JP2013157742A - Monitoring device, program and monitoring method - Google Patents

Abstract

PROBLEM TO BE SOLVED: To provide a monitoring device and others capable of monitoring a congestion state on a transmission side.SOLUTION: A monitoring device 1 obtains a segment length of a packet transmitted by monitoring object devices 2, 3 and 4. The device 1 compares the obtained segment length with a maximum segment length (MSS) which the monitoring object devices 2, 3 and 4 can receive, to calculate a rate of the number of packets having a segment length shorter than the MSS in a predetermined period to the number of packets transmitted by the monitoring object devices 2, 3 and 4 in the predetermined period. The device 1 compares the calculated rate with an above-mentioned rate previously obtained for the monitoring object devices 2, 3 and 4 in a normal state to determine whether or not the monitoring object devices 2, 3 and 4 are in a congestion state.

Description

  The present invention relates to a monitoring device, a program, and a monitoring method for monitoring a congestion state.

  Conventionally, in order to detect a congestion state and to obtain network performance, monitoring of the window size is performed in the TCP layer (for example, Patent Document 1 and Patent Document 2).

JP 2005-533449 A JP 2000-106557 A

  However, in the conventional technique, there is a protocol sequence for notifying the window size that can be received by the transmission side apparatus and notifying the congestion of the reception side apparatus, but it is not necessary to notify the reception side apparatus of the congestion of the transmission side apparatus. It cannot be monitored from the protocol sequence.

  In one aspect, an object of the present invention is to provide a monitoring device or the like that can monitor a congestion state of a transmission side device.

  The monitoring apparatus according to the present invention compares an acquisition unit that acquires a segment length of a packet transmitted by a monitoring target device, a segment length acquired by the acquisition unit, and a maximum segment length that can be received by the monitoring target device acquired in advance. A calculation unit that calculates a ratio of the number of packets having a segment length shorter than the maximum segment length within a predetermined time to the number of packets transmitted by the monitoring target device within the predetermined time; and a calculation performed by the calculation unit A notification unit that performs notification when the ratio is larger than a threshold value corresponding to the number of packets.

  According to an embodiment of the present application, it is possible to monitor the congestion state of the transmission side device.

It is explanatory drawing which shows an example of an information processing system. It is the graph which showed an example of the transmission interval distribution of a packet. It is a graph which shows an example of the relationship between CPU usage rate and the number of congestion packets. It is a block diagram which shows the hardware group of a monitoring apparatus. It is explanatory drawing which shows an example of the record layout of a connection information table. It is explanatory drawing which shows an example of the record layout of a MSS table. It is explanatory drawing which shows an example of the record layout of a normal / congestion packet number table. It is explanatory drawing which shows an example of the record layout of a congestion determination parameter table. It is a graph which shows an example of the relationship between the total packet number and the congestion packet number. It is a flowchart which shows the procedure of the observation process of the total packet number and the congestion packet number. It is the flowchart which showed the procedure of the monitoring process. It is a screen figure which shows an example of a congestion notification screen. It is a screen figure which shows an example of a congestion notification screen. It is a graph which shows the relationship between the total packet number and the congestion packet number. It is explanatory drawing which shows an example of the record layout of a line speed threshold value table. It is explanatory drawing which shows an example of the record layout of a network bandwidth table. It is a flowchart which shows the procedure of the observation process of the total packet number and the congestion packet number. It is the flowchart which showed the procedure of the monitoring process. It is explanatory drawing which shows an example of the record layout of a packet classification table immediately before. It is a flowchart which shows the procedure of the observation process of the total packet number and the congestion packet number. It is a functional block diagram which shows operation | movement of the monitoring apparatus which concerns on Embodiment 1-3. FIG. 10 is a block diagram illustrating a hardware group of a monitoring device according to a fourth embodiment.

Embodiment 1
Hereinafter, embodiments will be described with reference to the drawings. FIG. 1 is an explanatory diagram illustrating an example of an information processing system. The information processing system includes a monitoring device 1, a plurality of server computers 2, 3, 4, a switch 5 and a client computer 6.

  The switch 5 connects the monitoring device 1, the server computers 2, 3, 4 and the client computer 6 to each other. The monitoring device 1 is, for example, a server computer or a personal computer.

  The server computers 2, 3, and 4 receive services from the client computer 6 and provide services.

  In the transaction for providing the service, the server computers 2, 3, 4 exchange messages with the client computer 6 via the switch 5. Message transmission / reception between server computers is also performed via the switch 5. The monitoring device 1 monitors messages transmitted / received via the switch 5. The switch 5 has a function of mirroring data passing through the switch 5. Mirroring is a function for outputting the same data as data output to a certain port from other ports.

  The port to which the monitoring device 1 is connected is designated as the mirroring destination of the port to which the server computer 2 is connected, the port to which the server computer 3 is connected, and the port to which the server computer 4 is connected. As a result, packets addressed to the server computers 2, 3, and 4 are input to the destination server computers 2, 3, and 4 and also input to the monitoring device 1.

The monitoring apparatus 1 monitors whether or not the server computers 2, 3, and 4 to be monitored are in a congested state by monitoring the segment length of the packet transmitted by the server computers 2, 3, and 4 to be monitored. Judgment is made. The server computers 2, 3, and 4 are examples of monitoring devices.
Hereinafter, the server computers 2, 3, and 4 to be monitored are referred to as monitoring target devices 2, 3, and 4. In the following description, the processing performed by the monitoring device 1 is mainly described with respect to the monitoring target device 2, but it is assumed that the monitoring device 1 performs similar processing regarding the monitoring target devices 3 and 4.

  Next, a method for determining whether or not the monitoring target device 2 is in a congestion state will be described. FIG. 2 is a graph showing an example of a packet transmission interval distribution. The horizontal axis is the elapsed time. The unit is seconds. The vertical axis represents the packet transmission interval. The unit is milliseconds.

  In communication using TCP / IP (Transmission Control Protocol / Internet Protocol), when sending a message in the application layer, if the value exceeds the MTU (Max Transfer Unit) value in the IP layer, the message is divided into a plurality of packets and sent. . In this packet division, since the application makes a transmission request for each message to the kernel and then is divided and transmitted in the kernel, the packet transmission interval in the same message is relatively short, such as several microseconds.

  On the other hand, in the TCP layer, the maximum size MSS (Maximum Segment Size) of a receivable segment is notified to the other party when a connection is established, and a message to be transmitted is divided and transmitted in units of MSS. From the results shown in FIG. 2, the relationship between the segment length and the packet transmission interval can be found. When comparing the transmission interval between packets having the same segment length as the MSS and the transmission interval with the next packet after the packet whose segment length is shorter than the MSS, many of the latter transmission intervals are Is bigger than. That is, a packet having a large transmission interval has a segment length of MSS or less. MSS is an example of the maximum segment length that can be received.

  TCP congestion control on the transmission side has a congestion window (window size on the transmission side) and an advertising window (buffer capacity on the reception side), and the transmission side uses the smaller value of the two windows to determine the packet length. When the resource on the transmission side is depleted, for example, when the transmission buffer is depleted, the packet length to be transmitted is shortened and the packet is transmitted. Therefore, by monitoring the number of packets having a shorter segment length among the packets in the same message, it is possible to determine whether or not the monitored device 2 on the transmission side is in a congestion state.

  FIG. 3 is a graph showing an example of the relationship between the CPU usage rate and the number of congestion packets. The horizontal axis represents the CPU usage rate, and the unit is%. The vertical axis represents the number of congestion packets, and the unit is number / minute. FIG. 3 shows how the load is applied to the CPU with the same amount of request transmission load applied from the client to the server, and how the number of congestion packets changes depending on the CPU usage rate. As shown in FIG. 3, when the CPU usage rate is about 90% or less, the number of congested packets is almost constant, but when it exceeds 90%, it rapidly increases. Therefore, it is possible to detect that the CPU usage rate has approached 100% by increasing the number of congestion packets. Hereinafter, processing for detecting the congestion state of the monitoring target device 2 due to an increase in the number of congestion packets will be described in detail below.

  FIG. 4 is a block diagram illustrating a hardware group of the monitoring device 1. The monitoring device 1 includes a CPU (Central Processing Unit) 11 as a control unit, a RAM (Random Access Memory) 12, a storage unit 13, a clock unit 14, a communication unit 15, and the like. The CPU 11 is connected to each part of hardware via a bus. The CPU 11 controls each part of the hardware according to the control program 13P stored in the storage unit 13. The RAM 12 is, for example, SRAM (Static RAM), DRAM (Dynamic RAM), flash memory, or the like. The RAM 12 also functions as a storage unit, and temporarily stores various data generated when the CPU 11 executes various programs.

  The storage unit 13 is, for example, a hard disk or a large-capacity flash memory, and stores a control program 13P and a master table / data table 13T. The master table / data table 13T is composed of a plurality of tables storing data necessary for the packets transmitted and received by the monitoring target device 2. The clock unit 14 outputs the current time or an elapsed time starting from an arbitrary time point to the CPU 11.

  The communication unit 15 is a LAN (Local Area Network) card, for example, and transmits / receives information to / from the switch 5. The communication unit 15 uses the mirroring function of the switch 5 as described above. The communication unit 15 operates in promiscuous mode. The promiscuous mode is a mode in which all received packets are passed to higher-layer software regardless of the destination. By using the promiscuous mode, packets addressed to the monitoring target device 2 that are not addressed to the monitoring device 1 are not discarded and can be captured by the control program 13P.

  Next, the master table / data table 13T will be described. FIG. 5 is an explanatory diagram showing an example of the record layout of the connection information table 131. The connection information table 131 includes a connection destination IP address, a connection destination port number, a connection source IP address, a connection source port number, a connection number, and the like. The connection number is a primary key for specifying a connection. The connection destination IP address, the connection destination port number, the connection source IP address, and the connection source port number are acquired from a message exchanged when the connection is established and stored. The connection destination IP address and the connection source IP address can be acquired from the IP header, and the connection destination port number and the connection source port number can be acquired from the TCP header.

  FIG. 6 is an explanatory diagram showing an example of the record layout of the MSS table 132. The MSS table 132 includes a connection destination IP address, MSS, and the like. The MSS is obtained from a message SYN (SYNchronize) or SYN ACK (ACKnowledgement) exchanged when the connection is established. The MSS is stored in units of connection destination IP addresses.

  FIG. 7 is an explanatory diagram showing an example of the record layout of the normal / congested packet count table 133. The normal / congested packet count table 133 includes time, connection number, normal packet count, congestion packet count, and the like. The time is every unit time. As the number of normal packets and the number of congestion packets, the number of packets observed per unit time is stored for each connection number. In this embodiment, the unit time is 1 minute. The example shown in FIG. 7 indicates that 500,000 normal packets and 500 congestion packets were observed during communication of connection number 1 between 12:00:00 and 12:00:59. The above unit time is an example of a predetermined time. The ratio of the number of congestion packets to the number of normal packets stored in the normal / congestion packet number table 133 is an example of the ratio.

  FIG. 8 is an explanatory diagram showing an example of a record layout of the congestion determination parameter table 134. The congestion determination parameter table 134 includes a connection destination IP address, an approximate expression, an upper limit value 95% of a confidence interval, and the like. In the approximate expression, an approximate expression indicating the relationship between the total number of packets and the number of congested packets at normal time (load amount that does not cause congestion) is stored for each connection destination IP address. The stored approximate expression is obtained using the least square method. It is good also as memorize | storing not a formula but an inclination and an intercept. As the upper limit value 95%, a coefficient of the upper limit value 95% of the confidence interval is stored.

  FIG. 9 is a graph showing an example of the relationship between the total number of packets and the number of congestion packets. The solid line is a graph representing the approximate expression (y = kx + c) obtained by the least square method. The dotted line represents the upper limit 95% of the confidence interval as a graph. When the coefficient is α, the intercept of the upper limit 95% graph is c + α. The monitoring device 1 monitors the relationship between the total number of packets and the number of congested packets, and detects that the monitored device 2 is in a congested state when the area exceeds the upper limit of 95%. The ratio between the total number of packets and the number of congestion packets obtained from the graph shown in FIG. 9 is an example of the threshold value.

  Next, a method for observing the total number of packets and the number of congestion packets will be described. FIG. 10 is a flowchart showing the procedure for observing the total number of packets and the number of congestion packets. The CPU 11 of the monitoring device 1 monitors whether the communication unit 15 has received a packet (step S1). If not received (NO in step S1), the CPU 11 shifts the process to step S1. If received (YES in step S1), the CPU 11 shifts the process to step S2. The CPU 11 acquires connection information (step S2). The CPU 11 acquires the transmission source IP address and transmission destination IP address from the IP header of the received packet, and acquires the transmission source port number and transmission destination port number from the TCP header. The CPU 11 searches the connection information table 131 (step S3). It is checked whether or not the acquired connection information hits the search (step S4). When the search is hit, that is, when the acquired connection information is registered (YES in step S4), the CPU 11 determines that the search is upstream (step S5). If it is up, the process is terminated without doing anything.

  When the search does not hit, that is, when the acquired connection information is not registered (NO in step S4), the CPU 11 sets the transmission source IP address and the transmission destination IP address, the transmission source port number, and the transmission destination port number. Each is replaced, and the connection information table 131 is searched again (step S6). It is examined whether or not the search is hit (step S7). When the search is hit (YES in step S7), the CPU 11 acquires a connection number. If the search is hit, it is down, so the CPU 11 acquires the segment length (step S8). The CPU 11 compares the segment length with the MSS (step S9). The CPU 11 searches the MSS table 132 using the transmission source IP address as a key, and acquires the MSS. If the segment length is smaller than the MSS (YES in step S9), the CPU 11 adds 1 to the number of congestion packets. The CPU 11 adds 1 to the value in the congestion packet number column corresponding to the connection number in the normal / congestion packet number table 133 (step S10). When the record corresponding to the connection number does not exist in the normal / congestion packet number table 133, the CPU 11 adds a record to the normal / congestion packet number table 133. The CPU 11 ends the process. If the segment length is equal to the MSS (NO in step S9), the CPU 11 adds 1 to the number of normal packets. The CPU 11 adds 1 to the value in the normal packet number column corresponding to the connection number in the normal / congested packet number table 133 (step S11). The CPU 11 ends the process. The number of normal packets and the number of congestion packets are stored in the normal / congestion packet number table 133 in association with the time acquired from the clock unit 14. The number of congestion packets per unit time stored in the normal / congestion packet count table 133 is an example of the number of packets having a segment length shorter than the maximum segment length within a predetermined time. A value obtained by summing the number of normal packets and the number of congestion packets is an example of the number of packets transmitted by the monitoring target device.

  If the search is not hit (NO in step S7), the CPU 11 checks whether the packet is a connection establishment message (step S12). CUP 11 refers to the SYN flag and ACK flag of the TCP header. If it is not a connection establishment message (NO in step S12), since it is not a packet to be processed, the CPU 11 ends the process. In the case of a connection establishment message (YES in step S12), the CPU 11 performs connection connection confirmation (step S13).

  In the case of a SYN packet (SYN flag = 1, ACK flag = 0), the CPU 11 acquires a transmission source IP address and a transmission destination IP address from the IP header. The CPU 11 acquires a transmission source port number and a transmission destination port number from the TCP header. The CPU 11 stores the acquired transmission source IP address, transmission destination IP address, transmission source port number, transmission destination port number in a storage unit such as the RAM 12.

  In the case of a SYN / ACK packet (SYN flag = 1, ACK flag ≠ 0), the CPU 11 obtains the transmission source IP address and window size, that is, the MSS of the server, from the IP header. The CPU 11 stores the acquired MSS in the MSS table 132 in association with the server IP address (source IP address). If the record already exists in the MSS table 132, the CPU 11 overwrites the value of the MSS, and if not, the CPU 11 adds a new record to the MSS table 132.

  In the case of an ACK packet (SYN flag = 0, ACK flag ≠ 0), the CPU 11 acquires the transmission source IP address and transmission destination IP address from the IP header, and the transmission source port number and transmission destination port number from the TCP header. The CPU 11 acquires the transmission source IP address, the transmission destination IP address, the transmission source port number, and the transmission destination port number stored in the RAM 12 or the like. Confirm that the source port number and destination port number match, and confirm that the connection has been established.

  Since the connection is established, the CPU 11 sets connection information (step S14). The CPU 11 adds a record to the connection information table 131. The CPU 11 issues a connection number as a key. The CPU 11 stores the issued connection number, the acquired transmission source IP address, transmission destination IP address, transmission source port number, and transmission destination port number in the connection information table 131 in association with each other.

  Before the monitoring target device 2 is monitored, the above-described observation process is performed in a load amount that does not cause congestion, that is, in a normal state, and the number of normal packets and the number of congestion packets are observed. After the observation processing, for each connection IP address (the IP addresses of the monitored devices 2, 3, and 4), the least square method is used to calculate the total number of packets (the number of normal packets + the number of congestion packets) and the number of congestion packets. Find an approximate expression that shows the relationship. Further, the upper limit value 95% of the confidence interval of the approximate expression is obtained. The obtained approximate expression and the upper limit 95% of the confidence interval are stored in the congestion determination parameter table 134 in association with the connection destination IP address.

  Next, a procedure for monitoring processing of the monitoring target device 2 will be described. FIG. 11 is a flowchart showing the procedure of the monitoring process. In addition to the monitoring process shown in FIG. 11, the above-described process of FIG. 10 is performed, and the number of normal packets for each connection destination IP address (monitored devices 2, 3, and 4) in a predetermined time unit, for example, 1 minute unit. And the number of congestion packets are stored in the normal / congestion packet number table 133. The CPU 11 reads the latest normal packet number and congestion packet number associated with the monitoring target device 2 from the normal / congestion packet number table 133 (step S21). The CPU 11 acquires the approximate expression corresponding to the monitoring target device 2 and the upper limit value 95% of the confidence interval from the congestion determination parameter table 134, and from the number of normal packets acquired earlier, the approximate expression and the upper limit value 95% of the confidence interval. Then, the number of congestion packets is calculated (step S22).

  The CPU 11 compares the number of congestion packets acquired in step S21 with the number of congestion packets calculated in step S22 (step S23). When the acquired number of congestion packets is larger than the calculated number of congestion packets (YES in step S23), the CPU 11 notifies the administrator of the congestion. The notification is performed by displaying on a monitor connected to the monitoring apparatus 1 or by sending an e-mail to a designated e-mail address. In the above description, the observed number of congestion packets and the calculated number of congestion packets are compared. However, the ratio of the number of congestion packets to the total number of packets may be calculated and compared.

  12 and 13 are screen views showing examples of the congestion notification screen. In the window W1 shown in FIG. 12, the monitoring target device 2 in a congested state is displayed by a list. The column C1 displays the IP address of the monitoring target device 2 that has become congested, and the column C2 displays the time when the congestion state was observed. When the administrator moves the mouse pointer Mp to the column C1 or the column C2 and performs an operation such as click or double click, the CPU 11 may display the screen shown in FIG. In the window W2 shown in FIG. 13, the approximate expression G1 indicating the relationship between the total number of packets and the number of congestion packets in the normal state and the upper limit value 95% G2 of the confidence interval are shown in a graph, and the point P observed as the congestion state is displayed. Show. The administrator can visually determine the degree of congestion by referring to this screen. The window W1 or W2 described above is an example of notification.

  When the acquired number of congestion packets is equal to or less than the calculated number of congestion packets (NO in step S23), the CPU 11 shifts the process to step S25. The CPU 11 determines whether all the monitoring target devices 2, 3, and 4 have been processed (step S25). When the processes for all the monitoring target devices 2, 3, and 4 have been completed (YES in step S25), the CPU 11 ends the process. If there is an unprocessed monitoring target device (NO in step S25), the CPU 11 returns the process to step S21.

  As described above, in the present embodiment, the ratio of the congestion packets whose segment length is smaller than the MSS among all the packets transmitted by the monitoring target device 2 is observed. By detecting that the observed rate is higher than the normal rate, it is possible to detect the congestion state of the server on the data transmission side.

Embodiment 2
In the second embodiment, the congestion state of the monitoring target device 2 is more accurately detected in consideration of network line congestion. FIG. 14 is a graph showing the relationship between the total number of packets and the number of congestion packets. The horizontal axis is the total number of packets, and the vertical axis is the number of congestion packets. Both units are in pieces / minute. The load of gradually increasing the request transmission from the client to the monitored device 2 was applied, and the number of normal packets and the number of congestion packets were observed. Based on the observation result, the relationship between the total number of packets and the number of congestion packets is shown in a graph. As shown in FIG. 14, when the total number of packets is 3.1 million packets (dotted line), the network bandwidth becomes 900 Mbps (90% of the maximum line speed) or more, and reaches the physical limit (1 Gbps) of the line speed. At this time, the data transmission waiting state frequently occurs on the server side, so that there is a shortage of resources in the transmission buffer, and congestion packets are rapidly increasing. It can be seen that when the bandwidth used for the network line is 90% or more of the maximum line speed, the server becomes congested due to congestion of the network line. Therefore, in this embodiment, the bandwidth used for the network line is also measured, and CPU congestion and network line congestion are separated. Therefore, prior to monitoring, the physical line speed of the monitored device 2 is measured. Based on the measured line speed, a threshold value for determining that the line is congested is determined and stored.

  FIG. 15 is an explanatory diagram showing an example of the record layout of the line speed threshold table 135. The line speed threshold table 135 includes connection destination IP addresses (IP addresses of the monitoring target devices 2, 3, and 4), line speed thresholds, and the like. As described above, the line speed threshold value is set to a value of about 90% of the maximum speed by measuring the line speed before monitoring.

  FIG. 16 is an explanatory diagram showing an example of a record layout of the network bandwidth table 136. The network bandwidth table 136 includes time, connection number, number of bytes, and the like. The time is every unit time. For example, every minute. The number of bytes is obtained by adding the packet length of the downlink packet transmitted within the unit time for each connection number.

  FIG. 17 is a flowchart showing the procedure for observing the total number of packets and the number of congestion packets. It is almost the same as the processing procedure shown in FIG. 10, and is different in that step S15 is added. Hereinafter, step S15 will be mainly described. When receiving a downstream packet, the CPU 11 measures the packet length. The CPU 11 stores the packet length in the network bandwidth table 136 in association with the measurement time and the connection number (step S15). If there is already a record, the CPU 11 adds 1 to the stored packet length. If there is no record, the CPU 11 adds a record to the network bandwidth table 136 and stores the packet length. By this processing, the number of bytes of data transmitted for each monitored device and per unit time is stored.

  FIG. 18 is a flowchart showing the procedure of the monitoring process. Since it is almost the same as the processing procedure shown in FIG. 11, the different points will be mainly described. The CPU 11 compares the number of congestion packets acquired in step S21 with the number of congestion packets calculated in step S22 (step S23). If the acquired number of congestion packets is larger than the calculated number of congestion packets (YES in step S23), the CPU 11 compares the line speed with a threshold value (step S26). The line speed (communication speed) is calculated from the number of transmitted bytes per unit time stored in the network bandwidth table 136. As for the threshold value, the threshold value of the monitoring target device 2 is read from the line speed threshold value table. When the calculated line speed is equal to or less than the threshold (YES in step S26), the CPU 11 notifies the CPU congestion (step S241). When the calculated line speed is greater than the threshold (NO in step S26), the CPU 11 notifies the congestion of the line (step S242). The CPU 11 shifts the process to step S25. The CPU 11 determines whether all the monitoring target devices 2, 3, and 4 have been processed (step S25). If all the monitoring target devices 2, 3, and 4 have been processed (YES in step S25), the CPU 11 ends the processing. If there is an unprocessed monitoring target device (NO in step S25), the CPU 11 returns the process to step S21.

  As described above, in the present embodiment, the ratio of the congestion packets whose segment length is smaller than the MSS among all the packets transmitted by the monitoring target device 2 is observed. In addition, the line speed of the monitored device 2 is also observed. When it is detected that the rate of the observed congestion packet is higher than the normal rate, the observed line speed is compared with the line speed threshold to determine whether or not the line is in a congestion state. As a result, it is possible to determine whether the cause of the increase in congestion packets is congestion by the CPU of the monitoring target device 2 or line congestion. Therefore, it becomes possible to more appropriately detect the congestion state of the monitoring target device 2 on the data transmission side.

Embodiment 3
In Embodiment 1 or Embodiment 2 described above, all packets smaller than the MSS are handled as congestion packets. However, in the TCP layer, there is a case where the segment length is shortened by the congestion control (TCP Window Update, TCP Zero Window) on the reception side in addition to the congestion control on the transmission side. Here, the TCP Window Update is an ACK packet for changing the window size, and the TCP Zero Window is an ACK packet for temporarily stopping the packet from the transmission side by changing the window size to 0. In the case of congestion control on the receiving side, the window size changes, and the segment length of the transmission packet immediately before the TCP acknowledgment message (ACK) becomes shorter than the MSS. Therefore, ACK is detected, and even if the packet immediately before it is smaller than the MSS, it is excluded from the congestion packet. As a result, it is possible to detect only packets whose segment length is short by congestion control on the transmission side (monitored device 2 side). The above ACK packet is an example of a response confirmation message.

  FIG. 19 is an explanatory diagram showing an example of the record layout of the immediately preceding packet type table 137. The immediately preceding packet type table 137 includes connection numbers, packet types, and the like. As the connection number, the connection number set in the connection information table 131 is stored. The immediately preceding packet type stores the type of the latest packet transmitted by the monitored device 2.

  FIG. 20 is a flowchart showing the procedure for observing the total number of packets and the number of congestion packets. Here, differences from FIG. 17 described above will be mainly described. Steps S16 and S17 are added to the processing when a downstream packet is received. Steps S18 to S21 are added to the processing when an upstream packet is received. The CPU 11 compares the segment length of the packet (downstream packet) transmitted by the monitored device 2 with the MSS (step S9). If the segment length is smaller than the MSS (YES in step S9), the CPU 11 adds 1 to the number of congestion packets (step S10). The CPU 11 stores the packet type “congestion” in association with the connection number in the previous packet type table 137 (step S16). If there is a record corresponding to the connection number, the CPU 11 overwrites the record. If not, the CPU 11 adds the record to the immediately preceding packet type table 137 and stores the packet type “congestion”. If the segment length is greater than or equal to the MSS (NO in step S9), the CPU 11 adds 1 to the number of normal packets (step S11). The CPU 11 stores the packet type “normal” in association with the connection number in the previous packet type table 137 (step S17). If there is a record corresponding to the connection number, the CPU 11 overwrites the record. If not, the CPU 11 adds the record to the immediately preceding packet type table 137 and stores the packet type “normal”.

  After determining that the packet type is uplink, the CPU 11 checks whether there is an ACK bit (step S18). If there is no ACK bit (none in step S18), the CPU 11 ends the process. When there is an ACK bit (step S18), the CPU 11 reads the packet type corresponding to the connection number from the previous packet type table (step S19). If the packet type is congestion (congestion in step S19), the CPU 11 subtracts 1 from the number of congestion packets in the record corresponding to the connection number in the normal / congestion packet number table 133 (step S20). When the packet type is normal (normal in step S19), the CPU 11 subtracts 1 from the normal packet number of the record corresponding to the connection number in the normal / congested packet number table 133 (step S21). The CPU 11 ends the process. With the above processing, it is possible to exclude packets whose segment length has changed due to the congestion control on the receiving side.

  The determination as to whether or not the monitoring target device 2 is in a congested state is the same as in the first embodiment or the second embodiment, and thus the description thereof is omitted.

  As described above, in the third embodiment, since the packet whose segment length has been changed by the congestion control on the receiving side is excluded and the number of normal packets and the number of congestion packets are observed, the monitored device 2 is in a congestion state. It becomes possible to determine whether or not more accurately.

Embodiment 4
FIG. 21 is a functional block diagram illustrating the operation of the monitoring device 1 according to the first to third embodiments. When the CPU 11 executes the control program 13P and the like, the monitoring device 1 operates as follows. The acquisition unit 201 acquires the segment length of the packet transmitted by the monitoring target device 2. The calculation unit 202 compares the segment length acquired by the acquisition unit 201 with the maximum segment length that can be received by the monitoring target device 2 in advance, and the number of packets having a segment length shorter than the maximum segment length within a predetermined time. Calculates the ratio of the number of packets transmitted by the monitored device 2 within the predetermined time. The notification unit 203 notifies when the ratio calculated by the calculation unit 202 is larger than the threshold corresponding to the number of packets.

  The detection unit 204 detects that the monitoring target device 2 has received the confirmation response message. In addition, when the detection unit 204 detects reception of the confirmation response message, the calculation unit 202 excludes the packet transmitted by the monitoring target device 2 immediately before it from the calculation target of the ratio.

  The measuring unit 205 measures the communication speed of the monitoring target device 2. The line determination unit 206 compares the communication speed measured by the measurement unit 205 with the line speed of the network line used by the monitoring target device, and determines whether or not the network line is in a congestion state. The notification unit 203 also determines the notification content based on the determination result of the line determination unit 206.

  FIG. 22 is a block diagram illustrating a hardware group of the monitoring device 1 according to the fourth embodiment. A program for operating the monitoring device 1 reads a portable recording medium 1A such as a CD-ROM, a DVD (Digital Versatile Disc) disk, a memory card, or a USB (Universal Serial Bus) memory into a reading unit 10A such as a disk drive. May be stored in the storage unit 13. Further, a semiconductor memory 1B such as a flash memory storing the program may be mounted in the monitoring device 1. Further, the program can be downloaded from another server computer (not shown) connected via a communication network such as the Internet. The contents will be described below.

  The monitoring device 1 shown in FIG. 22 reads a program for executing the above-described various software processes from the portable recording medium 1A or the semiconductor memory 1B or downloads it from another server computer (not shown) via a communication network. To do. The program is installed as the control program 13P, loaded into the RAM 12, and executed. Thereby, it functions as the monitoring apparatus 1 mentioned above.

  The fourth embodiment is as described above, and the others are the same as those of the first to third embodiments. Therefore, the corresponding parts are denoted by the same reference numerals and detailed description thereof is omitted.

  In the first to fourth embodiments described above, the master table / data table 13T is stored in the storage unit 13 of the monitoring device 1. However, the present invention is not limited to this, and is stored in an external storage device. Data may be read or written directly from an external storage device or via a network.

In Embodiments 1 to 4 described above, the number of congestion packets, which is a criterion for determining whether or not there is a congestion state, is obtained from the approximate expression and the upper limit value 95% of the confidence interval each time. Absent. A plurality of values are determined as the total number of packets, a reference number of congestion packets corresponding to each total number of packets is calculated in advance, and the calculated number of congestion packets is stored in the storage unit 13 in association with the total number of packets. You can leave it. Further, a person may set the number of congestion packets corresponding to the total number of packets based on experience.
Furthermore, the number of congestion packets, which is a criterion for determining whether or not there is a congestion state, is determined for each monitored device. If there is no significant difference in performance among the monitored devices, a common approximate expression and a reliability An upper limit value of 95% or the like of the section may be used.

  Further, the above-described embodiment is an example in all respects, and should be considered not restrictive. The scope of the present invention is defined by the terms of the claims, rather than the meanings described above, and is intended to include any modifications within the scope and meaning equivalent to the terms of the claims.

  The following additional notes are further disclosed with respect to the embodiments including the first to fourth embodiments.

(Appendix 1)
An acquisition unit that acquires the segment length of the packet transmitted by the monitored device;
The segment length acquired by the acquisition unit and the maximum segment length that can be received by the monitoring target device acquired in advance are compared, and the number of packets having a segment length shorter than the maximum segment length is within the predetermined time. A calculation unit that calculates a ratio of the number of packets transmitted by the monitored device to
And a notification unit that performs notification when the ratio calculated by the calculation unit is greater than a threshold value corresponding to the number of packets.

(Appendix 2)
A measurement unit for measuring the communication speed of the monitored device;
A line determination unit that compares the communication speed measured by the measurement unit and the line speed of the network line used by the monitored device, and determines whether or not the network line is in a congested state;
The monitoring apparatus according to claim 1, wherein the notification unit determines notification contents based on a determination result of the line determination unit.

(Appendix 3)
A detection unit for detecting that the monitored device has received an acknowledgment message indicating that the arrival of the transmitted packet has been confirmed;
When the detection unit detects reception of the confirmation response message, the calculation unit excludes a packet transmitted by the monitoring target device immediately before reception of the confirmation response message from the calculation target of the ratio. The monitoring device described.

(Appendix 4)
Compare the segment length of the packet to be transmitted by the monitoring target device and the maximum segment length that can be received by the monitoring target device acquired in advance.
Calculating the ratio of the number of packets having a segment length shorter than the maximum segment length in a predetermined time to the number of packets transmitted by the monitored device within the predetermined time;
A program for causing a computer to execute a notification process when a calculated ratio is larger than a threshold corresponding to the number of packets.

(Appendix 5)
Measure the communication speed of the monitored device,
Compare the measured communication speed and the line speed of the network line used by the monitored device,
Determining whether the network line is congested;
The program according to appendix 4, wherein notification content is determined based on the determination result.

(Appendix 6)
Detecting that the monitored device has received an acknowledgment message indicating confirmation of arrival of the transmitted packet;
The program according to claim 4 or 5, wherein a packet transmitted by the monitoring target device immediately before receiving the confirmation response message is excluded from the ratio calculation target.

(Appendix 7)
A monitoring method executed by a computer,
The segment length of the packet transmitted by the monitoring target device is compared with the maximum segment length that can be received by the monitoring target device acquired in advance, and the number of packets having a segment length shorter than the maximum segment length within a predetermined time is Calculate the ratio of the number of packets sent by the monitored device in time,
A monitoring method for notifying when the calculated ratio is larger than a threshold corresponding to the number of packets.

(Appendix 8)
Measure the communication speed of the monitored device,
Compare the measured communication speed and the line speed of the network line used by the monitored device,
Determining whether the network line is congested;
The method according to appendix 7, wherein notification content is determined based on the determination result.

(Appendix 9)
Detecting that the monitored device has received an acknowledgment message indicating confirmation of arrival of the transmitted packet;
The method according to claim 7 or 8, wherein a packet transmitted immediately before reception of the confirmation response message is excluded from a target for calculating the ratio.

1 Monitoring device 11 CPU
12 RAM
13 Storage Unit 13P Control Program 13T Master Table / Data Table 131 Connection Information Table 132 MSS Table 133 Normal / Congestion Packet Number Table 134 Congestion Judgment Parameter Table 135 Line Speed Threshold Table 136 Network Bandwidth Table 137 Immediate Packet Type Table 14 Clock Unit 15 Communication Part 2, 3, 4 Server computer (monitored device)
5 switch 6 client

Claims (5)

An acquisition unit that acquires the segment length of the packet transmitted by the monitored device;
The segment length acquired by the acquisition unit and the maximum segment length that can be received by the monitoring target device acquired in advance are compared, and the number of packets having a segment length shorter than the maximum segment length is within the predetermined time. A calculation unit that calculates a ratio of the number of packets transmitted by the monitored device to
And a notification unit that performs notification when the ratio calculated by the calculation unit is greater than a threshold value corresponding to the number of packets. A measurement unit for measuring the communication speed of the monitored device;
A line determination unit that compares the communication speed measured by the measurement unit and the line speed of the network line used by the monitored device, and determines whether or not the network line is in a congested state;
The monitoring apparatus according to claim 1, wherein the notification unit determines notification contents based on a determination result of the line determination unit. A detection unit for detecting that the monitored device has received an acknowledgment message indicating that the arrival of the transmitted packet has been confirmed;
3. When the detection unit detects reception of the confirmation response message, the calculation unit excludes a packet transmitted by the monitoring target device immediately before reception of the confirmation response message from the calculation target of the ratio. The monitoring device described in 1. Compare the segment length of the packet to be transmitted by the monitoring target device and the maximum segment length that can be received by the monitoring target device acquired in advance.
Calculating the ratio of the number of packets having a segment length shorter than the maximum segment length in a predetermined time to the number of packets transmitted by the monitored device within the predetermined time;
A program for causing a computer to execute a notification process when a calculated ratio is larger than a threshold corresponding to the number of packets. A monitoring method executed by a computer,
The segment length of the packet transmitted by the monitoring target device is compared with the maximum segment length that can be received by the monitoring target device acquired in advance, and the number of packets having a segment length shorter than the maximum segment length within a predetermined time is Calculate the ratio of the number of packets sent by the monitored device in time,
A monitoring method for notifying when the calculated ratio is larger than a threshold corresponding to the number of packets.

Images