to measurement delay. In this case, the values of the octet counter before and after are linearly interpolated to calculate the value of the octet counter at each time.

Here, FIG. 3 is a diagram showing how data is interpolated. Of these, FIG. 3A is a diagram in the case of compensating for the measurement omission, and at any of the times t ₃₁ , t ₃₃ , t ₃₄ , and t ₃₅ , it was possible to make a measurement, but at the time t ₃₄ that is supposed to be measured, It shows the case where it could not be done. The dashed rectangle with indicating the value of the octet counter that can not be measured at time t _34, from each measured value of the octet counter at time _{t 33, t 35, "●} " to obtain an interpolated value represented by It indicates that On the other hand, FIG. 3 (b) is a diagram showing a state in which the octet counter value at a specific time is obtained in the case of statisticization, etc., and shows that the measurement can be performed at the times t ₃₆ , t ₃₇ , and t ₃₉ . There is. In this case, since it is desired to obtain the octet counter value at the time t ₃₈ , the time t ₃₆ and the time t ₃₇ are interpolated to obtain the interpolated value indicated by "●".

(4) Statistical calculation Based on the measured values at each time obtained above, 1
The statistical value is obtained by obtaining the difference from the octet counter before the time. Here, FIG. 4 is a diagram showing a state at the time of performing the statistic, in which the difference between the octet counter values at the times t ₄₁ and t ₄₂ separated by the statistic interval TS is obtained to obtain the statistic value S. There is.

To summarize the above, as described in [Problems to be Solved by the Invention], in the measuring means provided as standard in the network node, the item representing traffic is of the counter type. A problem with this is that it is greatly affected by the revolution and reset of the octet counter. Further, as a problem when acquiring data using SNMP, leakage of measurement data and delay of measurement occur due to high load of the router and congestion of the network.

On the other hand, according to the present invention, by using the following three methods, it is possible to accurately calculate the statistical value for a fixed time even if these problems occur. Even if the measurement is made for a certain long time interval (that is, the measurement interval in which the octet counter may make two or more laps between the measurements by SNMP), the lap count can be estimated by correcting the lap count of the octet counter. It is possible to correct the octet counter value. By the reset correction, it is possible to distinguish between the circulation of the octet counter and the phenomenon that the octet counter returns to "0" by the reset. Further, the traffic data can be corrected even if the octet counter returns to "0" by the reset. By interpolating the data required for the statistic, the data required for the statistic can be estimated even if the measurement data is leaked or the measurement delay occurs.

[0039]

As described above, according to the invention described in claim 1 or 4, based on the management information of the network node, the flow rate that has flowed between the previous measurement time and the current measurement time is estimated, and this flow rate and The number of laps of the counter type data between the previous and present measurement time points is estimated from the counter type data at the previous and present measurement points, and the counter type data at the present measurement point is corrected according to the number of laps. . As a result, it is possible to accurately calculate the statistical value at a constant time interval even if the measurement has a long time interval such that the counter type data may make two or more rounds between the measurements. The accuracy of traffic measurement from the node is improved.

Further, in the invention according to claim 2 or 5, based on the management information of the network node, a reset of the network node between the previous measurement time and the current measurement time is detected and reset to "0". We are correcting counter type data that has been lost. As a result, even if the network node is reset due to a down or the like, it is possible to accurately calculate the statistic value at a fixed time interval, so that the accuracy of the traffic measurement from the network node is improved. In the invention according to claim 3 or 6, the counter type data at a specific time is obtained by linearly interpolating the counter type data measured before and after the time. As a result, due to the high load of the network node or the congestion of the network, even in a situation where the management information to be measured is leaked or the measurement delay occurs, it is possible to accurately calculate the statistical value for a certain period of time. Improves accuracy when making traffic measurements from network nodes.

According to the seventh aspect of the invention, in a network environment in which a plurality of traffic measurement systems are provided, a plurality of ports of network nodes can be measured by one traffic measurement system.
Accordingly, when the traffic is measured from the network node, the measurement can be performed if the communication such as the SNMP can be performed between the traffic measurement system and the network node. Therefore, the processing load can be distributed by distributing the traffic measurement system. Moreover, since the management information measured between the traffic measurement systems can be backed up mutually, a highly reliable traffic measurement system can be constructed.

Further, in the invention according to claim 8, a management system is provided to monitor whether or not each traffic measurement system is operating normally, and when a failure occurs in a specific traffic measurement system, other than this. It switches to another traffic measurement system.
As a result, even if a traffic measurement system fails, another traffic measurement system can cover it, thereby improving fault tolerance. In addition, when adding or deleting network nodes to be measured or ports in network nodes, work such as hardware connection is not required, so even if a traffic measurement system fails, Therefore, it is possible to start the measurement with another traffic measurement system, and it is possible to construct a highly reliable traffic measurement system.

According to the ninth aspect of the invention, the traffic measuring system and the management system are provided in the management network separated from the general network by the network node. This makes it possible to separate management traffic and support free communication between traffic measurement systems and monitoring of the traffic measurement system by the management system.

[Brief description of drawings]

FIG. 1 is an explanatory diagram showing a state of a lap number correction method when a lap number is determined to occur between measurements as a result of calculation according to a lap number estimation method according to an embodiment of the present invention. is there.

FIG. 2 is an explanatory diagram showing a state in which an octet counter is corrected when it is determined that there is a reset as a result of determining the presence or absence of a reset according to the reset detection algorithm according to the same embodiment.

3A and 3B are explanatory diagrams showing a state of a data interpolation method according to the same embodiment, FIG. 3A is a diagram for explaining a method of purely supplementing a measurement omission, and FIG. It is a figure explaining the method of calculating | requiring the octet counter value in the time of.

FIG. 4 is an explanatory diagram showing a state of statisticalization in the same embodiment.

FIG. 5 is a block diagram showing a configuration of a computer network including the traffic measurement system according to the same embodiment.

[Explanation of symbols]

1, 2 router 3,4 measuring system 5 management system 6 General network 7 management network 11 MIB 12 SNMP agents 31 Measurement process 32 Statistics process 33 data files

─────────────────────────────────────────────────── ─── Continuation of the front page (56) Reference JP-A-7-250077 (JP, A) JP-A-10-28145 (JP, A) (58) Fields investigated (Int.Cl. ⁷ , DB name) H04L 29/14 G06F 13/00 353 H04L 12/24 H04L 12/26

Claims (9)

(57) [Claims] 1. A traffic data correction method for measuring traffic of a network, wherein management information including counter type data is periodically measured from among management information of a network node provided in the network, When calculating the difference between the counter-type data at the time of this measurement and the statistical value of the traffic at regular intervals, the flow rate between the time of the previous measurement and the time of this measurement is calculated based on the management information. Estimate and estimate the number of laps of the counter-type data between the previous measurement time and the current measurement time from the counter-type data at the previous measurement time and the current measurement time, and the flow rate. A traffic data correction method characterized in that correction is performed with a count value according to the number of turns. 2. A traffic data correction method for measuring traffic of a network, wherein management information including counter type data is periodically measured from among management information held by a network node provided in the network, And when obtaining the statistical value of the traffic at regular time intervals by obtaining the difference between the counter-type data at the current measurement time, based on the management information, the network node of the network node between the last measurement time and the current measurement time. The presence or absence of a reset is determined, and when the reset is detected, the counter type data at the time of the previous measurement is added to the counter type data at the time of the current measurement, and the counter type returned to “0” by the reset. A traffic data correction method characterized by correcting data. 3. The traffic data correction method according to claim 1, wherein the counter type data at a specific time is obtained by linearly interpolating the counter type data before and after the time. 4. A traffic measuring system for measuring traffic of a network, comprising: measuring means for measuring management information including counter type data from a network node provided in the network at regular intervals; and based on the management information, A flow rate estimating means for estimating the flow rate that has flowed between the previous measurement time point and the current measurement time point, and the counter type data and the flow rate at the previous measurement time point and the current measurement time point from the previous measurement time point to the current measurement time point. A round number estimating means for estimating the round number of the counter type data, a correcting means for correcting the counter type data at the time of the current measurement with a count value according to the round number, and a counter type data of the previous time and the corrected number. A calculator that calculates the statistical value of the traffic for each fixed time by calculating the difference between the counter type data at the time of this measurement Traffic measurement system, characterized by comprising and. 5. A traffic measuring system for measuring traffic of a network, wherein measuring means for measuring management information including counter type data from a network node provided in the network at regular intervals, and based on the management information, Detecting means for detecting a reset of the network node between the previous measurement time point and the current measurement time point, and by detecting the reset, the counter type data at the previous measurement time point is added to the counter type data at the current measurement time point. Correction means for correcting the counter type data that has returned to "0" by the reset, and the difference between the previous counter type data and the corrected counter type data at the current measurement time point is calculated to obtain the fixed time interval Traffic measuring means for calculating statistical values of System. 6. The method according to claim 4, further comprising an interpolating unit that obtains the counter type data at a specific time by linearly interpolating the counter type data before and after the time.
Alternatively, the traffic measurement system according to item 5. 7. A computer network having the traffic measurement system according to claim 4 and a network node, wherein a plurality of the traffic measurement systems are provided, and each of the traffic measurement systems is A computer network characterized by measuring a plurality of ports of a network node with a single traffic measurement system. 8. A system for monitoring the normal operation of each of the traffic measurement systems and detecting the occurrence of a failure in a specific measurement system, and detecting the failure in a measurement system other than the specific measurement system. The computer network according to claim 7, further comprising a management system that performs switching. 9. A computer network separated into a general network and a management network by the network node, wherein the traffic measurement system and the management system include:
The computer network according to claim 8, wherein the computer network is provided in the management network.