Delay prediction system, delay prediction device, and delay prediction method

Abstract

To provide a delay prediction system that estimates the predicted value of network delay.SOLUTION: A delay prediction system includes: a state determination unit that determines whether a network state is stable or fluctuating based on a cumulative value of an estimated network state and the difference between the latest sRTT value and a predicted delay value calculated from the probability distribution of multiple sRTT values at a predetermined time; and a delay estimation unit that, when the network state is stable, uses a predicted value of the round-trip time of communication calculated from an average predicted value of the multiple sRTT values at the predetermined time and the probability distribution of the multiple sRTT values at the predetermined time as a predicted delay value, and when the network state is fluctuating, uses the latest sRTT value as the predicted delay value.SELECTED DRAWING: Figure 1

Description

Technical Field

[0001] The present disclosure relates to a delay prediction system, a delay prediction device, and a delay prediction method.

Background Art

[0002] In recent years, due to the remarkable diversification of electronic attack means such as electronic interference and cyberattacks, it is necessary to continue communication even in an intense electronic warfare environment with disturbances that change every moment and provide a highly resilient network. As an example of related technology, for example, there is the network state estimation method described in Patent Document 1.

[0003] The network state estimation method described in Patent Document 1 includes steps of obtaining the RTT (Round Trip Time) of each of a first received packet and a second received packet received via a network, and calculating an RTT difference that is the difference between the RTT of the first received packet and the RTT of the second received packet.

[0004] The network state estimation method further includes steps of smoothing the RTT difference using a plurality of stages of Kalman filters connected in series to each other, specifying a threshold for the output of the final-stage Kalman filter among the plurality of stages of Kalman filters based on the output of the first-stage Kalman filter among the plurality of stages of Kalman filters, and comparing the output of the final-stage Kalman filter among the plurality of stages of Kalman filters with the specified threshold and estimating the state of the network based on the comparison result.

Prior Art Documents

Patent Documents

[0005]

Patent Document 1

Summary of the Invention

Problems to be Solved by the Invention

[0006] In the network state estimation method described in Patent Document 1 above, although the current network state can be estimated, there is a problem that a predicted value of future network delay cannot be estimated.

[0007] The present disclosure has been made in view of the above problems, and an exemplary object thereof is to provide a technique capable of estimating a predicted value of network delay.

Means for Solving the Problems

[0008] A delay prediction system according to an exemplary aspect of the present disclosure includes a state estimation unit that estimates a network state based on a comparison result between a first threshold value specified in smoothing performed on a difference between an sRTT value that is a round-trip time of communication related to a first received packet and an sRTT value of a second received packet and a value obtained by the smoothing, a cumulative value of the estimated network state, a difference between a predicted value of a round-trip time of communication obtained from a probability distribution of a plurality of sRTT values at a predetermined time and a latest sRTT value, and a state determination unit that determines whether the network state is a stable state or a fluctuating state, and a delay estimation unit that sets an average predicted value of the plurality of sRTT values at the predetermined time and a predicted value of a round-trip time of communication obtained from the probability distribution of the plurality of sRTT values at the predetermined time as delay predicted values when the network state is the stable state, and sets the latest sRTT value as the delay predicted value when the network state is the fluctuating state.

[0009] A delay prediction apparatus according to an exemplary aspect of the present disclosure includes a state estimation unit that estimates a network state based on a comparison result between a first threshold value specified in smoothing performed on a difference between an sRTT value that is a round-trip time of communication regarding a first received packet and an sRTT value of a second received packet and a value obtained by the smoothing; a state determination unit that determines whether the network state is a stable state or a fluctuating state based on a cumulative value of the estimated network state, a difference between a predicted value of a round-trip time of communication obtained from a probability distribution of a plurality of sRTT values at a predetermined time and a latest sRTT value; and a delay estimation unit that sets, as a delay prediction value, an average predicted value of the plurality of sRTT values at the predetermined time and a predicted value of a round-trip time of communication obtained from the probability distribution of the plurality of sRTT values at the predetermined time when the network state is the stable state, and sets the latest sRTT value as the delay prediction value when the network state is the fluctuating state.

[0010] A delay prediction method according to an exemplary aspect of the present disclosure includes estimating a network state based on a comparison result between a first threshold value specified in smoothing performed on a difference between an sRTT value that is a round-trip time of communication regarding a first received packet and an sRTT value of a second received packet and a value obtained by the smoothing; determining whether the network state is a stable state or a fluctuating state based on a cumulative value of the estimated network state, a difference between a predicted value of a round-trip time of communication obtained from a probability distribution of a plurality of sRTT values at a predetermined time and a latest sRTT value; and setting, as a delay prediction value, an average predicted value of the plurality of sRTT values at the predetermined time and a predicted value of a round-trip time of communication obtained from the probability distribution of the plurality of sRTT values at the predetermined time when the network state is the stable state, and setting the latest sRTT value as the delay prediction value when the network state is the fluctuating state.

Advantages of the Invention

[0011] According to an exemplary aspect of the present disclosure, there is an exemplary effect that a technique for estimating a predicted value of the round-trip time of network communication can be provided.

Brief Description of the Drawings

[0012]

Figure 1

Figure 2

Figure 3

Figure 4

Figure 5

Figure 6

Figure 7

Figure 8

Figure 9

Figure 10

Embodiments for Carrying Out the Invention

[0013] Hereinafter, embodiments of the present invention will be exemplified. However, the present invention is not limited to the following exemplary embodiments, and various modifications are possible within the scope indicated in the claims. For example, embodiments obtained by appropriately combining the technical means employed in the following exemplary embodiments may also be included in the scope of the present invention. Further, embodiments obtained by appropriately omitting a part of the technical means employed in the following exemplary embodiments may also be included in the scope of the present invention. Also, the effects mentioned in the following exemplary embodiments are merely examples of the effects expected in those exemplary embodiments and do not define the scope of the present invention. That is, embodiments that do not exhibit the effects mentioned in the following exemplary embodiments may also be included in the scope of the present invention.

[0014] 〔First Exemplary Embodiment〕 A first exemplary embodiment, which is an example of an embodiment of the present invention, will be described in detail with reference to the drawings. This exemplary embodiment is a basic form for each of the exemplary embodiments described later. Note that the scope of application of each technical means employed in this exemplary embodiment is not limited to this exemplary embodiment. That is, each technical means employed in this exemplary embodiment can be employed in other exemplary embodiments included in this disclosure as long as there are no particular technical obstacles. Also, each technical means shown in the drawings referred to for explaining this exemplary embodiment can be employed in other exemplary embodiments included in this disclosure as long as there are no particular technical obstacles.

[0015] (Configuration of Delay Prediction System 100) The configuration of the delay prediction system 100 will be described with reference to FIG. 1. FIG. 1 is a block diagram showing the configuration of the delay prediction system 100. As shown in FIG. 1, the delay prediction system 100 includes a state estimation unit 101, a state determination unit 102, and a delay estimation unit 103.

[0016] The state estimation unit 101, the state determination unit 102, and the delay estimation unit 103 are configured to be communicable via a communication network N as an example. Here, the specific configuration of the communication network N does not limit this exemplary embodiment, but as an example, a wireless LAN (Local Area Network), a wired LAN, a WAN (Wide Area Network), a public switched telephone network, a mobile data communication network, or a combination of these networks can be used.

[0017] Note that the state estimation unit 101, the state determination unit 102, and the delay estimation unit 103 may be implemented in one device or may be implemented in separate devices. Also, each unit may be distributed and arranged on the cloud (that is, on the communication network N). For example, when implemented in the cloud or in separate devices, information of each unit is transmitted and received via the communication network N and the processing proceeds.

[0018] The state estimation unit 101 estimates the state of the network based on a comparison result between a first threshold value specified in smoothing performed on the difference between the sRTT value of the first received packet and the sRTT value of the second received packet and a value obtained by the smoothing. Here, the sRTT value is a value obtained by smoothing the RTT value and represents the round-trip time of communication related to the corresponding received packet. For example, assuming that the RTT value of the received packet at time t is RTT_t and the sRTT value of the received packet at time (t - 1) is sRTT_(t - 1), the sRTT value sRTT_t of the received packet at time t can be expressed by the following formula (Formula 1). The recommended value of α is 0.9, but it is not limited to this.

[0019] sRTT_t = α × sRTT_(t - 1)+(1 - α)×RTT_t ···(Formula 1) More specifically, the state estimation unit 101 acquires the sRTT values of the first received packet and the second received packet received via the network N, and calculates the difference between these sRTT values. Then, the state estimation unit 101 performs smoothing on the difference of the sRTT values using a plurality of stages of Kalman filters connected in series to each other.

[0020] The state estimation unit 101 further specifies a threshold value for the output of the last-stage Kalman filter based on the output of the first-stage Kalman filter among the plurality of stages of Kalman filters. Then, the state estimation unit 101 compares the output of the last-stage Kalman filter with the specified threshold value, and estimates the state of the network N based on the comparison result. Note that the state estimation unit 101 may estimate the state of the network N using the RTT instead of the sRTT.

[0021] Hereinafter, the state of the network N estimated by the state estimation unit 101 will be referred to as State. State takes any one of the values of "0", "1", and "-1". As will be described later, State "0" indicates a state where the sRTT value of the received packet is stable. State "1" indicates a state where the sRTT value of the received packet is increasing. Also, State "-1" indicates a state where the sRTT value of the received packet is decreasing.

[0022] The state determination unit 102 determines whether the state of the network is a stable state or a fluctuating state based on the cumulative value of the estimated network state, the difference between the predicted value of the round-trip time of communication obtained from the probability distribution of a plurality of sRTT values at a predetermined time, and the latest sRTT value. The stable state means that neither the State "1" state nor the State "-1" state has occurred continuously, indicating that the state of the network N is stable. Also, the fluctuating state means that the State "1" state or the State "-1" state has occurred continuously, indicating that the state of the network N is unstable.

[0023] For example, the state determination unit 102 accumulates the number of times State becomes "1" as a cumulative value, and creates a probability distribution of the sRTT of received packets at a predetermined time x (seconds). In this state, since the sRTT value of the received packets is increasing, the delay estimation unit 103 calculates a deterioration prediction value as the predicted value of the delay in the probability distribution within the past x - second range.

[0024] If the delay value corresponding to the z% probability value in the probability distribution is defined as the deterioration prediction value (z%), the deterioration prediction value (z%) can be calculated using a normal probability distribution. Note that z% is a parameter and is a predetermined value.

[0025] Also, the state determination unit 102 accumulates the number of times State becomes "-1" as a cumulative value, and creates a probability distribution of the sRTT of received packets at a predetermined time x (seconds). In this state, since the sRTT value of the received packets is decreasing, the delay estimation unit 103 calculates an improvement prediction value as the predicted value of the round - trip time of communication in the probability distribution within the past x - second range.

[0026] If the delay value corresponding to the y% probability value in the probability distribution is defined as the improvement prediction value (y%), the improvement prediction value (y%) can be calculated using a normal probability distribution. Note that y% is a parameter and is a predetermined value.

[0027] The state determination unit 102 determines whether the state of the network N is a stable state or a fluctuating state based on the cumulative value (A) of State "1" or State "-1", the difference (B) between the deterioration prediction value (z%) or the improvement prediction value (y%) and the current sRTT value. For example, the state determination unit 102 calculates the weighted addition of the cumulative value (A) and the difference (B) as a score value. If the score value does not exceed the threshold, it determines that the state of the network N is a stable state; if the score value exceeds the threshold, it determines that the state of the network N is a fluctuating state.

[0028] The cumulative value (A) of State is used to reflect in the score value how many times the same State has been detected. This is because as the number of times of State increases, the possibility of being in a fluctuating state becomes higher. Also, the difference (B) is used to reflect in the score value how far the current sRTT value is from the deterioration prediction value or the improvement prediction value. This is because as the difference (B) increases, the possibility of being in a fluctuating state becomes higher.

[0029] When the state of the network is in a stable state, the delay estimation unit 103 uses the average prediction value of a plurality of sRTT values at a predetermined time and the predicted value of the round-trip time of communication obtained from the probability distribution of the plurality of sRTT values at the predetermined time as the delay prediction value. When the state of the network is in a fluctuating state, the latest sRTT value is used as the delay prediction value.

[0030] For example, when State is "1" and the state of network N is in a stable state, the delay estimation unit 103 uses the deterioration prediction value (z%) as the delay prediction value. When State is "-1" and the state of network N is in a stable state, the delay estimation unit 103 uses the improvement prediction value (y%) as the delay prediction value.

[0031] Also, when State is "0" and the state of network N is in a stable state, the delay estimation unit 103 uses the average prediction value of a plurality of sRTT values at a predetermined time as the delay prediction value. For example, the average prediction value is the weighted moving average value of a plurality of sRTT values at a predetermined time. Note that the average prediction value is not limited to the weighted moving average value, and may be a value calculated by other methods such as a simple average value.

[0032] When the state of network N is in a fluctuating state, the delay estimation unit 103 uses the latest sRTT value as the delay prediction value.

[0033] (Effect of the delay prediction system 100) As described above, in the delay prediction system 100, in the stable state, either the average predicted value of a plurality of sRTT values at a predetermined time or the predicted value of the round-trip time of communication obtained from the probability distribution of the plurality of sRTT values at a predetermined time is adopted as the delay predicted value. Therefore, according to the delay prediction system 100, the effect of being able to estimate the predicted value of the round-trip time of communication in the network N can be obtained.

[0034] (Flow of the delay prediction method) The flow of the delay prediction method S1 will be described with reference to FIG. 2. FIG. 2 is a flowchart showing the flow of the delay prediction method S1. As shown in FIG. 2, the delay prediction method S1 includes processes S11 to S13.

[0035] First, the state estimation unit 101 estimates the state of the network based on the comparison result between the first threshold value specified in the smoothing performed on the difference between the sRTT value, which is the round-trip time of communication regarding the first received packet, and the sRTT value of the second received packet, and the value obtained by the smoothing (S11).

[0036] Next, the state determination unit 102 determines whether the state of the network is a stable state or a fluctuating state based on the cumulative value of the state of the network estimated by the state estimation unit 101, the difference between the predicted value of the round-trip time of communication obtained from the probability distribution of a plurality of sRTT values at a predetermined time, and the latest sRTT value (S12).

[0037] The state determination unit 102 determines whether the state of the network N is a stable state or a fluctuating state based on the cumulative value (A) of State "1" or State "-1", and the difference (B) between the predicted value of deterioration (z%) or the predicted value of improvement (y%) and the current sRTT value. For example, the state determination unit 102 calculates the weighted addition of the cumulative value (A) and the difference (B) as a score value, and determines that the state of the network N is a stable state when the score value does not exceed the threshold value, and determines that the state of the network N is a fluctuating state when the score value exceeds the threshold value.

[0038] Finally, when the network state is in a stable state, the delay estimation unit 103 uses, as the delay prediction value, the average predicted value of a plurality of sRTT values at a predetermined time and the predicted value of the round-trip time of communication obtained from the probability distribution of the plurality of sRTT values at the predetermined time. When the network state is in a fluctuating state, the delay estimation unit 103 uses the latest sRTT value as the delay prediction value (S13).

[0039] For example, when State is "1" and the state of network N is in a stable state, the delay estimation unit 103 uses the deterioration prediction value (z%) as the delay prediction value. When State is "-1" and the state of network N is in a stable state, the delay estimation unit 103 uses the improvement prediction value (y%) as the delay prediction value.

[0040] Also, when State is "0" and the state of network N is in a stable state, the delay estimation unit 103 uses the average predicted value of a plurality of sRTT values at a predetermined time as the delay prediction value. For example, the average predicted value is the weighted moving average value of a plurality of sRTT values at a predetermined time.

[0041] When the state of network N is in a fluctuating state, the delay estimation unit 103 uses the latest sRTT value as the delay prediction value.

[0042] (Effect of the delay prediction method) As described above, in the delay prediction method S1, when in the stable state, a configuration is adopted in which either the average predicted value of a plurality of sRTT values at a predetermined time or the predicted value of the round-trip time of communication obtained from the probability distribution of the plurality of sRTT values at the predetermined time is used as the delay prediction value. Therefore, according to the delay prediction method S1, an effect can be obtained in that it becomes possible to estimate the predicted value of the round-trip time of communication of network N.

[0043] (Configuration of the delay prediction device 1) The configuration of the delay prediction device 1 will be described with reference to FIG. 3. FIG. 3 is a block diagram showing the configuration of the delay prediction device 1. As shown in FIG. 3, the delay prediction device 1 includes a state estimation unit 11, a state determination unit 12, and a delay estimation unit 13.

[0044] Based on the comparison result between the first threshold value specified in the smoothing performed on the difference between the sRTT value, which is the round-trip time of communication for the first received packet, and the sRTT value of the second received packet, and the value obtained by this smoothing, the state estimation unit 11 estimates the state of the network.

[0045] More specifically, the state estimation unit 11 acquires the sRTT value of the first received packet received via the network and the sRTT value of the second received packet, and calculates the difference between these sRTT values. Then, the state estimation unit 11 performs smoothing on the difference of the sRRT values using a plurality of stages of Kalman filters connected in series to each other.

[0046] The state estimation unit 11 further specifies a threshold value for the output of the last-stage Kalman filter based on the output of the first-stage Kalman filter among the plurality of stages of Kalman filters. Then, the state estimation unit 11 compares the output of the last-stage Kalman filter with the specified threshold value, and estimates the state of the network based on the comparison result. Note that the state estimation unit 11 may estimate the state of the network using RTT instead of sRTT.

[0047] Based on the cumulative value of the network state estimated by the state estimation unit 11, the difference between the predicted value of the communication round-trip time obtained from the probability distribution of a plurality of sRTT values at a predetermined time, and the latest sRTT value, the state determination unit 12 determines whether the state of the network is a stable state or a fluctuating state.

[0048] For example, the state determination unit 12 uses the number of times the State becomes "1" as the cumulative value, and creates a probability distribution of the sRTT of the received packet at a predetermined time x (seconds). In this state, since the sRTT value of the received packet is increasing, the delay estimation unit 13 calculates the deterioration prediction value (z%) as the predicted value of the delay in the probability distribution in the past x-second range.

[0049] Further, the state determination unit 12 accumulates the number of times the State becomes "-1" as a cumulative value, and creates a probability distribution of the sRTT of the received packets at a predetermined time x (seconds). In this state, since the sRTT value of the received packets is decreasing, the delay estimation unit 13 calculates a prediction value of improvement (y%) as the predicted value of the delay in the probability distribution in the past x-second range.

[0050] Based on the cumulative value (A) of State "1" or State "-1", the difference (B) between the predicted value of deterioration (z%) or the predicted value of improvement (y%) and the current sRTT value, the state determination unit 12 determines whether the state of the network is a stable state or a fluctuating state. For example, the state determination unit 12 calculates the weighted sum of the cumulative value (A) and the difference (B) as a score value. If the score value does not exceed the threshold, it is determined that the state of the network is a stable state. If the score value exceeds the threshold, it is determined that the state of the network is a fluctuating state.

[0051] When the state of the network is a stable state, the delay estimation unit 13 uses the average predicted value of a plurality of sRTT values at a predetermined time and the predicted value of the round-trip time of communication obtained from the probability distribution of a plurality of sRTT values at a predetermined time as the delay predicted value. When the state of the network is a fluctuating state, the latest sRTT value is used as the delay predicted value.

[0052] For example, when State is "1" and the state of the network is a stable state, the delay estimation unit 13 uses the predicted value of deterioration (z%) as the delay predicted value. When State is "-1" and the state of the network is a stable state, the delay estimation unit 13 uses the predicted value of improvement (y%) as the delay predicted value.

[0053] When State is "0" and the state of the network is a stable state, the delay estimation unit 13 uses the average predicted value of a plurality of sRTT values at a predetermined time as the delay predicted value. For example, the average predicted value is a weighted moving average value of a plurality of sRTT values at a predetermined time. Note that the average predicted value is not limited to the weighted moving average value, and may be a value calculated by other methods such as a simple average value.

[0054] When the state of the network is in a fluctuating state, the delay estimation unit 13 uses the current sRTT value as the delay prediction value.

[0055] (Effect of the delay prediction device 1) As described above, in the delay prediction device 1, in the stable state, a configuration is adopted in which either the average prediction value of a plurality of sRTT values in a predetermined time or the prediction value of the round-trip time of communication obtained from the probability distribution of a plurality of sRTT values in a predetermined time is used as the delay prediction value. For this reason, according to the delay prediction device 1, an effect that it becomes possible to estimate the prediction value of the round-trip time of network communication is obtained.

[0056] [Second exemplary embodiment] A second exemplary embodiment, which is an example of an embodiment of the present invention, will be described in detail with reference to the drawings. Components having the same functions as those described in the above-described exemplary embodiment are denoted by the same reference numerals, and the description thereof will be omitted as appropriate. Note that the scope of application of each technical means employed in this exemplary embodiment is not limited to this exemplary embodiment. That is, each technical means employed in this exemplary embodiment can be employed in other exemplary embodiments included in the present disclosure as long as there is no particular technical problem. In addition, each technical means shown in each drawing referred to for explaining this exemplary embodiment can be employed in other exemplary embodiments included in the present disclosure as long as there is no particular technical problem.

[0057] (Configuration of the delay prediction system 100A) The configuration of the delay prediction system 100A will be described with reference to FIG. 4. FIG. 4 is a block diagram showing the configuration of the delay prediction system 100A. The delay prediction system 100A is different only in that the state determination unit 102 and the delay estimation unit 103 included in the delay prediction system 100 are replaced with a state determination unit 102A and a delay estimation unit 103A.

[0058] FIG. 5 is a diagram for explaining the state transition in the delay prediction system 100A. As the states of the delay prediction system 100A, there are two types: a stable state and a fluctuating state. The state estimation unit 101 estimates the state (State value) of the network N, and the state determination unit 102A determines whether the state of the network N is a stable state or a fluctuating state.

[0059] As shown in FIG. 5, in the stable state, the State value is "-1" or "1", and if the current sRTT value is outside the threshold range of the probability distribution of a plurality of sRTT values, the state determination unit 102A estimates that the state of the network N has transitioned to the fluctuating state.

[0060] Also, when in the fluctuating state, if the State value becomes "0", the state determination unit 102A estimates that the state of the network N has transitioned to the stable state.

[0061] FIG. 6 is a diagram showing the state value of the network N and whether the state of the network N is a stable state or a fluctuating state. In the upper diagram of FIG. 6, the horizontal axis is the number of received packets, and the vertical axis is the State value. Also, in the lower diagram of FIG. 6, the horizontal axis is the number of received packets, and the vertical axis is the sRTT value.

[0062] As shown in FIG. 6, the stable state continues until the number of packets is around "201", but when the number of packets exceeds "201", the State value becomes "1", and the sRTT value of the received packets increases rapidly. The state determination unit 102A detects this state and estimates that the state has transitioned from the stable state to the fluctuating state.

[0063] After that, the State value becomes "0" and transitions to the stable state, and the stable state continues until the number of packets is around "501", but when the number of packets exceeds "501", the State value becomes "-1", and the sRTT value of the received packets decreases rapidly. The state determination unit 102A detects this state and estimates that the state has transitioned from the stable state to the fluctuating state.

[0064] FIG. 7 is a diagram for explaining the average predicted value, the improved predicted value, and the deteriorated predicted value. The horizontal axis represents the number of packets of received packets, and the vertical axis represents the sRTT value.

[0065] When the state estimation unit 101 estimates that the state of the network N is during an increase in the communication round-trip time (State "1"), and the state determination unit 102A determines that the state of the network N is in a stable state, the delay estimation unit 103A calculates a deteriorated predicted value, which is a value indicating the deterioration of the network state, from the probability distributions of a plurality of sRTT values at a predetermined time, and uses it as the delay predicted value.

[0066] As shown in FIG. 7, since the stable state continues until the number of packets is around "201", the state determination unit 102A creates a probability distribution of the sRTT of the received packets at a predetermined time x (seconds). If the delay value corresponding to the probability value of z% in the probability distribution is defined as the deteriorated predicted value (z%), the deteriorated predicted value (z%) can be calculated using a normal probability distribution. Note that z% is a parameter and is a predetermined value.

[0067] Also, when the state estimation unit 101 estimates that the state of the network N is during a decrease in the communication round-trip time (State "-1"), and the state determination unit 102A determines that the state of the network N is in a stable state, the delay estimation unit 103A calculates an improved predicted value, which is a value indicating the improvement of the network N state, from the probability distributions of a plurality of sRTT values at a predetermined time, and uses it as the delay predicted value.

[0068] If the delay value corresponding to the probability value of y% in the probability distribution is defined as the improved predicted value (y%), the improved predicted value (y%) can be calculated using a normal probability distribution. Note that y% is a parameter and is a predetermined value.

[0069] Further, when the state estimation unit 101 estimates that the state of the network N is during the stabilization of the communication round-trip time (State "0"), and the state determination unit 102A determines that the state of the network N is a stable state, the delay estimation unit 103A uses the weighted moving average value of a plurality of sRTTs at a predetermined time as the delay prediction value.

[0070] The state determination unit 102A calculates a weighted addition using the cumulative value of the number of times the state estimation unit 101 estimates that the state of the network N is during the increase of the communication round-trip time (State "1"), the difference between the deterioration prediction value obtained from the probability distribution of a plurality of sRTT values at a predetermined time and the current sRTT value, and determines whether the state of the network N is a stable state or a fluctuating state by comparing the result of the weighted addition with a second threshold value.

[0071] As shown in FIG. 7, the stable state continues until the number of packets is near "201", but when the number of packets exceeds "201", the State value becomes "1", and the sRTT value of the received packet increases rapidly. The state determination unit 102A calculates a score value from the cumulative value (A) of the number of times estimated as State "1", the difference (B) between the deterioration prediction value obtained from the probability distribution of a plurality of sRTT values at a predetermined time (x seconds) and the current sRTT value, according to the following formula (Formula 2).

[0072] Score value = (A) × m + (B) × n ··· (Formula 2) Here, m is the weight coefficient of the cumulative value (A), and n is the weight coefficient of the difference (B). m and n are used to make the cumulative value (A) and the difference (B) comparable by matching their dimensions, and to weight which of the cumulative value (A) and the difference (B) is prioritized.

[0073] The state determination unit 102A compares the score value with a predetermined threshold value, and if the score value does not exceed the threshold value, determines that the state of the network N is a stable state. Also, if the score value exceeds the threshold value, the state determination unit 102A determines that the state of the network N is a fluctuating state.

[0074] In FIG. 7, when the number of packets exceeds "201", the State value becomes "1", and the sRTT value of the received packets increases rapidly. The state determination unit 102A detects that the score value exceeds the threshold value and estimates that the state has transitioned from a stable state to a fluctuating state.

[0075] Further, the state determination unit 102A calculates a weighted sum using the cumulative value of the number of times the state of the network N is estimated to be in the decreasing round-trip time of communication (State "-1") by the state estimation unit 101, the difference between the improvement prediction value obtained from the probability distribution of a plurality of sRTT values at a predetermined time and the current sRTT value, and determines whether the state of the network N is a stable state or a fluctuating state by comparing the result of the weighted sum with a second threshold value.

[0076] As shown in FIG. 7, the stable state continues until the number of packets is near "501", but when the number of packets exceeds "501", the State value becomes "-1", and the sRTT value of the received packets decreases rapidly. The state determination unit 102A calculates the score value according to the above (Equation 2) from the cumulative value (A) of the number of times estimated as State "-1" and the difference (B) between the improvement prediction value obtained from the probability distribution of a plurality of sRTT values at a predetermined time (x seconds) and the current sRTT value.

[0077] The state determination unit 102A compares the score value with a predetermined threshold value. If the score value does not exceed the threshold value, it determines that the state of the network N is a stable state. Also, if the score value exceeds the threshold value, the state determination unit 102A determines that the state of the network N is a fluctuating state.

[0078] In FIG. 7, when the number of packets exceeds "501", the State value becomes "-1", and the sRTT value of the received packets decreases rapidly. The state determination unit 102A detects that the score value exceeds the threshold value and estimates that the state has transitioned from a stable state to a fluctuating state.

[0079] FIG. 8 is a diagram for explaining the process when the state of the network N is a fluctuating state. The horizontal axis represents the number of packets of received packets, and the vertical axis represents the sRTT value. In FIG. 8, the thick curve with a dark color is the deterioration prediction value, and the thin curve with a light color is the weighted moving average value of sRTT. Also, the thick straight line extending in the horizontal axis direction indicates the threshold value.

[0080] As shown in FIG. 8, the delay prediction system 100A continuously transmits measurement packets P(t1), P(t2), P(t3) ···, and continuously calculates the weighted moving average value and the probability distribution from the respective sRTT values.

[0081] As described above, the state determination unit 102A determines whether the state of the network N is a stable state or a fluctuating state based on the cumulative value of State and the deterioration prediction value (z%). When the score value exceeds the threshold value, it is determined that the state of the network N has transitioned to a fluctuating state. When the state determination unit 102A determines that the state of the network N is a fluctuating state, the delay estimation unit 103A initializes the weighted moving average value. In FIG. 8, it is assumed that z% is 99.9% and y% is 0.1%.

[0082] After the weighted moving average value is initialized, the delay prediction system 100A continuously transmits measurement packets P(t1’), P(t2’), P(t3’) ···, and continuously calculates the weighted moving average value and the probability distribution from the respective sRTT values.

[0083] FIG. 9 is a flowchart showing the flow of the delay prediction method S2 according to the present disclosure. First, the state estimation unit 101 inputs the sRTT value of the received packet (S21), and calculates the State value that is the state of the network N (S22).

[0084] Next, the state determination unit 102A determines whether the State value that is the state of the network N is "0" (S23). If the State value is "0" (S23, Yes), the process proceeds to step S28.

[0085] Also, if the State value is not "0" (S23, No), the state determination unit 102A determines whether the previous state was a stable state (S24). If the previous state was not a stable state (S24, No), that is, if the previous state was a fluctuating state, the process proceeds to step S31.

[0086] On the other hand, if the previous state was a stable state (S24, Yes), the state determination unit 102A calculates the cumulative value (A) of State (S25). Then, the state determination unit 102A calculates how much the deterioration prediction value or improvement prediction value obtained from the probability distribution of sRTT deviates from the current sRTT value (difference (B)) (S26). At this time, if the State value is "1", the state determination unit 102A calculates the difference (B) using the deterioration prediction value. If the State value is "-1", the state determination unit 102A calculates the difference (B) using the improvement prediction value.

[0087] Next, the state determination unit 102A calculates a score value from the cumulative value (A) and the difference (B) using the above (Equation 2), and determines whether the score value exceeds a threshold (S27). If the score value does not exceed the threshold (S27, No), the process proceeds to step S28. If the score value exceeds the threshold (S27, Yes), the process proceeds to step S31.

[0088] In step S28, the state determination unit 102A updates the state of the network N to a stable state and adds it to an sRTT value buffer (not shown) (S29). The sRTT values of the packets are stored in this sRTT value buffer, and the information stored in this sRTT value buffer is referred to when calculating the probability distribution of sRTT and the weighted moving average value of sRTT.

[0089] The delay estimation unit 103A refers to the sRTT values stored in the sRTT buffer, calculates the average prediction value (weighted moving average value) in the past x - second range, the improvement prediction value (y%) in the probability distribution, and the deterioration prediction value (z%), and sets one of the three values as the delay prediction value according to the State value (S30), and ends the process.

[0090] Also, in step S31, the state determination unit 102A updates the state of the network N to a variable state and clears an sRTT value buffer (not shown) (S32). Then, the delay estimation unit 103A sets the current sRTT value (sRTT measured value) as the delay prediction value (S33) and ends the process.

[0091] (Effect of the delay prediction system 100A) As described above, in the delay prediction system 100A, when the state estimation unit 101 estimates that the state of the network N is during an increase in the round-trip time of communication and the state determination unit 102A determines that the state of the network N is a stable state, the delay estimation unit 103A calculates a deterioration prediction value, which is a value indicating a deterioration of the network state, from the probability distribution of a plurality of sRTT values at a predetermined time and uses it as the delay prediction value. For this reason, according to the delay prediction system 100A, in addition to the effect exhibited by the delay prediction system 100, an effect can be obtained in that the deterioration prediction value when the state of the network N deteriorates can be estimated as the delay prediction value of the network N.

[0092] Also, in the delay prediction system 100A, the state determination unit 102A calculates a weighted sum using the cumulative value of the number of times the state estimation unit 101 estimates that the state of the network N is during an increase in the round-trip time of communication, the difference between the deterioration prediction value obtained from the probability distribution of a plurality of sRTT values at a predetermined time and the current sRTT value, and determines whether the state of the network N is a stable state or a variable state by comparing the result of the weighted sum with a second threshold value. For this reason, according to the delay prediction system 100A, in addition to the effect exhibited by the delay prediction system 100, an effect can be obtained in that it is possible to accurately determine whether the state of the network N is a stable state or a variable state when the state of the network N deteriorates.

[0093] In the delay prediction system 100A, when the state estimation unit 101 estimates that the state of the network N is during a decrease in the round-trip time of communication, and the state determination unit 102A determines that the state of the network N is in a stable state, the delay estimation unit 103A calculates a prediction value of improvement, which is a value indicating the improvement of the state of the network N, from the probability distribution of a plurality of sRTTs at a predetermined time and uses it as the delay prediction value. Therefore, according to the delay prediction system 100A, in addition to the effect exhibited by the delay prediction system 100, an effect can be obtained in that the prediction value of improvement when the state of the network N improves can be estimated as the delay prediction value of the network N.

[0094] Also, in the delay prediction system 100A, the state determination unit 102A calculates a weighted sum using the cumulative value of the number of times the state estimation unit 101 estimates that the state of the network N is during a decrease in the round-trip time of communication, the difference between the prediction value of improvement obtained from the probability distribution of a plurality of sRTT values at a predetermined time and the current sRTT value, and determines whether the state of the network N is in a stable state or a fluctuating state by comparing the result of the weighted sum with a second threshold value. Therefore, according to the delay prediction system 100A, in addition to the effect exhibited by the delay prediction system 100, an effect can be obtained in that it is possible to accurately determine whether the state of the network N is in a stable state or a fluctuating state when the state of the network N improves.

[0095] In the delay prediction system 100A, when the state estimation unit 101 estimates that the state of the network N is stable during the round-trip time of communication, and the state determination unit 102A determines that the state of the network N is in a stable state, the delay estimation unit 103A uses the weighted moving average value of a plurality of sRTT values at a predetermined time as the delay prediction value. Therefore, according to the delay prediction system 100A, in addition to the effect exhibited by the delay prediction system 100, an effect can be obtained in that the weighted moving average value when the state of the network N is stable can be estimated as the delay prediction value of the network N.

[0096] Also, in the delay prediction system 100A, when the state determination unit 102A determines that the state of the network N is a fluctuating state, the delay estimation unit 103A adopts a configuration of initializing the weighted moving average value. Therefore, according to the delay prediction system 100A, in addition to the effects exhibited by the delay prediction system 100, an effect that the weighted moving average value can be calculated more accurately is obtained.

[0097] 〔Example of Realization by Software〕 Some or all of the functions of the delay prediction systems 100 and 100A (hereinafter, also referred to as "each of the above systems") may be realized by hardware such as an integrated circuit (IC chip), or may be realized by software.

[0098] In the latter case, each of the above systems is realized by a computer that executes instructions of a program, which is software for realizing each function. An example of such a computer (hereinafter, referred to as computer C) is shown in FIG. 10. FIG. 10 is a block diagram showing the hardware configuration of the computer C that functions as each of the above systems.

[0099] The computer C includes at least one processor C1 and at least one memory C2. A program P for operating the computer C as each of the above systems is recorded in the memory C2. In the computer C, the processor C1 reads and executes the program P from the memory C2, whereby each function of each of the above systems is realized.

[0100] As the processor C1, for example, a CPU (Central Processing Unit), GPU (Graphic Processing Unit), DSP (Digital Signal Processor), MPU (Micro Processing Unit), FPU (Floating point number Processing Unit), PPU (Physics Processing Unit), TPU (Tensor Processing Unit), quantum processor, microcontroller, or a combination thereof can be used. As the memory C2, for example, a flash memory, HDD (Hard Disk Drive), SSD (Solid State Drive), or a combination thereof can be used.

[0101] Note that the computer C may further include a RAM (Random Access Memory) for expanding the program P during execution and temporarily storing various data. Further, the computer C may further include a communication interface for transmitting and receiving data to and from other devices. Further, the computer C may further include an input / output interface for connecting input / output devices such as a keyboard, mouse, display, and printer.

[0102] Also, the program P can be recorded on a non-transitory tangible recording medium M readable by the computer C. As such a recording medium M, for example, a tape, disk, card, semiconductor memory, or programmable logic circuit can be used. The computer C can obtain the program P via such a recording medium M. Also, the program P can be transmitted via a transmission medium. As such a transmission medium, for example, a communication network or broadcast wave can be used. The computer C can also obtain the program P via such a transmission medium.

[0103] 〔Supplementary Note 1〕 The present disclosure includes the technologies described in the following appendices. However, the present invention is not limited to the technologies described in the following appendices, and various modifications are possible within the scope indicated in the claims.

[0104] (Appendix 1) State estimation means for estimating the state of a network based on a comparison result between a first threshold value specified in smoothing performed on the difference between an sRTT value, which is the round-trip time of communication regarding a first received packet, and an sRTT value of a second received packet, and a value obtained by the smoothing; State determination means for determining whether the state of the network is a stable state or a fluctuating state based on the cumulative value of the estimated state of the network and the difference between a predicted value of the round-trip time of communication obtained from the probability distribution of a plurality of sRTT values at a predetermined time and the latest sRTT value; Delay estimation means for setting, as a delay prediction value, an average predicted value of the plurality of sRTT values at the predetermined time and a predicted value of the round-trip time of communication obtained from the probability distribution of the plurality of sRTT values at the predetermined time when the state of the network is the stable state, and setting the latest sRTT value as the delay prediction value when the state of the network is the fluctuating state; A delay prediction system comprising the above.

[0105] (Appendix 2) When the state of the network is estimated by the state estimation means to be during an increase in the round-trip time of communication and the state of the network is determined by the state determination means to be a stable state, the delay estimation means calculates a deterioration prediction value, which is a value indicating deterioration of the state of the network, from the probability distribution of the plurality of sRTT values at the predetermined time and sets it as the delay prediction value. The delay prediction system according to Appendix 1.

[0106] (Appendix 3) The state determination means calculates a weighted sum using the cumulative value of the number of times the state estimation means estimates that the state of the network is during an increase in the round-trip time of communication, the difference between the deterioration prediction value obtained from the probability distribution of the plurality of sRTT values at the predetermined time and the current sRTT value, and determines whether the state of the network is a stable state or a fluctuating state by comparing the result of the weighted sum with a second threshold value. The delay prediction system according to Supplementary Note 2.

[0107] (Supplementary Note 4) When the state estimation means estimates that the state of the network is during a decrease in the round-trip time of communication and the state determination means determines that the state of the network is a stable state, the delay estimation means calculates a prediction value of improvement, which is a value indicating an improvement in the state of the network, from the probability distribution of the plurality of sRTT values at the predetermined time and uses it as the delay prediction value. The delay prediction system according to any one of Supplementary Notes 1 to 3.

[0108] (Supplementary Note 5) The state determination means calculates a weighted sum using the cumulative value of the number of times the state estimation means estimates that the state of the network is during a decrease in the round-trip time of communication, the difference between the prediction value of improvement obtained from the probability distribution of the plurality of sRTT values at the predetermined time and the current sRTT value, and determines whether the state of the network is a stable state or a fluctuating state by comparing the result of the weighted sum with a second threshold value. The delay prediction system according to Supplementary Note 4.

[0109] (Supplementary Note 6) When the state estimation means estimates that the state of the network is stable during the round-trip time of communication and the state determination means determines that the state of the network is a stable state, the delay estimation means uses the weighted moving average value of the plurality of sRTT values at the predetermined time as the delay prediction value. The delay prediction system according to any one of Supplementary Notes 1 to 5.

[0110] (Appendix 7) When it is determined by the state determination means that the state of the network is a fluctuating state, the delay estimation means initializes the weighted moving average value. The delay prediction system according to Appendix 6.

[0111] (Appendix 8) Based on the comparison result between the first threshold value specified in the smoothing performed on the difference between the sRTT value, which is the round-trip time of communication for the first received packet, and the sRTT value of the second received packet, and the value obtained by the smoothing, state estimation means for estimating the state of the network; Based on the cumulative value of the estimated network state, the difference between the predicted value of the round-trip time of communication obtained from the probability distribution of a plurality of sRTT values at a predetermined time and the latest sRTT value, state determination means for determining whether the network state is a stable state or a fluctuating state; When the network state is the stable state, the average predicted value of the plurality of sRTT values at the predetermined time and the predicted value of the round-trip time of communication obtained from the probability distribution of the plurality of sRTT values at the predetermined time are used as the delay prediction values, and when the network state is the fluctuating state, the latest sRTT value is used as the delay prediction value, delay estimation means; A delay prediction device comprising:

[0112] (Appendix 9) Based on the comparison result between the first threshold value specified in the smoothing performed on the difference between the sRTT value, which is the round-trip time of communication for the first received packet, and the sRTT value of the second received packet, and the value obtained by the smoothing, estimating the state of the network; Based on the cumulative value of the estimated network state, the difference between the predicted value of the round-trip time of communication obtained from the probability distribution of a plurality of sRTT values at a predetermined time and the latest sRTT value, determining whether the network state is a stable state or a fluctuating state; When the state of the network is the stable state, the average predicted value of the plurality of sRTT values at the predetermined time and the predicted value of the round-trip time of communication obtained from the probability distribution of the plurality of sRTT values at the predetermined time are used as the delay prediction value. When the state of the network is the fluctuating state, the latest sRTT value is used as the delay prediction value. A delay prediction method including the above.

[0113] (Appendix 10) A delay prediction program for operating a computer as the delay prediction device described in Appendix 8, the delay prediction program causing the computer to function as each of the above means.

[0114] [Appendix Item 2] The present disclosure includes the technologies described in the following appendices. However, the present invention is not limited to the technologies described in the following appendices, and various modifications can be made within the scope shown in the claims.

[0115] (Appendix 1) Comprising at least one processor, the at least one processor Based on the comparison result between the first threshold value specified in the smoothing performed on the difference between the sRTT value, which is the round-trip time of communication for the first received packet, and the sRTT value of the second received packet, and the value obtained by the smoothing, a process of estimating the state of the network. Based on the cumulative value of the estimated network state, the difference between the predicted value of the round-trip time of communication obtained from the probability distribution of a plurality of sRTT values at a predetermined time and the latest sRTT value, a process of determining whether the network state is a stable state or a fluctuating state. When the state of the network is the stable state, the average predicted value of the plurality of sRTT values at the predetermined time and the predicted value of the round-trip time of communication obtained from the probability distribution of the plurality of sRTT values at the predetermined time are used as the delay prediction value. When the state of the network is the fluctuating state, the latest sRTT value is used as the delay prediction value, and execute the process. A delay prediction system.

[0116] Note that the delay prediction system may further include a memory. Further, a program for causing the at least one processor to execute each of the above processes may be stored in the memory.

[0117] (Appendix 2) In the process of estimating the state, when the state of the network is estimated to be during an increase in the round-trip time of communication, and in the process of determining the state, when the state of the network is determined to be in a stable state, the at least one processor calculates a deterioration prediction value, which is a value indicating deterioration of the network state, from the probability distribution of the plurality of sRTT values at the predetermined time, and uses the calculated value as the delay prediction value. The delay prediction system according to Appendix 1 (Appendix 3) In the process of determining the state, the at least one processor calculates a weighted sum using the cumulative value of the number of times the state of the network is estimated to be during an increase in the round-trip time of communication, the difference between the deterioration prediction value obtained from the probability distribution of the plurality of sRTT values at the predetermined time and the current sRTT value, and determines whether the state of the network is in a stable state or a fluctuating state by comparing the result of the weighted sum with a second threshold value. The delay prediction system according to Appendix 2.

[0118] (Appendix 4) In the process of estimating the state, when the state of the network is estimated to be during a decrease in the round-trip time of communication, and in the process of determining the state, when the state of the network is determined to be in a stable state, in the process of estimating the delay, the at least one processor calculates an improvement prediction value, which is a delay value indicating improvement of the network state, from the probability distribution of the plurality of sRTTs at the predetermined time, and uses the calculated value as the delay prediction value. The delay prediction system according to Appendix 1.

[0119] (Appendix 5) In the process of determining the state, the at least one processor calculates a weighted sum using the cumulative value of the number of times the state of the network is estimated to be during a decrease in the round-trip time of communication, and the difference between the predicted improvement value obtained from the probability distribution of the plurality of sRTT values at the predetermined time and the current sRTT value, and determines whether the state of the network is a stable state or a fluctuating state by comparing the result of the weighted sum with a second threshold value. The delay prediction system according to Supplementary Note 4.

[0120] (Supplementary Note 6) In the process of estimating the state, if the state of the network is estimated to be stable during the round-trip time of communication, and in the process of determining the state, if the state of the network is determined to be a stable state, in the process of estimating the delay, the at least one processor uses the weighted moving average value of the plurality of sRTTs at the predetermined time as the delay prediction value. The delay prediction system according to any one of Supplementary Notes 1 to 5.

[0121] (Supplementary Note 7) In the process of determining the state, if the state of the network is determined to be a fluctuating state, in the process of estimating the delay, the at least one processor initializes the weighted moving average value. The delay prediction system according to Supplementary Note 6.

Explanation of Signs

[0122] 1 Delay prediction device 11, 101 State estimation unit 12, 102, 102A State determination unit 13, 103, 103A Delay estimation unit 100, 100A Delay prediction system N Network

Claims

1. State estimation means for estimating the state of a network based on a comparison result between a first threshold value specified in smoothing performed on the difference between an sRTT value representing the round-trip time of communication related to a first received packet and an sRTT value representing the round-trip time of communication related to a second received packet, and a value obtained by the smoothing; State determination means for determining whether the state of the network is a stable state or a fluctuating state based on the cumulative value of the estimated state of the network and the difference between a predicted value of the round-trip time of communication obtained from the probability distribution of a plurality of sRTT values at a predetermined time and the latest sRTT value; Delay estimation means for setting, as a delay prediction value, an average prediction value of the plurality of sRTT values at the predetermined time and a predicted value of the round-trip time of communication obtained from the probability distribution of the plurality of sRTT values at the predetermined time when the state of the network is the stable state, and setting the latest sRTT value as the delay prediction value when the state of the network is the fluctuating state; A delay prediction system comprising the above.

2. When the state of the network is estimated by the state estimation means to be during an increase in the round-trip time of communication and the state of the network is determined by the state determination means to be a stable state, the delay estimation means calculates, as the delay prediction value, a deterioration prediction value, which is a value indicating deterioration of the state of the network, from the probability distribution of the plurality of sRTT values at the predetermined time. The delay prediction system according to Claim 1.

3. The state determination means calculates a weighted sum using the cumulative value of the number of times the state of the network is estimated by the state estimation means to be during an increase in the round-trip time of communication, and the difference between the deterioration prediction value obtained from the probability distribution of the plurality of sRTT values at the predetermined time and the current sRTT value, and determines whether the state of the network is a stable state or a fluctuating state by comparing the result of the weighted sum with a second threshold value. The delay prediction system according to Claim 2.

4. When the state of the network is estimated by the state estimation means to be during a decrease in the round-trip time of communication and the state of the network is determined by the state determination means to be a stable state, the delay estimation means calculates, as the delay prediction value, an improvement prediction value, which is a value indicating improvement of the state of the network, from the probability distribution of the plurality of sRTT values at the predetermined time. The delay prediction system according to claim 1.

5. The state determination means calculates a weighted sum using the cumulative value of the number of times the state estimation means estimates that the state of the network is during a decrease in the round-trip time of communication, the difference between the predicted improvement value obtained from the probability distribution of the plurality of sRTT values at the predetermined time and the current sRTT value, and determines whether the state of the network is a stable state or a fluctuating state by comparing the result of the weighted sum with a second threshold value. The delay prediction system according to claim 4.

6. When the state estimation means estimates that the state of the network is stable during the round-trip time of communication and the state determination means determines that the state of the network is a stable state, the delay estimation means uses the weighted moving average value of the plurality of sRTT values at the predetermined time as the delay prediction value. The delay prediction system according to any one of claims 1 to 5.

7. When the state determination means determines that the state of the network is a fluctuating state, the delay estimation means initializes the weighted moving average value. The delay prediction system according to claim 6.

8. A state estimation means for estimating the state of the network based on the comparison result between a first threshold value specified in smoothing performed on the difference between an sRTT value representing the round-trip time of communication for a first received packet and an sRTT value representing the round-trip time of communication for a second received packet and the value obtained by the smoothing; A state determination means for determining whether the state of the network is a stable state or a fluctuating state based on the cumulative value of the estimated state of the network and the difference between the predicted value of the round-trip time of communication obtained from the probability distribution of the plurality of sRTT values at a predetermined time and the latest sRTT value; A delay estimation means for using the average predicted value of the plurality of sRTT values at the predetermined time and the predicted value of the round-trip time of communication obtained from the probability distribution of the plurality of sRTT values at the predetermined time as the delay prediction value when the state of the network is the stable state, and using the latest sRTT value as the delay prediction value when the state of the network is the fluctuating state; A delay prediction apparatus comprising the above.

9. Estimating the state of a network based on a comparison result between a first threshold value specified in smoothing performed on a difference between an sRTT value representing a round-trip time of communication related to a first received packet and an sRTT value representing a round-trip time of communication related to a second received packet and a value obtained by the smoothing; Determining whether the state of the network is a stable state or a fluctuating state based on a cumulative value of the estimated network state, a difference between a predicted value of a round-trip time of communication obtained from a probability distribution of a plurality of sRTT values at a predetermined time and a latest sRTT value; When the state of the network is the stable state, setting an average predicted value of the plurality of sRTT values at the predetermined time and a predicted value of a round-trip time of communication obtained from the probability distribution of the plurality of sRTT values at the predetermined time as delay predicted values, and when the state of the network is the fluctuating state, setting the latest sRTT value as the delay predicted value; A delay prediction method including the above.

Images