JP6692336B2 - Management device, control method thereof, and program - Google Patents

Description

  The present invention relates to a management device for edge computing in a mobile network, a control method thereof, and a program.

  The European Telecommunications Standards Institute (ETSI) is proceeding with standardization of multi-access edge computing (MEC) (for example, Non-Patent Document 1), a mobile network, a Wi-Fi network, and a fixed network. The application of edge computing technology to various access networks including the above is being considered. The MEC provides IT services by using computing resources (CPU, memory, storage, etc.) arranged at a location (edge) near a user apparatus (UE: User Equipment). Therefore, compared to cloud computing, it can be expected to provide an effect of providing a service to a user with low delay and an effect of suppressing traffic flowing to a higher level network.

  When the MEC is applied to a mobile network, for example, a MEC system (MEC server) is arranged in a mobile base station, a base band unit (BBU) accommodating station, a group center (GC) station, or the like. To be done. In Patent Document 1, a MEC server arranged in a radio access network (RAN) sets radio resource requirements (delay, throughput, priority or mobility requirement) of UEs related to MEC to a mobile base station for radio resource management. Techniques for notifying (eNodeB) have been proposed. In Patent Document 1, a mobile base station manages radio resources on the basis of radio resource requirements notified from the MEC server, so that a service provided by an application operating on the MEC server can be reduced in latency. Have been described.

JP, 2017-17656, A

ETSI GS MEC 002 V1.1.1 (URL: http://www.etsi.org/deliver/etsi_gs/MEC/001_099/002/01.01.01_60/gs_MEC002v010101p.pdf)

  However, in regional stations such as mobile base stations, electric power and space are generally limited, and furthermore, it is difficult to obtain the statistical multiplex effect, and therefore it is difficult to judge capital investment in response to demand fluctuations. Therefore, in the MEC system, the available computing resources are limited as compared with the cloud system (cloud server). This means that when the number of UEs using the MEC application (edge application) increases and the MEC application is heavily loaded, the computing resources tend to be tight. If the MEC application is not provided with sufficient computing resources, processing by the MEC application will be delayed.

  The low latency, which is an advantage of MEC, depends not only on the transmission delay in the network but also on the processing time of the MEC application. When the processing time of the MEC application increases, the low delay property of the MEC cannot be maintained, and the quality of service provided by the MEC application may deteriorate. That is, the user may not be able to enjoy the advantages of MEC. In Patent Document 1 described above, even if low-delay communication can be realized in the network, it is not possible to deal with the delay due to the processing time of such an application.

  The present invention has been made in view of the above problems. An object of the present invention is to provide a technique for preventing a high load from being applied to an edge application (MEC application) operating on an edge server (MEC system).

  A management device according to an aspect of the present invention is an edge application that is connected to a cloud server that operates a cloud application and one or more cells in a wireless access network and that can be used by wireless terminals in the one or more cells. An edge server that operates the wireless terminal, and a management device that manages the edge server, and an acquisition unit that acquires position information indicating a position of each wireless terminal at a first time and application information indicating an application being used by each wireless terminal. First predicting means for predicting the number of wireless terminals using the edge application at a second time after the first time based on the position information and the application information acquired by the acquiring means, Based on the number of wireless terminals predicted by the first prediction means, at the second time Second prediction means for predicting the degree of tightness of the computing resources of the edge server used by the edge application, and a part of the radio using the edge application based on the prediction result by the second prediction means. Processing means for performing a first switching process for switching the connection destination of the terminal from the edge application to the cloud application.

  According to the present invention, it is possible to prevent a high load from being applied to an edge application (MEC application) operating on an edge server (MEC system). This makes it possible to prevent deterioration of the quality of service provided by the edge application.

A diagram showing a network configuration example including an MEC system and a cloud system Diagram showing a configuration example of a communication system Block diagram showing a hardware configuration example of a management system and an MEC system Block diagram showing a functional configuration example of the management system Flowchart showing the procedure of processing executed by the management system The figure which shows the example of the state transition model relevant to movement of UE between cells. The figure which shows the example of the change of computing resource usage with respect to the number of UE

  Hereinafter, exemplary embodiments of the present invention will be described with reference to the drawings. In each of the following drawings, components that are not necessary for explaining the embodiment are omitted from the drawings.

<Network configuration>
FIG. 1 is a diagram showing a network configuration example including an MEC system and a cloud system according to an embodiment of the present invention, and shows an application example of edge computing (MEC) to a mobile network. In this example, the MEC systems 21 and 22 for edge computing are arranged in the radio access network. Further, the cloud system 10 is arranged in an external network higher than the core network (for example, a packet data network (PDN) or the Internet).

  In the configuration example of FIG. 1, only two MEC systems 21 and 22 and one cloud system 10 exist, but any number of MEC systems can be arranged in the radio access network, and any MEC system can be arranged. A number of cloud systems can be placed on the external network. Further, assuming that LTE (Long Term Evolution) / LTE-Advanced is applied to a mobile network, the base stations 20a to 20c are eNodeBs, the radio access network is E-UTRAN (Evolved Universal Terrestrial Radio Access Network), and the core network is It corresponds to an EPC (Evolved Packet Core).

  The MEC system 21 can communicate with the base stations 20a and 20b without going through the core network, and provides a service to UEs (wireless terminals) in a cell formed by the base stations 20a and 20b. The MEC system 22 can communicate with the base station 20c without going through the core network and provides services to the UEs in the cell formed by the base station 20c. The cloud system 10 provides a service to UE via a core network and a radio access network.

  The MEC systems 21 and 22 are composed of a MEC platform and an MEC application that is an edge application that operates on an edge server. The MEC platform performs various processes such as name resolution, routing and packet filtering. The MEC platform includes a single or a plurality of hardware groups, a virtualization platform provided by using a virtualization technology on the hardware groups, and an MEC service. The MEC service is, for example, a service that enables acquisition of radio quality information, acquisition of UE position information, band control, authentication, and the like. The MEC application or an external system (for example, the management system 30 in FIG. 2) can use the MEC service to acquire the above information and perform the command.

  The MEC application runs on the MEC platform and serves UEs in a cell formed by the corresponding base station. The UE can use the MEC application of the MEC system 21 by connecting (accessing) the MEC system 21 via the base station 20a or 20b. Further, the UE can use the MEC application of the MEC system 22 by connecting to the MEC system 22 via the base station 20c.

  Since the UE moves every moment, it is necessary to ensure the portability of the MEC application by using a virtualization technology such as a virtual machine (VM) or a container (Non-Patent Document 1). In addition, the virtualization technology enables flexible allocation of computing resources to MEC applications. As described above, the MEC application can acquire the location information and the radio quality information of the UE by the MEC service provided by the MEC platform. The MEC application can control the service provided to the UE based on the information acquired by using the MEC service. Such control is expected to improve the quality of experience (QoE) for the user.

  The cloud system 10 is composed of a cloud platform and a cloud application. The cloud platform has more computing resources than the MEC platform. The cloud application runs on the cloud platform and provides services to the UE. The UE can use the cloud application by connecting to the cloud system 10 on the external network via the radio access network and the core network. The cloud application has the same function as the MEC application. Therefore, the UE can receive the service of the application regardless of whether the MEC application or the cloud application is used.

  By utilizing the MEC application instead of the cloud application, the UE can receive the service provision with low delay and can enjoy the advantages of the MEC. In addition, it is possible to reduce the traffic flowing to the host network. However, as described above, the MEC system has a limited available computing resource as compared with the cloud system. If the processing time of the MEC application increases due to the tight computing resources, the low latency of the MEC cannot be maintained, which leads to deterioration of the quality of service provided by the MEC application.

  Therefore, in this embodiment, the number of UEs that use the MEC application is predicted, and the degree of tightness of computing resources in the MEC system is predicted based on the predicted number of UEs. When the computing resources of the MEC platform are predicted to be tight, some UEs that are using the MEC application change the connected application to use the cloud application (ie, turn off the cloud application. Load). This prevents the MEC application from being overloaded due to an increase in the number of UEs that use the MEC application. That is, it is possible to prevent the computing resources from being tight and to maintain the quality of service provided by the MEC application.

  Hereinafter, a configuration example of a communication system for realizing such processing, a configuration example of each device (system) configuring the communication system, and an example of a specific processing procedure will be described.

<System configuration>
FIG. 2 is a diagram showing a configuration example of the communication system according to the present embodiment. The communication system of FIG. 2 includes one or more cloud systems (one cloud system 10 in this example), one or more MEC systems (three MEC systems 21 to 23 in this example), and a management system 30. Composed. In the present embodiment, the cloud system 10 is an example of a cloud server that operates a cloud application, and the MEC systems 21 to 23 are connected to one or more cells in the radio access network and Is an example of an edge server that operates an edge application that can be used by the wireless terminal. The management system 30 is an example of a management device that manages the cloud system 10 (cloud server) and the MEC system 21 (edge server).

  In the example of FIG. 2, there are an MEC system 21 connected to cells A and B, an MEC system 22 connected to cell C, and an MEC system 23 connected to cell D. The cells A to D are cells formed by the base stations 20a to 20d, respectively. The UE existing in each cell can wirelessly connect to the base station forming each cell and can access the MEC system connected to the base station (cell) via the connected base station. Note that there can be any number of UEs in each cell. The UE can use the MEC application or the cloud application while moving between cells. When the UE moves between cells connected to different MEC systems, the UE can switch the MEC application to be used to the MEC application operating on the MEC system connected to the movement destination cell.

  The management system 30 can communicate with the cloud system 10 and the MEC systems 21 to 23. The management system 30 may be connected to any of the wireless access network, the core network, and the external network. The management system 30 controls start and stop of applications (MEC application or cloud application) operating on the cloud system 10 and the MEC systems 21 to 23, respectively. The management system 30 also controls allocation of computing resources to applications in the cloud system 10 and the MEC systems 21 to 23.

  The management system 30 can acquire the location information and application information of each UE in the communication system by using the MEC service provided by the MEC platform of each MEC system. The application information is information indicating the application used by the UE, and includes, for example, information indicating the type of the application being used (MEC application or cloud application). The management system 30 can identify the UE using the MEC application of each MEC system based on the location information and the application information. In addition, the management system 30 can control the traffic rules applied by the MEC platform of each MEC system.

<Device configuration>
FIG. 3 is a block diagram showing a hardware configuration example of the management system 30 according to this embodiment. In this embodiment, the MEC systems 21 to 23 also have the hardware configuration shown in FIG. The cloud system 10 may also have the hardware configuration shown in FIG. The management system 30 has a CPU 31, a ROM 32, a RAM 33, an external storage device 34, and a communication device 35.

  In the management system 30, the CPU 31 executes a program, which is stored in any of the ROM 32, the RAM 33, and the external storage device 34 and realizes each function of the management system 30, for example. The CPU 31 may be replaced by one or more processors such as an ASIC (application specific integrated circuit), an FPGA (field programmable gate array), a DSP (digital signal processor), and the like. The management system 30 controls the communication device 35 by, for example, the CPU 31 to communicate with the MEC systems 21 to 23 and the cloud system 10.

  The management system 30 may include dedicated hardware that executes each function, or a part of the management system 30 may be executed by hardware and a computer that operates a program may execute other parts. Further, all the functions may be executed by the computer and the program.

  FIG. 4 is a block diagram showing a functional configuration example of the management system 30 according to the present embodiment. Each function of the management system 30 is a logical function implemented by, for example, the hardware in FIG. 3, and can be implemented by the CPU 31 executing a program stored in the ROM 32 or the like. In the present embodiment, the management system 30 includes a communication unit 41, an information acquisition unit 42, a UE number prediction unit 43, a resource usage prediction unit 44, and a switching processing unit 45.

  The communication unit 41 uses the communication device 35 to communicate with external devices such as the MEC systems 21 to 23 and the cloud system 10. The information acquisition unit 42 acquires the position information and the application information of each UE in the communication system at the first time (time t) from the external device via the communication unit 41. The UE number prediction unit 43 predicts the number of UEs that use the MEC application at the second time (time t + 1) after the first time based on the position information and the application information acquired by the information acquisition unit 42. ..

The resource usage prediction unit 44 predicts the degree of tightness of the computing resources of the MEC system used by the MEC application at the second time, based on the number of UEs predicted by the UE number prediction unit 43. Specifically, the resource usage prediction unit 44
An index value indicating the degree of tightness of computing resources, which corresponds to the number of UEs predicted by the UE number prediction unit 43, is determined. In this embodiment, the usage amount of the computing resource is used as the index value indicating the degree of tightness. However, as the index value indicating the degree of tightness, it is possible to use an index value other than the usage amount of the computing resource, such as the HTTP response time of the MEC application or the access processing time by the MEC application.

  The switching processing unit 45 switches the connection destinations of some UEs that are using the MEC application from the MEC application to the cloud application based on the prediction result of the degree of tightness of the computing resources by the resource usage prediction unit 44. I do. The switching processing unit 45 prevents the MEC application from being heavily loaded by switching the connection destinations of some UEs from the MEC application to the cloud application when it is predicted that the computing resources will be tight. .. As a result, it is possible to avoid the tightness of the computing resources, and thus it is possible to prevent the quality of service provided by the MEC application from being deteriorated.

  Furthermore, the switching processing unit 45 may perform a switching process of returning the connection destination of the UE offloaded to the cloud application to the MEC application when determining that the computing resource of the MEC system has a margin. As a result, it is possible to increase the number of UEs that can enjoy the advantages of MEC among the UEs that use the application (cloud application or MEC application).

<Processing procedure of management system>
FIG. 5 is a flowchart showing a procedure of processing executed by the management system 30. The management system 30 repeatedly executes the processing according to the procedure of FIG. 5 (for example, at fixed time intervals) for each of the MEC systems 21 to 23 to be managed. Here, the processing for the MEC system 21 will be described, but the management system 30 also executes the same processing for the MEC systems 22 and 23.

  First, in S1, the information acquisition unit 42 acquires the position information and application information of each UE at time t (first time). The position information includes information indicating the latitude, longitude and altitude at which the UE is located, or information indicating the cell to which the UE belongs. The position information may be acquired from the MEC systems 21 to 23, or may be acquired from an external system (not shown) that manages the position information of the UE. The application information also includes at least information indicating the type of application (MEC application or cloud application) that the UE is using at time t.

  Next, in S2, the UE number prediction unit 43 predicts the number of UEs using the MEC application at time t + 1 (second time) after time t based on the position information and the application information acquired in S1. To do. Specifically, for this number of UEs, the number of UEs that can use the MEC system 21 (MEC application) is predicted (first step), and the total number of UEs is multiplied by the application utilization rate (second step). Calculated by The application usage rate represents the ratio of the number of UEs that use the MEC application to the number of UEs that can use the MEC system 21.

The number of UEs that can use the MEC system 21 corresponds to the number of UEs in the cells (cell A and cell B) connected to the MEC system 21, and is related to the movement of the UEs between cells as shown in FIG. It can be predicted by a state transition model. In FIG. 6, q _X (t) indicates the number of UEs in the cell X at time t, and p _{X, Y} (t) indicates that the UE moves between cells (from cell X to cell Y) at time t + 1. Indicates the movement probability. For example, q _A (t) indicates the number of UEs in the cell A at the time t, and p _{A, B} (t) indicates the movement probability of the UE from the cell A to the cell B at the time t + 1. Hereinafter, the above-mentioned first and second steps will be described.

また、ＵＥ数予測部４３は、時刻ｔ＋１におけるセルＢ内のＵＥ数ｑ_B(ｔ＋１)についても同様に予測できる。 (First step)
The UE number prediction unit 43 predicts the number of UEs q _A (t + 1) and q _B (t + 1) in the cells A and B corresponding to the MEC system 21 at time t + 1. The number of UEs q _X (t) in each cell at time t is determined based on the position information of the UE acquired in S1. The movement probability p _{X, Y} (t) is acquired based on the statistical movement information of the UE before time t. The UE number prediction unit 43 can predict the number of UEs q _A (t + 1) in the cell A at time t + 1 by using the following equation using these pieces of information and the state transition model shown in FIG.

Further, the UE number prediction unit 43 can similarly predict the number of UEs q _B (t + 1) in the cell B at time t + 1.

(Second step)
The UE number prediction unit 43 calculates the number of UEs predicted to use the MEC application of the MEC system 21 at time t + 1. The number of UEs is calculated based on the number of UEs q _X (t + 1) calculated in the first step and the application utilization rate r _X (t + 1) at time t + 1. The number of UEs in cell X expected to utilize the MEC application can be calculated by multiplying q _X (t + 1) by r _X (t + 1). Therefore, the UE number prediction unit 43 can calculate the number of UEs predicted to use the MEC application of the MEC system 21 corresponding to the cell A and the cell B, for example, by the following formula.

Here, the application usage rate r _X (t + 1) may be the application usage rate r _X (t) at time t. In this case, the application usage rate can be determined based on the application information acquired in S1. Alternatively, a statistical arithmetic result (for example, a weighted average value) may be used as r _X (t + 1), or an arithmetic result using the application utilization rate for the neighboring MEC system (MEC system 22 or 23) may be used. It may be used or a fixed value may be used.

  Next, in S3, the resource usage prediction unit 44 uses the amount of computing resources of the MEC system 21 used (consumed) by the MEC application at time t + 1 based on the number of UEs predicted by the UE number prediction unit 43. Predict. Thereby, the resource usage prediction unit 44 predicts the tightness of the computing resources used by the MEC application at the time t + 1. The resource usage prediction unit 44 determines a predicted value of the usage amount of the computing resource corresponding to the number of UEs by using a statistical process such as linear regression as shown in FIG. 7, for example. In the example of FIG. 7, linear regression is performed based on the usage amount of computing resources with respect to the number of UEs measured before time t.

  After that, in S4, the switching processing unit 45 determines (evaluates) whether or not the computing resources are tight in the MEC system 21 based on the prediction result by the resource usage prediction unit 44. In this example, the switching processing unit 45 determines that the computing resource is tight if the predicted value of the computing resource usage amount exceeds the threshold value, and the computing resource is tight if the predicted value does not exceed the threshold value. It is determined not to. Note that a predetermined fixed value may be used for this threshold value, or a value that varies depending on the amount of computing resources used by another application may be used. When it is determined that the computing resources are tight, the switching processing unit 45 proceeds to S5, and when it is determined that the computing resources are not tight, the switching processing unit 45 proceeds to S6.

  In S5, the switching processing unit 45 performs switching processing (first switching processing) for switching the connection destinations of some UEs that are using the MEC application from the MEC application to the cloud application. First, the switching processing unit 45 determines the number of UEs that need to switch the connection destination from the MEC application to the cloud application in order to prevent the MEC system 21 from having a tight computing resource (to avoid a tight situation).

  For example, the switching processing unit 45 may determine the number of UEs that switch the connection destination to the cloud application, depending on the predicted tightness of the computing resources. Specifically, the switching processing unit 45 may determine the number of UEs that switch the connection destination to the cloud application according to the predicted value of the computing resource usage amount. In this case, the switching processing unit 45 may increase the number of UEs that switch the connection destination to the cloud application as the predicted value of the usage amount of the computing resource is higher (that is, as the tightness is higher). Note that the number of UEs that switch the connection destination to the cloud application may be predetermined in association with the predicted value of the usage amount of the computing resources.

  Furthermore, the switching processing unit 45 switches the connection destinations of the number of UEs determined as described above from the MEC application to the cloud application. The switching processing unit 45 may notify the MEC system 21 of information indicating the UE that is the target of switching the connection destination. In that case, the MEC system 21 may instruct the target UE existing in the corresponding cells A and B to switch the application of the connection destination via the base stations 20a and 20b.

In this example, the switching processing unit 45 determines the UE to be the switching target of the connection destination according to the priority set for each UE. As an example, this priority is
(1) Start time of use of application (MEC application or cloud application) by UE,
(2) Duration of application usage by the UE,
It is determined according to at least one of (3) the duration of the connection to the MEC application and (4) the number of inter-cell handovers in a predetermined period. The above times (1) to (3) can be specified based on the application information acquired in S1. Further, the number of times of handover in (4) above can be specified based on the position information acquired in S1.

  For example, in the case of the above (1), the switching processing unit 45 may preferentially switch the connection destination to the cloud application by prioritizing the UE whose application start time is early (or late). In the case of (2) above, the UE that has a long (or short) duration of application use may be prioritized to switch the connection destination to the cloud application. In the case of (3) above, the UE to which the duration of the connection to the MEC application is long (or short) may be prioritized to switch the connection destination to the cloud application. In the case of the above (4), the UE to which the number of handovers is large (or small) may be prioritized to switch the connection destination to the cloud application.

  When the switching process in S5 is completed, the management system 30 ends the process. On the other hand, when the process proceeds from S4 to S6 (when it is determined that the computing resources are not tight), the switching processing unit 45 does not perform the switching process of S5. In this case, in S6, the switching processing unit 45 determines whether or not there is a UE that is using the cloud application. If it does not exist, the process ends, and if it does, the process proceeds to S7. Proceed.

  In S7, the switching processing unit 45 performs switching processing (second switching processing) for switching the connection destination of the UE that is using the cloud application from the cloud application to the MEC application. This makes it possible to increase the number of UEs using the MEC application when it is predicted that the MEC system 21 will not have a tight computing resource. That is, the number of UEs that can enjoy the advantages of MEC can be increased.

  Specifically, the switching processing unit 45 can switch the connection destination to the MEC application based on the predicted value of the usage amount of the computing resource, within a range where the computing resource is not tight in the MEC system 21. Determine the number of. Furthermore, the switching processing unit 45 switches the connection destination of the determined number of UEs from the cloud application to the MEC application.

In this example, similarly to the switching process in S5, the switching processing unit 45 determines the UE to be the switching target of the connection destination according to the priority set for each UE. As an example, this priority is
(1) Start time of use of application (MEC application or cloud application) by UE,
(2) Duration of application usage by the UE,
It is determined according to at least one of (3) the duration of the connection to the cloud application and (4) the number of inter-cell handovers in a predetermined period. The above times (1) to (3) can be specified based on the application information acquired in S1. In addition, the number of handovers in (4) above can be specified based on the position information acquired in S1.

  For example, in the case of (1), the switching processing unit 45 may preferentially switch the connection destination to the MEC application by prioritizing the UE whose start time of using the application is early (or late). In the case of the above (2), the UE to which the duration of use of the application is long (or short) may be prioritized to switch the connection destination to the MEC application. In the case of (3) above, the UE to which the duration of the connection to the cloud application is long (or short) may be prioritized to switch the connection destination to the MEC application. In the case of (4) above, the UE to which the number of handovers is large (or small) may be prioritized to switch the connection destination to the MEC application. When the switching process in S7 is completed, the management system 30 ends the process.

  As described above, the management system 30 of the present embodiment predicts the number of UEs using the MEC application at the subsequent second time based on the position information and the application information of each UE at the first time. The management system 30 further predicts the tightness of the computing resources of the MEC systems 21 to 23 used by the MEC application at the second time based on the predicted number of UEs. The management system 30 performs a switching process for switching the connection destinations of some UEs that are using the MEC application from the MEC application to the cloud application based on the prediction result of the degree of tightness of computing resources. This makes it possible to prevent a high load from being applied to the MEC application (edge application) operating on the MEC systems 21 to 23 (edge server). As a result, it becomes possible to prevent deterioration of the quality of service provided by the MEC application.

  Further, in the present embodiment, the usage amount of the computing resources is sequentially measured, and the load on the MEC application is not reduced based on the measurement result, but based on the prediction result of the tightness of the computing resource, The load on the MEC application is reduced in advance. This makes it possible to deal with a situation in which the number of UEs using the MEC application rapidly increases.

  Various modifications can be made to the above-described embodiment. For example, in the above-described embodiment, the case where a single MEC application operates on the MEC system has been described as an example, but even when a plurality of MEC applications operate on the MEC system, the processing in the above-described embodiment is performed. It is applicable as well.

  The management system (management device) according to the present embodiment can be realized by a computer program that causes a computer to function as a management device. The computer program is stored in a computer-readable storage medium and can be distributed, or can be distributed via a network.

10: Cloud systems 20a, 20b, 20c: Base stations 21-23: MEC system 30: Management system 41: Communication unit, 42: Information acquisition unit, 43: UE number prediction unit 44: Resource usage prediction unit 45: Switching process Department

Claims (15)

Management for managing a cloud server that operates a cloud application and an edge server that is connected to one or more cells in a wireless access network and that operates an edge application that can be used by wireless terminals in the one or more cells A device,
An acquisition unit that acquires position information indicating the position of each wireless terminal at the first time and application information indicating an application being used by each wireless terminal;
First predicting means for predicting the number of wireless terminals using the edge application at a second time after the first time, based on the position information and the application information acquired by the acquiring means,
Second prediction means for predicting the degree of tightness of the computing resources of the edge server used by the edge application at the second time, based on the number of wireless terminals predicted by the first prediction means;
Processing means for performing a first switching process for switching the connection destinations of some wireless terminals that are using the edge application from the edge application to the cloud application based on the prediction result by the second prediction means;
A management device comprising: The second prediction means determines an index value indicating the degree of tightness, which corresponds to the predicted number of wireless terminals,
The management device according to claim 1, wherein the processing unit performs the first switching process based on the index value. The management apparatus according to claim 2, wherein the index value is a usage amount of the computing resource, a response time of the edge application, or an access processing time by the edge application. The processing means is
Based on the prediction result by the second prediction means, it is determined whether the computing resources are tight,
The first switching process is not performed when it is determined that the computing resource is not tight, and the first switching process is performed when it is determined that the computing resource is not tight. The management device according to any one of 1 to 3. When the processing unit determines that the computing resource is tight, the processing unit determines the number of wireless terminals that need to switch the connection destination from the edge application to the cloud application because the computing resource is not tight. The management device according to claim 4. The processing unit determines the number of wireless terminals that switch the connection destination from the edge application to the cloud application according to the degree of tightness predicted by the second predicting unit. The management device according to any one of 1. The management device according to claim 6, wherein the processing unit increases the number of wireless terminals that switch the connection destination from the edge application to the cloud application as the degree of tightness increases. The processing unit determines a wireless terminal that switches a connection destination from the edge application to the cloud application in the first switching processing according to a priority determined for each wireless terminal. The management device according to any one of 1 to 7. The priority depends on at least one of the start time of application use, the duration of application use, the duration of connection to the edge application, and the number of handovers between cells in a predetermined period by the terminal device. The management device according to claim 8, wherein the management device is determined. When the processing means determines that the computing resource is not tight, it performs a second switching process for switching the connection destination of the wireless terminal using the cloud application from the cloud application to the edge application. The management device according to claim 4 or 5. The processing unit determines a wireless terminal that switches a connection destination from the cloud application to the edge application in the second switching processing according to a priority set for each wireless terminal. 10. The management device according to 10. The priority depends on at least one of the start time of application use, the duration of application use, the duration of connection to the cloud application, and the number of handovers between cells in a predetermined period by the terminal device. The management device according to claim 11, wherein the management device is determined. The first prediction means,
Determining the number of wireless terminals in each cell at the first time based on the location information;
Determining the usage rate of the edge application at the first time based on the application information,
Calculate the number of wireless terminals in the one or more cells at the second time, based on the determined number of wireless terminals in each cell and the probability of wireless terminals moving between cells,
13. The number of wireless terminals predicted to use the edge application is calculated based on the calculated number of wireless terminals and the determined utilization rate. Management device described in. Management for managing a cloud server that operates a cloud application and an edge server that is connected to one or more cells in a wireless access network and that operates an edge application that can be used by wireless terminals in the one or more cells A method for controlling a device,
An acquisition step of acquiring position information indicating the position of each wireless terminal at the first time and application information indicating an application being used by each wireless terminal;
A first predicting step of predicting the number of wireless terminals using the edge application at a second time after the first time based on the position information and the application information acquired in the acquiring step,
A second prediction step of predicting the degree of tightness of the computing resources of the edge server used by the edge application at the second time, based on the number of wireless terminals predicted in the first prediction step;
A processing step of performing a first switching process for switching the connection destinations of some wireless terminals that are using the edge application from the edge application to the cloud application based on the prediction result in the second prediction step;
A method for controlling a management device, comprising:   A program for causing a computer to execute each step of the control method of the management device according to claim 14.

Images