JP2006140761A - Ethernet load tester - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To provide an ethernet load tester which accurately and easily measures pose response time which is a kind of an IEEE802.3x flow control function mounted on a switch or the like belonging to a layer 2 particularly regarding the ethernet load tester for measuring an inter-frame gap. <P>SOLUTION: The ethernet load tester is provided with a transmission means for repeatedly transmitting a test frame to an object to be measured having a memory, a reception means for receiving the test frame looped back from the memory, and a pose frame transmission means for transmitting a pose frame for determining the time to store the test frame in the memory. An inter-frame gap measurement means for measuring the inter-frame gap enlarged by the pose frame is provided. <P>COPYRIGHT: (C)2006,JPO&NCIPI

Description

  The present invention relates to an Ethernet (registered trademark) load tester that measures an inter-frame gap (hereinafter also referred to as “IFG”), and is particularly a kind of IEEE 802.3x flow control function implemented in a switch or the like belonging to layer 2. The present invention relates to an Ethernet load tester that accurately and easily measures pause response time.

  In general, for the development of network devices (media converters, LAN switches, routers, transmission devices) equipped with Ethernet, performance evaluation at the time of manufacturing, network throughput verification for IP network construction, measurement of throughput, delay time measurement A measurement device called an Ethernet load tester is required.

  One of the functions of such an Ethernet load tester is pause response time measurement. Pause response time measurement means that the Ethernet load tester outputs a command not to output the set time test frame to the network device (hereinafter also referred to as pause frame), and the network device that receives this command follows the command. It is to test whether it works. As a technique for measuring such an interframe gap, there is a technique as described in Patent Document 1.

JP 2004-120504 A

  Hereinafter, the usage method of the Ethernet load tester will be described with reference to FIG. 3 which is a connection example of the Ethernet load tester. The Ethernet load tester 1 is a device that measures a pause response. The measurement target 200 is a network device such as a switch, includes a buffer (memory) (not shown), and is connected to the Ethernet load tester 1 via a communication line such as a LAN cable. The port a of the Ethernet load tester 1 is connected to the port n of the measurement target 200, and the port b is connected to the port m.

  Next, the configuration of the Ethernet load tester 1 in FIG. 3 will be described in detail with reference to FIG. The Ethernet load tester 1 includes a control unit 2, a transmission CPU / IF 3, a transmission frame control unit 4, a transmission frame generation unit 5, a transmission frame pattern memory 6, a physical layer device 7, a reception frame termination unit 8, and reception frame statistics. Unit 9, a receiving CPU / IF 10a, a capture memory 11, and a capture control unit 12. 4 is a block diagram on the side of the port a in FIG. 3. Since the configuration of FIG. 4 exists for each of the ports a, b,... In FIG. Although it exists on the port b side, the description is omitted.

  The control unit 2 controls the entire Ethernet load tester 1. The transmission CPU / IF 3 is an interface between the control unit 2 and the transmission frame control unit 4, and transmits a command output from the control unit 2 to the transmission frame control unit 4. The transmission frame control unit 4 selects a frame to be transmitted from the test frames and pause frames stored in the transmission frame pattern memory 6 with respect to the transmission frame generation unit 5 based on an instruction from the control unit 2.

  The transmission frame generation unit 5 reads various frames stored in the transmission frame pattern memory 6, adds predetermined information to the frames, and transmits the frames to the measurement target 200 via the physical layer device 7. Here, the predetermined information includes, for example, the transmission time of the test frame. The transmission frame pattern memory 6 stores a test frame, a pause frame, and the like.

  The physical layer device 7 terminates the test frame at the physical layer level. The port a corresponds to the port a in FIG. 3, and a test frame is output from the port a to the port n of the measurement target 200 in FIG. On the other hand, the port x also corresponds to the port a in FIG. 3, but this is the port to which the test frame output from the port n of the measurement target 200 in FIG. 3 is input. Further, a test frame is output from the port b of FIG. 4 corresponding to the port b of FIG. 3 to the port m of the measurement target 200, and a test signal is transmitted from the port m of the measurement target 200 to the reception port y of FIG. Although it is input, since the configuration is the same as described above, the illustration is omitted.

  The reception frame termination unit 8 detects normality / abnormality (error frame) of the test frame input via the physical layer device 7. The received frame statistics unit 9 adds up the number of normal frames and the number of abnormal frames (number of error frames) of the test frames input from the received frame termination unit 8 so that the control unit 2 can determine whether the frame is normal or abnormal. Output to the receiving CPU / IF 10a.

  The reception CPU / IF 10 a notifies the control unit 2 of the statistical result calculated by the reception frame statistical unit 9. The capture memory 11 captures and stores the test frame input via the reception frame termination unit 8, and outputs the test frame to the control unit 2 via the reception CPU / IF 10a based on a command from the capture control unit 12. . The capture control unit 12 controls the capture operation of the capture memory 11.

  Next, the operation of FIG. 3 will be described with reference to FIG. 4 by dividing it into “test frame transmission operation”, “pause frame transmission operation”, and “test frame reception operation”.

  First, the “test frame transmission operation” will be described. The Ethernet load tester 1 repeatedly outputs a test frame from the port a. Here, for example, the test frame has a transmission rate of 100% as shown in FIG. 5A (that is, 960 ns when the interframe gap is 12 bytes and the speed is 100 Mbps).

  A command for setting the test frame transmission rate to 100% is input from a control terminal PC (not shown) connected to the control unit 2, and this command is transmitted via the control unit 2, the transmission CPU / IF 3, and the transmission frame control unit 4. It is output to the frame generation unit 5. The transmission frame generation unit 5 selects a test frame input by the control terminal PC from the test frames stored in the transmission frame pattern memory 6 and outputs a test frame having a transmission rate of 100% from the port a of the physical layer device 7. Output. This test frame is represented as shown in FIG. 5A, and is repeatedly transmitted while maintaining an interframe gap interval (for example, 12 bytes).

  This test frame is input to the port n of the measurement target 200 in FIG. 3 and stored in the internal buffer. After a certain period of time, the test frame is output from the port m and input to the port b of the Ethernet load tester 1. Hereinafter, the measurement target 200 stops outputting the test frame by receiving the pause frame from the Ethernet load tester. The “pause frame transmission operation” for setting the pause time will be described with reference to FIG. To do.

  The pause frame used for the “pause frame transmission operation” has a 2-byte field for specifying the pause time (stop time). The Ethernet load tester uses this pause time as a minimum of 0 slot time (pause release). To a maximum of 65535 slot times (corresponding to the stop time, 335.5 ms).

  The pause time is set by inputting a pause frame from the port b in FIG. First, a command for setting the pause frame setting time to, for example, 335.5 ms is input from the control terminal PC to a control unit (not shown) corresponding to the control unit 2 in FIG. This situation is shown in FIG. 5A, in which a pause frame is transmitted between pattern c and pattern d while maintaining an interframe gap between test frames transmitted repeatedly.

  Then, this command is output to a transmission frame generation unit (not shown) via the CPU / IF 3 and the transmission frame control unit (not shown) corresponding to the transmission CPU / IF 3 and the transmission frame control unit 4. The transmission frame generation unit selects the port frame input by the control terminal PC from the pause frames stored in the transmission frame pattern memory (not shown), and the set time from the port b of the physical layer device is 335.5 ms. Output the pause frame. In this way, the pause frame transmission operation ends.

  Next, the “test frame reception operation” of FIG. 3 will be described with reference to FIG. When the test frame output from the port m of the measurement target 200 receives the pause frame described above, the transfer of the test frame is stopped. When the pause time (335.5 ms) elapses, the measurement target 200 is output from the port m to the port b (that is, the port y corresponding to the port x in FIG. 4). Hereinafter, the “reception operation” of the test frame input to the port b will be described with reference to FIG. However, since the reception system corresponding to the port y is not shown in the figure, the block diagram (reference numerals 7, 8, 9, 10, 11, 12) of the reception system including the port x having the same configuration will be described here. To do.

  The test frame is input to the reception frame termination unit 8 via the port x of the physical layer 7. The test frame is detected as normal / abnormal (error frame) by the reception frame termination unit 8 and is output to the reception frame statistics unit 9. The reception frame statistics unit 9 determines the number of normal frames and the number of abnormal frames (number of error frames). Accumulated.

  On the other hand, the test frame branched from the reception frame termination unit 8 is captured by the capture memory 11 under the control of the capture control unit 12, and is output to the control unit 2 via the reception CPU / IF 10a.

  The captured information is displayed on a display unit of a control terminal PC (not shown) connected to the control unit 2, and the operator visually confirms the interframe gap. In FIG. 5 (c), the interframe gap between pattern 3 and pattern 4 is expanded by the “pause frame transmission operation”, so the interframe gap (960 ns) of the test frame that is repeatedly received depends on the pause frame. The operator confirms the value (time) obtained by subtracting the frame length of the test frame from the value obtained by adding the set time (335.5 ms) on the display unit of the control terminal PC.

  Theoretically, if the capacity of the capture memory 11 is large, a large number of test frames can be captured. Therefore, all test frames are captured, and Δ (delta) time of the test frame (second time from the time when the first test frame arrives) The interframe gap can be measured by subtracting the time of the test frame length from the time until the test frame arrives.

  However, the capture memory 11 may have a small capacity due to hardware restrictions. Therefore, when the capacity of the capture memory 11 is exceeded, a desired interframe gap that is enlarged by the pause frame may not be measured.

  In particular, when a test frame having a transmission rate of 100% (IFG: 12 bytes) capable of measuring the performance of the measurement target 200 with high accuracy is selected and measured, the number of test frames transferred at a predetermined time is determined. In many cases, the capacity of the capture memory 11 becomes insufficient and measurement cannot be performed.

  Specifically, in a test frame having a communication speed of 100 M, a test frame length of 64 bytes, and a transmission rate of 100%, test frames of 148,809 frames are transmitted per second, but the number of test frames that can be captured by the capture memory 11 Is sufficiently small, and the capacity of the memory 11 becomes insufficient when all the test frames are captured.

  Further, the result captured by the capture memory 11 can be confirmed by a display unit or the like of the control terminal PC (not shown), but it is necessary to visually check a test frame in which the interframe gap is widened by a pause frame. However, there is a problem that work efficiency is poor because information on a large number of test frames must be analyzed.

  Furthermore, since the pause frame has a 2-byte field in which the pause time (stop time) can be specified, the pause time (which means the time during which the test frame is stopped in the buffer of the measurement target 200 by the pause frame) is the minimum value. It can be set arbitrarily from 0 stop time (referred to as the time interval for sending data) to 65535 slot time (335.5 ms in terms of time), which is the maximum value, but the transmission rate is 10 ms or less. In some cases, the pause response time may not be measured.

  That is, for example, when it is desired to measure a pause response time of 10 ms at the port m of the measurement target 200, a test frame with a transmission rate of 100% (IFG: 12 bytes) is set in the transmission frame of the port a of the Ethernet load test. However, it is necessary to set a pause frame for the transmission frame of port b, but if the transmission rate of the pause frame is set to 10 ms or less, a new pause frame is transmitted before the pause time ends. As the measurement target 200, the flow control cannot be canceled, and the pause response time cannot be measured.

  Therefore, when the pause response time of 10 ms is measured, the transmission rate transmitted from the port b in FIG. 3 needs to be 10 ms or more, and it is necessary to set the measurement method and measuring instrument considering the pause time. Is complicated.

  The present invention relates to an Ethernet load tester for measuring an interframe gap, and in particular, an Ethernet load test for accurately and easily measuring a pause response time, which is a kind of IEEE 802.3x flow control function implemented in a switch or the like belonging to layer 2. The purpose is to provide a vessel.

In order to achieve such a problem, the invention described in claim 1
Transmitting means for repeatedly transmitting a test frame to a measurement target including a memory;
Receiving means for receiving the test frame looped back from the memory;
An Ethernet load tester comprising: a pause frame transmitting means for transmitting a pause frame for determining a time for storing the test frame in the memory;
Inter-frame gap measuring means for measuring an inter-frame gap enlarged by the pause frame.

The invention according to claim 2 is the invention according to claim 1,
The inter-frame gap measuring means obtains a time from the time when the receiving means receives the first test frame to the time when the second test frame is received, and from this time, the first or second test frame The interframe gap is calculated by subtracting the transmission time.

The invention according to claim 3 is the invention according to claim 1,
The inter-frame gap measuring unit includes a counter for obtaining a time at which the receiving unit receives the first test frame from a time at which the transmitting unit transmits the first test frame.
The interframe gap is calculated by subtracting the transmission time of the first or test frame from the time counted by the counter.

The invention according to claim 4 is the invention according to claim 1,
The transmission times of the first and second test frames are added to the test frame, the reception means reads these times to calculate Δ time, and the transmission time of the first test frame is calculated from the Δ time. The interframe gap is calculated by subtraction.

Furthermore, the invention according to claim 5 is the invention according to any one of claims 1 to 4,
The receiving means starts receiving the test frame when the transmitting means transmits a pause frame.

  The present invention has the following effects. Since the IFG measurement unit for measuring the interframe gap is provided, the interframe gap can be measured accurately and easily. That is, the interframe gap can be measured accurately and easily by utilizing the timer function and the frame identification function of the IFG measurement unit.

  Since the interframe gap is measured after a predetermined time has elapsed by utilizing the timer function of the IFG measurement unit, a desired interframe gap can be easily obtained without acquiring an interframe gap that is not required for pause response measurement. It can be measured accurately by operation. Furthermore, the response time of the ARP request frame with respect to the ARP reply frame and the response time of the PING request frame with respect to the PING reply frame that can roughly calculate the time from transmission of the pause frame to reception can be accurately measured with a simple operation. it can.

  Hereinafter, embodiments of the present invention will be described with reference to the drawings. Since the connection example of the Ethernet load tester of the present invention is the same as that shown in FIG. A configuration example of the Ethernet load tester according to the present invention will be described with reference to FIG. 1, and the same configurations as those in FIG.

  In FIG. 1, an Ethernet load tester 100 includes a control unit 2, a transmission CPU / IF 3, a transmission frame control unit 4, a transmission frame generation unit 5, a transmission frame pattern memory 6, a physical layer device 7, and a reception frame termination unit 80. The reception frame statistics unit 9 and the reception CPU / IF 10 are provided. That is, the reception frame termination unit 80 is provided instead of the reception frame termination unit 8 of FIG. 4 which is a block diagram of the prior art, and the capture memory 11 and the capture control unit 12 are unnecessary.

  The reception CPU / IF 10 outputs the signal input from the reception frame statistics unit 9 to the control unit 2. The reception frame termination unit 80 includes an IFG measurement unit 81. The reception frame termination unit 80 detects normality / abnormality (error frame) of a frame input via the physical layer device 7. In order to measure the interframe gap, the IFG measurement unit 81 has a function of identifying predetermined information included in the test frame (also referred to as “frame identification function”) in addition to a timer function.

  The interframe gap measured by the IFG measurement unit 81 is displayed on the display unit of the control terminal PC (not shown) connected to the control unit 2 via the reception frame statistics unit 9 and the reception CPU / IF 10, and the interframe gap is measured. The measurement is performed as follows.

  Next, the operation of FIG. 1 will be described. Among the “test frame transmission operation”, “pause frame transmission operation”, and “reception operation” described in the prior art, “test frame transmission operation” and “pause frame transmission operation”. "Is the same as the prior art, so the description is omitted, and only the" reception operation "is described.

  As a first method, the time from the time when the IFG measurement unit 81 receives the first test frame to the time when the second test frame is received is obtained, and the transmission time of the first test frame is subtracted from this time. To calculate the inter-frame gap.

  Specifically, for example, the time when the test frame of pattern 3 and the test frame of pattern 4 in FIG. 5C are received is obtained by a timer function, and the time difference Δ calculated from these test frames is 335.5 ms + 960 ns. To do. When the frame length of the first test frame is 64 bytes and the time is 5.76 μs in terms of time, the interframe gap between pattern 3 and pattern 4 is calculated to be 335.494 ms. The

  As a second method, the timer function which is the function of the IFG measuring unit 81 is utilized, the control unit 2 starts counting from the time when the test frame is output to the transmission frame generating unit 5, and the IFG measuring unit 81 performs this test. When a frame is received, the count is stopped, and a value (time) obtained by subtracting the frame length of the test frame from the count value (time) during this period is calculated as an interframe gap.

  As a second method, a function for identifying predetermined information added to the test frame is utilized, and the transmission time of the test frame is added to the first and second test frames by the transmission frame generation unit 5 in the “transmission operation”. The IFG measurement unit 81 receives these test frames, calculates Δ time from the additional information, and calculates a value (time) obtained by subtracting the frame length of the test frame from the Δ time as an inter frame gap.

  Thus, since the IFG measuring unit 81 for measuring the interframe gap is provided, the interframe gap can be measured accurately and easily. That is, the interframe gap can be measured accurately and easily by utilizing the timer function and the frame identification function of the IFG measurement unit.

  Next, a second embodiment of the present invention will be described with reference to FIG. However, the same components as those in FIG. 4 which is a block diagram of the conventional example or FIG. 1 which is a block diagram of the first embodiment are denoted by the same reference numerals and description thereof is omitted. The Ethernet load tester 110 includes a control unit 2, a transmission CPU / IF 3, a transmission frame control unit 4a, a transmission frame generation unit 5, a transmission frame pattern memory 6, a physical layer device 7, a reception frame termination unit 80, and reception frame statistics. Unit 9, a receiving CPU / IF 10, a capture memory 11, and a capture control unit 12a.

  The transmission frame control unit 4a outputs a command to output a pause frame stored in the transmission frame pattern memory 6 to the transmission frame control unit 5 and simultaneously outputs insert frame information to the capture control unit 12a. In addition to controlling the capture operation of the capture memory 11, the capture control unit 12a receives insert frame information from the transmission frame control unit 4a. Further, the capture control unit 12a outputs a control signal to the IFG measurement unit 82 based on the insert frame information input from the transmission frame control unit 4a. Here, the IFG measurement unit 82 has the same function as the IFG measurement unit 81 except that it operates based on this control signal.

  In such a configuration, for example, when the above-described capture function is used, the inter-frame gap can be similarly measured even when the capture control is performed using the insert frame information as a trigger. That is, when insert frame information is output from the transmission frame control unit 4a to the capture control unit 12a in FIG. 2, the capture control unit 12a outputs a timing signal for capturing a test frame to the capture memory 11, and starts a capture operation. . In this way, if a pause frame is set in the transmission frame pattern memory 6 as an insert frame, it is possible to measure the interframe gap using the capture function.

  In recent years, a network such as a wide area Ethernet network has been constructed, and the network path is made redundant in the backbone. Under such circumstances, for example, when a network path switching occurs due to a failure caused by a network device, it can be used for measuring a switching time by measuring an interframe gap.

  That is, for example, in the case where the IFG time is set as the capture trigger condition in FIG. 2, when an IFG time longer than the time set by the IFG measurement unit 82 is detected, the IFG measurement unit 82 outputs a control signal to the capture control unit 12a. At the same time, a timing signal is output to the capture memory 11 to control the start / stop of the capture operation.

  In this way, since the interframe gap is measured after a predetermined time has elapsed by utilizing the timer function of the IFG measurement unit 82, for example, the interframe gap between the patterns 1 and 2 in FIG. In this way, it is possible to accurately measure a desired interframe gap with a simple operation without acquiring an interframe gap that is unnecessary for the measurement of pause response.

  In addition, due to such effects, the response time of the ARP request frame to the ARP reply frame and the response time of the PING request frame to the PING reply frame can be calculated by the same simple operation. It can be measured accurately.

FIG. 3 is a block diagram of an Ethernet load tester according to the present invention. It is the block diagram which showed the 2nd Example of this invention. 2 is a connection diagram of an Ethernet load tester and a measurement target 200. FIG. It is a block diagram of the conventional Ethernet load tester. It is an example of the transmission timing of a test frame and a pause frame.

Explanation of symbols

2 Control unit 3 CPU / IF for transmission
5 Transmission Frame Generation Unit 6 Transmission Frame Pattern Memory 80 Reception Frame Termination Unit 81 IFG Measurement Unit 82 IFG Measurement Unit 100 Ethernet Load Tester 110 Ethernet Load Tester

Claims (5)

Transmitting means for repeatedly transmitting a test frame to a measurement target including a memory;
Receiving means for receiving the test frame looped back from the memory;
An Ethernet load tester comprising: a pause frame transmitting means for transmitting a pause frame for determining a time for storing the test frame in the memory;
An Ethernet load tester comprising an interframe gap measuring means for measuring an interframe gap enlarged by the pause frame.   The inter-frame gap measuring means obtains a time from the time when the receiving means receives the first test frame to the time when the second test frame is received, and from this time, the first or second test frame The Ethernet load tester according to claim 1, wherein the interframe gap is calculated by subtracting a transmission time. The inter-frame gap measuring unit includes a counter for obtaining a time at which the receiving unit receives the first test frame from a time at which the transmitting unit transmits the first test frame.
2. The Ethernet load tester according to claim 1, wherein the interframe gap is calculated by subtracting the transmission time of the first or second test frame from the time counted by the counter.   The transmission times of the first and second test frames are added to the test frame, and these times are read by the receiving means, and Δ time (the second test frame arrives from the time when the first test frame arrives) 2. The Ethernet load tester according to claim 1, wherein the interframe gap is calculated by calculating a time until the first frame is calculated and subtracting a transmission time of the first test frame from the Δ time. 5. The Ethernet load tester according to claim 1, wherein the receiving unit starts receiving the test frame when the transmitting unit transmits a pause frame.

Images