Communication control device and communication system

Abstract

The present invention achieves greater power savings in a radio access network. This communication control device comprises a demand prediction unit, a calculation unit, a correction unit, and a resource allocation unit. The demand prediction unit creates a first demand model representing a prediction of first demand corresponding to first traffic and a second demand model representing a prediction of second demand corresponding to second traffic. The calculation unit calculates an allowable delay on the basis of the quality required by the first traffic. The correction unit corrects the first demand model within a range in which the shift amount of the first demand model does not exceed the allowable delay, so that a non-allocated section in which neither the first traffic nor the second traffic is allocated resources increases. The resource allocation unit allocates a first resource to the first traffic on the basis of the corrected first demand model, and allocates a second resource to the second traffic on the basis of the second demand model.

Description

Communication control device and communication system

[0001] The present invention relates to a communication device, a communication control device, a communication system, and a communication control method for wireless communication.

[0002] Standardization organizations such as 3GPP (registered trademark) (3rd Generation Partnership Project) and O-RAN (Open RAN) alliances are currently discussing network energy saving (NES) for radio access networks (RANs). For example, in O-RAN, an O-RAN architecture based on 3GPP specifications and incorporating a RAN Intelligent Controller (RIC) is expected to achieve efficient power saving by collecting configuration information and statistical information about the RAN and user equipment (UE) via a standard interface and implementing control according to the RAN situation.

[0003] Carrier and Cell Switch Off / On (CSC) and Advanced Sleep Mode (ASM) are being discussed as use cases for power saving in radio access networks. Carrier and Cell Off / On control allows some cells (or carriers) to be shut down depending on the communication situation in a network where multiple cells (or carriers) cover the same service area. This achieves power saving. ASM control reduces power consumption by putting some of the functions of a radio device into a sleep state during unused periods (or discontinuous periods) of radio resources on a symbol / slot / frame basis in response to reduced traffic. The length of the discontinuous period is several microseconds for symbol units (micro sleep), several milliseconds for slot units (light sleep), and 100 milliseconds or more for frame units (deep sleep). Note that a technique for extending the sleep time of a radio device is described, for example, in Non-Patent Document 1. Related techniques are also described in Non-Patent Documents 2 and 3.

[0004] O-RAN.WG3.UCR-R003-v04.00.053GPP TS23.501 V18.4.0 (2023-12)3GPP TS28.541 V18.5.0 (2023-09)

[0005] In ASM control, power consumption is reduced by putting some of the functions of the wireless device into a sleep state during periods when wireless resources are not in use. Therefore, the longer the period when wireless resources are not in use, the further power consumption is reduced. However, conventional technologies have not been able to sufficiently extend the period when wireless resources are not in use. In other words, further power saving in wireless access networks is expected.

[0006] An object of one aspect of the present invention is to extend the unused period of radio resources in power saving of a radio access network.

[0007] A communication control device according to one aspect of the present invention comprises a demand prediction unit that creates a first demand model that represents a prediction of a first demand corresponding to a first traffic and a second demand model that represents a prediction of a second demand corresponding to a second traffic; a calculation unit that calculates an allowable delay for the first traffic based on the quality required by the first traffic; a correction unit that corrects the first demand model so that a non-allocated section in which no resources are allocated to either the first traffic or the second traffic becomes larger, within a range in which the shift amount of the first demand model does not exceed the allowable delay; a resource allocation unit that allocates a first resource to the first traffic based on the corrected first demand model and allocates a second resource to the second traffic based on the second demand model; and a resource control unit that notifies the first resource and the second resource to a communication device that processes the first traffic and the second traffic.

[0008] According to the above-described aspect, further power saving in the radio access network is achieved.

[0009] FIG. 1 is a diagram showing an example of a communication system according to an embodiment of the present invention. FIG. 1 is a diagram showing an example of ASM control. FIG. 2 is a diagram showing an example of functional deployment related to NES / ASM control in O-RAN architecture. FIG. 3 is a diagram showing an example of correction of a demand model. FIG. 4 is a diagram showing another example of correction of a demand model. FIG. 5 is a diagram showing an example of a method for solving the problem in the case shown in FIG. 5. FIG. 6 is a diagram showing an example of processing for shifting a communication section on a time axis. FIG. 7 is a diagram showing an example of the functional configuration of a communication system according to an embodiment of the present invention. FIG. 8 is a diagram showing an example of processing for allocating resources to traffic based on a demand model. A flowchart showing an example of a communication control method according to an embodiment of the present invention. FIG. 9 is a diagram showing the configuration of a communication system according to an example of the present invention. FIG. 10 is a diagram showing an example of traffic demand predicted in advance. FIG. 11 is a diagram showing an example of traffic load prediction. FIG. 12 is a diagram explaining RRM Policy Ratio. FIG. 13 is a diagram showing an example of a sequence of communication control according to an embodiment of the present invention. FIG. 14 is a diagram showing an example of the hardware configuration of a real-time RIC operating as a communication control device.

[0010] 1 shows an example of a communication system according to an embodiment of the present invention. In this example, the communication system 1 includes a communication control device 2, a communication device 3, and a radio device 4 to configure a radio access network. The communication control device 2, the communication device 3, and the radio device 4 configure a base station system. The base station system is not particularly limited, but may be, for example, a gNB (next generation NodeB) or NR (New Radio) base station.

[0011] The communication control device 2 controls communication by the communication device 3 and the radio device 4 by executing one or more application programs. The communication device 3 transmits data to a terminal UE (User Equipment) using the radio device 4, and receives data from the terminal UE using the radio device 4. The communication device 3 may also measure parameters (e.g., delay time) related to communication in the radio access network and notify the communication control device 2. In this case, the communication control device 2 may control communication in the radio access network based on the measurement by the communication device 3. The radio device 4 includes a radio circuit and an antenna, and performs radio communication with the terminal UE via a radio link. In the O-RAN architecture, the communication device 3 corresponds to an E2 node including an O-CU (O-RAN Central Unit) and an O-DU (O-RAN Distributed Unit), and the radio device 4 corresponds to an O-RU (O-RAN Radio Unit).

[0012] The communication system 1 provides a cell for accommodating a terminal UE. The communication system 1 can set multiple RAN slices within the cell. The multiple RAN slices are virtual networks independent of each other. Each RAN slice is identified by, for example, S-NSSAI (Single Network Slice Selection Assistance Information). The communication system 1 can set a different quality of service (QoS) for each RAN slice. For example, an allowable delay time is set for each RAN slice. In the following description, a RAN slice may be simply referred to as a "slice."

[0013] The communication control device 2 controls the communication devices 3 and the wireless devices 4, and also performs control related to network energy saving (NES) of the wireless access network. Specifically, the communication control device 2 executes advanced sleep mode (ASM) control as one of the NES controls. When the traffic volume decreases, the ASM control puts some of the functions of the wireless devices 4 into a sleep state in symbol / slot / frame units in a section (or an intermittent section) where the wireless resources of the communication system 1 are not allocated to any traffic. This procedure reduces the power consumption of the wireless devices 4.

[0014] 2 shows an example of ASM control. In this example, slices A and B are set within a cell. The communication control device 2 then predicts the traffic demand for each slice. In this example, as shown in FIG. 2A, it is predicted that traffic for slice A will occur in the interval T1-T2, and traffic for slice B will occur in the interval T3-T4.

[0015] The communication control device 2 allocates resources to each slice based on traffic demand prediction. In the example shown in Fig. 2B, a required number of physical resource blocks (PRBs) are allocated to each of slices A and B in response to the traffic demand of each slice. One physical resource block is composed of, for example, 12 subcarriers (frequencies).

[0016] 2C shows the power consumption of the wireless device 4 for the traffic demands shown in FIG. 2A. The standby power is generated when physical resource blocks (PRBs) are active. That is, the standby power is proportional to the number of active physical resource blocks (PRBs). In addition, power is consumed to process user traffic. That is, power is consumed to transmit or receive radio signals. Thus, power is consumed to process user traffic for slice A during the interval T1-T2, and power is consumed to process user traffic for slice B during the interval T3-T4.

[0017] Here, the communication control device 2 deactivates each physical resource block PRB in intervals where there is no traffic demand through ASM control. In this embodiment, each physical resource block PRB is deactivated in intervals T0 to T1 and intervals T2 to T3. When a physical resource block PRB is deactivated, at least some of the circuits or functions of the wireless equipment 4 are stopped. That is, in intervals where a physical resource block PRB is deactivated, at least some of the circuits or functions of the wireless equipment 4 are stopped. As a result, power consumption of the wireless equipment 4 is reduced in intervals where a physical resource block PRB is deactivated. In the example shown in FIG. 2D , standby power is reduced in intervals T0 to T1 and intervals T2 to T3.

[0018] In this way, the communication control device 2 reduces the power consumption of the wireless device 4 by suspending some of the circuits or functions of the wireless device 4 through ASM control during sections where there is no traffic demand. Here, the longer the section where there is no traffic demand, the greater the amount of power consumption reduction. Therefore, in an embodiment of the present invention, communication is controlled to lengthen the section where there is no traffic demand, taking service quality into consideration. Note that, during sections where there is no traffic demand, there is no need to allocate resources to each slice. Therefore, in the following description, sections where there is no traffic demand may be referred to as "non-traffic sections" or "non-allocation sections."

[0019] 3 shows an example of functional deployment related to NES / ASM control in an O-RAN architecture. In the embodiment shown in FIG. 3, a communication system 1 related to NES / ASM control includes a Service Management Orchestration (SMO) 10, an E2 node 22, and an O-RU 23.

[0020] The SMO 10 is an example of a communication management device, and manages each device or function in the O-RAN architecture. The SMO 10 is connected to the E2 node 22 and the O-RU 23 via an O1 interface. The SMO 10 is also connected to the O-RU 23 via an Open Fronthaul interface. The SMO 10 can collect configuration information and statistical information of the radio access network via the O1 interface or the Open Fronthaul interface. The SMO 10 can also send instructions to change the configuration of the E2 node 22 via the O1 interface. The SMO 10 can also control the operation of the O-RU 23 via the Open Fronthaul interface.

[0021] The O-RAN architecture includes a RAN Intelligent Controller (RIC). The RIC provides services to each device or function within the O-RAN architecture. The RIC includes a non-real-time RIC 11 and a near-real-time RIC 21. The non-real-time RIC 11 is implemented inside the SMO 10, and the real-time RIC 21 is provided outside the SMO 10. The non-real-time RIC 11 and the real-time RIC 21 are connected via an A1 interface. The A1 interface includes A1-P and A1-EI.

[0022] The non-real-time RIC 11 includes a non-real-time RIC framework. Applications related to NES control run on the non-real-time RIC framework. In this example, an NES application and a QoS guaranteed application are implemented. Note that application programs running on the non-real-time RIC 11 are sometimes referred to as "rApp."

[0023] The NES application acquires the configuration information and statistical information of the radio access network collected by the SMO 10 and performs learning processing by providing this information to an AI (Artificial Intelligence) / ML (Machine Learning) model. The NES application also updates the AI / ML model. Furthermore, the NES application uses the AI / ML model to estimate optimal configuration information for the E2 node. The QoS application guarantees the quality of each slice.

[0024] In this way, the non-real-time RIC 11 collects various data stored in the E2 node 22 via the O1 interface. The non-real-time RIC 11 also determines optimal parameters based on the wireless environment and traffic demand using the AI / ML model, and provides these to the E2 node via the O1 interface. Furthermore, the non-real-time RIC 11 generates policies related to the control of the wireless access network, and notifies the real-time RIC 21 of these policies via the A1 interface.

[0025] 1, and collects and analyzes configuration information and statistical information of the radio access network from the E2 node 22 via the E2 interface. The real-time RIC 21 also controls the E2 node in accordance with the policy notified by the non-real-time RIC 11.

[0026] The real-time RIC 21 is implemented with an ASM application and a QoS-guaranteed application. The ASM application acquires configuration information and statistical information of the radio access network and controls the radio access network by providing this information to the AI / ML model. This AI / ML model is updated, for example, by the non-real-time RIC 11, although not particularly limited thereto. The ASM application then uses the AI / ML model to predict traffic demand for each slice and allocates resources to each slice based on the prediction. The QoS-guaranteed application allocates resources while taking into account the quality (e.g., delay) of each slice. Note that application programs running on the real-time RIC 21 are sometimes referred to as "xApps."

[0027] The E2 node 22 is an example of the communication device 3 shown in FIG. 1 and includes an O-CU and an O-DU. The E2 node 22 provides radio link control, medium access control, PHY-High functions, and the like, and processes signals from the O-RU 23 in higher layers. Note that multiple O-RUs 23 can be connected to the O-DU. The E2 node 22 may also acquire configuration information and statistical information of the radio access network and provide the information to the real-time RIC 21.

[0028] The O-RU 23 is an example of the radio device 4 shown in FIG. 1 , and is equipped with a radio circuit and can accommodate multiple terminals UE. The O-RU 23 transmits reports related to power saving of the radio access network to the SMO 10 via the O1 interface or the M plane of the Open fronthaul interface. The O-RU 23 performs on / off control of cells or carriers in accordance with NES instructions received from the SMO 10 via the O1 interface or the M plane of the Open fronthaul interface. The O-RU 23 then performs wireless communication in accordance with instructions provided by the E2 node (here, the O-DU). At this time, the O-RU 23 may suspend the function of part of the radio circuit in non-traffic sections.

[0029] The A1 interface connects the non-real-time RIC 11 and the real-time RIC 21. The E2 interface connects the real-time RIC 21 and the E2 node 22. The R1 interface connects the non-real-time RIC framework and the rApp.

[0030] In the O-RAN architecture configured as described above, the real-time RIC 21, which operates as a communication control device, performs power saving (here, ASM) of the radio access network based on the traffic demand prediction of each slice. At this time, the real-time RIC 21 expands the non-traffic section while taking into account the quality of each slice. This reduces the power consumption of the O-RU 23.

[0031] FIG. 4 shows an example of correcting the demand model. In this example, as shown in FIG. 4A, user traffic is predicted to occur in the interval T1-T2. Other traffic is predicted to occur after time T3. That is, no traffic demand is predicted in the intervals T0-T1 and T2-T3. In this case, standby power and power for processing user traffic are predicted to be consumed in the intervals T1-T2. In contrast, ASM control is used to place some of the circuits or functions of the wireless device 4 into a sleep state in the intervals T0-T1 and T2-T3. As a result, power consumption, represented by the shaded area in FIG. 4A, is reduced. Power consumption after time T3 is not shown.

[0032] In an embodiment of the present invention, the demand model is corrected so that the communication interval corresponding to the predicted traffic demand is shortened. In this example, as shown in FIG. 4B, the demand model is corrected so that user traffic is not transmitted in the interval T1-TX and is transmitted in the interval TX-T2. Here, the interval TX-T2 shown in FIG. 4B is shorter than the interval T1-T2 shown in FIG. 4A. Furthermore, the total traffic volume in the interval T1-T2 shown in FIG. 4A and the total traffic volume in the interval TX-T2 shown in FIG. 4B are the same. Therefore, the amount of resources used in the interval TX-T2 shown in FIG. 4B is greater than the amount of resources used in the interval T1-T2 shown in FIG. 4A.

[0033] Note that using more resources can shorten the communication interval. However, the resources that the communication system 1 can provide have a predetermined upper limit (Capability Limit). Therefore, the communication system 1 may correct the demand model so that the resources allocated to user traffic approach the predetermined upper limit.

[0034] In the communication system 1, ASM control is executed as described above. As a result, in the case shown in FIG. 4A, the intervals T0-T1 and T2-T3 are non-traffic intervals, and power consumption by the wireless device is reduced in these intervals. In contrast, in the case shown in FIG. 4B, in addition to the intervals T0-T1 and T2-T3, the intervals T1-TX are also non-traffic intervals. That is, power consumption by the wireless device is reduced in the intervals T0-TX and T2-T3. In this way, power consumption is further reduced by correcting the predicted traffic demand model.

[0035] FIG. 5 shows another example of the correction of the demand model. In this example, as shown in FIG. 5A, the traffic demand for slice A is predicted in the section T1-T2, and the traffic demand for slice B is predicted in the section T3-T4. Here, in order to increase the amount of power consumption reduction achieved by ASM control, it is necessary to expand the non-traffic section. To expand the non-traffic section, it is preferable that the communication section corresponding to the traffic of slice A and the communication section corresponding to the traffic of slice B overlap as much as possible.

[0036] 5B, the demand model is corrected so that the communication interval corresponding to the traffic of slice A overlaps with the communication interval corresponding to the traffic of slice B. In this embodiment, the traffic demand model of slice A is corrected so that the length of the communication interval corresponding to the traffic of slice A is the same as the length of the communication interval corresponding to the traffic of slice B. Then, the traffic demand of slice A is placed in intervals T3 to T4.

[0037] 5C, in addition to the interval T0-T1 and the interval T2-T3, the interval T1-T2 also becomes a non-traffic interval. Therefore, ASM control reduces the power consumption of the wireless device in the interval T0-T3.

[0038] In this manner, in an embodiment of the present invention, the non-traffic section is expanded by overlapping the traffic demands of multiple slices. However, the traffic volume cannot exceed the upper limit of the communication system. For example, when the traffic demands of slices A and B are overlapped, as shown in FIG. 6A, the sum of the traffic volumes is not allowed to exceed the upper limit of the communication system. Therefore, the demand model for traffic A is corrected so that the sum of the traffic volumes of slices A and B does not exceed the upper limit. In this case, the traffic demand of slice A is assigned to the section TX to T4.

[0039] As described above, in order to improve the NES efficiency (i.e., the efficiency of power saving), it is necessary to extend the non-traffic section. To extend the non-traffic section, in the ASM control shown in Figures 5 and 6, the communication section corresponding to the traffic demand of one of the slices (slice A in this embodiment) is shifted by a predetermined time on the time axis.

[0040] On the other hand, to guarantee quality of service (QoS), a maximum delay time is specified for each slice depending on the traffic type, and the communication section corresponding to the traffic demand is shifted on the time axis taking into account the specified maximum delay time.

[0041] The maximum delay time of traffic is expressed, for example, as a packet delay budget (PDB), although not limited thereto. Here, the maximum delay time of traffic is required by a service level agreement (SLA) corresponding to the traffic type. Traffic is classified, for example, as guaranteed bit rate (GBR) traffic, non-GBR traffic, or delay-critical GBR traffic. The allowable packet delay PDB for GBR traffic is 50 to 500 milliseconds. The allowable packet delay PDB for non-GBR traffic is 10 to 1100 milliseconds. The allowable packet delay PDB for delay-critical GBR traffic is 5 to 30 milliseconds. The allowable packet delay PDB is an example of a maximum delay time specified for traffic or a slice, and is expressed by a 5QI value (3GPP TS23.501 V18.4.0, 5.7.4).

[0042] As such, the maximum delay time (or tolerable packet delay PDB) of GBR traffic and non-GBR traffic is relatively large. Therefore, in an embodiment of the present invention, it is assumed that the communication intervals corresponding to GBR traffic and non-GBR traffic can be shifted under ASM control. Note that in the following description, GBR traffic or non-GBR traffic may be referred to as "GBR / non-GBR traffic." In contrast, the maximum delay time of delay-critical GBR traffic is relatively small. Therefore, it is assumed that the communication interval corresponding to delay-critical GBR traffic is not shifted under ASM control.

[0043] 7 shows an example of a process for shifting communication intervals corresponding to traffic on the time axis. In this example, as shown in FIG. 7A, the demand for traffic A is predicted in intervals T1 to T2, and the demand for traffic B is predicted in intervals T3 to T4. Traffic A is GBR / Non-GBR traffic, and traffic B is delay-critical GBR traffic.

[0044] Here, in order to extend the non-traffic interval, it is preferable to lengthen the time during which the communication interval of traffic A and the communication interval of traffic B overlap. For example, in the case shown in Figure 7B, the communication interval of traffic A is shifted on the time axis so that the end of the communication interval of traffic A coincides with the end of the communication interval of traffic B. In this case, the interval T0 to TX becomes the non-traffic interval.

[0045] However, the shift amount ΔT is set so as not to exceed the allowable delay for NES control, which is determined based on the maximum delay time (or allowable packet delay PDB) of traffic A. Here, the allowable delay for NES control NES_D is expressed by the following formula (1) or (2): NES_D=PDB−maximum measured delay value−α (1) NES_D=PDB−average measured delay value−β (2)

[0046] PDB is the allowable packet delay time, and is predetermined according to the type of traffic. The delay measurement value represents the packet propagation delay time between the terminal UE and the UPF (User Plane Function), or the packet propagation delay time between the terminal UE and the O-CU. The packet propagation delay time is measured, for example, by the O-CU. The maximum delay measurement value represents the maximum value of the measured packet propagation delay times. The average delay measurement value represents the average value of the measured packet propagation delay times. α and β each represent a margin that takes into account fluctuations in prediction using the AI / ML model.

[0047] In an embodiment of the present invention, the communication interval of traffic is shifted on the time axis within a range that does not exceed the above-mentioned NES control allowable delay NES_D. In the example shown in FIG. 7C, the communication interval corresponding to traffic A is shifted by the NES control allowable delay NES_D. In this case, the interval T0 to TY becomes a non-traffic interval. That is, ASM control reduces power consumption in the interval T0 to TY. At this time, the delay of traffic A does not exceed the allowable packet delay PDB, even when taking into account the packet propagation delay time and a predetermined margin. Therefore, the service quality required for traffic A is guaranteed.

[0048] 8 shows an example of the functional configuration of a communication system 1 according to an embodiment of the present invention. The communication system 1 includes an SMO 10, a real-time RIC 21, an O-CU 22a, an O-DU 22b, and an O-RU 23. The SMO 10 includes the non-real-time RIC 11 shown in FIG. 3. The O-CU 22a and the O-DU 22b constitute the E2 node shown in FIG. 3.

[0049] The SMO 10 includes a measurement information collection unit 31, a configuration information storage unit 32, and an AI / ML model learning unit 33. The measurement information collection unit 31, the configuration information storage unit 32, and the AI / ML model learning unit 33 are realized by, for example, a non-real-time RIC.

[0050] The measurement information collector 31 collects measurement information related to the radio access network via the O1 interface from the O-CU 22 a, the O-DU 22 b, and the O-RU 23. This measurement information includes, for example, information indicating the load of the cell and a measurement report related to power saving.

[0051] The configuration information storage unit 32 stores configuration information of each slice set in the wireless access network. This configuration information includes identification information for identifying each slice and information related to the SLA (Service Level Agreement) of each slice. The information related to the SLA may include the above-mentioned allowable packet delay PDB.

[0052] The AI / ML model learning unit 33 updates the AI / ML model for predicting traffic demand for each slice. Here, the AI / ML model learning unit 33 updates the AI / ML model by using the measurement information collected by the measurement information collecting unit 31 and the configuration information stored in the configuration information storing unit 32. The updated AI / ML model is deployed and provided to the real-time RIC 21 via the O1 / A1 interface.

[0053] The real-time RIC 21 includes a measurement information collection unit 41, an allowable delay calculation unit 42, a demand prediction unit 43, a demand model correction unit 44, a resource allocation unit 45, a resource control unit 46, and an allowable delay notification unit 47. The real-time RIC 21 may further include other functions not shown in FIG. 8.

[0054] The measurement information collector 41 collects measurement information related to each slice of the radio access network from the O-CU 22a via the E2 interface. This measurement information includes information indicating the packet propagation delay time between the terminal UE and a UPF (User Plane Function) or the packet propagation delay time between the terminal UE and the O-CU 22a. The packet propagation delay time is measured by the O-CU 22a.

[0055] Measurement of packet propagation delay is described in 3GPP TS 28.552. Downlink packet delay from UPF to radio access network is described in 5.1.1.1.8.2. Downlink packet delay from O-CU to terminal UE is described in 5.1.1.1.6. Uplink packet delay from terminal UE to O-CU is described in 5.1.1.1.7.

[0056] The allowable delay calculation unit 42 calculates the allowable delay for NES control for each slice. Here, in the case where one traffic is transmitted using one slice, the allowable delay for NES control is calculated for each traffic. The allowable delay for NES control (NES_D) is calculated based on the following equation (1) or (2). NES_D=PDB-maximum delay measurement value-α (1) NES_D=PDB-average delay measurement value-β (2)

[0057] The allowable packet delay PDB for each slice is stored in the configuration information storage unit 32. Here, the allowable packet delay PDB for each slice is predetermined for the type of traffic transmitted using that slice. The maximum delay measurement value or the average delay measurement value is calculated based on continuous information collected by the measurement information collection unit 41. That is, the maximum value of multiple delay measurement values ​​or the average value of multiple delay measurement values ​​is calculated. α and β each represent a margin that takes into account fluctuations in demand forecasts using an AI / ML model. Therefore, α and β are determined in advance, for example, by simulation or the like.

[0058] The demand forecasting unit 43 creates a demand model that represents a prediction of traffic demand in each slice. Specifically, the demand forecasting unit 43 creates a demand model for each slice using the AI / ML model learned in the SMO 10. At this time, the demand forecasting unit 43 may create a demand model for each slice using measurement information collected by the measurement information collecting unit 41. The demand model includes information representing a communication section for transmitting (or receiving) traffic and information representing the traffic volume within that communication section. The communication section may be represented by a communication start time and a communication end time.

[0059] The demand model correction unit 44 corrects the demand model created by the demand forecasting unit 43. At this time, the demand model correction unit 44 corrects the demand model by referring to the allowable delay for NES control calculated by the allowable delay calculation unit 42.

[0060] Specifically, when the demand prediction unit 43 creates a first demand model representing a prediction of a first demand corresponding to the first traffic and a second demand model representing a prediction of a second demand corresponding to the second traffic, the demand model correction unit 44 corrects the first demand model so as to increase a non-allocated section in which resources are not allocated to either the first traffic or the second traffic. Note that a "non-allocated section in which resources are not allocated to either the first traffic or the second traffic" is equivalent to a "non-traffic section in which neither the first traffic nor the second traffic exists." Here, the correction of the demand model includes a process of increasing the traffic volume per unit time while maintaining the total traffic volume and / or a process of delaying the time at which a communication section corresponding to a traffic demand appears.

[0061] For example, in the example shown in FIG. 4, a demand model is created that indicates that user traffic will occur in the section T1 to T2. The demand model correction unit 44 then performs a correction to increase the traffic volume per unit time while maintaining the total traffic volume of this demand model. In the example shown in FIG. 7, a demand model corresponding to traffic A and a demand model corresponding to traffic B are created. The demand model correction unit 44 then performs a correction to delay the demand model corresponding to traffic A. At this time, the demand model correction unit 44 corrects the demand model corresponding to traffic A so that the shift amount of the demand model corresponding to traffic A does not exceed the NES control allowable delay calculated for traffic A.

[0062] The demand model correction unit 44 corrects the demand model in accordance with the following policy. In the following description, it is assumed that demand model A is created for traffic A, and demand model B is created for traffic B. It is also assumed that demand model B appears later than demand model A on the time axis. Furthermore, it is assumed that demand model A is corrected without correcting demand model B.

[0063] (1) Expand the non-allocation section where resources are not allocated to either traffic A or traffic B. Alternatively, lengthen the section on the time axis where demand model A and demand model B overlap. (2) Increase the traffic volume of traffic A per unit time while maintaining the total traffic volume of traffic A. Alternatively, shorten the communication time of traffic A while maintaining the total traffic volume of traffic A. (3) In the section where demand model A and demand model B overlap, the sum of the traffic volumes of traffic A and traffic B does not exceed the physical upper limit of the communication capacity of the communication system 1. (4) The shift amount of demand model A does not exceed the allowable delay for NES control calculated for traffic A.

[0064] The resource allocation unit 45 allocates resources to traffic based on each demand model. At this time, the resource allocation unit 45 allocates resources to traffic for each slice. Therefore, the resource allocation unit 45 allocates resources to each traffic by referring to the configuration information of each slice stored in the configuration information storage unit 32. Note that for traffic whose corresponding demand model has been corrected by the demand model correction unit 44, resources are allocated to that traffic based on the corrected demand model. Also, for traffic whose corresponding demand model has not been corrected by the demand model correction unit 44, resources are allocated to that traffic based on the demand model created by the demand prediction unit 43.

[0065] 9 shows an example of a process for allocating resources to each traffic based on a demand model. In this example, as shown in FIG. 9A, the demand forecasting unit 43 creates a demand model A that represents a forecast of demand corresponding to traffic A and a demand model B that represents a forecast of demand corresponding to traffic B. Demand model A represents the demand for traffic A in sections T1 to T4 as "traffic volume = 20." Demand model B represents the demand for traffic B as "traffic volume = 15" in sections T8 to T9.

[0066] The demand model correction unit 44 corrects the demand model A. At this time, the demand model correction unit 44 corrects the demand model A in accordance with the above-mentioned policies (1) to (4). As a result, the corrected demand model A indicates that the demand of traffic A is "traffic volume = 40" in the section T8 to T9.

[0067] The resource allocation unit 45 allocates resources to each traffic based on the demand model B and the corrected demand model A. Specifically, to traffic B, a physical resource block equivalent to "resources = 15" is allocated in the interval T8, and a physical resource block equivalent to "resources = 15" is allocated in the interval T9. Furthermore, to traffic A, a physical resource block equivalent to "resources = 40" is allocated in the interval T8, and a physical resource block equivalent to "resources = 40" is allocated in the interval T9.

[0068] The resource control unit 46 sets the resource allocation plan created by the resource allocation unit 45 in the O-DU 22b. The O-DU 22b then notifies the O-RU 23 of this resource allocation plan. The O-RU 23 then processes the radio signals of each traffic type in accordance with the notified resource allocation plan.

[0069] 10 is a flowchart showing an example of a communication control method according to an embodiment of the present invention. In this example, a trained AI / ML model for predicting traffic demand is deployed in the demand forecasting unit 43. Furthermore, slice configuration information stored in the configuration information storage unit 32 is provided to the allowable delay calculation unit 42 and the resource allocation unit 45. Note that the processes of S1 to S8 are repeatedly executed.

[0070] In S1, the measurement information collector 41 collects measurement information from the O-CU 22a via the E2 interface. Here, the O-CU 22a measures the packet propagation delay time between the terminal UE and the UPF, or the packet propagation delay time between the terminal UE and the O-CU 22a. Then, the measurement information collector 41 collects at least information indicating the packet propagation delay time measured by the O-CU 22a.

[0071] In S2, the allowable delay calculation unit 42 calculates the allowable delay for NES control for each slice (or each traffic). As described above, the allowable delay for NES control is calculated based on the allowable packet delay PDB, packet propagation delay time, and margin parameter specified for each slice. The allowable packet delay PDB is included in the slice configuration information stored in the configuration information storage unit 32. The packet propagation delay time is a maximum delay measurement value or an average delay measurement value, and is calculated from the measurement information collected by the measurement information collection unit 41. The margin parameter is determined in advance depending on the fluctuations in demand forecast using the AI / ML model.

[0072] In S3, the allowable delay notifying unit 47 notifies the O-DU 22b of the allowable delay for NES control calculated by the allowable delay calculating unit 42. The operation of the O-DU 22b will be described later.

[0073] In S4, the real-time RIC 21 determines whether the O-DU 22b will comply with ASM control by the RIC. Whether the O-DU 22b will comply with ASM control by the RIC is assumed to be predetermined. That is, information indicating whether the O-DU 22b will comply with ASM control by the RIC is set in advance in the real-time RIC 21 or the SMO 10. If the O-DU 22b will comply with ASM control by the RIC, the real-time RIC 21 executes the processes of S5 to S8.

[0074] In S5, the demand forecasting unit 43 forecasts traffic demand for each slice and creates a demand model. The traffic demand is forecast based on, for example, the type and amount of user data to be transmitted to the terminal UE. At this time, the demand forecasting unit 43 may forecast traffic demand taking into account the measurement information collected by the measurement information collecting unit 41. The demand model includes information representing a communication interval in which traffic is transmitted (or received) and information representing the traffic volume per unit time within that communication interval. The communication interval may be represented by a communication start time and a communication end time.

[0075] In S6, the demand model correction unit 44 corrects at least one of the multiple demand models created in S5. For example, when two demand models are located close to each other on the time axis, the demand model located at the earlier time is corrected. In this case, the demand model correction unit 44 may increase the traffic volume per unit time while maintaining the total traffic volume of the earlier demand model. Alternatively, the demand model correction unit 44 may correct the earlier demand model by delaying it on the time axis so that the two demand models overlap each other. However, the earlier demand model is corrected to the extent that the shift (i.e., delay) on the time axis relative to the earlier demand model does not exceed the allowable delay for NES control calculated in S2.

[0076] In S7, the resource allocation unit 45 creates a resource allocation plan for each slice based on the demand model. At this time, if the demand model has been corrected in S6, the resource allocation plan is created based on the corrected demand model. The resource allocation plan may represent, for example, the number of physical resource blocks for each slice and each time slot.

[0077] In S8, the resource control unit 46 controls the O-DU 22b for each slice based on the resource allocation plan. At this time, the resource control unit 46 may set the resource allocation plan created by the resource allocation unit 45 in the O-DU 22b. The O-DU 22b then notifies the O-RU 23 of this resource allocation plan. The O-RU 23 then processes the radio signal in accordance with the notified resource allocation plan. The E2 node includes a receiving unit that receives the resource allocation plan created by the real-time RIC 21 and a processing unit that processes each traffic according to the resource allocation plan.

[0078] If the O-DU 22b does not follow the ASM control by the RIC (S4: No), the real-time RIC 21 does not need to execute the processes of S5 to S8. In this case, the O-DU 22b may execute ASM control independently. That is, when the real-time RIC 21 notifies the O-DU 22b of the allowable delay for NES control in S3, the O-DU 22b may execute ASM control using the allowable delay for NES control.

[0079] In this case, it is preferable that the O-DU 22b includes a demand forecasting unit 43, a demand model correcting unit 44, and a resource allocating unit 45. The O-DU 22b also creates a resource allocation plan based on ASM control by executing the processes of S5 to S7. The O-DU 22b then issues necessary instructions to the O-RU 23 based on this resource allocation plan.

[0080] In the above-described embodiment, the real-time RIC 21 executes ASM control, but the embodiment of the present invention is not limited to this configuration. For example, the non-real-time RIC 11 implemented in the SMO 10 may execute ASM control as a communication control device. In this case, the non-real-time RIC 11 may include a measurement information collection unit 41, an allowable delay calculation unit 42, a demand prediction unit 43, a demand model correction unit 44, a resource allocation unit 45, a resource control unit 46, and an allowable delay notification unit 47, as shown in FIG. 8 .

[0081] 11 shows the configuration of a communication system according to an embodiment of the present invention. In this embodiment, slices A and B are set in a cell of communication system 1. Slice A is configured as follows: Traffic type: Non-GBR 5QI: 6 Allowable packet delay PDB: 300 ms Service: Video communication, streaming Slice B is configured as follows: Traffic type: Delay-critical GBR 5QI: 86 Allowable packet delay PDB: 5 ms Service: V2X message

[0082] 12 shows an example of traffic demand predicted in advance. This prediction represents traffic demand between 0:00 and 24:00, and is created based on actual traffic volumes in the past.

[0083] In this example, traffic demand is predicted to be low between 2:00 AM and 5:00 AM. Here, during times of low traffic demand, it is easy to shift the demand model on the time axis. Therefore, ASM control according to an embodiment of the present invention is easy to execute during these times. However, during times of very low traffic demand (e.g., between 3:00 AM and 4:00 AM), conventional carrier / cell switch off / on control alone is sufficient to reduce power consumption.

[0084] Therefore, the communication system 1 may refer to the traffic demand forecast and perform ASM control during time periods when traffic demand is relatively low. For example, the communication system 1 may perform ASM control between 2:00 AM and 3:00 AM and between 4:00 AM and 5:00 AM.

[0085] 13 shows an example of traffic demand prediction obtained by the demand prediction unit 43. In this example, the traffic demand of slice A is continuous and is 10 percent of the transmission capacity of the communication system 1. The traffic demand of slice B occurs every 300 milliseconds, and the length of each communication interval is 100 milliseconds. The traffic demand of each communication interval of slice B is 20 percent of the transmission capacity of the communication system 1.

[0086] Here, the traffic type in slice A is Non-GBR, and the allowable packet delay PDB is large. Therefore, the communication system 1 corrects the demand model for slice A in the ASM control.

[0087] 10, the measurement information collector 41 collects measurement information representing the packet propagation delay time of each slice. Then, the average delay measurement value of each slice is calculated. In this example, the average delay measurement value of slice A and slice B is assumed to be 10 milliseconds.

[0088] The allowable delay calculation unit 42 calculates the allowable delay for NES control of slice A in S2 of Fig. 10. Here, the margin parameter β, which takes into account fluctuations in demand forecast using the AI / ML model, is assumed to be 10 milliseconds. Then, the allowable delay for NES control of slice A, NES_D, is calculated using the following formula: NES_D = 300 - 10 - 10 = 280 [milliseconds]

[0089] In S6 of Fig. 10, the demand model correction unit 44 corrects the demand model of slice A shown in Fig. 13. In this example, as shown in Fig. 14A, in the section T0 to T2, the demand of slice A is 10 percent of the transmission capacity of the communication system 1. Note that A1 to A3 shown in Fig. 4 constitute the demand model of slice A. Furthermore, in the section T1 to T2, the demand of slice B is 20 percent of the transmission capacity of the communication system 1.

[0090] Here, if the demand model for slice A is corrected as shown in FIG. 14B, the section where neither slice A traffic nor slice B traffic exists (i.e., the non-traffic section) is maximized. Note that in the case shown in FIG. 14B, the demand (A1-A2) in the section T0-T1 is shifted (i.e., delayed) to the section T1-T2. However, in this case, the delay of the traffic for slice A may exceed the NES control allowable delay NES_D. Therefore, in the case shown in FIG. 14B, the quality of service cannot be guaranteed.

[0091] Therefore, the demand model correction unit 44 corrects the demand model of slice A so that the traffic delay of slice A does not exceed the NES control allowable delay NES_D. Specifically, as shown in FIG. 14C, the demands (A1 to A2) within the section T0 to T1 are arranged in the section T1 to TX. TX represents a time shifted by 280 milliseconds from time T0. Note that the traffic demand of slice A in the section T1 to T2 shown in FIG. 14B is 30 percent of the transmission capacity of the communication system 1. In contrast, the traffic demand of slice A in the section T1 to TX shown in FIG. 14C is 35 percent of the transmission capacity of the communication system 1.

[0092] In the case shown in Figure 14C, the traffic delay in slice A is equal to or less than the NES control allowable delay NES_D. Therefore, quality of service is guaranteed. Furthermore, the non-traffic intervals in which neither slice A traffic nor slice B traffic exists are substantially the same in Figures 14B and 14C. Therefore, the power consumption reduction efficiency achieved by ASM control is substantially the same in Figures 14B and 14C.

[0093] The resource allocation unit 45 allocates resources to each slice based on the corrected demand model in S7 of Fig. 10. Furthermore, the resource control unit 46 sets a resource allocation plan for the O-DU 22b in S8 of Fig. 10.

[0094] The O-RU 23 performs wireless communication in accordance with instructions provided by the E2 node (here, O-DU 22b). At this time, the O-RU 23 suspends the function of part of its radio circuit during non-traffic intervals. Here, a non-traffic interval is an unallocated interval in which resources are not allocated to user traffic. Therefore, the O-RU 23 can suspend the circuit for transmitting or receiving radio signals during this interval. In the embodiment shown in FIGS. 13 and 14, at least part of the radio circuit of the O-RU 23 is set to a sleep state during intervals T0 to T1. However, in reality, there are turn-on periods during which the radio circuit of the O-RU 23 transitions from a sleep state to a normal operating state, and turn-off periods during which the radio circuit of the O-RU 23 transitions from a normal operating state to a sleep state.

[0095] The O-DU 22b uses the RRMPolicyRatio when controlling the wireless communication of the O-RU 23. The RRMPolicyRatio includes three parameters, as shown in FIG. 15. rRMPolicyDedicatedRatio indicates the ratio of resources that are exclusively secured regardless of whether or not a service is used. rRMPolicyMinRatio indicates the resources that can be used preferentially when a service is executed. rRMPolicyMaxRatio indicates the upper limit of resources that can be used for the service. The rRMPolicyMaxRatio is determined as follows based on the demand model shown in FIG. 14C. Slice A: Section T0 to T1: 0 percent Section T1 to TX: 35 percent Section TX to T2: 10 percent Slice B: Section T0 to T1: 0 percent Section T1 to TX: 20 percent Section TX to T2: 20 percent

[0096] 16 shows an example of a communication control sequence according to an embodiment of the present invention. In this example, the communication devices (O-CU, O-DU) transmit measurement information representing measured values ​​of packet propagation delay to the communication control device (RT-RIC). The communication control device calculates the allowable delay for NES control for each slice based on the allowable packet delay and the measurement information. The communication control device transmits the allowable delay for NES control to the communication devices as necessary.

[0097] The communication control device executes S5 to S7 in FIG. 10. That is, the communication control device creates a demand model representing traffic demand for each slice. The communication control device corrects the demand model using the NES control allowable delay so that non-traffic sections (i.e., sections where resources are not allocated) become larger. The communication control device creates a resource allocation plan by allocating resources to each slice or each traffic based on the corrected demand model. Then, the communication control device notifies the communication device of the resource allocation plan.

[0098] A communication device sets resources to be used by a radio unit (O-RU) according to a resource allocation plan. The radio unit performs wireless communication using the set resources. At this time, the radio unit reduces power consumption by stopping at least a portion of its radio circuitry in non-traffic sections (i.e., sections where resources are not allocated) through ASM control. As described above, according to the communication control of the embodiment of the present invention, the non-traffic sections are expanded by correcting the demand model, thereby improving power saving efficiency.

[0099] 17 shows an example of the hardware configuration of a real-time RIC 21 operating as a communication control device. The communication control device is realized by a computer 100 including a processor 101, a memory 102, a storage device 103, an input / output device 104, a recording medium reader 105, and a communication interface 106.

[0100] The processor 101 realizes the real-time RIC 21 by executing a communication control program stored in the storage device 103. Here, the communication control program includes various xApps shown in FIG. 3. Therefore, when the processor 101 executes this communication control program, the functions of the measurement information collection unit 41, the allowable delay calculation unit 42, the demand prediction unit 43, the demand model correction unit 44, the resource allocation unit 45, the resource control unit 46, and the allowable delay notification unit 47 shown in FIG. 8 are provided. The memory 102 is used as a work area for the processor 101. The storage device 103 stores the above-mentioned communication control program and other programs.

[0101] The input / output device 104 may include input devices such as a keyboard, a mouse, a touch panel, and a microphone. The input / output device 104 may also include output devices such as a display device and a speaker. The recording medium reader 105 can acquire data and information recorded on the recording medium 110. The recording medium 110 is a removable recording medium that can be attached to and detached from the computer 100. The recording medium 110 may be realized, for example, by a semiconductor memory, a medium that records signals optically, or a medium that records signals magnetically. The communication control program may be provided to the computer 100 from the recording medium 110. The communication interface 106 provides a function for connecting to a network. When the communication control program is stored in the program server 120, the computer 100 may acquire the communication control program from the program server 120.

[0102] DESCRIPTION OF SYMBOLS 10 SMO (Service Management Orchestration) 11 Non-Real Time RIC 21 Real-time RIC (Near-Real Time RIC) 22 E2 node 22a: O-CU (O-RAN Central Unit) 22b: O-DU (O-RAN Distributed Unit) 23 O-RU (O-RAN Radio Unit) 41 Measurement information collection unit 42 Allowable delay calculation unit 43 Demand forecasting unit 44 Demand model correction unit 45 Resource allocation unit 46 Resource control unit 47 Allowable delay notification unit

Claims

1. A traffic control device comprising: a demand prediction unit that creates a first demand model representing a prediction of a first demand corresponding to a first traffic and a second demand model representing a prediction of a second demand corresponding to a second traffic; a calculation unit that calculates an allowable delay of the first traffic based on the quality required by the first traffic; a correction unit that corrects the first demand model within a range where a shift amount of the first demand model does not exceed the allowable delay so that a non-allocation period during which no resources are allocated to either the first traffic or the second traffic becomes large; a resource allocation unit that allocates a first resource to the first traffic based on the corrected first demand model and allocates a second resource to the second traffic based on the second demand model; and a resource control unit that notifies the first resource and the second resource to a communication device that processes the first traffic and the second traffic.

2. The traffic control device according to claim 1, wherein the correction unit delays a period during which the first demand model is arranged on a time axis.

3. The traffic control device according to claim 1, wherein the correction unit increases a traffic demand represented by the first demand model.

4. The traffic control device according to claim 1, further comprising a collection unit that collects measurement information representing a propagation delay time of the first traffic measured by the communication device, wherein the calculation unit calculates the allowable delay of the first traffic based on the quality required by the first traffic and the propagation delay time represented by the measurement information.

5. The traffic control device according to claim 1, wherein the correction unit corrects the first demand model so that a sum of the first demand and the second demand does not exceed a predetermined upper limit value in a section where the first demand model and the second demand model overlap each other.

6. The traffic control device according to claim 5, wherein the correction unit corrects the first demand model so that the sum of the first demand and the second demand approaches the upper limit value in a section where the first demand model and the second demand model overlap each other.

7. A receiving unit that receives resource allocation information generated by a communication control device, and a processing unit that processes first traffic and second traffic according to the resource allocation information, wherein the communication control device includes: a demand prediction unit that creates a first demand model representing a prediction of a first demand corresponding to the first traffic and a second demand model representing a prediction of a second demand corresponding to the second traffic; a calculation unit that calculates an allowable delay of the first traffic based on the quality required by the first traffic; a correction unit that corrects the first demand model within a range where a shift amount of the first demand model does not exceed the allowable delay so that a non-allocation interval in which no resource is allocated to either the first traffic or the second traffic becomes large; and a resource allocation unit that allocates a first resource to the first traffic based on the corrected first demand model and allocates a second resource to the second traffic based on the second demand model, and the processing unit processes the first traffic and the second traffic according to the resource allocation information representing the first resource and the second resource. A communication device characterized by the above.

8. A communication system comprising a communication device that processes first traffic and second traffic, and a communication control device that controls the communication device, wherein the communication control device includes: a demand prediction unit that creates a first demand model representing a prediction of a first demand corresponding to the first traffic and a second demand model representing a prediction of a second demand corresponding to the second traffic; a calculation unit that calculates an allowable delay of the first traffic based on the quality required by the first traffic; a correction unit that corrects the first demand model within a range where a shift amount of the first demand model does not exceed the allowable delay so that a non-allocation section in which no resources are allocated to either the first traffic or the second traffic becomes large; and a resource allocation unit that allocates a first resource to the first traffic based on the corrected first demand model and allocates a second resource to the second traffic based on the second demand model, and the communication device processes the first traffic and the second traffic using the first resource and the second resource.

Images