Transmission system, transmission method and transmission program
A transmission system includes a calculation unit that calculates a jitter pattern sequence of a plurality of packets transmitted by a transmission device via a network using a transmission interval of the plurality of packets and a difference between arrival times at which the packets arrive at a reception device, a generation unit that inputs the jitter pattern sequence to a first learning model to generate a state index indicating a state of the network, and a buffer estimation unit that determines a buffer size of a reception buffer of the reception device based on the state index.
1 . A transmission system comprising:
a calculation unit, including one or more processors, configured to calculate a jitter pattern sequence of a plurality of packets transmitted by a transmission device via a network using a transmission interval of the plurality of packets and a difference between arrival times at which the packets arrive at a reception device;
a generation unit, including one or more processors, configured to input the jitter pattern sequence to a first learning model to generate a state index indicating a state of the network; and
a buffer estimation unit, including one or more processors, configured to input the state index to a second learning model to determine a buffer size of a reception buffer of the reception device based on the state index.
2 . The transmission system according to claim 1 , further comprising:
a control unit, including one or more processors, configured to update the buffer size of the reception buffer to the buffer size determined by the buffer estimation unit.
3 . The transmission system according to claim 1 , wherein
the calculation unit is configured to acquire a predetermined constant transmission interval or a plurality of transmission times when the transmission device transmits a plurality of the packets from the transmission device, and
in a case of acquiring the transmission time, the calculation unit is configured to calculate the jitter pattern sequence using a difference between the transmission times.
4 . The transmission system according to claim 1 , wherein the first learning model is a neural network generated by performing unsupervised learning using a slack time zone of the network as learning data, or by performing unsupervised learning in which labels are attached to the transmission interval collected from the transmission device and the arrival time collected from the reception device.
5 . A transmission method performed by a transmission system, the method comprising:
calculating a jitter pattern sequence of a plurality of packets transmitted by a transmission device via a network using a transmission interval of the plurality of packets and a difference between arrival times at which the packets arrive at a reception device,
inputting the jitter pattern sequence to a first learning model to generate a state index indicating a state of the network, and
acquiring an observation value output from a reception buffer, and determining a buffer size of the reception buffer of the reception device based on the state index by using a deviation between the observation value or an estimation value estimated from the observation value and the state index.
6 . The transmission method according to claim 5 , further comprising:
updating the buffer size of the reception buffer to the determined buffer size.
7 . The transmission method according to claim 5 , further comprising:
acquiring a predetermined constant transmission interval or a plurality of transmission times when the transmission device transmits a plurality of the packets from the transmission device, and
in a case of acquiring the transmission time, calculating the jitter pattern sequence using a difference between the transmission times.
8 . The transmission method according to claim 5 , wherein the first learning model is a neural network generated by performing unsupervised learning using a slack time zone of the network as learning data, or by performing unsupervised learning in which labels are attached to the transmission interval collected from the transmission device and the arrival time collected from the reception device.
9 . A non-transitory computer-readable storage medium storing a transmission program for causing a computer to perform operations comprising:
calculating a jitter pattern sequence of a plurality of packets transmitted by a transmission device via a network using a transmission interval of the plurality of packets and a difference between arrival times at which the packets arrive at a reception device,
inputting the jitter pattern sequence to a first learning model to generate a state index indicating a state of the network, wherein the first learning model is a neural network generated by performing unsupervised learning using a slack time zone of the network as learning data, or by performing unsupervised learning in which labels are attached to the transmission interval collected from the transmission device and the arrival time collected from the reception device, and
determining a buffer size of a reception buffer of the reception device based on the state index.
10 . The non-transitory computer-readable storage medium according to claim 9 , wherein the operations further comprise:
updating the buffer size of the reception buffer to the determined buffer size.
11 . The non-transitory computer-readable storage medium according to claim 9 , wherein the operations further comprise:
acquiring a predetermined constant transmission interval or a plurality of transmission times when the transmission device transmits a plurality of the packets from the transmission device, and
in a case of acquiring the transmission time, calculating the jitter pattern sequence using a difference between the transmission times.
12 . The non-transitory computer-readable storage medium according to claim 9 , wherein the operations further comprise:
inputting the state index to a second learning model to determine the buffer size of the reception buffer.
13 . The non-transitory computer-readable storage medium according to claim 9 , wherein the operations further comprise:
acquiring an observation value output from the reception buffer, and determining the buffer size of the reception buffer using a deviation between the observation value or an estimation value estimated from the observation value and the state index.