(19)
(11)EP 3 531 621 B1

(12)EUROPEAN PATENT SPECIFICATION

(45)Mention of the grant of the patent:
09.09.2020 Bulletin 2020/37

(21)Application number: 17879733.8

(22)Date of filing:  21.07.2017
(51)International Patent Classification (IPC): 
H04L 12/24(2006.01)
H04L 12/26(2006.01)
(86)International application number:
PCT/CN2017/093885
(87)International publication number:
WO 2018/107756 (21.06.2018 Gazette  2018/25)

(54)

METHOD AND APPARATUSES FOR DELAY MEASUREMENT ON A NETWORK NODE

VERFAHREN UND VORRICHTUNGEN ZUR VERZÖGERUNGSMESSUNG AN EINEN NETZWERKKNOTEN

PROCÉDÉ ET DISPOSITIF DE MESURE DE RETARD SUR UN APPAREIL DE NOEUD DE RÉSEAU


(84)Designated Contracting States:
AL AT BE BG CH CY CZ DE DK EE ES FI FR GB GR HR HU IE IS IT LI LT LU LV MC MK MT NL NO PL PT RO RS SE SI SK SM TR

(30)Priority: 15.12.2016 CN 201611159723

(43)Date of publication of application:
28.08.2019 Bulletin 2019/35

(73)Proprietor: Huawei Technologies Co., Ltd.
Longgang District Shenzhen, Guangdong 518129 (CN)

(72)Inventors:
  • ZHANG, Xiquan
    Shenzhen Guangdong 518129 (CN)
  • SONG, Jianmin
    Shenzhen Guangdong 518129 (CN)
  • WEI, Jiahong
    Shenzhen Guangdong 518129 (CN)

(74)Representative: Isarpatent 
Patent- und Rechtsanwälte Behnisch Barth Charles Hassa Peckmann & Partner mbB Friedrichstrasse 31
80801 München
80801 München (DE)


(56)References cited: : 
WO-A1-2004/017221
WO-A1-2016/075924
CN-A- 101 848 118
US-A1- 2009 089 092
WO-A1-2004/017221
CN-A- 101 834 901
CN-A- 105 376 112
  
      
    Note: Within nine months from the publication of the mention of the grant of the European patent, any person may give notice to the European Patent Office of opposition to the European patent granted. Notice of opposition shall be filed in a written reasoned statement. It shall not be deemed to have been filed until the opposition fee has been paid. (Art. 99(1) European Patent Convention).


    Description

    TECHNICAL FIELD



    [0001] The present invention relates to the field of communications technologies, and more specifically, to a delay measurement method of a network node device, an apparatus, and the network node device.

    BACKGROUND



    [0002] It is learned from current research and development of 5G (5th Generation) communications technologies that a 5G network architecture is a sliced network, and there is a low-latency service slice. Therefore, network operation services of an end-to-end delay and a point-to-point delay impose a higher requirement for reliability of entire-network delay performance.

    [0003] With application and development of a delay-sensitive service such as virtual reality or automated driving, an end-to-end delay measurement technology such as TWAMP (Two-Way Active Measurement Protocol) or IP FPM (IP Flow Performance Measurement) in the prior art is usually used in an entire-network path planning and deployment process of the delay-sensitive service. In a process of performing delay performance measurement, the TWAMP is used to perform sampling measurement on a path based on a constructed measurement packet, and a measurement result of the TWAMP reflects a delay result of an end-to-end full path between a transmit end and a receive end of the path. The IP FPM is used to perform sampling recording on an actual packet and measure an actual forwarding time of the packet in a network, and an obtained measurement result also reflects a delay result of an end-to-end full path between a transmit end and a receive end of the path.

    [0004] The two end-to-end delay measurement technologies in the prior art both reflect the delay result of the end-to-end full path from the transmit end to the receive end. If a communication link between the transmit end and the receive end includes at least one network node, a delay result of an end-to-end full path obtained by using the end-to-end delay measurement technology in the prior art cannot reflect a node delay generated when the network node processes a packet. Consequently, delay reliability of the network node on the communication link cannot be quantitatively evaluated and monitored.

    [0005] In the published patent prior art, WO 2016/075 924 A1 discloses a delay measurement method comprising the steps of obtaining, by a delay measurement apparatus, a node delay value of a current packet when a preset time period starts, wherein the node delav value is used to indicate a delay generated when the current packet is forwarded from an inqress port of the network node device to an egress port of a network node device.

    SUMMARY



    [0006] In view of this, the object of the present invention is to provide a delay measurement method of a network node device, an apparatus, and a network node device, to measure a node delay generated when a network node on a communication link between a transmit end and a receive end processes a packet, and quantitatively evaluate and monitor delay reliability of the network node on the communication link. This is object is solved by the respective independent claims and further advantageous embodiments and improvements are listed in the dependent claims. Aspect which contribute to the understanding of the invention are listed below.

    [0007] A first aspect of the present invention discloses a delay measurement method of a network node device, where the network node device includes a delay measurement apparatus, and the delay measurement method includes:

    obtaining, by the delay measurement apparatus, a node delay value of a current packet when a preset time period starts, where the node delay value is used to indicate a delay generated when the current packet is forwarded from an ingress port of the network node device to an egress port of the network node device;

    determining, by the delay measurement apparatus based on M+1 boundary values of preset M delay intervals, a delay interval in which the node delay value is located, and adding 1 to a value of a delay segment counter corresponding to the delay interval in which the node delay value is located, where M>1, M is a positive integer, the delay interval in which the node delay value is located belongs to the M delay intervals, and each delay interval corresponds to one delay segment counter;

    returning, by the delay measurement apparatus, and obtaining a node delay value of a next packet until the preset time period ends; and

    calculating, by the delay measurement apparatus, a cumulative probability Pacc.j, determining a cumulative probability Pacc.j that is greater than or equal to a target probability, and obtaining, as a delay threshold, a node delay value that meets the target probability, where the cumulative probability Pacc.j =Cacc.j /Ctotal, Cacc.j is a sum obtained by accumulating from a value of a first delay segment counter to a value of a jth delay segment counter, 1<j≤M, j is a positive integer, and Ctotal is a sum Ctotal of values of M delay segment counters in the preset time period.



    [0008] In the present invention, the delay measurement apparatus performs corresponding delay measurement on the node delay value generated when the packet is transmitted from the ingress port of the network node device to the egress port of the network node device. Serviceability, availability, and reliability of delay performance of the network node device can be evaluated based on obtained delay measurement data. Therefore, precise quantitative evaluation and monitoring of delay reliability of the network node device can be completed based on comprehensive delay measurement data.

    [0009] In a first implementation disclosed in the first aspect of the present invention, the determining a cumulative probability Pacc.j that is greater than or equal to a target probability, and obtaining, as a delay threshold, a node delay value that meets the target probability includes:

    when obtaining a cumulative probability Pacc.j that is equal to the target probability, obtaining, by the delay measurement apparatus as the delay threshold, the node delay value that meets the target probability; or

    when obtaining a cumulative probability Pacc.j that is greater than the target probability, obtaining, by the delay measurement apparatus, the jth delay segment counter when calculating the cumulative probability Pacc.j, and determining boundary values Dj-1 and Dj of a delay interval corresponding to the jth delay segment counter; and

    searching, by the delay measurement apparatus between the boundary value Dj-1 and the boundary value Dj, for a node delay value that meets the target probability, and using, as the delay threshold, the node delay value that meets the target probability.



    [0010] In the present invention, the delay measurement apparatus compares the target probability with a dynamically calculated cumulative probability, and obtains the node delay value corresponding to the target probability as the delay threshold when obtaining the cumulative probability that is equal to the target probability. When obtaining the cumulative probability that is greater than the target probability, the delay measurement apparatus needs to search, within the determined delay interval, for a node delay value corresponding to the target probability and use the node delay value corresponding to the target probability as the delay threshold. A probability distribution of a node delay value obtained by a currently measured network node device is determined by using the delay threshold, and the delay threshold may be further used to evaluate delay performance of the network node device.

    [0011] In a second implementation disclosed in the first aspect of the present invention, the method further includes:

    determining, by the delay measurement apparatus, a maximum node delay value and a minimum node delay value in node delay values obtained in the preset time period;

    determining, by the delay measurement apparatus, a delay interval between the maximum node delay value and the minimum node delay value, and obtaining a quantity of valid delay segment counters corresponding to the delay interval, where the valid delay segment counters are delay segment counters that perform counting in the preset time period; and

    obtaining, by the delay measurement apparatus, a probability density curve f(x) through fitting calculation based on a distribution function and values of a determined quantity of valid delay segment counters, where x is a random variable and varies with the values of the determined quantity of valid delay segment counters.



    [0012] In the present invention, the delay measurement apparatus determines the maximum node delay value and the minimum node delay value in the preset time period, determines the valid delay segment counters based on the maximum node delay value and the minimum node delay value, uses the values in the valid delay segment counters to obtain the probability density curve through fitting calculation, and uses the probability density curve to make subsequent delay probability measurement more precise.

    [0013] In a third implementation disclosed in the first aspect of the present invention based on the second implementation, the determining a cumulative probability Pacc.j that is greater than or equal to a target probability, and obtaining, as a delay threshold, a node delay value that meets the target probability includes:

    when obtaining a cumulative probability Pacc.j that is equal to the target probability, obtaining, by the delay measurement apparatus as the delay threshold, the corresponding node delay value that meets the target probability; or

    when obtaining the cumulative probability Pacc.j that is greater than the target probability, determining, by the delay measurement apparatus on the probability density curve f(x), a node delay value that meets the target probability, and using, as the delay threshold, the node delay value that meets the target probability.



    [0014] In the present invention, the delay measurement apparatus may obtain, according to the probability density curve, a more precise node delay value that meets the target probability.

    [0015] In a fourth implementation disclosed in the first aspect of the present invention based on the second implementation, the method further includes:

    calculating, by the delay measurement apparatus, an average delay value of node delay values that are of all packets and that are obtained in the preset time period; and

    determining, by the delay measurement apparatus, a node delay value that is corresponding to the average delay value and that is on the probability density curve f(x), where the node delay value corresponding to the average delay value is used to measure node delay performance of the network node device.



    [0016] In a fifth implementation disclosed in the first aspect of the present invention, the method further includes: classifying, by the delay measurement apparatus, the node delay value of the current packet based on a forwarding priority of the current packet, an ID number of a service flow in which the current packet is located, or a port number of a port through which the current packet is input to the network node device.

    [0017] In the present invention, the delay measurement apparatus may perform category-based measurement on the node delay value of the network node according to different application requirements. With category-based measurement on the node delay value, a node delay value that meets a corresponding application requirement may be collected more pertinently, so that in a process of precisely evaluating and monitoring serviceability, availability, and reliability of a network node delay, pertinence can be improved, and power consumption can be reduced.

    [0018] In a sixth implementation disclosed in the first aspect of the present invention, the method further includes: presetting a delay interval;

    obtaining, by the delay measurement apparatus, a basic delay value range, a maximum node delay value, and a theoretically minimum node delay value of the network node device, where the basic delay value range is determined based on a physical structure of the network node device, the maximum node delay value is a maximum node delay value allowed by the network node device in a congestion case, and the theoretically minimum node delay value is a minimum node delay value of the network node device in a theoretical state;

    determining, by the delay measurement apparatus, a to-be-divided node delay value range based on the theoretically minimum node delay value and the maximum node delay value; and

    dividing, by the delay measurement apparatus, the to-be-divided node delay value range based on a long-tailed distribution and the basic delay value range, to obtain M delay intervals and M+1 delay boundary values, and configuring one delay segment counter for each delay interval, where

    the long-tailed distribution indicates that a divided length of a delay interval that is closer to and within the basic delay value range is shorter, and a divided length of a delay interval that is farther away from the basic delay value range is longer.



    [0019] In a seventh implementation disclosed in the first aspect of the present invention, the method further includes:
    obtaining a link delay value with another network node in the preset time period for statistics collection.

    [0020] In the present invention, the delay measurement apparatus can complete better quantitative evaluation and monitoring of the serviceability, the availability, and the reliability of the network node delay after obtaining and combining the link delay value between network node devices and a node delay value of each network node.

    [0021] A delay measurement apparatus disclosed in a second aspect of the present invention includes:

    a delay distribution measurement module, configured to: obtain a node delay value of a current packet when a preset time period starts, where the node delay value is used to indicate a delay generated when the current packet is forwarded from an ingress port of a network node device to an egress port of the network node device; determine, based on M+1 boundary values of preset M delay intervals, a delay interval in which the node delay value is located, and add 1 to a value of a delay segment counter corresponding to the delay interval in which the node delay value is located, where M>1, M is a positive integer, the delay interval in which the node delay value is located belongs to the M delay intervals, and each delay interval corresponds to one delay segment counter; return and obtain a node delay value of a next packet until the preset time period ends; and send a value of each delay segment counter to a delay probability measurement module; and

    the delay probability measurement module, configured to: calculate a cumulative probability Pacc.j, determine a cumulative probability Pacc.j that is greater than or equal to a target probability, and obtain, as a delay threshold, a node delay value that meets the target probability, where the cumulative probability Pacc.j =Cacc.j /Ctotal, Cacc.j is a sum obtained by accumulating from a value of a first delay segment counter to a value of a jth delay segment counter, 1<j≤M, j is a positive integer, and Ctotal is a sum Ctotal of values of M delay segment counters in the preset time period.



    [0022] In a first implementation disclosed in the second aspect of the present invention, the delay probability measurement module includes:

    a first obtaining unit, configured to: when the delay probability measurement module obtains a cumulative probability Pacc.j that is equal to the target probability, obtain, as the delay threshold, the node delay value that meets the target probability; and

    a second obtaining unit, configured to: when the delay probability measurement module obtains a cumulative probability Pacc.j that is greater than the target probability, obtain the jth delay segment counter when calculating the cumulative probability Pacc.j, and determine boundary values Dj-1 and Dj of a delay interval corresponding to the jth delay segment counter; and search, between the boundary value Dj-1 and the boundary value Dj, for a node delay value that meets the target probability, and use, as the delay threshold, the node delay value that meets the target probability.



    [0023] In a second implementation disclosed in the second aspect of the present invention, the delay measurement apparatus further includes a delay feature measurement module; and
    the delay feature measurement module is configured to: determine a maximum node delay value and a minimum node delay value in node delay values obtained in the preset time period; determine a delay interval between the maximum node delay value and the minimum node delay value, and obtain a quantity of valid delay segment counters corresponding to the delay interval, where the valid delay segment counters are delay segment counters that perform counting in the preset time period; and obtain a probability density curve f(x) through fitting calculation based on a distribution function and values of a determined quantity of valid delay segment counters, and send the probability density curve f(x) to the delay probability measurement module, where x is a random variable and varies with the values of the determined quantity of valid delay segment counters.

    [0024] In a third implementation disclosed in the second aspect of the present invention, the delay probability measurement module includes:

    a first obtaining unit, configured to: when the delay probability measurement module obtains a cumulative probability Pacc.j that is equal to the target probability, obtain, as the delay threshold, the node delay value that meets the target probability; and

    a third obtaining unit, configured to: when the delay probability measurement module obtains a cumulative probability Pacc.j that is greater than the target probability, determine, on the probability density curve f(x), a node delay value that meets the target probability, and use, as the delay threshold, the node delay value that meets the target probability.



    [0025] In a fourth implementation disclosed in the second aspect of the present invention, the delay measurement apparatus further includes:
    a classification module, configured to classify the calculated node delay value of the current packet based on a forwarding priority of the current packet, an ID number of a service flow in which the current packet is located, or a port number of a port through which the current packet is input to the network node device.

    [0026] In a fifth implementation disclosed in the second aspect of the present invention, the delay measurement apparatus further includes:
    a preset module, configured to: obtain a basic delay value range of the network node device, a maximum node delay value of the network node device in a congestion case, and a theoretically minimum node delay value; determine a to-be-divided node delay value range based on the theoretically minimum node delay value and the maximum node delay value; divide the to-be-divided node delay value range based on a long-tailed distribution and the basic delay value range, to obtain M delay intervals and M+1 delay boundary values; and configure one delay segment counter for each delay interval.

    [0027] A delay measurement apparatus disclosed in a third aspect of the present invention includes a memory and a processor that communicates with the memory, where

    the memory is configured to store program code for delay measurement; and

    the processor is configured to invoke the program code for delay measurement in the memory to perform the delay measurement method disclosed in the first aspect of the present invention.



    [0028] A network node device disclosed in a fourth aspect of the present invention includes the delay measurement apparatus disclosed in the second aspect or the third aspect of the present invention, where the delay measurement apparatus is disposed inside an egress port or outside the egress port of the network node device.

    BRIEF DESCRIPTION OF DRAWINGS



    [0029] To describe the technical solutions in the embodiments of the present invention more clearly, the following briefly describes the accompanying drawings required for describing the embodiments. Apparently, the accompanying drawings in the following description show merely the embodiments of the present invention, and a person of ordinary skill in the art may still derive other drawings from these accompanying drawings without creative efforts.

    FIG. 1 is a schematic diagram of an application scenario of a delay measurement method according to an embodiment of the present invention;

    FIG. 2A and FIG. 2B are a schematic flowchart of a delay measurement method according to Embodiment 1 of the present invention;

    FIG. 3 is a schematic diagram of delay interval division according to Embodiment 1 of the present invention;

    FIG. 4A, FIG. 4B, and FIG. 4C are a schematic flowchart of a delay measurement method according to Embodiment 2 of the present invention;

    FIG. 5 is a schematic diagram of a probability density curve f(x) according to Embodiment 2 of the present invention;

    FIG. 6 is a schematic structural diagram of a delay measurement apparatus according to the present invention;

    FIG. 7 is a schematic structural diagram of another delay measurement apparatus according to the present invention;

    FIG. 8 is a schematic diagram of an application scenario of another delay measurement apparatus according to the present invention; and

    FIG. 9 is a schematic structural diagram of a delay measurement apparatus according to the present invention.


    DESCRIPTION OF EMBODIMENTS



    [0030] Full names and explanations of English abbreviations used in specific embodiments of the present invention are as follows:

    5G: 5th Generation;

    TWAMP: Two-Way Active Measurement Protocol;

    IP FPM: IP Flow Performance Measurement;

    PHY: Physical Layer;

    MAC: Media Access Control; and

    MIB: Management Information Base.



    [0031] It can be learned from the background that, for a 5G communications technology, an end-to-end delay measurement technology such as the TWAMP or IP FPM is mainly used in an existing delay measurement manner for a delay-sensitive service. For example, when the TWAMP is used for delay measurement, delay measurement is performed on a path between a transmit end and a receive end by using a constructed measurement packet, to obtain a delay result that reflects an end-to-end full path. For at least one network node device included on a communication link between the transmit end and the receive end, a node delay of the network node device on the communication link cannot be reflected by using the delay result. Therefore, a TWAMP or IP FPM measurement technology cannot quantitatively evaluate and monitor reliability of a network node delay.

    [0032] The present invention provides a delay measurement technical solution of a network node device. A delay measurement apparatus separately performs corresponding delay measurement on a node delay value generated when a packet is transmitted from an ingress port of the network node device to an egress port of the network node device. Serviceability, availability, and reliability of delay performance of the network node device can be evaluated based on obtained delay measurement data. Therefore, precise quantitative evaluation and monitoring of delay reliability of the network node device can be completed based on comprehensive delay measurement data.

    [0033] The delay measurement apparatus mentioned in the present invention may be a physical device or a function module disposed on a physical device. The delay measurement apparatus may be disposed inside the egress port of the network node device, or may be disposed outside the egress port of network node device. The network node device may be a server, a routing device, or a switching device.

    Embodiment 1



    [0034] FIG. 1 is a schematic diagram of an application scenario of a delay measurement technical solution of a network node device according to Embodiment 1 of the present invention. Two network node devices located between a transmit end and a receive end on a path are used as an example. A network node device R1 and a network node device R2 are included in the application scenario.

    [0035] An ingress port of the network node device R1 is P1, an egress port of the network node device R1 is P2, and a delay measurement apparatus A is disposed at the egress port P2 of the network node device R1. An ingress port of the network node device R2 is P3, an egress port of the network node device R2 is P4, and a delay measurement apparatus A is disposed at the egress port of the network node device R2. The egress port P2 of the network node device R1 is connected to the ingress port P3 of the network node device R2 by using a communication link.

    [0036] The communication link that connects the network node device R1 and the network node device R2 and that is disclosed in this embodiment of the present invention may be a 10GE communication link, or a 100GE communication link, but is not merely limited thereto.

    [0037] For packet transmission between the network node device R1 and the network node device R2, when a packet is input from a PHY of the ingress port P1 of the network node device R1, in this case, a timestamp T1 is added to the packet. After being processed by the network node device R1, the packet is output from a PHY of the egress port P2. In this case, a timestamp T2 is added to the packet. A timestamp is a moment at which a packet is correspondingly processed. In the network node device R1, a node delay value DUi of the packet from the ingress port P1 to the egress port P2 is shown in formula (1):



    [0038] The network node device R1 outputs the packet from the PHY of the egress port P2, and inputs the packet from a PHY of the ingress port P3 of the network node device R2 through the communication link. In this case, the network node device R2 adds a timestamp T3 to the packet. After being processed by the network node device R2, the packet is output from a PHY of the egress port P4. In this case, the network node device R2 adds a timestamp T4 to the packet. In the network node device R2, a node delay value DUi' of the packet from the ingress port P3 to the egress port P4 is shown in formula (2):



    [0039] A link delay generated by the communication link between the network node device R1 and the network node device R2 is DL.

    [0040] For a measurement process performed for a node delay value of the network node device, a delay measurement apparatus disposed at the egress port P2 of the network node device R1 is used as an example. A specific implementation process of implementing the delay measurement technical solution of the present invention by the delay measurement apparatus is described in detail in the following embodiment.

    [0041] FIG. 2A and FIG. 2B are a schematic flowchart of a delay measurement method of a network node device according to an embodiment of the present invention. The delay measurement apparatus first divides a to-be-divided node delay value range, including the following steps.

    [0042] S101. A delay measurement apparatus obtains a basic delay value range t of a network node device R1, a maximum node delay value Dmax of the network node device R1 in a congestion case, and a theoretically minimum node delay value Dmin.

    [0043] The basic delay value range mentioned in S101 is set based on a physical structure of the network node device, and belongs to an inherent delay value range of the network node device. In this embodiment of the present invention, an example in which the basic delay value range t of the network node device R1 is from 30 µs (microsecond) to 45 µs is used for description. However, as the network node device is optimized, the inherent delay value range becomes increasingly small.

    [0044] The maximum node delay value Dmax mentioned in S101 is a maximum delay value that can be generated when a packet is transmitted in the network node device R1. In this embodiment of the present invention, an example in which the maximum node delay value Dmax of the network node device R1 is 100 ms (millisecond) is used for description.

    [0045] In S101, in this embodiment of the present invention, an example in which theoretically a delay value is not generated in a process of transmitting a packet in the network node device is used, and the theoretically minimum node delay value Dmin is 0.

    [0046] S102. The delay measurement apparatus determines a to-be-divided node delay value range T based on the theoretically minimum node delay value Dmin and the maximum node delay value Dmax.

    [0047] The theoretically minimum node delay value Dmin is used as a minimum value of the to-be-divided node delay value range, the maximum node delay value Dmax is used as a maximum value of the to-be-divided node delay value range, and it is determined that the to-be-divided node delay value range T is from Dmin to Dmax. The example given in S101 is used. The to-be-divided node delay value range is from 0 ms to 100 ms.

    [0048] The basic delay value range t is within the to-be-divided node delay value range.

    [0049] S103. The delay measurement apparatus divides the to-be-divided node delay value range based on a long-tailed distribution (Long-Tailed Distributions) and the basic delay value range, to obtain M delay intervals and M+1 delay boundary values, and configures one delay segment counter for each delay interval.

    [0050] In S103, M>1, and M is a positive integer.

    [0051] The long-tailed distribution refers to a logarithmic distribution (Logarithmic Distribution), an exponential distribution (Exponential Distribution), a Poisson distribution (Poisson Distribution), or the like.

    [0052] In a process of transmitting a packet by using the network node device, a node delay value generated in the network node device usually falls within the basic delay value range.

    [0053] In this embodiment of the present invention, the to-be-divided node delay value range is divided based on a long-tailed distribution manner, and a length of each obtained delay interval is described as follows: A delay interval with a larger probability of occurrence of the node delay value has a shorter length, and a delay interval with a smaller probability of occurrence of the node delay value has a longer length.

    [0054] In most cases, the node delay value falls within the basic delay value range. Therefore, in a process of delay interval division, for a part in which the to-be-divided node delay value range overlaps the basic delay value range, a length of a delay interval obtained through division is shorter, and a length of a divided delay interval that is far away from the overlapping part is longer. As shown in FIG. 3, it is assumed that 0 ms to 100 ms are divided into 32 grids (segments), and 32 delay intervals and 33 boundary values are obtained. Specific division is as follows:

    [0055] The first delay interval to the tenth delay interval are divided from 0 µs to 100 µs. Each delay interval occupies 10 µs. Obtained boundary values include: D0=0 µs, D1=10 µs, D2=20 µs, ..., and D10=100 µs.

    [0056] The eleventh delay interval to the fifteenth delay interval are divided from 100 µs to 200 µs. Each delay interval occupies 20 µs. Obtained boundary values include: D11=120 µs, D12=140 µs, ..., and D15=200 µs.

    [0057] The sixteenth delay interval to the twentieth delay interval are divided from 200 µs to 400 µs. Each delay interval occupies 40 µs. Obtained boundary values include: D16=240 µs, D17=280 µs, ..., and D20=400 µs.

    [0058] The twenty-first delay interval to the twenty-fifth delay interval are divided from 400 µs to 800 µs. Each delay interval occupies 80 µs. Obtained boundary values include: D21=480 µs, D22=560 µs, ..., and D25=800 µs.

    [0059] The twenty-sixth delay interval to the thirtieth delay interval are divided from 800 µs to 1.6 ms. Each delay interval occupies 160 µs. Obtained boundary values include: D26=960 µs, D27=1120 µs=1.12 ms, ..., and D30=1.6 ms.

    [0060] The thirty-first delay interval is divided from 1.6 ms to 3.2 ms, and occupies 1.6 ms. An obtained boundary value is D31=3.2 ms.

    [0061] The thirty-second delay interval is divided from 3.2 ms to 100 ms, and occupies 96.8 ms. An obtained boundary value is D32=100 ms.

    [0062] One delay segment counter is configured for each delay interval. For example, 32 delay segment counters are configured in total for the foregoing 32 delay intervals obtained through division, and are numbered by using corresponding delay intervals. For example, the first delay interval corresponds to the first delay segment counter, and the second delay interval corresponds to the second delay segment counter. By analogy, the thirty-second delay interval corresponds to the thirty-second delay segment counter.

    [0063] In the present invention, the delay measurement apparatus divides the node delay range in the network node device into a plurality of delay intervals, and configures one delay segment counter for each delay interval. A boundary value of each delay interval may be flexibly configured according to a requirement in a division process, or may be subsequently reconfigured.

    [0064] As shown in FIG. 2A and FIG. 2B, the delay measurement apparatus performs, based on the M delay intervals and the M+1 boundary values that are obtained in advance, delay distribution measurement and delay probability measurement on a node delay value obtained by the network node device through calculation in a preset time period.

    [0065] That the delay measurement apparatus performs delay distribution measurement on the obtained node delay value starting from the preset time period includes the following:
    S104. Obtain a node delay value of a current packet, determine, based on the M+1 boundary values of the preset M delay intervals, a delay interval in which the node delay value is located, and add 1 to a value of a delay segment counter corresponding to the delay interval.

    [0066] S105. Determine whether the preset time period ends. If the preset time period does not end, continue to perform S104; or if the preset time period ends, perform delay probability measurement.

    [0067] In this embodiment of the present invention, according to the foregoing example, 32 delay intervals and 33 boundary values are obtained in advance. An example in which the delay measurement apparatus obtains in total 100 node delay values: a node delay value DU1 to a node delay value DU100 in the preset time period is used to describe S104 and S105.

    [0068] It is assumed that the node delay value DU1 to a node delay value DU20 that are obtained through calculation are each 29 µs, a node delay value DU21 to a node delay value DU55 that are obtained through calculation are each 32 µs, a node delay value DU56 that is obtained through calculation is 62 µs, a node delay value DU57 that is obtained through calculation is 35 µs, a node delay value DU58 that is obtained through calculation is 65 µs, and a node delay value DU59 to the node delay value DU100 that are obtained through calculation are each 41 µs.

    [0069] When the preset time period starts, the delay measurement apparatus obtains the node delay value DU1 of the current packet, determines, based on the 33 boundary values of the preset 32 delay intervals, that the node delay value DU1 is located in the third delay interval, and adds 1 to a value of the third delay segment counter corresponding to the third delay interval.

    [0070] There are a plurality of manners of determining a delay interval in which the node delay value DU1 is located.

    [0071] For example, the 33 boundary values: the boundary value D0 to the boundary value D32 are traversed, and a difference is obtained between each of the 33 boundary values: the boundary value D0 to the boundary value D32 and the node delay value DU1. If a difference obtained previously is positive, and a difference obtained currently is negative, the node delay value is between the two boundary values. It is assumed that the node delay value DU1=29 µs, and a difference is obtained between the node delay value DU1 and each of the boundary value D0 to the boundary value D32, to obtain DU1-D2>0 and DU1-D3<0. It is determined, based on the obtained differences, that the node delay value DU1 is between the boundary value D2 and the boundary value D3. Therefore, it is determined that the node delay value DU1 is located in the third delay interval.

    [0072] Optionally, a comparison manner may alternatively be used. For example, if the node delay value DU1 is compared with each of the boundary value D0 to the boundary value D32 to obtain that the node delay value DU1 is greater than the boundary value D2 and less than the boundary value D3, it is determined that the node delay value DU1 is between the boundary value D2 and the boundary value D3, and is located in the third delay interval.

    [0073] It should be noted that, if a node delay value obtained through calculation is the same as a boundary value, the delay measurement apparatus may determine, based on two delay intervals in which the boundary value is located, that the node delay value is located in either of the delay intervals.

    [0074] By analogy, in a process in which the delay measurement apparatus obtains remaining 99 node delay values: the node delay value DU2 to the node delay value DU100 in real time, each time the delay measurement apparatus obtains one node delay value, the delay measurement apparatus determines, in a same manner as that of the node delay value DU1, a delay interval in which each node delay value is located.

    [0075] Finally, it is obtained that the node delay value DU1 to the node delay value DU20 are located in the third delay interval, and a value of the third delay segment counter is 20. The node delay value DU21 to the node delay value DU55, and the node delay value DU57 are located in the fourth delay interval, and a value of the fourth delay segment counter is 36. The node delay value DU56 to the node delay value DU58 are located in the sixth delay interval, and a value of the sixth delay segment counter is 2. The node delay value DU59 to the node delay value DU100 are located in the fifth delay interval, and a value of the fifth delay segment counter is 42.

    [0076] A sum of values of all delay segment counters is equal to a quantity of node delay values received in the preset time period.

    [0077] Optionally, after the delay interval in which the node delay value is located is determined, and the value of the corresponding delay segment counter is added by 1, the node delay value may be discarded.

    [0078] As shown in FIG. 2A and FIG. 2B, the delay measurement apparatus performs delay probability measurement based on values that are of delay segment counters and that are obtained by performing delay distribution measurement.

    [0079] S106. The delay measurement apparatus calculates a sum Ctotal of values of M delay segment counters in the preset time period, calculates a cumulative probability Pacc.j according to formula (3), and compares, with a target probability, the cumulative probability Pacc.j obtained based on values that are of delay segment counters and that are accumulated each time. If a cumulative probability Pacc.j that is less than the target probability is obtained, continue to accumulate and calculate the values of the delay segment counters. If a cumulative probability Pacc.j that is equal to the target probability is obtained, perform S107. If a cumulative probability Pacc.j that is greater than the target probability is obtained, perform S108.



    [0080] Cacc.j is a sum obtained by accumulating from a value of the first delay segment counter to a value of a jth delay segment counter, 1 <j≤M, and j is a positive integer. Ctotal is the sum of the values of the M delay segment counters in the preset time period.

    [0081] In a process of calculating the sum of the values of the M delay segment counters, after the values of the M delay segment counters are obtained, one addition operation may be performed to obtain the sum Ctotal of the values of the M delay segment counters in the preset time period.

    [0082] Alternatively, after S104 is performed once, one delay segment counter may be obtained to perform a counting operation, and 1 is accumulated once until the preset time period ends, to obtain the sum Ctotal of the accumulated values of the M delay segment counters.

    [0083] In formula (3), Cacc.j is a variable. In a process of calculating the cumulative probability Pacc.j, Cacc.j is the sum obtained by accumulating from the value of the first delay segment counter to the value of the jth delay segment counter.

    [0084] The target probability is a preset probability, may be set by a technician, and may be 99%, 99.5%, or 99.999%. This is not limited in the present invention.

    [0085] S107. When obtaining a cumulative probability Pacc.j that is equal to the target probability, the delay measurement apparatus obtains, as a delay threshold, a node delay value that meets the target probability.

    [0086] Based on the foregoing example, the 32 delay intervals and the 33 boundary values that are obtained in advance are used as an example. It is assumed that the delay measurement apparatus obtains in total 900 node delay values: a node delay value D1 to a node delay value D900 in the preset time period. In a process of performing delay distribution measurement by the delay measurement apparatus, each delay segment counter performs a corresponding operation of adding 1 based on delay intervals in which the node delay value D1 to the node delay value D900 are located. After the preset time period ends, the delay measurement apparatus obtains that a value of the third delay segment counter is 120, a value of the fourth delay segment counter is 436, a value of the fifth delay segment counter is 340, a value of the sixth delay segment counter is 1, a value of the ninth delay segment counter is 1, and a sum of values of all delay segment counters is 900.

    [0087] It is assumed that the target probability is 99.5%. When the value of the third delay segment counter is accumulated, a cumulative probability Pacc.3=120/900=13.3% is obtained according to formula (3), and the cumulative probability Pacc.3 is less than the target probability. Therefore, the value of the fourth delay segment counter continues to be accumulated.

    [0088] When the value of the fourth delay segment counter is accumulated, a cumulative probability Pacc.4=556/900=61.8% is obtained according to formula (3), and the cumulative probability Pacc.4 is less than the target probability. Therefore, a value of a delay segment counter continues to be accumulated.

    [0089] When the value of the fifth delay segment counter is accumulated, a cumulative probability Pacc.5=896/900=99.5% is obtained according to formula (3), and the cumulative probability Pacc.5 is equal to the target probability. Therefore, the delay measurement apparatus determines a node delay value corresponding to the target probability as the delay threshold.

    [0090] S108. When obtaining a cumulative probability Pacc.j that is greater than the target probability, the delay measurement apparatus obtains a jth delay segment counter when calculating the cumulative probability Pacc.j, and determines boundary values Dj-1 and Dj of a delay interval corresponding to the jth delay segment counter.

    [0091] S109. The delay measurement apparatus searches, between the boundary value Dj-1 and the boundary value Dj, for a node delay value that meets the target probability, and uses, as a delay threshold, the node delay value that meets the target probability.

    [0092] In this embodiment of the present invention, the 32 delay intervals and the 33 boundary values that are obtained in advance are used as an example. It is assumed that the delay measurement apparatus obtains in total 1000 node delay values: a node delay value D1 to a node delay value D1000 in the preset time period. In a process of performing delay distribution measurement by the delay measurement apparatus, each delay segment counter performs a corresponding operation of adding 1 based on delay intervals in which the node delay value D1 to the node delay value D1000 are located. After the preset time period ends, the delay measurement apparatus obtains that a value of the fourth delay segment counter is 652, a value of the fifth delay segment counter is 346, a value of the ninth delay segment counter is 2, and a sum of values of all delay segment counters is 1000.

    [0093] It is assumed that the target probability is 99.5%. When the value of the fourth delay segment counter is accumulated, a cumulative probability Pacc.4=652/1000=65.2% is obtained according to formula (3), and the cumulative probability Pacc.4 is less than the target probability. Therefore, the value of the fifth delay segment counter continues to be accumulated.

    [0094] When the value of the fifth delay segment counter is accumulated, a cumulative probability Pacc.5=998/1000=99.8% is obtained according to formula (3), and the cumulative probability Pacc.5 is greater than the target probability. Therefore, it is obtained that Cacc.j is accumulated to the fifth delay segment counter when the cumulative probability is calculated.

    [0095] The delay measurement apparatus determines the fifth delay interval corresponding to the fifth delay segment counter. The delay measurement apparatus determines a boundary value D4 and a boundary value D5 of the fifth delay interval, searches, between the boundary value D4 and the boundary value D5, for a node delay value that meets the target probability, and uses, as the delay threshold, the node delay value that meets the target probability.

    [0096] In this embodiment of the present invention, optionally, a value that is of each delay segment counter and that is obtained by performing delay distribution measurement by the delay measurement apparatus, and the cumulative probability and the delay threshold that are obtained by performing delay probability measurement are stored in a unified format. For example, when the delay measurement apparatus includes an MIB, the value of each delay segment counter, the cumulative probability, and the delay threshold are stored in the MIB in an MIB format.

    [0097] According to the delay measurement method disclosed in Embodiment 1 of the present invention, delay distribution measurement and delay probability measurement are separately performed on the node delay value received in the preset time period, and serviceability, availability, and reliability of delay performance of the network node device are evaluated based on obtained delay measurement data. Further, delay measurement data that is provided in an entire network when each network node device forwards a packet is obtained. Precise quantitative evaluation and monitoring of serviceability, availability, and reliability of a network node delay can be completed based on obtained comprehensive delay measurement data.

    Embodiment 2



    [0098] Based on the application scenario shown in FIG. 1, Embodiment 2 of the present invention further discloses a delay measurement method of a network node device, and a flowchart of the delay measurement method is shown in FIG. 4A, FIG. 4B, and FIG. 4C. A delay measurement apparatus first divides a to-be-divided node delay value range, including the following steps.

    [0099] S201. A delay measurement apparatus obtains a basic delay value range t of a network node device R1, a maximum node delay value Dmax of the network node device R1 in a congestion case, and a theoretically minimum node delay value Dmin.

    [0100] S202. The delay measurement apparatus determines a to-be-divided node delay value range T based on the theoretically minimum node delay value Dmin and the maximum node delay value Dmax.

    [0101] S203. The delay measurement apparatus divides the to-be-divided node delay value range based on a long-tailed distribution (Long-Tailed Distributions) and the basic delay value range, to obtain M delay intervals and M+1 delay boundary values, and configures one delay segment counter for each delay interval.

    [0102] In Embodiment 2 of the present invention, a principle and a process based on which the delay measurement apparatus performs S201 to S203 are the same as those of S101 to S103 in Embodiment 1 of the present invention. For specific content, refer to S101 to S103 in Embodiment 1 of the present invention. Details are not described herein again.

    [0103] In Embodiment 2 of the present invention, delay measurement performed by the delay measurement apparatus on a node delay value of the network node device includes delay distribution measurement, delay feature measurement, and delay probability measurement.

    [0104] As shown in FIG. 4A, FIG. 4B, and FIG. 4C, that the delay measurement apparatus performs delay distribution measurement on an obtained node delay value starting from a preset time period includes the following:
    S204. Obtain a node delay value of a current packet, determine, based on the M+1 boundary values of the preset M delay intervals, a delay interval in which the node delay value is located, and add 1 to a value of a delay segment counter corresponding to the delay interval.

    [0105] S205. Determine whether the preset time period ends. If the preset time period does not end, continue to perform S204; or if the preset time period ends, perform delay probability measurement.

    [0106] In Embodiment 2 of the present invention, a principle and a process based on which the delay measurement apparatus performs S204 and S205 are the same as those of S104 and S105 in Embodiment 1 of the present invention. For specific content, refer to S104 and S105 in Embodiment 1 of the present invention. Details are not described herein again.

    [0107] A difference between Embodiment 2 of the present invention and Embodiment 1 lies in that delay feature measurement is further added before the delay measurement apparatus performs delay probability measurement. Optionally, the delay measurement apparatus may simultaneously perform delay distribution measurement and delay feature measurement.

    [0108] As shown in FIG. 4A, FIG. 4B, and FIG. 4C, that the delay measurement apparatus performs delay feature measurement on the obtained node delay value starting from the preset time period includes the following:
    S206. The delay measurement apparatus obtains the node delay value of the current packet, and calculates an average delay value according to formula (4):

    where
    i indicates a sum of currently obtained node delay values, a value of i is from 1 to N, and N indicates a maximum quantity of node delay values that are obtained by the delay measurement apparatus in the preset time period. For example, in the preset time period, if the delay measurement apparatus receives 100 node delay values, N is 100, and the value of i is from 1 to 100.

    [0109] DUi indicates a currently obtained ith node delay value. Dav.Ui indicates an average delay value of currently obtained i node delay values. Dav.Ui-1 indicates an average delay value of the obtained first i-1 node delay values.

    [0110] The current average delay value is calculated based on formula (4), and there is no need to calculate an average value after all node delay values are obtained. This is very practical and occupies only a few resources.

    [0111] S207. The delay measurement apparatus determines a maximum node delay value and a minimum node delay value in currently obtained node delay values, and saves the maximum node delay value and the minimum node delay value.

    [0112] Each time one node delay value is obtained, the node delay value is separately compared with a maximum node delay value that is previously determined and stored in a maximum delay register and a minimum node delay value that is stored in a minimum delay register. If the currently obtained node delay value is greater than the maximum node delay value, the original maximum node delay value is replaced with the currently obtained node delay value, and the currently obtained node delay value is stored in the maximum delay register as a new maximum node delay value. If the currently obtained node delay value is less than the minimum node delay value, the original minimum node delay value is replaced with the currently obtained node delay value, and the currently obtained node delay value is stored in the minimum delay register as a new minimum node delay value.

    [0113] S208. The delay measurement apparatus determines whether the preset time period ends; and if the preset time period does not end, returns and continues to perform S206; or if the preset time period ends, determines a final maximum node delay value and a final minimum node delay value, and performs S209.

    [0114] An example in which the delay measurement apparatus obtains in total 100 node delay values: a node delay value DU1 to a node delay value DU100 in the preset time period is used to describe S206 to S208.

    [0115] It is assumed that the node delay value DU1 to a node delay value DU20 that are obtained through calculation are each 29 µs, a node delay value DU21 to a node delay value DU55 that are obtained through calculation are each 32 µs, a node delay value DU56 that is obtained through calculation is 62 µs, a node delay value DU57 that is obtained through calculation is 35 µs, a node delay value DU58 that is obtained through calculation is 65 µs, and a node delay value DU59 to the node delay value DU100 that are obtained through calculation are each 41 µs.

    [0116] When the first node delay value DU1 is obtained, Dav.U1 =DU1/1=29 µs is obtained according to formula (4).

    [0117] In this case, there is only one node delay value DU1. Therefore, the node delay value DU1 is stored in the maximum delay register as a current maximum node delay value, and the node delay value DU1 is also stored in the minimum delay register as a current minimum node delay value.

    [0118] Because the preset time period does not end, the second node delay value DU2 continues to be obtained, and Dav.U2 =Dav.U1 (1/2)+DU2/2=29 µs is obtained according to formula (4).

    [0119] In this case, the node delay value DU2 is compared with the node delay value DU1 stored in the maximum delay register, and it is determined that the two values are the same, and no replacement is performed. The node delay value DU2 is compared with the node delay value DU1 stored in the minimum delay register, and it is determined that the two values are the same, and no replacement is performed.

    [0120] By analogy, when the twenty-first node delay value DU21 is obtained, Dav.U21 =Dav.U20 (20/21)+DU21/21=29.1 µs is obtained according to formula (4).

    [0121] In this case, the node delay value DU21 is compared with the node delay value DU1 stored in the maximum delay register. Because the node delay value DU21 is 32 µs, and the node delay value DU1 is 29 µs, it is determined that the node delay value DU21 is greater than the node delay value DU1. Then, the node delay value DU21 is used to replace the node delay value DU1, and the node delay value DU21 is stored in the maximum delay register. The node delay value DU1 is still stored in the minimum delay register.

    [0122] By analogy, when the last node delay value DU100 in the preset time period is obtained, Dav.U100 =Dav.U99 99/100+DU100/100=44 µs is obtained according to formula (4).

    [0123] In this case, the maximum node delay value stored in the maximum delay register is the node delay value DU58, and the minimum node delay value stored in the minimum delay register is the node delay value DU1.

    [0124] Optionally, after obtaining all node delay values in the preset time period, the delay measurement apparatus may alternatively make a comparison, or obtain a difference between each two of the node delay values, to determine the maximum node delay value and the minimum node delay value in the preset time period.

    [0125] S209. The delay measurement apparatus determines, based on the preset M delay intervals, a delay interval between the maximum node delay value and the minimum node delay value, and obtains a quantity of valid delay segment counters corresponding to the delay interval, where the valid delay segment counters are delay segment counters that perform counting in the preset time period.

    [0126] The M delay intervals are not evenly divided, and a subsequent fitted curve is relatively complicated by using values in delay segment counters corresponding to all delay intervals. Therefore, optionally, a value in a delay segment counter corresponding to the delay interval between the maximum node delay value and the minimum node delay value may be selected to construct the subsequent fitted curve.

    [0127] S210. The delay measurement apparatus obtains a probability density curve f(x) through fitting calculation based on a distribution function and values of a determined quantity of valid delay segment counters.

    [0128] The distribution function herein includes a distribution function such as a normal distribution (Normal) function, a lognormal distribution (Lognormal) function, a Gamma distribution (Gamma) function, a Weibull (Weibull) function, an exponential distribution function, or a Poisson distribution function.

    [0129] The probability density curve f(x) is obtained through fitting calculation based on the values in the determined valid delay segment counters and the distribution function. In the probability density curve f(x) obtained through fitting calculation, x is a random variable, and varies with the values of the determined quantity of valid delay segment counters. A larger value in a valid delay segment counter corresponds to a larger peak value in the probability density curve f(x). The probability density curve f(x) is used for subsequent delay probability measurement.

    [0130] Optionally, the probability density curve f(x) may be obtained through fitting calculation by using a distribution function. Probability density curves may be alternatively obtained separately through fitting calculation by using a plurality of distribution functions. Then, the probability density curves obtained by separately fitting the plurality of distribution functions are evaluated by using a sum of squared errors (SSE, Sum of Squared Errors) and an R-square (R-square) index, so as to determine a probability density curve closest to an actual delay distribution case, and use the probability density curve as the probability density curve f(x) for subsequent delay probability measurement.

    [0131] Optionally, after the delay interval between the maximum node delay value and the minimum node delay value is determined, the obtained delay interval may be re-divided by using a preset minimum granularity, so that delay intervals obtained through re-division are evenly divided. In addition, the delay segment counter corresponding to the original delay interval is also re-divided based on the divided delay intervals.

    [0132] For example, the preset minimum granularity is 10 µs. There are three delay intervals between the maximum node delay value and the minimum node delay value: the third delay interval, the fourth delay interval, and the fifth delay interval respectively. Delay lengths occupied by the delay intervals are respectively as follows: The third delay interval occupies 10 µs, the fourth delay interval occupies 20 µs, and the fifth delay interval occupies 30 µs. Delay segment counters corresponding to the delay intervals are respectively as follows: A value of the third delay segment counter is 15, a value of the fourth delay segment counter is 30, and a value of the fifth delay segment counter is 15.

    [0133] The third delay interval, the fourth delay interval, and the fifth delay interval are re-divided based on the preset minimum granularity. The third delay interval is not changed, the fourth delay interval is divided into two delay intervals, the fifth delay interval is divided into three delay intervals, and a total of six delay intervals are obtained. Based on the foregoing division, a value of the delay segment counter corresponding to the original delay interval is also divided, and values of delay segment counters corresponding to the six delay intervals are successively: 15, 30, 30, 15, 15, and 15. FIG. 5 is a probability density curve f(x) fitted by using this example as an example. In the figure, D3 indicates the third delay interval, D4 indicates the fourth delay interval, and D5 indicates the fifth delay interval.

    [0134] Fitting calculation is performed by using the delay interval determined through even division and a value of a delay segment counter corresponding to each re-divided delay interval, and an obtained probability density curve f(x) is more precise.

    [0135] As shown in FIG. 4A, FIG. 4B, and FIG. 4C, after performing delay distribution measurement to obtain the value of the delay segment counter corresponding to each delay interval, and performing delay feature measurement to obtain the average delay value and the probability density curve f(x), the delay measurement apparatus performs delay probability measurement based on a result obtained through delay distribution measurement and a result obtained through delay feature measurement.

    [0136] S211. The delay measurement apparatus calculates a sum Ctotal of values of M delay segment counters in the preset time period, calculates a cumulative probability Pacc.j according to formula (3), and compares, with a target probability, the cumulative probability Pacc.j obtained based on values that are of delay segment counters and that are accumulated each time. If a cumulative probability Pacc.j that is less than the target probability is obtained, continue to accumulate and calculate the values of the delay segment counters. If a cumulative probability Pacc.j that is equal to the target probability is obtained, perform S212. If a cumulative probability Pacc.j that is greater than the target probability is obtained, perform S213.

    [0137] In Embodiment 2 of the present invention, an execution process and a principle of S211 are the same as those of S106 in Embodiment 1 of the present invention. For specific content, refer to S106. Details are not described herein again.

    [0138] S212. When obtaining a cumulative probability Pacc.j that is equal to the target probability, the delay measurement apparatus obtains, as a delay threshold, a corresponding node delay value that meets the target probability.

    [0139] In Embodiment 2 of the present invention, an execution process and a principle of S212 are the same as those of S107 in Embodiment 1 of the present invention. For specific content, refer to S107. Details are not described herein again.

    [0140] S213. When obtaining a cumulative probability Pacc.j that is greater than the target probability, the delay measurement apparatus determines, on the probability density curve f(x), a node delay value that meets the target probability, and use, as a delay threshold, the node delay value that meets the target probability.

    [0141] In this embodiment of the present invention, 32 delay intervals and 33 boundary values that are obtained in advance are used as an example. It is assumed that the delay measurement apparatus obtains in total 1000 node delay values: a node delay value D1 to a node delay value D1000 in the preset time period. In a process of performing delay distribution measurement by the delay measurement apparatus, each delay segment counter performs a corresponding operation of adding 1 based on delay intervals in which the node delay value D1 to the node delay value D1000 are located. After the preset time period ends, the delay measurement apparatus obtains that a value of the fourth delay segment counter is 652, a value of the fifth delay segment counter is 346, a value of a ninth delay segment counter is 2, and a sum of values of all delay segment counters is 1000.

    [0142] It is assumed that the target probability is 99.5%. When the value of the fourth delay segment counter is accumulated, a cumulative probability Pacc.4=652/1000=65.2% is obtained according to formula (3), and the cumulative probability Pacc.4 is less than the target probability. Therefore, the value of the fifth delay segment counter continues to be accumulated.

    [0143] When the value of the fifth delay segment counter is accumulated, a cumulative probability Pacc.5=998/1000=99.8% is obtained according to formula (3), and the cumulative probability Pacc.5 is greater than the target probability. Therefore, it is obtained that Cacc.j is accumulated to the fifth delay segment counter when the cumulative probability is calculated.

    [0144] The delay measurement apparatus determines the fifth delay interval corresponding to the fifth delay segment counter, and determines a line segment that is on the probability density curve f(x) and that is corresponding to the fifth delay interval. The delay measurement apparatus calculates a slope of the line segment, determines, in a step-by-step integration manner based on an area formed by the line segment and the X-axis, the node delay value that meets the target probability, and uses, as the delay threshold, the node delay value that meets the target probability.

    [0145] Optionally, the delay measurement apparatus may alternatively calculate a delay value in a specified probability distribution based on the probability density curve f(x), for example, calculate a node delay value in a case of a probability distribution 20%, or calculate a node delay value in a case of a probability distribution 99%.

    [0146] Optionally, the delay measurement apparatus may alternatively obtain, based on the probability density curve f(x), a node delay value corresponding to the average delay value, and measure delay performance of the network node device by using the node delay value corresponding to the average delay value.

    [0147] In this embodiment of the present invention, optionally, a value that is of each delay segment counter and that is obtained by performing delay distribution measurement by the delay measurement apparatus, the maximum node delay value, the minimum node delay value, and the average delay value that are obtained by performing delay feature measurement, and the cumulative probability and the delay threshold that are obtained by performing delay probability measurement are stored in a unified format. For example, when the delay measurement apparatus includes an MIB, results obtained by performing delay distribution measurement, delay feature measurement, and delay probability measurement are stored in an MIB format.

    [0148] According to the delay measurement method disclosed in Embodiment 2 of the present invention, delay distribution measurement, delay feature measurement, and delay probability measurement are separately performed on the node delay value received in the preset time period, and serviceability, availability, and reliability of delay performance of the network node device are evaluated based on delay measurement data. Further, delay measurement data that is of each network node device and that is provided in an entire network is obtained. Precise quantitative evaluation and monitoring of serviceability, availability, and reliability of a network node delay can be completed based on obtained comprehensive delay measurement data.

    [0149] Optionally, based on the delay measurement methods disclosed in Embodiment 1 and Embodiment 2 of the present invention, in a process of performing delay measurement, a link delay between network node devices may be further collected, for example, a link delay value DL between a network node device R1 and a network node device R2 shown in FIG. 1. Optionally, when the network node devices are connected by using an optical fiber, a link delay generated when a packet is transmitted through the optical fiber is a fixed latency.

    [0150] Further, better quantitative evaluation and monitoring of the serviceability, the availability, and the reliability of the network node delay can be completed after the link delay value between the network node devices and a node delay value of each network node are obtained and combined.

    [0151] Optionally, based on the delay measurement methods disclosed in Embodiment 1 and Embodiment 2 of the present invention, when the preset time period starts, the node delay value obtained through calculation may be first classified, and the node delay value obtained through calculation is measured based on a classification in the delay measurement apparatus. Alternatively, a node delay value corresponding to a particular type in the classification may be measured according to a user requirement.

    [0152] The node delay value is generated when a packet is transmitted from an ingress port of the network node device to an egress port of the network node device. The packet further carries port number information when being input to the network node device, a forwarding priority of the packet, or an ID number of a service flow in which the packet is located. The port number information of the network node device indicates a number of each port on the network node device. The forwarding priority of the packet indicates a priority classified according to importance of the packet. For example, forwarding priorities of packets may be classified in ascending order into 0 to 7, a total of eight levels. The ID number of the service flow in which the packet is located indicates a number of the service flow.

    [0153] For example, classification is performed according to the port number information when the packet is input to the network node device, and a current delay measurement apparatus processes only a node delay value that is generated by a packet input from a second port. If the network node device receives a packet input from a first port, and obtains a node delay value of the packet in the network node device through calculation, the delay measurement apparatus does not measure the node delay value because the delay measurement apparatus processes only the node delay value that is generated by the packet input from the second port. If the node delay value is the node delay value that is generated by the packet input from the second port, the delay measurement apparatus measures the node delay value.

    [0154] For example, classification is performed according to the forwarding priority of the packet, and a current delay measurement apparatus processes only a node delay value that is generated by a packet whose forwarding priority is 7. If the network node device receives a packet whose forwarding priority is 5, and obtains a node delay value of the packet in the network node device through calculation, the delay measurement apparatus does not measure the node delay value because the delay measurement apparatus processes only the node delay value that is generated by the packet whose forwarding priority is 7. If the node delay value is the node delay value that is generated by the packet whose forwarding priority is 7, the delay measurement apparatus measures the node delay value.

    [0155] Further, the delay measurement apparatus may perform category-based measurement on the node delay value of the network node according to different application requirements. With category-based measurement on the node delay value, a node delay value that meets a corresponding application requirement may be collected more pertinently, so that in a process of precisely evaluating and monitoring the serviceability, the availability, and the reliability of the network node delay, pertinence can be improved, and power consumption can be reduced.

    [0156] Based on the foregoing delay measurement technical solutions disclosed in Embodiment 1 and Embodiment 2 of the present invention, the present invention further discloses a corresponding delay measurement apparatus.

    [0157] FIG. 6 is a schematic structural diagram of a delay measurement apparatus according to the present invention. The delay measurement apparatus 10 includes a preset module 11, a delay distribution measurement module 12, and a delay probability measurement module 13.

    [0158] The preset module 11 is configured to: obtain a basic delay value range of a network node device, a maximum node delay value of the network node device in a congestion case, and a theoretically minimum node delay value; determine a to-be-divided node delay value range based on the theoretically minimum node delay value and the maximum node delay value; divide the to-be-divided node delay value range based on a long-tailed distribution and the basic delay value range, to obtain M delay intervals and M+1 delay boundary values; and configure one delay segment counter for each delay interval.

    [0159] The delay distribution measurement module 12 is configured to: obtain a node delay value of a current packet when a preset time period starts, where the node delay value is used to indicate a delay generated when the current packet is forwarded from an ingress port of the network node device to an egress port of the network node device; determine, based on the M+1 boundary values of the preset M delay intervals, a delay interval in which the node delay value is located, and add 1 to a value of a delay segment counter corresponding to the delay interval in which the node delay value is located, where M>1, M is a positive integer, the delay interval in which the node delay value is located belongs to the M delay intervals, and each delay interval corresponds to one delay segment counter; return and obtain a node delay value of a next packet until the preset time period ends; and send a value of each delay segment counter to the delay probability measurement module 13.

    [0160] The delay probability measurement module 13 includes a calculation unit 131, a first obtaining unit 132, and a second obtaining unit 133.

    [0161] The calculation unit 131 is configured to calculate a cumulative probability Pacc.j. For a process of calculating the cumulative probability Pacc.j, refer to formula (3).

    [0162] The first obtaining unit 132 is configured to: when it is determined that a cumulative probability Pacc.j that is equal to a target probability is obtained, obtain, as a delay threshold, a node delay value that meets the target probability.

    [0163] The second obtaining unit 133 is configured to: when it is determined that a cumulative probability Pacc.j that is greater than the target probability is obtained, obtain a jth delay segment counter when calculating the cumulative probability Pacc.j, and determine boundary values Dj-1 and Dj of a delay interval corresponding to the jth delay segment counter; and search, between the boundary value Dj-1 and the boundary value Dj, for a node delay value that meets the target probability, and use, as a delay threshold, the node delay value that meets the target probability.

    [0164] FIG. 7 is a schematic structural diagram of another delay measurement apparatus according to the present invention. The delay measurement apparatus 20 includes a preset module 21, a delay distribution measurement module 22, a delay feature measurement module 23, and a delay probability measurement module 24.

    [0165] The preset module 21 is the same as the preset module 11 shown in FIG. 6, and the delay distribution measurement module 22 is the same as the delay distribution measurement module 12 shown in FIG. 6. Details are not described herein again.

    [0166] The delay feature measurement module 23 is configured to: determine a maximum node delay value and a minimum node delay value in node delay values obtained in a preset time period; and determine a delay interval between the maximum node delay value and the minimum node delay value, and obtain a quantity of valid delay segment counters corresponding to the delay interval, where the valid delay segment counters are delay segment counters that perform counting in the preset time period; and obtain a probability density curve f(x) through fitting calculation based on a distribution function and values of a determined quantity of valid delay segment counters, and send the probability density curve f(x) to the delay probability measurement module 24, where x is a random variable and varies with the values of the determined quantity of valid delay segment counters.

    [0167] The delay probability measurement module 24 includes a calculation unit 241, a first obtaining unit 242, and a third obtaining unit 243.

    [0168] The calculation unit 241 is the same as the calculation unit 131 shown in FIG. 6, and the first obtaining unit 242 is the same as the first obtaining unit 132 shown in FIG. 6. Details are not described herein again.

    [0169] The third obtaining unit 243 is configured to: when the delay probability measurement module obtains a cumulative probability Pacc.j that is greater than a target probability, determine, on the probability density curve f(x), a node delay value that meets the target probability, and use, as a delay threshold, the node delay value that meets the target probability.

    [0170] Based on each of the delay measurement apparatuses separately shown in FIG. 6 and FIG. 7 of the present invention, optionally, the delay measurement apparatus may further include a classification module. The classification module is configured to classify the calculated node delay value of the current packet based on a forwarding priority of the current packet, an ID number of a service flow in which the current packet is located, or a port number of a port through which the current packet is input to the network node device. The delay measurement apparatus further performs corresponding measurement based on a category. For details, refer to a corresponding part of the delay measurement method disclosed in the present invention.

    [0171] Optionally, each of the delay measurement apparatuses separately shown in FIG. 6 and FIG. 7 may be disposed inside the egress port of the network node device, or may be independently disposed outside the egress port of the network node device.

    [0172] Optionally, each of the delay measurement apparatuses separately shown in FIG. 6 and FIG. 7 may be alternatively disposed in a network controller, and the network controller is connected to each network node device. For example, in an application scenario shown in FIG. 8, a delay measurement apparatus 31 is disposed in a network controller 30, and the network controller is connected to a network node device R1 and a network node device R2. The network node device R1 and the network node device R2 separately send, to the network controller 30, respective node delay values that are obtained through calculation, and the delay measurement apparatus 31 disposed in the network controller performs corresponding measurement. Optionally, the delay measurement apparatus may be alternatively split into two parts: One part is disposed in the network controller, and the other part is disposed in the network node device. This is not limited in the present invention.

    [0173] With reference to the delay measurement method disclosed in the embodiment of the present invention, the delay measurement apparatus disclosed in this embodiment of the present invention may also be implemented directly by using hardware, a memory executed by a processor, or a combination thereof.

    [0174] Therefore, the present invention further discloses another delay measurement apparatus 40 corresponding to the delay measurement method disclosed in the embodiment of the present invention. As shown in FIG. 9, the delay measurement apparatus 40 includes a memory 41 and a processor 42 that communicates with the memory 41.

    [0175] The memory 41 has a storage medium, and the storage medium stores an operation procedure of performing delay measurement on a node delay value.

    [0176] For the operation procedure of performing delay measurement on the node delay value, the operation procedure may include program code, and the program code may include a series of operating instructions sorted in a specific order. The processor may be a central processing unit CPU, or an application-specific integrated circuit, or one or more integrated circuits configured to implement this embodiment of the present invention.

    [0177] The memory may include a high-speed RAM memory, and may further include a non-volatile memory, such as at least one magnetic disk memory.

    [0178] The processor is connected to the memory by using a bus. When delay measurement needs to be performed on a node delay value of a network node device, the processor invokes the operation procedure of delay measurement that is stored in the memory. For the operation procedure, refer to measurement on the node delay value of the network node device in the delay measurement method disclosed in Embodiment 1 of the present invention or Embodiment 2 of the present invention. Details are not described herein again.

    [0179] In conclusion, according to the delay measurement technical solutions disclosed in the present invention, delay measurement is performed on the node delay value received in the preset time period, and serviceability, availability, and reliability of delay performance of the network node device are evaluated based on delay measurement data. Therefore, comprehensive delay measurement data is obtained. Based on this, precise quantitative evaluation of serviceability, availability, and reliability of network node delays across an entire network can be completed.

    [0180] Further, node delay data corresponding to a specific probability in the network node device may be provided.

    [0181] Still further, delay measurement data of each network node device may be further used to monitor a rule that the network node delay varies with port traffic of the network node device, so that the rule is used to direct network expansion and upgrade.


    Claims

    1. A delay measurement method of a network node device, wherein the network node device comprises a delay measurement apparatus, and the delay measurement method comprises the following steps:

    a) obtaining (S104, S204), by the delay measurement apparatus, a node delay value of a current packet when a preset time period starts, wherein the node delay value is used to indicate a delay generated when the current packet is forwarded from an ingress port of the network node device to an egress port of the network node device;

    b) determining (S104, S204), by the delay measurement apparatus based on M+1 boundary values of preset M delay intervals, a delay interval in which the node delay value is located, and adding (S104, S204) 1 to a value of a delay segment counter corresponding to the delay interval in which the node delay value is located, wherein M>1, M is a positive integer, the delay interval in which the node delay value is located belongs to the M delay intervals, and each delay interval corresponds to one delay segment counter;

    c) repeating (S105, S205), by the delay measurement apparatus, the above steps for subsequent packets forwarded from the ingress port of the network node device to the egress port of the network node device until the preset time period ends; and

    d1) calculating (S106, S211), by the delay measurement apparatus, a cumulative probability Pacc.j, determining (S106, S211) the cumulative probability Pacc.j that is greater than or equal to a target probability, wherein the determining is done in such a manner that the calculated cumulative probability Pacc.j is greater than or equal to the target probability, and obtaining (S106, S211), as a delay threshold, a node delay value satisfying the target probability, wherein the cumulative probability Pacc.j =Cacc.j /Ctotal, Cacc.j is a sum obtained by accumulating from a value of a first delay segment counter to a value of a jth delay segment counter, 1<j≤M, j is a positive integer, and Ctotal is a sum Ctotal of values of M delay segment counters in the preset time period; wherein

    d2) the values of the delay segment counters are obtained based on the packets forwarded from the ingress port to the egress port of the network node device during the preset time period, whereas the values of the delay segment counters are accumulated until the cumulative probability Pacc.j is greater than or equal to the target probability.


     
    2. The delay measurement method according to claim 1, wherein the determining (S106) the cumulative probability Pacc.j that is greater than or equal to a target probability, and obtaining (S 106), as a delay threshold, a node delay value satisfying the target probability comprises:

    when obtaining the cumulative probability Pacc.j that is equal to the target probability, obtaining (S 107), by the delay measurement apparatus as the delay threshold, the node delay value satisfying the target probability; or

    when obtaining the cumulative probability Pacc.j that is greater than the target probability, obtaining (S108), by the delay measurement apparatus, the jth delay segment counter when calculating the cumulative probability Pacc.j, and determining boundary values Dj-1 and Dj of a delay interval corresponding to the jth delay segment counter; and

    searching (S109), by the delay measurement apparatus between the boundary value Dj-1 and the boundary value Dj, for a node delay value satisfying the target probability, and using, as the delay threshold, the node delay value satisfying the target probability.


     
    3. The delay measurement method according to claim 1, further comprising:

    determining (S207), by the delay measurement apparatus, a maximum node delay value and a minimum node delay value in node delay values obtained in the preset time period;

    determining (S209), by the delay measurement apparatus, a delay interval between the maximum node delay value and the minimum node delay value, and obtaining a quantity of valid delay segment counters corresponding to the delay interval, wherein the valid delay segment counters are delay segment counters that perform counting in the preset time period; and

    obtaining (S210), by the delay measurement apparatus, a probability density curve f(x) through fitting calculation based on a distribution function and values of a determined quantity of valid delay segment counters, wherein x is a random variable and varies with the values of the determined quantity of valid delay segment counters.


     
    4. The delay measurement method according to claim 3, wherein the determining (S211) the cumulative probability Pacc.j that is greater than or equal to a target probability, and obtaining (S211), as a delay threshold, a node delay value satisfying the target probability comprises:

    when obtaining the cumulative probability Pacc.j that is equal to the target probability, obtaining (S212), by the delay measurement apparatus as the delay threshold, the corresponding node delay value satisfying the target probability; or

    when obtaining the cumulative probability Pacc.j that is greater than the target probability, determining (S213), by the delay measurement apparatus on the probability density curve f(x), a node delay value satisfying the target probability, and using, as the delay threshold, the node delay value satisfying the target probability.


     
    5. The delay measurement method according to claim 3, further comprising:

    calculating (S206), by the delay measurement apparatus, an average delay value of node delay values that are of all packets and that are obtained in the preset time period; and

    determining (S213), by the delay measurement apparatus, a node delay value corresponding to the average delay value on the probability density curve f(x), wherein the node delay value corresponding to the average delay value is used to measure node delay performance of the network node device.


     
    6. The delay measurement method according to any one of claims 1 to 5, further comprising:
    classifying, by the delay measurement apparatus, the node delay value of the current packet based on a forwarding priority of the current packet, an ID number of a service flow in which the current packet is located, or a port number of a port through which the current packet is input to the network node device.
     
    7. A delay measurement apparatus, wherein a network node device comprises the delay measurement apparatus, the delay measurement apparatus comprises:

    a) a delay distribution measurement module (12, 22), configured to: obtain a node delay value of a current packet when a preset time period starts, wherein the node delay value is used to indicate a delay generated when the current packet is forwarded from an ingress port of the network node device to an egress port of the network node device; determine, based on M+1 boundary values of preset M delay intervals, a delay interval in which the node delay value is located, and add 1 to a value of a delay segment counter corresponding to the delay interval in which the node delay value is located, wherein M>1, M is a positive integer, the delay interval in which the node delay value is located belongs to the M delay intervals, and each delay interval corresponds to one delay segment counter; repeat the above steps for subsequent packets forwarded from the ingress port of the network node device to the egress port of the network node device until the preset time period ends; and send a value of each delay segment counter to a delay probability measurement module (13, 24); and

    b) the delay probability measurement module (13, 24), configured to: calculate a cumulative probability Pacc.j, determine the cumulative probability Pacc.j that is greater than or equal to a target probability, wherein the determining is done in such a manner that the calculated cumulative probability Pacc.j is greater than or equal to the target probability, and obtain, as a delay threshold, a node delay value satisfying the target probability, wherein the cumulative probability Pacc.j =Cacc.j /Ctotal, Cacc.j is a sum obtained by accumulating from a value of a first delay segment counter to a value of a jth delay segment counter, 1<j≤M, j is a positive integer, and Ctotal is a sum Ctotal of values of M delay segment counters in the preset time period; wherein

    c) the values of the delay segment counters are obtained based on the packets forwarded from the ingress port to the egress port of the network node device during the preset time period, whereas the values of the delay segment counters are accumulated until the cumulative probability Pacc.j is greater than or equal to the target probability.


     
    8. The delay measurement apparatus according to claim 7, wherein the delay probability measurement module (13, 24) comprises:

    a first obtaining unit (132), configured to: when the delay probability measurement module (13, 24) obtains the cumulative probability Pacc.j that is equal to the target probability, obtain, as the delay threshold, the node delay value satisfying the target probability; and

    a second obtaining unit (133), configured to: when the delay probability measurement module (13, 24) obtains the cumulative probability Pacc.j that is greater than the target probability, obtain the jth delay segment counter when calculating the cumulative probability Pacc.j, and determine boundary values Dj-1 and Dj of a delay interval corresponding to the jth delay segment counter; and search, between the boundary value Dj-1 and the boundary value Dj, for a node delay value satisfying the target probability, and use, as the delay threshold, the node delay value satisfying the target probability.


     
    9. The delay measurement apparatus according to claim 7, further comprising a delay feature measurement module (23); wherein
    the delay feature measurement module (23) is configured to: determine a maximum node delay value and a minimum node delay value in node delay values obtained in the preset time period; determine a delay interval between the maximum node delay value and the minimum node delay value, and obtain a quantity of valid delay segment counters corresponding to the delay interval, wherein the valid delay segment counters are delay segment counters that perform counting in the preset time period; and obtain a probability density curve f(x) through fitting calculation based on a distribution function and values of a determined quantity of valid delay segment counters, and send the probability density curve f(x) to the delay probability measurement module (13, 24), wherein x is a random variable and varies with the values of the determined quantity of valid delay segment counters.
     
    10. The delay measurement apparatus according to claim 9, wherein the delay probability measurement module (13, 24) comprises:

    a first obtaining unit (242), configured to: when the delay probability measurement module (13, 24) obtains the cumulative probability Pacc.j that is equal to the target probability, obtain, as the delay threshold, the node delay value having the target probability; and

    a third obtaining unit (243), configured to: when the delay probability measurement module (13, 24) obtains the cumulative probability Pacc.j that is greater than the target probability, determine, on the probability density curve f(x), a node delay value having the target probability, and use, as the delay threshold, the node delay value having the target probability.


     
    11. A delay measurement apparatus (40), comprising a memory (41) and a processor (42) that communicates with the memory (41), wherein

    i) the memory (41) is configured to store program code for delay measurement; and

    ii) the processor (42) is configured to invoke the program code for delay measurement in the memory to perform the delay measurement method according to any one of claims 1 to 6.


     
    12. A network node device, comprising the delay measurement apparatus according to any one of claims 7 to 11, wherein the delay measurement apparatus is disposed inside an egress port or outside the egress port of the network node device.
     


    Ansprüche

    1. Verzögerungsmessverfahren einer Netzwerkknotenvorrichtung, wobei die Netzwerkknotenvorrichtung eine Verzögerungsmessvorrichtung umfasst und das Verzögerungsmessverfahren die folgenden Schritte umfasst:

    a) Erhalten (S104, S204), durch die Verzögerungsmessvorrichtung, eines Knotenverzögerungswerts eines aktuellen Pakets, wenn ein vorgegebener Zeitraum startet, wobei der Knotenverzögerungswert verwendet wird, um eine Verzögerung anzugeben, die erzeugt wird, wenn das aktuelle Paket von einem Eingangsport der Netzwerkknotenvorrichtung zu einem Ausgangsport der Netzwerkknotenvorrichtung weitergeleitet wird;

    b) Bestimmen (S104, S204), durch die Verzögerungsmessvorrichtung basierend auf M+1 Grenzwerten von vorgegebenen M Verzögerungsintervallen, eines Verzögerungsintervalls, in dem sich der Knotenverzögerungswert befindet, und Addieren (S104, S204) von 1 zu einem Wert eines Verzögerungssegmentzählers entsprechend dem Verzögerungsintervall, in dem sich der Knotenverzögerungswert befindet, wobei M>1, M eine positive Ganzzahl ist, das Verzögerungsintervall, in dem sich der Knotenverzögerungswert befindet, zu den M Verzögerungsintervallen gehört und jedes Verzögerungsintervall einem Verzögerungssegmentzähler entspricht;

    c) Wiederholen (S105, S205), durch die Verzögerungsmessvorrichtung, der vorstehenden Schritte für nachfolgende Pakete, die von dem Eingangsport der Netzwerkknotenvorrichtung zu dem Ausgangsport der Netzwerkknotenvorrichtung weitergeleitet werden, bis der vorgegebene Zeitraum endet; und

    d1) Berechnen (S106, S211), durch die Verzögerungsmessvorrichtung, einer kumulativen Wahrscheinlichkeit Pacc.j, Bestimmen (S106, S211) der kumulativen Wahrscheinlichkeit Pacc.j, die größer oder gleich einer Zielwahrscheinlichkeit ist, wobei das Bestimmen derart durchgeführt wird, dass die berechnete kumulative Wahrscheinlichkeit Pacc.j größer oder gleich der Zielwahrscheinlichkeit ist, und Erhalten (S106, S211), als einen Verzögerungsschwellwert, eines Knotenverzögerungswerts, der die Zielwahrscheinlichkeit erfüllt, wobei die kumulative Wahrscheinlichkeit Pacc.j=Cacc.j/Cgesamt, Cacc.j eine Summe ist, die durch Akkumulieren von einem Wert eines ersten Verzögerungssegmentzählers zu einem Wert eines j. Verzögerungssegmentzählers erhalten wird, 1<j≤M, j eine positive Ganzzahl ist und Cgesamt eine Summe Cgesamt von Werten von M Verzögerungssegmentzählern in dem vorgegebenen Zeitraum ist; wobei

    d2) die Werte der Verzögerungssegmentzähler basierend auf den Paketen erhalten werden, die während des vorgegebenen Zeitraums von dem Eingangsport zu dem Ausgangsport der Netzwerkknotenvorrichtung weitergeleitet werden, während die Werte der Verzögerungssegmentzähler akkumuliert werden, bis die kumulative Wahrscheinlichkeit Pacc.j größer oder gleich der Zielwahrscheinlichkeit ist.


     
    2. Verzögerungsmessverfahren nach Anspruch 1, wobei das Bestimmen (S106) der kumulativen Wahrscheinlichkeit Pacc.j, die größer oder gleich einer Zielwahrscheinlichkeit ist, und Erhalten (S 106), als einen Verzögerungsschwellwert, eines Knotenverzögerungswerts, der die Zielwahrscheinlichkeit erfüllt, umfasst:

    beim Erhalten der kumulativen Wahrscheinlichkeit Pacc.j, die gleich der Zielwahrscheinlichkeit ist, Erhalten (S 107), durch die Verzögerungsmessvorrichtung als den Verzögerungsschwellwert, des Knotenverzögerungswerts, der die Zielwahrscheinlichkeit erfüllt; oder

    beim Erhalten der kumulativen Wahrscheinlichkeit Pacc.j, die größer als die Zielwahrscheinlichkeit ist, Erhalten (S 108), durch die Verzögerungsmessvorrichtung, des j. Verzögerungssegmentzählers beim Berechnen der kumulativen Wahrscheinlichkeit Pacc.j und Bestimmen der Grenzwerte Dj-1 und Dj eines Verzögerungsintervalls entsprechend dem j. Verzögerungssegmentzähler; und

    Suchen (S 109), durch die Verzögerungsmessvorrichtung zwischen dem Grenzwert Dj-1 und dem Grenzwert Dj, eines Knotenverzögerungswerts, der die Zielwahrscheinlichkeit erfüllt, und Verwenden, als den Verzögerungsschwellwert, des Knotenverzögerungswerts, der die Zielwahrscheinlichkeit erfüllt.


     
    3. Verzögerungsmessverfahren nach Anspruch 1, ferner umfassend:

    Bestimmen (S207), durch die Verzögerungsmessvorrichtung, eines maximalen Knotenverzögerungswerts und eines minimalen Knotenverzögerungswerts in Knotenverzögerungswerten, die in dem vorgegebenen Zeitraum erhalten werden;

    Bestimmen (S209), durch die Verzögerungsmessvorrichtung, eines Verzögerungsintervalls zwischen dem maximalen Knotenverzögerungswert und dem minimalen Knotenverzögerungswert und Erhalten einer Menge von gültigen Verzögerungssegmentzählern entsprechend dem Verzögerungsintervall, wobei die gültigen Verzögerungssegmentzähler Verzögerungssegmentzähler sind, die das Zählen in dem vorgegebenen Zeitraum durchführen; und

    Erhalten (S210), durch die Verzögerungsmessvorrichtung, einer Wahrscheinlichkeitsdichtekurve f(x) durch eine Fitting-Berechnung basierend auf einer Verteilungsfunktion und Werten einer bestimmten Menge von gültigen Verzögerungssegmentzählern, wobei x eine Zufallsgröße ist und mit den Werten der bestimmten Menge von gültigen Verzögerungssegmentzählern variiert.


     
    4. Verzögerungsmessverfahren nach Anspruch 3, wobei das Bestimmen (S211) der kumulativen Wahrscheinlichkeit Pacc.j, die größer oder gleich einer Zielwahrscheinlichkeit ist, und Erhalten (S211), als einen Verzögerungsschwellwert, eines Knotenverzögerungswerts, der die Zielwahrscheinlichkeit erfüllt, umfasst:

    beim Erhalten der kumulativen Wahrscheinlichkeit Pacc.j, die gleich der Zielwahrscheinlichkeit ist, Erhalten (S212), durch die Verzögerungsmessvorrichtung als den Verzögerungsschwellwert, des entsprechenden Knotenverzögerungswerts, der die Zielwahrscheinlichkeit erfüllt; oder

    beim Erhalten der kumulativen Wahrscheinlichkeit Pacc.j, die größer als die Zielwahrscheinlichkeit ist, Bestimmen (S213), durch die Verzögerungsmessvorrichtung bei der Wahrscheinlichkeitsdichtekurve f(x), eines Knotenverzögerungswerts, der die Zielwahrscheinlichkeit erfüllt, und Verwenden, als den Verzögerungsschwellwert, des Knotenverzögerungswerts, der die Zielwahrscheinlichkeit erfüllt.


     
    5. Verzögerungsmessverfahren nach Anspruch 3, ferner umfassend:

    Berechnen (S206), durch die Verzögerungsmessvorrichtung, eines durchschnittlichen Verzögerungswerts von Knotenverzögerungswerten, die von allen Paketen sind und die in dem vorgegebenen Zeitraum erhalten werden; und

    Bestimmen (S213), durch die Verzögerungsmessvorrichtung, eines Knotenverzögerungswerts entsprechend dem durchschnittlichen Verzögerungswert bei der Wahrscheinlichkeitsdichtekurve f(x), wobei der Knotenverzögerungswert entsprechend dem durchschnittlichen Verzögerungswert verwendet wird, um die Knotenverzögerungsleistung der Netzwerkknotenvorrichtung zu messen.


     
    6. Verzögerungsmessverfahren nach einem der Ansprüche 1 bis 5, ferner umfassend:
    Klassifizieren, durch die Verzögerungsmessvorrichtung, des Knotenverzögerungswerts des aktuellen Pakets basierend auf einer Weiterleitungspriorität des aktuellen Pakets, einer ID-Nummer eines Dienstflusses, in dem sich das aktuelle Paket befindet, oder einer Port-Nummer eines Ports, durch den das aktuelle Paket in die Netzwerkknotenvorrichtung eingegeben wird.
     
    7. Verzögerungsmessvorrichtung, wobei eine Netzwerkknotenvorrichtung die Verzögerungsmessvorrichtung umfasst, wobei die Verzögerungsmessvorrichtung umfasst:

    a) ein Verzögerungsverteilungsmessmodul (12, 22), das ausgelegt ist zum: Erhalten eines Knotenverzögerungswerts eines aktuellen Pakets, wenn ein vorgegebener Zeitraum startet, wobei der Knotenverzögerungswert verwendet wird, um eine Verzögerung anzugeben, die erzeugt wird, wenn das aktuelle Paket von einem Eingangsport der Netzwerkknotenvorrichtung zu einem Ausgangsport der Netzwerkknotenvorrichtung weitergeleitet wird; Bestimmen, basierend auf M+1 Grenzwerten von vorgegebenen M Verzögerungsintervallen, eines Verzögerungsintervalls, in dem sich der Knotenverzögerungswert befindet, und Addieren von 1 zu einem Wert eines Verzögerungssegmentzählers entsprechend dem Verzögerungsintervall, in dem sich der Knotenverzögerungswert befindet, wobei M>1, M eine positive Ganzzahl ist, das Verzögerungsintervall, in dem sich der Knotenverzögerungswert befindet, zu den M Verzögerungsintervallen gehört und jedes Verzögerungsintervall einem Verzögerungssegmentzähler entspricht; Wiederholen der vorstehenden Schritte für nachfolgende Pakete, die von dem Eingangsport der Netzwerkknotenvorrichtung zu dem Ausgangsport der Netzwerkknotenvorrichtung weitergeleitet werden, bis der vorgegebene Zeitraum endet; und Senden eines Werts jedes Verzögerungssegmentzählers an ein Verzögerungswahrscheinlichkeitsmessmodul (13, 24); und

    b) das Verzögerungswahrscheinlichkeitsmessmodul (13, 24), das ausgelegt ist zum: Berechnen einer kumulativen Wahrscheinlichkeit Pacc.j, Bestimmen der kumulativen Wahrscheinlichkeit Pacc.j, die größer oder gleich einer Zielwahrscheinlichkeit ist, wobei das Bestimmen derart durchgeführt wird, dass die berechnete kumulative Wahrscheinlichkeit Pacc.j größer oder gleich der Zielwahrscheinlichkeit ist, und Erhalten, als einen Verzögerungsschwellwert, eines Knotenverzögerungswerts, der die Zielwahrscheinlichkeit erfüllt, wobei die kumulative Wahrscheinlichkeit Pacc.j=Cacc.j/Cgesamt, Cacc.j eine Summe ist, die durch Akkumulieren von einem Wert eines ersten Verzögerungssegmentzählers zu einem Wert eines j. Verzögerungssegmentzählers erhalten wird, 1<j≤M, j eine positive Ganzzahl ist und Cgesamt eine Summe Cgesamt von Werten von M Verzögerungssegmentzählern in dem vorgegebenen Zeitraum ist; wobei

    c) die Werte der Verzögerungssegmentzähler basierend auf den Paketen erhalten werden, die während des vorgegebenen Zeitraums von dem Eingangsport zu dem Ausgangsport der Netzwerkknotenvorrichtung weitergeleitet werden, während die Werte der Verzögerungssegmentzähler akkumuliert werden, bis die kumulative Wahrscheinlichkeit Pacc.j größer oder gleich der Zielwahrscheinlichkeit ist.


     
    8. Verzögerungsmessvorrichtung nach Anspruch 7, wobei das Verzögerungswahrscheinlichkeitsmessmodul (13, 24) umfasst:

    eine erste Erhaltungseinheit (132), die ausgelegt ist zum: wenn das Verzögerungswahrscheinlichkeitsmessmodul (13, 24) die kumulative Wahrscheinlichkeit Pacc.j, die gleich der Zielwahrscheinlichkeit ist, erhält, Erhalten, als den Verzögerungsschwellwert, des Knotenverzögerungswerts, der die Zielwahrscheinlichkeit erfüllt; und

    eine zweite Erhaltungseinheit (133), die ausgelegt ist zum: wenn das Verzögerungswahrscheinlichkeitsmessmodul (13, 24) die kumulative Wahrscheinlichkeit Pacc.j, die größer als die Zielwahrscheinlichkeit ist, erhält, Erhalten des j. Verzögerungssegmentzählers beim Berechnen der kumulativen Wahrscheinlichkeit Pacc.j, und Bestimmen der Grenzwerte Dj-1 und Dj eines Verzögerungsintervalls entsprechend dem j. Verzögerungssegmentzähler und Suchen, zwischen dem Grenzwert Dj-1 und dem Grenzwert Dj, eines Knotenverzögerungswerts, der die Zielwahrscheinlichkeit erfüllt, und Verwenden, als den Verzögerungsschwellwert, des Knotenverzögerungswerts, der die Zielwahrscheinlichkeit erfüllt.


     
    9. Verzögerungsmessvorrichtung nach Anspruch 7, ferner umfassend ein Verzögerungsmerkmalsmessmodul (23); wobei
    das Verzögerungsmerkmalsmessmodul (23) ausgelegt ist zum: Bestimmen eines maximalen Knotenverzögerungswerts und eines minimalen Knotenverzögerungswerts in Knotenverzögerungswerten, die in dem vorgegebenen Zeitraum erhalten werden; Bestimmen eines Verzögerungsintervalls zwischen dem maximalen Knotenverzögerungswert und dem minimalen Knotenverzögerungswert und Erhalten einer Menge von gültigen Verzögerungssegmentzählern entsprechend dem Verzögerungsintervall, wobei die gültigen Verzögerungssegmentzähler Verzögerungssegmentzähler sind, die das Zählen in dem vorgegebenen Zeitraum durchführen; und Erhalten einer Wahrscheinlichkeitsdichtekurve f(x) durch eine Fitting-Berechnung basierend auf einer Verteilungsfunktion und Werten einer bestimmten Menge von gültigen Verzögerungssegmentzählern und Senden der Wahrscheinlichkeitsdichtekurve f(x) an das Verzögerungswahrscheinlichkeitsmessmodul (13, 24), wobei x eine Zufallsgröße ist und mit den Werten der bestimmten Menge von gültigen Verzögerungssegmentzählern variiert.
     
    10. Verzögerungsmessvorrichtung nach Anspruch 9, wobei das Verzögerungswahrscheinlichkeitsmessmodul (13, 24) umfasst:

    eine erste Erhaltungseinheit (242), die ausgelegt ist zum: wenn das Verzögerungswahrscheinlichkeitsmessmodul (13, 24) die kumulative Wahrscheinlichkeit Pacc.j, die gleich der Zielwahrscheinlichkeit ist, erhält, Erhalten, als den Verzögerungsschwellwert, des Knotenverzögerungswerts, der die Zielwahrscheinlichkeit aufweist; und

    eine dritte Erhaltungseinheit (243), die ausgelegt ist zum: wenn das Verzögerungswahrscheinlichkeitsmessmodul (13, 24) die kumulative Wahrscheinlichkeit Pacc.j, die größer als die Zielwahrscheinlichkeit ist, erhält, Bestimmen, bei der Wahrscheinlichkeitsdichtekurve f(x), eines Knotenverzögerungswerts, der die Zielwahrscheinlichkeit aufweist, und Verwenden, als den Verzögerungsschwellwert, des Knotenverzögerungswerts, der die Zielwahrscheinlichkeit aufweist.


     
    11. Verzögerungsmessvorrichtung (40), umfassend einen Speicher (41) und einen Prozessor (42), der mit dem Speicher (41) kommuniziert, wobei

    i) der Speicher (41) dazu ausgelegt ist, Programmcode zur Verzögerungsmessung zu speichern; und

    ii) der Prozessor (42) dazu ausgelegt ist, den Programmcode zur Verzögerungsmessung in dem Speicher aufzurufen, um das Verzögerungsmessverfahren nach einem der Ansprüche 1 bis 6 durchzuführen.


     
    12. Netzwerkknotenvorrichtung, umfassend die Verzögerungsmessvorrichtung nach einem der Ansprüche 7 bis 11, wobei die Verzögerungsmessvorrichtung innerhalb eines Ausgangsports oder außerhalb des Ausgangsports der Netzwerkknotenvorrichtung angeordnet ist.
     


    Revendications

    1. Procédé de mesure de retard d'un dispositif de nœud de réseau, le dispositif de nœud de réseau comprenant un appareil de mesure de retard et le procédé de mesure de retard comprenant les étapes suivantes :

    a) obtenir (S 104, S204), par l'appareil de mesure de retard, une valeur de retard de nœud d'un paquet actuel quand une période de temps prédéfinie commence, la valeur de retard de nœud étant utilisée pour indiquer un retard généré quand le paquet actuel est transféré d'un port d'entrée du dispositif de nœud de réseau à un port de sortie du dispositif de nœud de réseau ;

    b) déterminer (S 104, S204), par l'appareil de mesure de retard sur la base de M+1 valeurs limites de M intervalles de retard prédéfinis, un intervalle de retard où se trouve la valeur de retard de nœud et ajouter (S104, S204) 1 à une valeur d'un compteur de segments de retard correspondant à l'intervalle de retard où se trouve la valeur de retard de nœud, M>1, M étant un entier positif, l'intervalle de retard où se trouve la valeur de retard de nœud appartenant aux M intervalles de retard et chaque intervalle de retard correspondant à un compteur de segments de retard ;

    c) répéter (S 105, S205), par l'appareil de mesure de retard, les étapes précédentes pour des paquets ultérieurs transférés du port d'entrée du dispositif de nœud de réseau au port de sortie du dispositif de nœud de réseau jusqu'à ce que la période de temps prédéfinie se termine ; et

    d1) calculer (S106, S211), par l'appareil de mesure de retard, une probabilité cumulée Pacc.j, déterminer (S 106, S211) la probabilité cumulée Pacc.j qui est supérieure ou égale à une probabilité cible, la détermination étant réalisée de sorte que la probabilité cumulée Pacc.j calculée soit supérieure ou égale à la probabilité cible, et obtenir (S106, S211) comme seuil de retard une valeur de retard de nœud satisfaisant à la probabilité cible, la probabilité cumulée Pacc.j = Cacc.j/Ctotal, Cacc.j étant une somme obtenue par cumul d'une valeur d'un premier compteur de segments de retard à une valeur d'un jème compteur de segments de retard, 1<j≤M, j étant un entier positif et Ctotal étant une somme Ctotal de valeurs de M compteurs de segments de retard dans la période de temps prédéfinie ;

    d2) les valeurs des compteurs de segments de retard étant obtenues sur la base des paquets transférés du port d'entrée au port de sortie du dispositif de nœud de réseau pendant la période de temps prédéfinie, les valeurs des compteurs de segments de retard étant cumulées jusqu'à ce que la probabilité cumulée Pacc.j soit supérieure ou égale à la probabilité cible.


     
    2. Procédé de mesure de retard selon la revendication 1, dans lequel la détermination (S 106) de la probabilité cumulée Pacc.j qui est supérieure ou égale à une probabilité cible et l'obtention (S 106) comme seuil de retard d'une valeur de retard de nœud satisfaisant à la probabilité cible comprennent les étapes consistant à :

    lors de l'obtention de la probabilité cumulée Pacc.j qui est égale à la probabilité cible, obtenir (S 107) comme seuil de retard, par l'appareil de mesure de retard, la valeur de retard de nœud satisfaisant à la probabilité cible ; ou

    lors de l'obtention de la probabilité cumulée Pacc.j qui est supérieure à la probabilité cible, obtenir (S 108), par l'appareil de mesure de retard, le jème compteur de segments de retard lors du calcul de la probabilité cumulée Pacc.j et déterminer des valeurs limites Dj-1 et Dj d'un intervalle de retard correspondant au jème compteur de segments de retard ; et

    rechercher (S 109), par l'appareil de mesure de retard entre la valeur limite Dj-1 et la valeur limite Dj, une valeur de retard de nœud satisfaisant à la probabilité cible et utiliser comme seuil de retard la valeur de retard de nœud satisfaisant à la probabilité cible.


     
    3. Procédé de mesure de retard selon la revendication 1, comprenant en outre les étapes consistant à :

    déterminer (S207), par l'appareil de mesure de retard, une valeur maximale de retard de nœud et une valeur minimale de retard de nœud parmi des valeurs de retard de nœud obtenues dans la période de temps prédéfinie ;

    déterminer (S209), par l'appareil de mesure de retard, un intervalle de retard entre la valeur maximale de retard de nœud et la valeur minimale de retard de nœud et obtenir un nombre de compteurs de segments de retard valides correspondant à l'intervalle de retard, les compteurs de segments de retard valides étant des compteurs de segments de retard qui réalisent un décompte dans la période de temps prédéfinie ; et

    obtenir (S210), par l'appareil de mesure de retard, une courbe de densité de probabilité f(x) par l'intermédiaire d'un calcul d'ajustement basé sur une fonction de distribution et des valeurs d'un nombre déterminé de compteurs de segments de retard valides, x étant une variable aléatoire qui varie selon les valeurs du nombre déterminé de compteurs de segments de retard valides.


     
    4. Procédé de mesure de retard selon la revendication 3, dans lequel la détermination (S211) de la probabilité cumulée Pacc.j qui est supérieure ou égale à une probabilité cible et l'obtention (S211) comme seuil de retard d'une valeur de retard de nœud satisfaisant à la probabilité cible comprennent les étapes consistant à :

    lors de l'obtention de la probabilité cumulée Pacc.j qui est égale à la probabilité cible, obtenir (S212) comme seuil de retard, par l'appareil de mesure de retard, la valeur de retard de nœud correspondante satisfaisant à la probabilité cible ; ou

    lors de l'obtention de la probabilité cumulée Pacc.j qui est supérieure à la probabilité cible, déterminer (S213), par l'appareil de mesure de retard sur la courbe de densité de probabilité f(x), une valeur de retard de nœud satisfaisant à la probabilité cible et utiliser comme seuil de retard la valeur de retard de nœud satisfaisant à la probabilité cible.


     
    5. Procédé de mesure de retard selon la revendication 3, comprenant en outre les étapes consistant à :

    calculer (S206), par l'appareil de mesure de retard, une valeur de retard moyenne de valeurs de retard de nœud qui sont de tous les paquets et qui sont obtenues dans la période de temps prédéfinie ; et

    déterminer (S213), par l'appareil de mesure de retard, une valeur de retard de nœud correspondant à la valeur de retard moyenne sur la courbe de densité de probabilité f(x), la valeur de retard de nœud correspondant à la valeur de retard moyenne étant utilisée pour mesurer des performances de retard de nœud du dispositif de nœud de réseau.


     
    6. Procédé de mesure de retard selon l'une quelconque des revendications 1 à 5, comprenant en outre l'étape consistant à :
    classer, par l'appareil de mesure de retard, la valeur de retard de nœud du paquet actuel sur la base d'une priorité de transfert du paquet actuel, d'un numéro d'ID d'un flux de service où se trouve le paquet actuel ou d'un numéro de port d'un port par lequel le paquet actuel est entré dans le dispositif de nœud de réseau.
     
    7. Appareil de mesure de retard, un dispositif de nœud de réseau comprenant l'appareil de mesure de retard, l'appareil de mesure de retard comprenant :

    a) un module de mesure de distribution de retard (12, 22), configuré pour : obtenir une valeur de retard de nœud d'un paquet actuel quand une période de temps prédéfinie commence, la valeur de retard de nœud étant utilisée pour indiquer un retard généré quand le paquet actuel est transféré d'un port d'entrée du dispositif de nœud de réseau à un port de sortie du dispositif de nœud de réseau ; déterminer, sur la base de M+1 valeurs limites de M intervalles de retard prédéfinis, un intervalle de retard où se trouve la valeur de retard de nœud et ajouter 1 à une valeur d'un compteur de segments de retard correspondant à l'intervalle de retard où se trouve la valeur de retard de nœud, M> 1, M étant un entier positif, l'intervalle de retard où se trouve la valeur de retard de nœud appartenant aux M intervalles de retard et chaque intervalle de retard correspondant à un compteur de segments de retard ; répéter les étapes précédentes pour des paquets ultérieurs transférés du port d'entrée du dispositif de nœud de réseau au port de sortie du dispositif de nœud de réseau jusqu'à ce que la période de temps prédéfinie se termine ; et envoyer une valeur de chaque compteur de segments de retard à un module de mesure de probabilité de retard (13, 24) ; et

    b) le module de mesure de probabilité de retard (13, 24), configuré pour : calculer une probabilité cumulée Pacc.j, déterminer la probabilité cumulée Pacc.j qui est supérieure ou égale à une probabilité cible, la détermination étant réalisée de sorte que la probabilité cumulée Pacc.j calculée soit supérieure ou égale à la probabilité cible, et obtenir comme seuil de retard une valeur de retard de nœud satisfaisant à la probabilité cible, la probabilité cumulée Pacc.j = Cacc.j/Ctotal, Cacc.j étant une somme obtenue par cumul d'une valeur d'un premier compteur de segments de retard à une valeur d'un jème compteur de segments de retard, 1<j≤M, j étant un entier positif et Ctotal étant une somme Ctotal de valeurs de M compteurs de segments de retard dans la période de temps prédéfinie ;

    c) les valeurs des compteurs de segments de retard étant obtenues sur la base des paquets transférés du port d'entrée au port de sortie du dispositif de nœud de réseau pendant la période de temps prédéfinie, les valeurs des compteurs de segments de retard étant cumulées jusqu'à ce que la probabilité cumulée Pacc.j soit supérieure ou égale à la probabilité cible.


     
    8. Appareil de mesure de retard selon la revendication 7, dans lequel le module de mesure de probabilité de retard (13, 24) comprend :

    une première unité d'obtention (132), configurée pour : quand le module de mesure de probabilité de retard (13, 24) obtient la probabilité cumulée Pacc.j qui est égale à la probabilité cible, obtenir comme seuil de retard la valeur de retard de nœud satisfaisant à la probabilité cible ; et

    une deuxième unité d'obtention (133), configurée pour : quand le module de mesure de probabilité de retard (13, 24) obtient la probabilité cumulée Pacc.j qui est supérieure à la probabilité cible, obtenir le jème compteur de segments de retard lors du calcul de la probabilité cumulée Pacc.j et déterminer des valeurs limites Dj-1 et Dj d'un intervalle de retard correspondant au jème compteur de segments de retard ; et rechercher, entre la valeur limite Dj-1 et la valeur limite Dj, une valeur de retard de nœud satisfaisant à la probabilité cible et utiliser comme seuil de retard la valeur de retard de nœud satisfaisant à la probabilité cible.


     
    9. Appareil de mesure de retard selon la revendication 7, comprenant en outre un module de mesure d'attribut de retard (23) ;
    le module de mesure d'attribut de retard (23) étant configuré pour : déterminer une valeur maximale de retard de nœud et une valeur minimale de retard de nœud parmi des valeurs de retard de nœud obtenues dans la période de temps prédéfinie ; déterminer un intervalle de retard entre la valeur maximale de retard de nœud et la valeur minimale de retard de nœud et obtenir un nombre de compteurs de segments de retard valides correspondant à l'intervalle de retard, les compteurs de segments de retard valides étant des compteurs de segments de retard qui réalisent un décompte dans la période de temps prédéfinie ; et obtenir une courbe de densité de probabilité f(x) par l'intermédiaire d'un calcul d'ajustement basé sur une fonction de distribution et des valeurs d'un nombre déterminé de compteurs de segments de retard valides et envoyer la courbe de densité de probabilité f(x) au module de mesure de probabilité de retard (13, 24), x étant une variable aléatoire qui varie selon les valeurs du nombre déterminé de compteurs de segments de retard valides.
     
    10. Appareil de mesure de retard selon la revendication 9, dans lequel le module de mesure de probabilité de retard (13, 24) comprend :

    une première unité d'obtention (242), configurée pour : quand le module de mesure de probabilité de retard (13, 24) obtient la probabilité cumulée Pacc.j qui est égale à la probabilité cible, obtenir comme seuil de retard la valeur de retard de nœud ayant la probabilité cible ; et

    une troisième unité d'obtention (243), configurée pour : quand le module de mesure de probabilité de retard (13, 24) obtient la probabilité cumulée Pacc.j qui est égale à la probabilité cible, déterminer, sur la courbe de densité de probabilité f(x), une valeur de retard de nœud ayant la probabilité cible et utiliser comme seuil de retard la valeur de retard de nœud ayant la probabilité cible.


     
    11. Appareil de mesure de retard (40), comprenant une mémoire (41) et un processeur (42) qui communique avec la mémoire (41),

    i) la mémoire (41) étant configurée pour stocker un code de programme pour une mesure de retard ; et

    ii) le processeur (42) étant configuré pour appeler le code de programme pour une mesure de retard dans la mémoire pour réaliser le procédé de mesure de retard selon l'une quelconque des revendications 1 à 6.


     
    12. Dispositif de nœud de réseau comprenant l'appareil de mesure de retard selon l'une quelconque des revendications 7 à 11, l'appareil de mesure de retard étant disposé à l'intérieur d'un port de sortie ou à l'extérieur du port de sortie du dispositif de nœud de réseau.
     




    Drawing









































    Cited references

    REFERENCES CITED IN THE DESCRIPTION



    This list of references cited by the applicant is for the reader's convenience only. It does not form part of the European patent document. Even though great care has been taken in compiling the references, errors or omissions cannot be excluded and the EPO disclaims all liability in this regard.

    Patent documents cited in the description