JP7546093B2 - Control device, control method, and program - Google Patents

Description

The present invention relates to a control device and control method for a base station system that applies network slicing, and a program for the control device.

In the fifth generation (5G) mobile communication system, service types are broadly divided into three categories: large capacity (eMBB: enhanced Mobile BroadBand), ultra-low latency (URLLC: Ultra-Reliable and Low Latency Communications), and multiple connections (mMTC: massive Machine Type Communications), each with different service requirements. Network slicing is being considered to provide services with such different requirements economically and flexibly.

In addition, in 5G, the first edition of the specifications was formulated as 3GPP (registered trademark) Release 15, a base station is composed of a logical node called DU (Distributed Unit) and a logical node called CU (Central Unit), and the base station functions are divided between the DU and CU. The DU is equipped with lower layer functions of the base station functions (from the lower layer: RF (Radio Frequency layer), PHY (Physical layer), MAC (Media Access Control layer), RLC (Radio Link Control layer), PDCP (Packet Data Convergence Protocol layer)), and the CU is equipped with other higher layer functions. In addition to the CU and DU, a radio unit (RU: Radio Unit) having functions such as RF and PHY can also be provided.

When the above-mentioned network slicing is applied to a base station system in which functions are divided into DUs and CUs in this way, a scheduler that schedules radio resources for wireless terminals in each slice is placed in the DU corresponding to each slice. In a configuration in which one RU is shared by multiple DUs, multiple schedulers will perform scheduling using common radio resources within the same base station area, so it is necessary to achieve isolation of radio resources between slices. In addition, when allocating radio resources used for scheduling for each slice to each slice (each scheduler), it is required to increase the satisfaction of the SLA (service level agreement) of each slice (improve the service quality provided by each slice).

Non-Patent Document 1 proposes a technique for allocating radio resources for scheduling to each slice based on a predetermined allocation rate so as to ensure isolation between slices when network slicing is applied. In addition, Non-Patent Document 2 proposes a technique for allocating radio resources for scheduling to each slice so as to increase the satisfaction of the SLA of each slice when network slicing is applied.

A. Ksentini and N. Nikaein "Toward Enforcing Network Slicing on RAN:Flexibility and Resources Abstraction," IEEE Communications Magazine, Jun. 2017, vol. 55, issue 6, pp. 102-108. B. Khodapanah, A. Awada, I. Viering, D. Oehmann, Meryem Simsek, and G. P. Fettweis, "Fulfillment of Service Level Agreements via Slice-Aware Radio Resource Management in 5G Networks," 2018 IEEE 87th Vehicular Technology Conference (VTC Spring ), pp. 1-6, Jun. 2018.

However, the above-mentioned conventional technology is not able to allocate (assign) radio resources for scheduling to each slice in a way that increases the satisfaction of the SLA while ensuring isolation. For example, in Non-Patent Document 2, in an attempt to uniformly increase the satisfaction of the SLA of multiple slices, excessive radio resources are allocated to a slice in which a wireless terminal with a poor wireless environment exists, which may result in a shortage of radio resources to be allocated to other slices. In this case, isolation between slices cannot be ensured.

In addition, when network slicing is applied to a base station system in which functions are divided into DU and CU, multiple schedulers corresponding to different slices may be placed in different locations (sites). However, the above-mentioned conventional technology does not assume that multiple schedulers will be placed in different locations, and is therefore not applicable in such a case.

The present invention has been made in consideration of the above-mentioned problems. The present invention aims to provide a technology for allocating radio resources to each slice in a base station system to which network slicing is applied, so as to improve the quality of service provided by each slice while ensuring isolation between slices.

The control device according to one aspect of the present invention is a control device that allocates radio resources used for scheduling for each of a plurality of slices, and is characterized by comprising: a determination means for determining in advance a provisional allocation amount of radio resources for each of the plurality of slices; a first allocation means for allocating radio resources to each of the plurality of slices with the provisional allocation amount as an upper limit based on a target value of service quality set for each slice; and a second allocation means for allocating surplus radio resources not allocated to slices for which the target value is met, to one or more slices among the plurality of slices for which the amount of radio resources allocated by the first allocation means has reached the upper limit without the target value being met.

According to the present invention, in a base station system to which network slicing is applied, it is possible to allocate radio resources to each slice so as to improve the quality of service provided by each slice while ensuring isolation between the slices.

A diagram showing an example of the functional configuration of a CU, DU, and RU in a base station system. FIG. 1 is a diagram showing an example of a basic configuration of a base station system. FIG. 1 is a diagram showing an example of the configuration of a base station system. FIG. 2 is a block diagram showing an example of the hardware configuration of a RAN controller. FIG. 2 is a block diagram showing an example of the functional configuration of a RAN controller. FIG. 13 is a diagram showing an example of the timing of allocation of radio resources to multiple slices. FIG. 13 is a diagram showing an example of an AMC (Adaptive Modulation and Coding) mapping table for determining a predicted throughput. 11A and 11B are diagrams showing examples of radio resource allocation results by the first and second allocation processes;

Below, exemplary embodiments of the present invention will be described with reference to the drawings. Note that in each of the following figures, components that are not necessary for explaining the embodiments are omitted from the figures.

[First embodiment]
<Dividing base station functions>
A base station (base station system) generally has a plurality of functions (RF, . . . , PDCP) ranging from lower layer functions to higher layer functions, and these functions are divided and arranged in a DU and a CU.

Figure 1 is a diagram showing an example of a configuration in which multiple functions of different layers of a base station are divided into a CU, DU, and RU. As shown in Figure 1, in this embodiment, in addition to the CU and DU, an RU having functions such as RF and PHY is provided. The base station (base station system) in Figure 1 is composed of a CU10, a DU20, and a RU30, with the CU10 being connected to a core network (5GC (5G Core), EPC (Evolved Packet Core), etc.), and the DU20 being connected between the CU10 and the RU30.

The DU20 has at least one of the base station functions, a scheduling function for performing scheduling to allocate radio resources to wireless terminals. The CU10 has a higher layer function than the functions of the connected DU20, among the base station functions. The RU30 has at least one of the base station functions, an RF function equivalent to the radio wave transmission and reception function.

In the configuration example of FIG. 1, DU20 has not only High MAC function, which is equivalent to a scheduling function, but also RLC and Low MAC functions, and CU10 has SDAP (Service Data Adaptation Protocol layer)/RRC (Radio Resource Control layer) and PDCP functions, which are higher layer functions than the functions of DU20. Furthermore, RU30 has not only RF function, which is equivalent to radio wave transmission and reception function, but also PHY function. Note that only some of the PHY functions may be implemented in RU30, and the remaining PHY functions may be implemented in DU20.

The following describes in detail the configuration and operation of the base station system according to this embodiment, as well as the control of the base station system, using the functional configuration shown in FIG. 1 as an example.

<Base Station System Configuration>
2 is a diagram showing a basic configuration example of a base station system in which slices 1 to 3 corresponding to mMTC, URLLC, and eMBB are generated as service types. The CU 10 and DU 20 are controlled and managed by a RAN (Radio Access Network) controller 40, and each slice is generated by the RAN controller 40.

The base station system is composed of multiple CUs 10 (10a, 10b, 10c), multiple DUs 20 (20a, 20b, 20c), and one RU 30. CU 10a and DU 20a correspond to slice 1, CU 10b and DU 20b correspond to slice 2, and CU 10c and DU 20c correspond to slice 3. In this way, the multiple CUs 10 each correspond to a different slice, and the same goes for the multiple DUs 20 corresponding to the multiple CUs 10. Note that each of the multiple CUs 10 may correspond to one or more slices. Also, each of the multiple DUs 20 may correspond to one or more slices.

DU20 has at least the radio resource scheduling function (e.g., High MAC function) among the functions of a base station. DU20 is placed at an antenna site or between an antenna site and a data center (at a regional accommodating station).

Each CU10 is disposed between a different one of the DUs 20 and the core network, and has base station functions (e.g., SDAP/RRC and PDCP functions) that are higher layer functions than the functions of the connected DU. In the example of FIG. 2, CUs 10a, 10b, and 10c are connected to different DUs 20a, 20b, and 20c, respectively.

A single RU 30 has at least the radio wave transmission and reception function (e.g., RF function) among the functions of a base station. The RU 30 is placed at an antenna site and connected to multiple DUs 20. As a result, multiple slices 1 to 3 provided via multiple DUs 20 are provided within the same cell formed by the RU 30.

In this way, the base station system of this embodiment is composed of CU10 and DU20 corresponding to slices 1 to 3, respectively, and a single RU 30 connected to the multiple DUs 20 and located at the antenna site. In other words, the base station system is configured to connect multiple DUs and CUs to a single RU (i.e., the RU is shared for multiple DUs and CUs) without providing an RU for each slice. This makes it possible for a single RU to accommodate multiple services (slices).

In this embodiment, a RAN controller 40, which is a control device that controls the functions of the RAN, is provided in the core network or RAN. The RAN controller 40 is communicably connected to a base station system on the RAN. The RAN controller 40 sets (generates) slices 1 to 3 corresponding to service requirements for multiple CUs 10 (10a, 10b, 10c) and multiple DUs 20 (20a, 20b, 20c) on the RAN. In this embodiment, the RAN controller 40 functions as an example of a control device that allocates radio resources used for scheduling for each slice to each of the multiple slices.

In the configuration example of FIG. 2, a 5G network configuration is assumed as an example, and the 5GC CPF (5GC Control Plane Function) 60 is a group of control processing functions of the 5G core network. The 5GC UPF (5G Core User Plane Function) 50 (50a, 50b, 50c) is a group of data processing functions of the 5G core network, and is provided for each slice. 5GC UPF 50a corresponds to slice 1, 5GC UPF 50b corresponds to slice 2, and 5GC UPF 50c corresponds to slice 3.

In the configuration example of FIG. 2, the arrangement of the corresponding CU10 and DU20 differs depending on the slice (service). Depending on the arrangement of the CU10 and DU20, the performance of the base station-to-base station cooperation (inter-cell cooperation), the amount of delay given to the application, and the network utilization efficiency differ. For this reason, in the configuration example of FIG. 2, the CU10 and DU20 are arranged appropriately for each slice (service).

For slice 1 (mMTC slice), CU10a is placed in the data center where the core network is located, and DU20a is placed at the antenna site. This is to enable efficient use of the computing resources of the data center through the statistical multiplexing effect.

For slice 2 (URLLC slice), CU10b is placed in a regional storage station, and DU20b is placed in an antenna site. This allows the introduction of MEC (Multi-Access Edge Computing), achieving low latency. In this embodiment, CU10b is connected to Edge App (Edge Application Server) 70, which is an edge server having an application for providing low latency services and is placed at an edge site. Note that the edge site where Edge App 70 is placed may be the regional storage station where CU10b is placed.

For slice 3 (eMBB slice), both CU10c and DU20c are deployed in a regional accommodating station. This allows DU20c to be connected to multiple RUs 30 that are deployed at different antenna sites. In the configuration example of FIG. 2, DU20c is connected to multiple RUs 30 that form different cells, and performs processing for inter-cell coordination (e.g., CoMP (Coordinated Multi-Point Transmission/reception)) between the connected RUs. In this way, by enabling inter-cell coordination, it is possible to improve wireless communication quality.

Figure 3 is a diagram showing a more specific example of the configuration of the base station system according to this embodiment. DUs 20a, 20b, and 20c each include schedulers 21a, 21b, and 21c that perform scheduling to allocate wireless resources to wireless terminals. In the example configuration of Figure 3, schedulers 21a and 21b corresponding to slices 1 and 2 are placed at the antenna site, and scheduler 21c corresponding to slice 3 is accommodated in a regional accommodation station. In this way, schedulers 21a and 21b are placed in different locations from scheduler 21c.

DUs 20a, 20b, and 20c, which include schedulers 21a, 21b, and 21c, respectively, are connected to a common RU 30, as described above. Schedulers 21a, 21b, and 21c perform scheduling for wireless terminals that use corresponding slices, using common wireless resources that can be used in the same cell formed by the common RU 30. The wireless resources used for scheduling by each scheduler 21a, 21b, and 21c are assigned in advance by the RAN controller 40.

<Device Configuration>
RAN controller 40 has, as an example, a hardware configuration as shown in Fig. 4. Specifically, RAN controller 40 has a CPU 101, a ROM 102, a RAM 103, an external storage device 104 such as a HDD, and a communication device 105.

In the RAN controller 40, the CPU 101 executes a program that realizes each function of the RAN controller 40, which is stored, for example, in the ROM 102, the RAM 103, or the external storage device 104. The CPU 101 may be replaced by one or more processors such as an ASIC (application specific integrated circuit), an FPGA (field programmable gate array), or a DSP (digital signal processor).

The communication device 105 is a communication interface for communicating with external devices such as the controlled CU 10 and DU 20 under the control of the CPU 101. The RAN controller 40 may have multiple communication devices 105, each connected to a different destination.

The RAN controller 40 may be provided with dedicated hardware to execute each function described below, or some functions may be executed by hardware and other functions may be executed by a computer running a program. Alternatively, all functions may be executed by a computer and a program.

The CU10, DU20, and RU30 may also have the hardware configuration shown in FIG. 4. However, the RU30, as a communication device 105, is provided with a communication interface for communication with each DU20, as well as a wireless communication interface for wireless communication with a wireless terminal.

Figure 5 is a block diagram showing an example of the functional configuration of the RAN controller 40. The RAN controller 40 of this embodiment has a tentative allocation determination unit 51, a first allocation processing unit 52, a second allocation processing unit 53, and an allocation notification unit 54. In this embodiment, these functional units are realized on the CPU 101 by the execution of a control program by the CPU 101, but dedicated hardware for realizing each functional unit may be provided. Note that, of the functions of the RAN controller 40, only the functional units related to the allocation of radio resources to each scheduler 21 (allocation of radio resources used for scheduling for each slice to each of multiple slices) are shown in Figure 5, and functional units related to other functions are omitted.

The provisional allocation determination unit 51 provisionally allocates radio resources to each of the slices for isolation between slices regarding radio resources used for scheduling in each slice. Specifically, the provisional allocation determination unit 51 determines a provisional allocation (provisional allocation amount) R _tmp of radio resources to each of the slices in accordance with a predetermined allocation rate. The provisional allocation determination unit 51 outputs the determined provisional allocation R _tmp to the first allocation processing unit 52.

The first allocation processing unit 52 performs a first allocation process of allocating radio resources to a plurality of slices with the tentative allocation R _tmp as an upper limit. The first allocation processing unit 52 outputs the allocation R _iso of the radio resources determined for the plurality of slices to the second allocation processing unit 53.

The second allocation processing unit 53 performs a second allocation process to allocate (reallocate) surplus wireless resources that have not been allocated by the first allocation processing unit 52, among all wireless resources available for allocation by the first allocation processing unit 52, to one or more slices. The second allocation processing unit 53 outputs the allocation R of wireless resources determined for multiple slices as the final allocation.

The allocation notification unit 54 transmits a notification indicating the allocation of radio resources available for scheduling to the schedulers 21 (21a, 21b, 21c) corresponding to each slice based on the allocation R determined by the second allocation processing unit 53. This completes the allocation of radio resources to multiple slices.

<Timing of Allocation of Radio Resources to Each Slice>
In this embodiment, the RAN controller 40 allocates (distributes) radio resources to each scheduler 21 (each slice) at a predetermined time interval. Fig. 6 is a diagram showing an example of the timing of allocation of radio resources to each scheduler.

Here, in the architecture of a base station system such as this embodiment, the schedulers can be placed at different locations (sites). In the configuration examples of Figures 2 and 3, scheduler 21c is placed at a location different from schedulers 21a and 21b. When multiple schedulers corresponding to different slices are placed at different locations, it is generally difficult to update the allocation of radio resources to each scheduler (each slice) for each TTI (transmission time interval), which is the smallest time unit for scheduling.

For this reason, as shown in FIG. 6, the RAN controller 40 allocates (updates) radio resources to each scheduler at each time interval ΔT that is longer than a TTI (e.g., 100 TTIs). At that time, the RAN controller 40 allocates (allocates) radio resources to multiple slices in the next specified period (time interval ΔT) based on input parameters obtained in one or more past specified periods (time intervals ΔT). An example of the input parameters is a target value (target throughput) corresponding to the throughput defined in the SLA (service level agreement) of each slice, and the SINR (signal-to-interference and noise ratio) for each wireless terminal using each slice. Another example of the input parameters is the average traffic requirement in each slice.

<First allocation process>
As described above, in the scheduling of radio resources when network slicing is applied, it is necessary to realize isolation between slices so that the slices do not affect each other. Specifically, the schedulers 21a, 21b, and 21c need to use different radio resources to schedule radio terminals that use corresponding slices. The RAN controller 40 of this embodiment allocates radio resources to each scheduler 21 (each slice) by the tentative allocation determination unit 51 and the first allocation processing unit 52 so as to ensure such isolation. Hereinafter, the number of slices set by the RAN controller 40 for the multiple CUs 10 and the multiple DUs 20 is set to L.

ｒ_tmp,sは、第１割当処理部５２による割り当てに使用可能な無線リソースの総量を基準とした、スライスｓ（ｓ＝１，２，...，Ｌ）に対する無線リソースの配分率を表す。各スライスに対する配分率は、例えば、複数のスライスのそれぞれに対する、ＳＬＡで定められた目標スループット（後述する予測スループットの目標値）Ｔ_tgt,sの比として予め定められる。この場合、スライスに対して定められた目標スループットが高いほど、より大きな配分率が当該スライスに対して定められ、目標スループットが低いほど、より小さな配分率が当該スライスに対して定められる。 The provisional allocation determination unit 51 determines a provisional allocation amount R _tmp of radio resources for each of a plurality of (L) slices in accordance with a predetermined allocation rate. R _tmp is expressed as follows.

r _tmp,s represents an allocation rate of wireless resources for slice s (s=1, 2, ..., L) based on the total amount of wireless resources available for allocation by the first allocation processing unit 52. The allocation rate for each slice is determined in advance as, for example, a ratio of a target throughput (a target value of a predicted throughput to be described later) T _tgt,s determined in the SLA for each of the multiple slices. In this case, the higher the target throughput determined for a slice, the larger the allocation rate determined for that slice, and the lower the target throughput, the smaller the allocation rate determined for that slice.

In this embodiment, the provisional allocation determination unit 51 determines the provisional allocation amount R tmp for each slice so that the total amount of radio resources available for allocation by the first allocation processing unit ₅₂ is allocated to each slice according to a predetermined allocation rate.

ｒ_iso,sは、第１割当処理部５２による割り当てに使用可能な無線リソースの総量を基準とした、スライスｓ（ｓ＝１，２，...，Ｌ）に対する、第１割当処理における無線リソースの配分率を表す。 The first allocation processing unit 52 determines a radio resource allocation R _iso by allocating radio resources to each of a plurality of (L) slices with the tentative allocation R _tmp as an upper limit. R _iso is expressed as follows.

r _iso,s represents the allocation rate of the wireless resources in the first allocation process for slice s (s=1, 2, . . . , L) based on the total amount of wireless resources available for allocation by the first allocation processing unit 52.

Specifically, the first allocation processing unit 52 calculates a predicted communication capacity (predicted throughput) for each slice while gradually increasing the amount of wireless resources r _iso,s to be allocated to the slice from 0. The predicted throughput is calculated for each slice based on the amount of wireless resources to be allocated to the slice and the communication quality for the slice. The predicted throughput T _s of slice s (s=1, 2, ..., L) is calculated as follows.

ここで、Ｎ_sは、スライスｓ内のユーザ数を表す。予測スループットＴ_sは、シャノンの定理に基づいて、次式により求められる。

ここで、ｎは、第１割当処理部５２による割り当てに使用可能な物理リソースブロック（ＰＲＢ）の総数、Ｂは、１ＰＲＢ（単位無線リソース）当たりの周波数帯域幅［Ｈｚ］、γⁱ _sは、スライスｓ内のユーザｉのＳＩＮＲを表す。なお、上式で用いたシャノンの定理は通信路限界を求めるものであるので、実際には、図７に示されるようなテーブルを用いて、単位周波数当たりのビットレートが求められる。 First, among users who use slice s, allocation r ⁱ _iso,s of radio resources to user i is expressed by the following equation when round robin is used.

Here, N _s represents the number of users in slice s. The predicted throughput T _s is calculated based on Shannon's theorem by the following formula.

Here, n is the total number of physical resource blocks (PRBs) available for allocation by the first allocation processing unit 52, B is the frequency bandwidth [Hz] per PRB (unit radio resource), and γ ⁱ _s is the SINR of user i in slice s. Note that, since Shannon's theorem used in the above formula is used to find the communication path limit, in practice, the bit rate per unit frequency is found using a table such as that shown in FIG.

In this manner, the first allocation processing unit 52 of this embodiment obtains the predicted throughput T _s by using the SINR ^{γ i} _s for each wireless terminal using a slice s as the communication quality for that slice.

The first allocation processing unit 52 increases the amount of wireless resources r _iso,s to be allocated to each slice s (s = 1, 2, ..., L) from 0, and determines whether the predicted throughput T _s obtained as described above satisfies the target value (target throughput T _tgt,s ) set for the slice (T _s ≧ T _tgt,s ).

When the predicted throughput T _s reaches the target throughput T _tgt,s while increasing the amount of radio resources allocated to the slice s, the first allocation processing unit 52 stops the allocation of radio resources to the slice. Also, when the amount r _{iso,s of radio resources allocated to the slice s reaches the tentative allocation amount r tmp,s} , which is the upper limit, without the predicted throughput T _s reaching the target throughput T _{tgt,s ,} the first allocation processing unit 52 stops the allocation of radio resources to the slice.

FIG. 8A is a diagram showing an example of the result of the allocation of wireless resources by the first allocation processing unit 52. The example shows the case where the number of slices L=4. In the example of FIG. 8A, for slices 1 and 3, the predicted throughput T _s reaches the target throughput T _tgt,s before the r _iso,s of the wireless resources to be allocated to each slice reaches the upper limit (i.e., the SLA is satisfied). In this case, the wireless resources are not allocated up to the tentative allocation amount r _tmp,s , which is the upper limit, and surplus wireless resources (r _tmp,s -r _iso,s ) are generated. This surplus wireless resource is used in the second allocation processing by the second allocation processing unit 53, and is allocated (reallocated) to other slices in order to increase the satisfaction of the SLA for the other slices.

8A, for slices 2 and 4, the predicted throughput _Ts does not reach the target throughput _Ttgt,s , and the radio resources r _iso,s allocated to each slice reaches the tentative allocation amount r _tmp,s , which is the upper limit (i.e., r _iso,s =r _tmp,s ). In this case, the surplus radio resources for slices 1 and 3 are further allocated to slices 2 and 4, which have reached the upper limit of the allocation without satisfying the target throughput, by the second allocation process performed by the second allocation processing unit 53.

In this way, the first allocation processing unit 52 allocates radio resources to each slice with the provisional allocation amount r tmp,s as an upper limit until the predicted communication capacity (predicted throughput T _s ) calculated based on the amount of radio resources allocated to the slice and the communication quality (SINR) for the slice satisfies the target value (target throughput T tgt _{, s} ) set for the slice. In this way, the first allocation processing unit 52 determines the allocation R _iso of radio resources based on the provisional allocation R _tmp and outputs it to the second allocation processing unit 53.

<Second allocation process>
The second allocation processing unit 53 performs a second allocation process to allocate (reallocate) surplus wireless resources that were not allocated by the first allocation processing unit 52 out of all wireless resources available for allocation by the first allocation processing unit 52, to one or more slices that lack wireless resources to satisfy the SLA.

In this embodiment, the second allocation processing unit 53 allocates surplus radio resources that were not allocated to slices for which the target value is met (slices 1 and ₃ in the example of Figure 8 ) to one or more slices (in the example of Figure 8 , slices 2 and 4) among the multiple (L) slices for which the target value (target throughput T _tgt,s ) is not met and the amount of radio resources allocated by the first allocation processing unit 52 has reached the upper limit, the tentative allocation amount r tmp,s.

なお、第１割当処理部５２による第１割当処理において、各スライスに割り当てる無線リソースｒ_iso,sが、仮割り当て量ｒ_tmp,sに達したスライス（図８の例では、スライス１及び３）については、ｒ_iso,s＝ｒ_tmp,sであるため、リソースプールＰに加えられる無線リソースの量は０となる。 Specifically, the second allocation processing unit 53 first adds the wireless resources that have not been allocated by the first allocation processing unit 52 to the resource pool P, as shown in the following equation.

In addition, in the first allocation process by the first allocation processing unit 52, for slices (slices 1 and 3 in the example of Figure 8) in which the radio resources r _iso,s allocated to each slice have reached the provisional allocation amount r _tmp,s, since r _iso,s = r _tmp,s , the amount of radio resources added to the resource pool P is 0.

ｒ'_sは、第１割当処理部５２による割り当てに使用可能な無線リソースの総量を基準とした、スライスｓ（ｓ＝１，２，...，Ｌ）に対する、第２割当処理における無線リソースの配分率を表す。 Here, the allocation rate R′ of the radio resources for each of the multiple (L) slices for reallocating the radio resources included in the resource pool P is defined by the following equation.

r _'s represents an allocation rate of the wireless resources in the second allocation process for slice s (s=1, 2, . . . , L) based on the total amount of wireless resources available for allocation by the first allocation processing unit 52.

The second allocation processing unit 53 allocates surplus wireless resources (wireless resources included in the resource pool P) to one or more slices for which the target throughput T _{tgt, s} was not satisfied in the first allocation processing, so as to maximize the satisfaction degree of the predicted throughput with respect to the target throughput T tgt,s (i.e., the satisfaction degree of the SLA).

ただし、以下を条件とする。

Specifically, the second assignment processing unit 53 solves an optimization problem to obtain R' that minimizes the evaluation function Φ(R') as shown in the following equation.

However, the following conditions are met:

ここで、ｗ_sは、スライスｓ（ｓ＝１，２，...，Ｌ）に対して適用される重み、Φ_s(Ｒ')は、次式により定義される、各スライスに対応するコスト関数を表す。

このコスト関数は、対応するスライスｓについての目標スループットＴ_tgt,sに対する予測スループットＡ_s(Ｒ')の割合が所定値（例えば、１）に近いほど出力値が小さくなる。 The above-mentioned evaluation function Φ(R′) is defined by the following equation using a cost function Φ _s (R′) corresponding to slice s (s=1, 2, . . . , L).

Here, w _s is a weight applied to slice s (s=1, 2, . . . , L), and Φ _s (R′) is a cost function corresponding to each slice, defined by the following equation:

This cost function has an output value that is smaller the closer the ratio of the predicted throughput A _s (R') to the target throughput T _tgt,s for the corresponding slice s is to a predetermined value (for example, 1).

In the above-mentioned evaluation function Φ(R'), the weight _ws applied to the cost function _Φs (R') corresponding to each slice is determined according to, for example, the target throughput _Ttgt,s determined in the SLA. The evaluation function Φ(R') is defined as the sum of the cost functions _Φs (R') weighted according to the weight _ws . By weighting the cost function _Φs (R') according to the weight _ws , it is possible to ensure fairness between slices in terms of the satisfaction of the SLA in the allocation of wireless resources in the second allocation process.

この予測スループットＡ_s(Ｒ')は、第１割当処理において求められる予測スループットＴ_sと同様、図７に示されるようなテーブルを用いて求められる。上式におけるｒ_sは、次式のように、スライスｓに対する最終的なリソース割り当てを表す。

The predicted throughput of slice s, which defines the cost function Φ _s (R′) included in the above-mentioned evaluation function Φ(R′), is calculated by the following equation.

This predicted throughput A _s (R') is obtained using a table such as that shown in Fig. 7, similar to the predicted throughput T _s obtained in the first allocation process. In the above equation, r _s represents the final resource allocation for slice s, as shown in the following equation.

Here, when the target throughput _Ttgt,s defined in the SLA for slice s is satisfied at the completion stage of the first allocation process ( _Ts ≧ _Ttgt,s ), i.e., when the SLA is satisfied, the amount _rs of radio resources finally allocated to slice s is the amount _rs allocated in the first allocation process ( _rs = _rs ). On the other hand, when the target throughput _Ttgt,s defined in the SLA for slice s is not satisfied at the completion stage of the first allocation process ( _Ts < _Ttgt,s ), i.e., when the SLA is not satisfied, _rs is the amount obtained by adding the amount _rs allocated in the second allocation process to the amount _rs allocated in the first allocation process ( _rs = _rs + _rs ).

The second allocation processing unit 53 reallocates the radio resources included in the resource pool P to one or more slices for which the target throughput T _tgt,s was not satisfied in the first allocation processing, so as to minimize the output value of the above-mentioned evaluation function Φ(R'). That is, the second allocation processing unit 53 determines R' so as to minimize the output value of the evaluation function Φ(R'). In this way, the final resource allocation r _s for each slice s (s=1, 2, ..., L) is determined.

The second allocation processing unit 53 determines an allocation R of radio resources to a plurality of (L) slices, expressed by the following equation, as a final allocation, and outputs the allocation R to the allocation notification unit 54.

Figure 8 (B) is a diagram showing an example of the result of wireless resource allocation by the second allocation processing unit 53. In the example of Figure 8 (B), the surplus wireless resources for slices 1 and 3 shown in Figure 8 (A) are reallocated to slices 2 and 4 by the above-mentioned second allocation process. That is, the surplus wireless resources for slices 1 and 3, whose SLAs are satisfied by the first allocation process, are reallocated to the remaining slices 2 and 4 so as to increase the satisfaction of the SLAs for those slices.

As described above, the RAN controller 40 of this embodiment determines the provisional allocation amount of radio resources for each of the multiple slices according to a predetermined allocation rate. For each slice, the RAN controller 40 allocates radio resources to the slice up to the provisional allocation amount until the predicted communication capacity determined for that slice meets the target value set for that slice. Furthermore, the RAN controller 40 allocates surplus radio resources that were not allocated to the slices for which the target value is met to one or more slices for which the amount of allocated radio resources has reached the upper limit without meeting the target value.

According to this embodiment, in the first allocation process, the RAN controller 40 allocates radio resources to each slice with the provisional allocation amount determined for each slice as the upper limit. As a result, when increasing the radio resources to be allocated to each slice, independent allocation of radio resources is performed between slices within the range of the provisional allocation amount until the SLA is satisfied (until the target throughput _Ts is satisfied). The RAN controller 40 stops the allocation of radio resources to slices that satisfy the SLA, and further allocates the surplus radio resources to slices that do not satisfy the SLA in the second allocation process. In this way, it is possible to allocate radio resources to each slice by the second allocation process so as to improve the service quality provided by each slice (i.e., to improve the satisfaction of the SLA) while ensuring isolation between the slices.

[Second embodiment]
In the second embodiment, an example will be described in which surplus radio resources are allocated (redistributed) based on preset priorities for a plurality of slices as the second allocation process by the second allocation processing unit 53. The following describes differences from the first embodiment.

In this embodiment, the second allocation process by the second allocation processing unit 53 is different from that in the first embodiment. The second allocation processing unit 53 allocates wireless resources from surplus wireless resources to each slice in descending order of pre-set priority among one or more slices for which the target throughput (target value of predicted communication capacity) was not satisfied in the first allocation process. In this case, the second allocation processing unit 53 obtains a predicted throughput A _s (R') of each slice (slice s), and allocates wireless resources from surplus wireless resources (resource pool P) until the predicted throughput satisfies the target throughput T _tgt,s .

More specifically, the second allocation processing unit 53 allocates radio resources from the surplus radio resources to each of the target slices in descending order of priority until the surplus radio resources are exhausted. For example, in the example of FIG. 8(A), the second allocation process is performed for slices 2 and 4, as in the first embodiment. Here, if slice 2 has a higher preset priority than slice 4, the surplus radio resources are allocated to slice 4 after the allocation of the surplus radio resources to slice 2 is completed. This allocation (reallocation) of radio resources is performed until the surplus radio resources are exhausted.

In this way, the second allocation process based on priority not only achieves the effect of the first embodiment, but also makes it possible to allocate radio resources to each slice while increasing the likelihood that slices with high priority will satisfy the SLA.

The priority that is preset for each slice described above can be set in various ways, as described below.

Priority setting based on SINR (communication quality) The second allocation processing unit 53 can set priorities for the slices so that the higher the communication quality for each slice, the higher the priority. In this case, for example, the average SINR (i.e., the average value of the SINR for all PRBs of all users in the target slice) in the immediately previous predetermined period (time interval ΔT) is used as the communication quality for setting the priority. By performing the second allocation process based on such priority setting, it is possible to further improve the frequency utilization efficiency.

Priority setting based on allowable delay The second allocation processing unit 53 may set priorities for the slices such that the smaller the allowable delay for each slice, the higher the priority. By performing the second allocation processing based on such priority setting, it is possible to shorten the queue required for a slice with a small allowable delay, such as a slice corresponding to a URLLC.

Priority setting by PF (Proportional Fairness) method The second allocation processing unit 53 may set priorities for multiple slices using the PF method. In this case, the second allocation processing unit 53 sets priorities for multiple slices such that the higher the ratio of the predicted throughput per unit radio resource in the immediately preceding predetermined period (ΔT) to the average value of the predicted throughput (predicted communication capacity) per unit radio resource (1 PRB) in multiple past predetermined periods (ΔT) for each of the multiple slices, the higher the priority.

By performing the second allocation process based on such priority settings, it becomes possible to allocate surplus wireless resources in the second allocation process while taking into consideration communication quality and fairness between slices. As a result, it becomes possible to provide an opportunity to allocate surplus wireless resources even to slices in which wireless terminals with low SINR exist.

Priority setting based on SLA achievement rate The second allocation processing unit 53 can set priorities for multiple slices such that the smaller the ratio of the predicted throughput at the end of allocation by the first allocation processing unit 52 to the target throughput (target value of predicted communication capacity) (i.e., the achievement rate of SLA), the higher the priority. By performing the second allocation process based on such priority setting, it is possible to increase fairness between slices in the allocation of surplus wireless resources from the viewpoint of the achievement rate of SLA.

[Third embodiment]
In the third embodiment, an example will be described in which surplus radio resources are allocated (redistributed) in accordance with weights respectively determined for a plurality of slices as the second allocation process by the second allocation processing unit 53. The following describes differences from the first and second embodiments.

In this embodiment, the second allocation process by the second allocation processing unit 53 differs from the first and second embodiments. The second allocation processing unit 53 pre-sets weights for multiple slices as parameters, and allocates wireless resources from surplus wireless resources (resource pool) in the second allocation process according to the ratio of the weights for each slice. Specifically, the second allocation processing unit 53 allocates surplus wireless resources to one or more slices for which the target throughput (target value of predicted communication capacity) was not met in the first allocation process according to the weights set for the respective slices.

In this way, according to the second allocation process based on weights, it becomes possible to distribute surplus wireless resources evenly according to the weights to one or more slices for which the target throughput was not met in the first allocation process. This makes it possible to avoid a situation in which surplus wireless resources are not allocated to slices with low priority when the wireless resources in the resource pool P are exhausted due to allocation to slices with high priority, unlike the second embodiment.

The weights that are preset for each slice described above can be set in various ways, as explained below.

Weight setting by PF method The second allocation processing unit 53 may set weights for multiple slices by using the PF method. In this case, the second allocation processing unit 53 sets, for each of the multiple slices, a ratio of a predicted throughput per unit radio resource in a predetermined period (ΔT) immediately before to an average value of a predicted throughput (predicted communication capacity) per unit radio resource (1 PRB) in multiple past predetermined periods (ΔT) as a weight for the slice.

By performing the second allocation process based on such weight settings, it becomes possible to allocate (distribute) surplus wireless resources in the second allocation process while taking into consideration communication quality and fairness between slices. As a result, it becomes possible to provide an opportunity to allocate surplus wireless resources even to slices in which wireless terminals with low SINR exist.

● Weighting based on the achievement rate of SLA For each of the multiple slices, the second allocation processing unit 53 sets the inverse of the ratio of the predicted throughput at the end of the allocation by the first allocation processing unit 52 to the target throughput (target value of predicted communication capacity) (i.e., the degree of achievement of the SLA) as a weight for that slice.
By performing the second allocation process based on such weight settings, it is possible to improve fairness between slices in the allocation of surplus radio resources from the perspective of the degree of achievement of the SLA.

[Fourth embodiment]
In the fourth embodiment, an example will be described in which, as the second allocation process by the second allocation processing unit 53, provisional allocation of wireless resources performed in the first allocation process and allocation of wireless resources based on the provisional allocation are repeated until there are no surplus wireless resources. The following describes differences from the first to third embodiments.

In this embodiment, the second allocation processing unit 53 repeats the allocation of radio resources included in the resource pool P to one or more slices for which the target throughput was not satisfied in the first allocation process, based on the tentative allocation R _tmp determined by the tentative allocation determination unit 51.

Specifically, the second allocation processing unit 53 first performs a tentative allocation of surplus wireless resources (resource pool P) to each of the one or more slices according to a predetermined allocation rate ( _Rtmp ). Furthermore, the second allocation processing unit 53 allocates wireless resources to each slice with the amount of the tentatively allocated wireless resources as an upper limit until the predicted throughput for the slice satisfies the target throughput. When the second allocation processing unit 53 completes the allocation with the tentative allocation amount as an upper limit, similar to the first allocation processing, it again configures the resource pool P with the surplus resources at that time. The second allocation processing unit 53 repeats the above-mentioned tentative allocation and the allocation of the wireless resources included in the resource pool P until there are no more surplus wireless resources.

According to this embodiment, the allocation of the radio resources included in the resource pool P is repeated based on the tentative allocation R _tmp until there are no surplus radio resources, thereby making it possible to improve the fairness of the allocation of the radio resources among the slices.

[Other embodiments]
The wireless communication device according to the above-described embodiment can be realized by a computer program for causing a computer to function as the wireless communication device. The computer program can be stored in a computer-readable storage medium and distributed, or can be distributed via a network.

10: CU, 20: DU, 21: Scheduler, 30: RU, 40: RAN controller, 50: 5GC UPF, 60: 5GC CPF, 70: Edge application, 101: CPU, 102: ROM, 103: RAM, 104: External storage device, 105: Communication device

Claims (27)

A control device that allocates radio resources used for scheduling for each of a plurality of slices,
a determining means for determining a provisional allocation amount of radio resources for each of the plurality of slices;
a first allocation means for allocating radio resources to each of the slices based on a target value of service quality determined for each slice, with the provisional allocation amount being set as an upper limit;
a second allocation means for allocating, to one or more slices among the plurality of slices, the amount of the radio resources allocated by the first allocation means having reached the upper limit without the target value being satisfied, surplus radio resources that have not been allocated to the slices for which the target value has been satisfied;
A control device comprising: The control device according to claim 1, characterized in that the target value is determined by the SLA (service level agreement) of each slice. The control device described in claim 1 or 2, characterized in that the first allocation means allocates radio resources to each of the multiple slices, with the provisional allocation amount as an upper limit, until the predicted communication capacity for that slice satisfies the target value set for that slice. The control device according to claim 3 , wherein the predicted communication capacity is calculated for each slice based on an amount of radio resources allocated to the slice and a communication quality for the slice. The control device according to claim 1 , wherein the determining means determines the provisional allocation amounts for the plurality of slices in accordance with a predetermined allocation ratio. the first allocation means, while increasing the amount of the radio resources to be allocated to each slice, stops allocating the radio resources to the slice when the amount of the radio resources allocated to the slice reaches the provisional allocation amount, which is the upper limit, or when the predicted communication capacity reaches the target value;
The control device according to claim 3 or 4, characterized in that the surplus radio resources are radio resources in an amount obtained by subtracting the total amount of radio resources allocated to the multiple slices by the first allocation means from the total amount of the provisional allocation amount for the multiple slices. The control device according to claim 6, characterized in that the second allocation means allocates the surplus radio resources to the one or more slices so that a degree of satisfaction of the predicted communication capacity with respect to the target value for the one or more slices is maximized. an evaluation function is defined using a cost function corresponding to each of the one or more slices, the cost function having an output value that decreases as a ratio of the predicted communication capacity to the target value for the corresponding slice approaches a predetermined value;
The control device according to claim 6 or 7, wherein the second allocation means allocates the surplus radio resources to the one or more slices so as to minimize an output value of the evaluation function. The cost function corresponding to each of the one or more slices is weighted according to the target value;
The control device according to claim 8 , wherein the evaluation function is defined as a sum of the weighted cost functions respectively corresponding to the one or more slices. The control device according to claim 6, characterized in that the second allocation means allocates radio resources from the surplus radio resources to each slice, in descending order of a predetermined priority, among the one or more slices, so that the predicted communication capacity satisfies the target value. The control device according to claim 10, wherein the second allocation means allocates radio resources from the surplus radio resources to each slice in order from the slice with the highest priority until the surplus radio resources are exhausted. The control device according to claim 10 or 11, further comprising a setting means for setting the priorities for the plurality of slices such that the higher the communication quality for each of the plurality of slices, the higher the priority. The control device according to claim 10 or 11, further comprising a setting unit configured to set the priorities for the plurality of slices such that the smaller the allowable delay for each of the plurality of slices, the higher the priority. The control device according to claim 10 or 11, further comprising a setting means for setting the priority for the plurality of slices so that the priority is higher as the ratio of the predicted communication capacity per unit radio resource in the immediately preceding specified period to the average value of the predicted communication capacity per unit radio resource in the past plurality of specified periods for each of the plurality of slices is larger. The control device according to claim 10 or 11, further comprising a setting means for setting the priority for each of the plurality of slices so that the priority is higher for each of the plurality of slices, the smaller the ratio of the predicted communication capacity at the end of allocation by the first allocation means to the target value. The control device according to claim 6 , wherein the second allocation means allocates the surplus radio resources to the one or more slices in accordance with weights determined for the one or more slices. 17. The control device according to claim 16, further comprising a setting means for setting, for each of the plurality of slices, a ratio of a predicted communication capacity per unit radio resource in a specified period immediately preceding the slice to an average value of a predicted communication capacity per unit radio resource in a plurality of specified periods in the past as the weight for the slice. The control device according to claim 16, further comprising a setting means for setting, for each of the plurality of slices, the inverse of a ratio of the predicted communication capacity at the end of allocation by the first allocation means to the target value as the weight for that slice. The control device according to claim 6, characterized in that the second allocation means performs a provisional allocation of the surplus radio resources to each of the one or more slices in accordance with a predetermined allocation rate, and repeats the process of allocating radio resources to the slice with the amount of the provisionally allocated radio resources as an upper limit until the predicted communication capacity satisfies the target value, and until the surplus radio resources are exhausted. The control device according to any one of claims 1 to 19, characterized in that the determination means determines a provisional allocation amount for each slice so as to allocate a total amount of radio resources available for allocation by the first allocation means to each slice according to a predetermined allocation rate. The control device according to any one of claims 5, 19 and 20, characterized in that the predetermined allocation rate is determined as a ratio of the target value for each of the plurality of slices. The control device according to claim 4 , wherein the first allocation means calculates the predicted communication capacity by using, as the communication quality, a signal-to-interference and noise ratio (SINR) for each wireless terminal using a corresponding slice. The control device according to claim 1 , wherein the target value is a target value corresponding to a throughput defined in a service level agreement (SLA) of each slice. the first and second allocation means allocate radio resources to a plurality of schedulers corresponding to the plurality of slices, respectively;
The control device according to any one of claims 1 to 23, characterized in that each of the multiple schedulers performs scheduling to allocate radio resources to radio terminals using corresponding slices, using the radio resources allocated by the first and second allocation means, for each predetermined period within the same cell formed by one radio unit. The plurality of schedulers are arranged in one or more distributed units (DUs),
The wireless unit and the one or more DUs constitute a base station system;
The wireless unit is disposed at an antenna site of the base station system;
The control device according to claim 24 , wherein the control device is communicatively connected to the base station system. A control method executed by a control device that allocates radio resources used for scheduling for each of a plurality of slices, the control method comprising:
determining a provisional allocation amount of radio resources for each of the plurality of slices;
a first allocation step of allocating radio resources to each of the slices based on a target value of service quality determined for each slice, with the provisional allocation amount being set as an upper limit;
a second allocation step of allocating surplus wireless resources that have not been allocated to the slices for which the target value is satisfied, to one or more slices among the plurality of slices for which the amount of wireless resources allocated in the first allocation step has reached the upper limit without the target value being satisfied;
A control method comprising: A program for causing a computer included in a control device to execute the control method described in claim 26.

Images