CN100459529C - Time delay characteristic measuring method in data communication network - Google Patents

Abstract

The present invention discloses a method of measuring the delay characteristics of a data communication network. The process method comprising: generating at least a transmitting end measurement message, a first mark by one transmitting timestamp by microcode / hardware, and the giving of such measurement message; receiving end separated measurement message, and one by microcode / hardware receiving a first time stamp mark; receiving end with its microcode / hardware isolated individually labeled second measurement message sending time stamp, and the above-described measurement message to the sending terminal via a data communications network; separating the transmitting end the measurement message, and use their own microcode / hardware individually labeled second receiving timestamp; a first transmission time stamp based on the measured labeled packets, receiving a first timestamp, the second timestamp and transmitting a second reception time stamp calculated delay characteristics between the transmitting end and a receiving end. This method of the present invention is simple, does not affect the actual operations, and can obtain a good measurement accuracy desired network.

Description

The method of measuring delay characteristics of a data communication network

FIELD

The present invention relates to a data communication network technology, particularly to measurement delay characteristics of a data communication network.

Background technique

As noted in the number of data communication networks, triple play of voice and video, voice and video data transmission higher and higher requirements on the network load, wherein the delay characteristic is an important indicator of network performance data bearing plant. It refers to the characteristic delay associated with the data transmission time in a network characteristics, comprising

Delay (Latency) and jitter (Jitter) and other indicators. Delay is the time required to transfer data packets in a network; a plurality of delay jitter refers to change in the packet transmission network with a period of time spent.

Delay characteristics is of great significance to measure existing data bearing plant network performance. For example, in the next generation network voice services, delay and jitter in real time these two important indicators, can provide reliable number noted to diagnose network problems. In addition, the delay characteristics but also provide the basis for fast and accurate performance analysis of user traffic. According to the delay characteristics of the measurement results online, you can find the network performance bottlenecks, optimize network operations, improve maintainability and manageability of the network, and ultimately improve customer satisfaction. After obtaining the delay characteristics, it can be applied to enhance the NTP packet processing, thereby improving the accuracy of the data network 2 to 3 times orders of magnitude.

Based on the above application, the delay characteristic on how to measure the data communication network packet line becomes a problem to be solved. In actual network operation, because the packet transceiver may be geographically separated by thousands of kilometers or more, it is a laboratory test instrument measuring the above is impractical. For this reason, the prior art has also employed a number of other practical ways to measure delay characteristic.

The method provided by a global positioning system (GPS) or other time of the whole network can be unified device HS clock accuracy of the order of the two different locations of the test instrument is maintained at an absolute time error) L1S the order, by recording the originating and terminating time of the time data packet, data packet transmission delay characteristic is obtained over the data communication network. This method is more general method for measuring the current delay characteristics.

This method has the following disadvantages:

(A) providing measurements require the use of GPS system clock and the GPS system supporting packet generator and the receiver. Since GPS systems are expensive, so the measurement method of the high cost of the measuring instrument requires relatively harsh.

(Ii) to reserve a dedicated test port in the device under test, by measuring the measured data transmission port, and the port can not be used to measure the actual service data packet transmission, on the one hand it reduces the utilization efficiency of the apparatus; on the other hand , delay characteristics can not be measured in the actual traffic flow in the business environment, is off-line measurement.

Method two, in general made by the network device diagnostic command "ping", the transmission and the reception time recorded by the software package, resulting in delay characteristics.

Since the operating time of the recording method by means of software, it is vulnerable to software execution efficiency factors, as well as system task scheduling tasks heavy load, so that a large error occurs measurement. In general, the delay characteristics obtained by this method, the measurement accuracy of the order of ms voice network supporting planting almost unusable.

SUMMARY

There Quan this, the main object of the present invention to provide a method of measuring the delay time characteristics of a data communication network, a control measurement error, measurement accuracy meet the requirement of practical application. To achieve the above object, the technical solution of the present invention is specifically implemented as follows: Measurement delay characteristics of a data communication network, the method comprising the steps of: when a, at the transmitting end to the receiving end transmits data packets, generating at least a measurement message, through self

Body microcode / hardware token sending time stamp is the first measurement message generated by one, and the geodetic

The amount of packets sent; receiving end separated from the received data packet to the measurement message, and use their own microcode / hardware isolated measurement message received by one of the first marking stamp;

B, the receiving end with its microcode / hardware isolated individually labeled second measurement message sending time stamp, and the above-described measurement message to the sending terminal via a data communications network; sending end from the received data packet separating said measurement message, and use their own microcode / hardware packet individually labeled second receiving time stamp is separated measurement message;

cl, sender and / or destination address is the receiving end generates its own loopback measurement packet, the loopback flag loopback measurement message transmitted by a time stamp microcode / hardware, and the measurement message sent from the ring;

c2, the sender and / or receiver end separated from the received data packets out of the ring from the measurement message, and the microcode / hardware isolated loopback flag loopback measurement message received timestamp;

c3, additional delay;

C, measured in accordance with a first transmission time stamp packets labeled, receiving a first timestamp, the second timestamp and transmitting a second receiving time stamp, calculating a time delay characteristic between the transmitting end and the receiving end of the time delay ; excess delay by subtracting the delay, the delay compensation correction is obtained.

Step a measurement message generated with measurement message identifier;

The method of measurement message is separated into: The measurement message identifier, using the method of separating the traffic classifier measurement message.

The stream classification method with three matching TCAM achieved.

The characteristics include latency time difference, the step c further comprising delay time calculation characteristic difference method is: receiving a first time stamp is added to each measurement message and the second sending time stamp, a first subtracting receiving a second transmission time stamp added and the time stamp, the obtained difference divided by two; then the results of all measurement packets were averaged to obtain the time difference between the transmitting end and the receiving end.

Said delay characteristics comprises a delay, the delay calculation method of step c of the delay characteristics between the transmitting end and the receiving end is: receiving a first time stamp is added to each second measurement message received time

The stamp and subtracting the first timestamp plus the transmission time stamp and the second transmission, the obtained difference divided by two to give

To each of the delay measurement packet; then all packet delay measurements averaged to obtain the sending and receiving end delay.

The method of calculating the additional delay c3 steps of:

The transmitting side is subtracted from the time stamp received from the ring loop transmission timestamp, additional delay;

Alternatively, the receiving end receives the time stamp from the ring transmission timestamp is subtracted from the loop, an additional delay is obtained;

Alternatively, the transmitting side receiving timestamp is subtracted from the first ring delay additional loopback transmission time stamp obtained from the receiving end and the receiving ring by subtracting the second timestamp from the additional ring delay averaged transmission time stamp obtained , additional delay.

The delay jitter characteristic further comprises the step of generating a transmitting end in two or more measurement message over a period of time;

C further comprises the step of delay variation calculation method of the delay characteristics, as follows: all delay measurement message generated by the period of time from the selected maximum delay and minimum delay, maximum delay and Save minimum delay to obtain the sending and receiving end delay jitter.

Seen from the above technical solution, this delay characteristics measurement data communication network according to the present invention, the simulation software through the upper measurement message, and the measurement packets microcode or hardware network devices (microcode / hardware) timestamping, to obtain a time measurement packet transmission network, and ultimately calculate the network delay characteristics measured.

This method generates a measurement message in a real business environment, its delay parameter measurements obtained reflect the traffic flow in case of delay the actual environment, belong to online measurement. Further, the method utilizes a network device microcode / hardware recording time, without the help of auxiliary equipment additional measuring instruments and special features, it is a simple and time recorded in accuracy is much higher than the upper layer software.

Further, the method may loopback calculated by the end of this additional delay, delay characteristics obtained by the measurement of the offset correction is performed, resulting in higher measurement accuracy can be obtained 10jus delay of the order of packets in the Ethernet 100M accuracy.

BRIEF DESCRIPTION

Figure 1 a preferred embodiment implemented delay characteristic measuring process of the present invention; FIG. 2 of the present invention is a process for the error compensation according to the preferred embodiment.

Detailed ways

For purposes of this invention, the technical solution and merits thereof more apparent, with reference to the following Examples affixed both embodiments, the present invention is further described in detail.

Is the embodiment, the measurement process in the delay characteristics of the packet network a certain preferred embodiment of the present invention shown in FIG. Assuming that the originating device is a network segment A, B is a terminating device, the originating device client device A and B are receiving any network devices across the network, such as routers and the like. The router may be implemented using a network processor, the network processors are programmable data communication apparatus a data frame processing means, typically dozens to several hundreds of processing units operate in parallel, to complete the look-up table, the frame processing, packets and forwarding the drop operation, and has hardware and microcode functions. The method includes the following steps:

Step 101, the upper layer of software to simulate actual originating device A generates at least one traffic flow measurement message, a first transmitting timestamp tl individually labeled on the generated measurement packet by microcode / hardware, and mark the measured time stamp packet to the terminating apparatus B.

In this step, the upper layer software is able to generate any source address, destination address, packet length, priority, protocol number and port number of the measurement message, this simulation method commonly used in the prior art, are not described herein again. Further, when the upper layer software generates a measurement packet, labeled packet to the measurement, the measurement on a specific packet ID.

The microcode (Picocode / Microcode) binary code is a proprietary network processor. Microcode in a timestamp defined action that can be sent to the microcode in the upper layer software on measurement message timestamp. In the actual system, the hardware functionality can also be used to complete the digital logic circuit.

Step 102, measuring the message is sent to the terminating device B via a data communications network, the terminating device B separated by its own microcode / hardware from a number of packets the above-mentioned measurement message, and the separated measurement message on The first mark by one receiving timestamp t2. The data communications network includes Internet,

National backbone network, provincial backbone network, metropolitan area networks, access networks, enterprise networks, and various private networks.

In this step, microcode / hardware isolated measurement message according to the measurement packet ID. The separation process using the conventional flow classification matching method, such as using a ternary content addressable memory

(TCAM) realization, not be repeated here.

Step 103, the terminating device B one by one marker packet time stamp t3, the second transmission of the measurement message received by its own microcode / hardware, then measurement message sent to the originating device through the port A.

Terminating in a port device B, a measurement message from time-stamped to leave a last bit of the last byte out of the port, the processing time is 20 ^ f Ao seconds, the processing time and said substantially constant, so the method the delay measurement accuracy can reach O.lms, meet the accuracy requirements Cheng planted voice networks.

Step 104, after the measurement message via a data communication network back to the originating transmitting apparatus A, microcode / hardware originating device A is separated by a traffic classifier measurement message, for receiving a second timestamp T4 individually marked, then measurement message Wen sent to the originating device a top software.

Step 105, the originating device A stamp upper software Remove all measurement message recorded, the calculated network delay characteristics between the transceiver terminal device.

In the above process, the upper layer software can calculate the delay characteristics of the time difference, delay and delay jitter timestamp based on all records.

1, usually, the time is inconsistent with the network device. Time difference refers to the time difference between the same time, the originating device A and the terminating device B marked, such as time certain time the originating device A is IL: 00 times the terminating device B is 9:00, the time difference between the two was 2 hours.

The method of calculating the time difference as follows: the upper layer software individually determines whether the go path measurement message recorded in the same, if the back and forth same path, (1) calculate the time difference z each measurement message recording according to the formula, then all the measurement message the results were averaged to obtain the time difference between transmission and reception equipment.

<Formula> formula see original document page 9 </ formula> (1)

In IP networks, ICMP packet can follow the path, to obtain a packet round-trip path; In the MPLS network, the MPLS LSP established using MPLS TE signaling path of the recording

Way to get messages back and forth path.

When the packet round-trip path certain measures are inconsistent, according to the formula (1) all the calculated reference value of time difference measurement message and then averaged, to obtain the time delay between transmission and reception characteristics of the apparatus also has.

2, according to equation (2) calculated for each delay measurement packet x, then packet delay all measurements were averaged to obtain transceiver delay.

<Formula> formula see original document page 10 </ formula> 3, delay jitter calculation. Jitter (Jitter) refers to the difference between the delay of a period of time, the maximum packet delay and the minimum delay of packets. Since the delay jitter from packet delay, i.e. the time difference from both ends of the flow of packets and influence sorting process, so that the measurement accuracy of the delay measurement accuracy higher than 1-2 orders of magnitude.

In the flow shown in FIG. 1, the microcode / hardware token in the message time stamps provided by the measurement hardware clock. The hardware clock is generated using the local oscillator signal sources may be selected to achieve 64-bit counter, accuracy of 10ppm. Delay Characteristics in the measurement process, when the apparatus A is inconsistent with the originating terminating device B clock frequency can be adjusted between the announcement factor A, B by the time the device prior to the data communication network protocol, both the clock frequency be consistent.

In the high-speed Ethernet, a packet enters ingress port equipment from a first bit of the first byte, flow classification performing matching, the time-stamped, the whole process is completed by the microcode / hardware processing time of 20 microseconds the following. Since the actual time stamping service packet is not, so the above additional processing time is introduced, the measurement error increases.

In the measurement accuracy is not required for supporting planting voice network, the measuring method of the above embodiment has been able to meet the accuracy requirements. Further, to improve the measurement accuracy of the delay characteristics, to meet the needs of other network applications may also be performed in the local loop of the present measurement, thereby calculating the microcode / hardware timestamp and incorporated in the process of Classifier Tz is the additional time delay, a correction to compensate the final result of the measurement, the specific procedure shown in Figure 2:

Step 201, the upper layer software to generate the originating device A loopback destination address is its own address measurement message, and transmit timestamp t5 loopback flag in said loopback measurement message by microcode / hardware, and

Issuing the loopback measurement message.

Step 202, since the ring measurement message via a data communication network to return the originating device A, microcode / hardware loop originating from the measuring device A packet flow classification by matching the separation, its marked loopback receive timestamp T6, and then from loop measurement message to the originating device a of upper layer software.

Step 203, the upper layer software transmission time stamp and the extra time delay t5 receive timestamp T6 Tz loopback, loopback measurement calculated loopback packet according to the record. Wherein, Tz = t6-t5.

Step 204, the upper layer software to compensate for the correction of the delay characteristics obtained by calculation according to the additional time delay Tz.

For example, the delay for x, after the compensation correction delay Tx-x-Tz. In the above process, may be calculated by the receiving apparatus B via the local loopback Tz is an additional time delay; or calculated from the originating device TZL A first additional delay, the terminating device B calculates a second additional delay Tz2 and then taking the average of the first and second additional delay Tzl Tz2 additional delay, the additional delay Tz.

Experimental testing showed that after the delay compensation correction characteristic obtained reaches 10jas level precision. Further, the upper layer software may also send multiple packets measure and record the time required, then the system

The method to obtain the count of the corresponding network more accurate delay characteristics.

Seen from the above embodiments, this delay characteristics measured side data communication network according to the present invention

Method, the upper layer software generated by the microcode / hardware of the network device measurement time stamped packets,

To calculate the delay characteristics of the network. The method is simple and does not affect the actual operations, and

Good measurement accuracy can be obtained the desired network.

Claims (7)

1, a delay characteristic measurement data communication network, characterized in that, the method comprising the steps of: a, transmission end transmits a packet to the receiving side, generating at least one measurement message, through its own microcode / hardware token sending time stamp is the first measurement message generated by one, and the issue of the measurement packet; receiving end separated from the received data packets in the measurement message, and use their own microcode / hardware isolated measuring a first mark by one packet receiving time stamp; B, the receiving end with its own microcode / hardware isolated individually labeled second measurement message sending time stamp, a data communications network and the packet to the above measurement sending end; isolated transmission side from the received data packet out the measurement message, and use their own microcode / hardware isolated individually labeled second measurement message received timestamp; C1, sender and / or receiver generating a destination address is its own end of the loopback measurement message, by microcode / hardware of the loopback flag loopback measurement message sending time stamp, and the measurement message sent from the ring; C2, transmission Isolated and / or a receiver from the received data packets out of the ring from the measurement message, and the microcode / hardware isolated loopback flag loopback measurement message received timestamp; C3, the transmitter side and / loopback or receiving terminal for receiving the time stamp and the time stamp calculated from the additional ring transmit delay; C, measured in accordance with a first transmission time stamp packets labeled, receiving a first time stamp, and the second receiving a second transmitting timestamp timestamp calculated delay characteristics between the transmitting end and the receiving end delay; delay by subtracting the additional delay, delay after the offset correction is obtained.

2. The method according to claim 1, wherein, in step a measurement message generated with measurement message identifier; separation process is measured as packets: packet identification based on the measurement, using flow classification matching method isolated measurement message.

3. The method according to claim 2, wherein said flow classification matching method using ternary content addressable memory implementation.

4. The method of claim 1 speed claims, characterized in that the child, including the speed of the delay time difference transplanted, step c further comprises the method of calculating the time difference in the delay characteristics, as follows: each of the first measurement message a second receiving time stamp plus the transmission time stamp and subtracting the first timestamp plus a second transmitting and receiving time stamp, the obtained difference divided by two; then the results of all measurement packets were averaged to give time difference between the transmitting end and the receiving end.

5. The method of claim 1, wherein said step c calculated delay time delay characteristic between the transmitting end and the receiving end is: each first measurement message receiving a time stamp and time stamp plus a second receiver, subtracting the first timestamp plus the transmission time stamp and the second transmission, the obtained difference divided by two, to obtain each delay measurement packet; then all the measurement message text delay are averaged to obtain the time delay of the transmission and reception ends.

6. The method as claimed in claim 5, characterized in that said step c3 method of calculating the additional delay is: on the transmitting side is subtracted from the time stamp received from the ring loop transmission timestamp, an additional delay is obtained; or the receiving end receives the time stamp from the ring transmission timestamp is subtracted from the loop, an additional delay is obtained; Alternatively, the transmitting side receiving timestamp is subtracted from the first ring delay additional loopback transmission time stamp obtained from the receiving end, and subtracting the second timestamp receiving additional ring delay loop transmission time stamp obtained from averaging, additional delay.

7. The method as claimed in claim 5, wherein said delay characteristic further comprises a delay jitter, step a transmitting side generates two or more at a time measurement message; the step c further the method comprises calculating delay variation in delay characteristics, as follows: all delay measurement message generated by the period of time from the selected maximum delay and minimum delay, the maximum delay and the minimum delay to salty , to obtain the sending end and receiving end delay jitter.