Control device, control method, and computer program
The control device optimizes application relocation in MEC systems by prioritizing applications that fail to meet delay conditions, reducing costs and maintaining service quality through strategic host selection.
Patent Information
- Application Number
- JP2021141420
- Authority / Receiving Office
- JP · JP
- Patent Type
- Patents
- Current Assignee / Owner
- Filing Date
- 2021-08-31
- Publication Date
- 2024-11-05
- Estimated Expiration
- 2041-08-31
AI Technical Summary
Existing application relocation techniques in MEC systems do not adequately address applications that fail to meet service request delay conditions, leading to inefficient relocation processes and increased costs.
A control device and method that prioritize the relocation of applications not meeting service request delay conditions by calculating and minimizing relocation costs, ensuring applications are moved to optimal hosts while maintaining quality of service.
Reduces relocation costs and maintains service quality by preferentially relocating applications that exceed delay thresholds, optimizing host selection based on cost and resource usage.
Smart Images

Figure 0007577630000001 
Figure 0007577630000002 
Figure 0007577630000003
Abstract
Description
[Technical field]
[0001] The present invention relates to a control device, a control method, and a computer program. [Background technology]
[0002] Conventionally, in a mobile network, a multi-access edge computing (MEC) technology is known for providing a service by an application to a user terminal (UE: User Equipment) at a location closer to the user terminal (for example, see Non-Patent Document 1). Non-Patent Document 1 describes an application relocation technology that, in a distributed cloud system consisting of multiple MEC hosts and a central cloud host, in a situation where an application for realizing a service to be provided to a certain user terminal is located in a certain host, re-accommodates the user terminal in an application located in another host.
[0003] In the application relocation technology described in Non-Patent Document 1, a combination of a relocation target application and a relocation destination host, which is to be relocated from a relocation source host where an application (threshold exceeding application) for which end-to-end (E2E) service delay conditions are not satisfied, to another host, is selected in consideration of the fluctuation in E2E delay of each application accompanying the relocation and the cost of the relocation. The candidate combination of the relocation target application and the relocation destination host is a combination that satisfies each constraint condition (request delay of each process, resource capacity upper limit of each host and link) after the relocation.
[0004] According to the application relocation technology described in Non-Patent Document 1, applications to be relocated are not limited to applications exceeding a threshold, but are selected from all applications that satisfy each constraint condition in the relocation source host. This increases the success rate of relocation. In addition, when selecting a combination of an application to be relocated and a relocation destination host, the E2E delay fluctuation of each application due to relocation and the cost of relocation are taken into consideration, so that the efficiency of the relocation process is improved in terms of the number of applications to be relocated and the cost. [Prior art documents] [Non-patent literature]
[0005] [Non-Patent Document 1] Tetsu Joh, Takayuki Warabino, Masaki Suzuki, Yusuke Suzuki, Tomohiro Otani, “Application Relocation Method for Distributed Cloud Environment Considering E2E Delay and Cost Variation”, Advanced Information Networking and Applications. AINA 2021, Lecture Notes in Networks and Systems, vol 226., pp 445-457, 2021. Summary of the Invention [Problem to be solved by the invention]
[0006] In the application relocation technology described in Non-Patent Document 1, an application to be relocated is selected from among all applications that satisfy each constraint condition in the relocation source host. For this reason, the case where an application that does not satisfy the service request delay condition even if relocated from the relocation source host is relocated to another host, and then another application is further relocated from the relocation destination host is not fully considered.
[0007] The present invention has been made in consideration of these circumstances, and its purpose is to reduce the costs associated with application relocation by giving priority to the relocation of applications that do not satisfy the service request delay conditions. [Means for solving the problem]
[0008] (1) One aspect of the present invention is a control device that determines an information processing device that executes an application for realizing a service to be provided to a user in an information processing system including a plurality of information processing devices, This is the end-to-end service delay requirement that a user requires from an application. Applications to be relocated that do not satisfy the service request delay conditions No. 1 A list of relocation candidate information processing devices that are candidates for relocation destinations is created, and for each of the relocation candidate information processing devices in the list, The relocation cost is composed of a first relocation cost required for the first relocation of the relocation target application and a second relocation cost required for the second relocation of the application on the relocation candidate information processing device to another information processing device, the second relocation cost being one or more times of the second relocation. An expected relocation cost is calculated, and the relocation candidate information processing device having the smallest expected relocation cost is determined as a relocation destination information processing device for the application to be relocated, No. 1 When an application that does not satisfy a service request delay condition occurs in the relocation destination information processing device to be relocated, the application is relocated to the relocation destination information processing device. Service request delay condition is not met The application is designated as a relocation candidate application and the application is transferred from the relocation destination information processing device to another information processing device. Second A given relocation calculation for relocation The process is repeated recursively until no application that does not satisfy the service request delay condition occurs in the relocation destination information processing device or until the upper limit of a predetermined number of repetitions is reached. The control device includes a placement calculation unit and a placement control unit that executes application placement control based on the re-placement content determined by the placement calculation unit. (2) In one aspect of the present invention, the relocation candidate information processing device is No. 1 The control device according to (1) above is an information processing device in which, when rearranged, the maximum value of the end-to-end delay rate in the application on each of the information processing devices does not deteriorate. ( 3 In one aspect of the present invention, the first relocation cost is No. 1 The fluctuation of resource consumption cost when relocating and the No. 1 The cost required for executing the relocation is calculated based on the above ( 1 ) control device. ( 4 In one aspect of the present invention, the second relocation cost is a cost for relocating an application to be relocated to the relocation candidate information processing device to another information processing device. Second The expected value of the change in the resource consumption cost when the relocation is performed and the expected value of the change in the resource consumption cost when the relocation is performed and the expected value of the change in the resource consumption cost when the relocation is performed Second The expected value of the required cost in executing the relocation and the relocation candidate information processing device are transferred from the relocation candidate information processing device to another information processing device in order to eliminate the application that does not satisfy the service request delay condition generated in the relocation candidate information processing device. Second The expected number of applications to be relocated, calculated based on the above ( 1 ) or ( 3 ) control device.
[0009] ( 5 One aspect of the present invention is a control method for determining an information processing device that executes an application for realizing a service to be provided to a user in an information processing system configured with a plurality of information processing devices, the method comprising: This is the end-to-end service delay requirement that a user requires from an application. Applications to be relocated that do not satisfy the service request delay conditions No. 1 A list of relocation candidate information processing devices that are candidates for relocation destinations is created, and for each of the relocation candidate information processing devices in the list, The relocation cost is composed of a first relocation cost required for the first relocation of the relocation target application and a second relocation cost required for the second relocation of the application on the relocation candidate information processing device to another information processing device, the second relocation cost being one or more times of the second relocation.An expected relocation cost is calculated, and the relocation candidate information processing device having the smallest expected relocation cost is determined as a relocation destination information processing device for the application to be relocated, No. 1 When an application that does not satisfy a service request delay condition occurs in the relocation destination information processing device to be relocated, the application is relocated to the relocation destination information processing device. Service request delay condition is not met The application is designated as a relocation candidate application and the application is transferred from the relocation destination information processing device to another information processing device. Second A given relocation calculation for relocation The process is repeated recursively until no application that does not satisfy the service request delay condition occurs in the relocation destination information processing device or until the upper limit of a predetermined number of repetitions is reached. The control method includes a placement calculation step, and a placement control step in which the control device executes application placement control based on the re-placement content determined in the placement calculation step.
[0010] ( 6 In one aspect of the present invention, in an information processing system including a plurality of information processing devices, a computer of a control device that determines an information processing device that executes an application for realizing a service to be provided to a user is provided with: This is the end-to-end service delay requirement that a user requires from an application. Applications to be relocated that do not satisfy the service request delay conditions No. 1 A list of relocation candidate information processing devices that are candidates for relocation destinations is created, and for each of the relocation candidate information processing devices in the list, The relocation cost is composed of a first relocation cost required for the first relocation of the relocation target application and a second relocation cost required for the second relocation of the application on the relocation candidate information processing device to another information processing device, the second relocation cost being one or more times of the second relocation. An expected relocation cost is calculated, and the relocation candidate information processing device having the smallest expected relocation cost is determined as a relocation destination information processing device for the application to be relocated, No. 1 When an application that does not satisfy a service request delay condition occurs in the relocation destination information processing device to be relocated, the application is relocated to the relocation destination information processing device. Service request delay condition is not met The application is designated as a relocation candidate application and the application is transferred from the relocation destination information processing device to another information processing device. Second A given relocation calculation for relocation The process is repeated recursively until no application that does not satisfy the service request delay condition occurs in the relocation destination information processing device or until the upper limit of a predetermined number of repetitions is reached.The computer program is for causing the computer to execute a placement calculation step and a placement control step of executing application placement control based on the re-placement content determined in the placement calculation step. Effect of the Invention
[0011] According to the present invention, by giving priority to the relocation of applications that do not satisfy the service request delay condition, it is possible to obtain an effect of suppressing costs associated with the relocation of applications. [Brief description of the drawings]
[0012] [Figure 1] 1 is a block diagram showing an example of the configuration of a distributed cloud system according to an embodiment. [Diagram 2] FIG. 2 is a block diagram showing a configuration example of a control device according to an embodiment. [Diagram 3] FIG. 1 is an explanatory diagram illustrating an outline of a rearrangement calculation method according to an embodiment. [Figure 4] 1 is a flowchart showing an example of a procedure of a rearrangement calculation method according to an embodiment. [Diagram 5] FIG. 11 is an explanatory diagram illustrating an example of rearrangement according to an embodiment. [Figure 6] FIG. 11 is an explanatory diagram illustrating an example of rearrangement according to an embodiment. [Figure 7] 10 is a flowchart illustrating an example of a procedure of a control method for new deployment of an application according to an embodiment. [Figure 8] 11 is a flowchart illustrating an example of a procedure of a control method for relocation of an existing application according to an embodiment. DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS
[0013] Hereinafter, an embodiment of the present invention will be described with reference to the drawings. Fig. 1 is a block diagram showing an example of the configuration of a distributed cloud system according to an embodiment. The distributed cloud system 100 (information processing system) shown in Fig. 1 includes a plurality of MEC hosts 110m (information processing devices) and each central cloud host 110c (information processing device) included in the central cloud system 5. These hosts 110m and 110c are connected via a communication network such as a mobile backhaul network MBN or the Internet INW. Hereinafter, the MEC host 110m and the central cloud host 110c will be referred to as "host 110" when there is no particular distinction between them. An application may also be referred to as App.
[0014] The MEC hosts 110m are distributed within the mobile communication network at locations close to the base station gNB to which the user equipment UE is connected, on the mobile backhaul network MBN side, etc. The central cloud host 110c is disposed on the Internet INW side, which is a communication network external to the mobile communication network.
[0015] The MEC host 110m and the central cloud host 110c are information processing devices that execute applications for implementing services to be provided to a user equipment UE (user). The MEC host 110m is provided closer to the user equipment UE than the central cloud host 110c, and therefore has a smaller communication delay than the central cloud host 110c. In the example of FIG. 1, each MEC host 110m (areas 1, 2, 3) is provided close to the base station gNB of each of areas 1, 2, 3 of the radio access network MNW in the mobile communication network. As a result, for example, the MEC host 110m provided close to the base station gNB of area 1 provides an application service to the user equipment UE present in area 1, thereby reducing service delay and communication load. On the other hand, the central cloud host 110c has a larger processing capacity than the MEC host 110m. For this reason, the central cloud host 110c is responsible for, for example, application services with less strict service request delay conditions and application services with a large processing load on the MEC host 110m. In the distributed cloud system 100, applications are provided according to the characteristics of each host 110.
[0016] In this embodiment, deploying an application refers to deploying an application for realizing a service to be provided to a certain user equipment UE in the MEC host 110m or the central cloud host 110c. In other words, deploying an application refers to assigning a certain user equipment UE to an application executed on the MEC host 110m or the central cloud host 110c.
[0017] For example, an application corresponding to a head mounted display to which AR (Augmented Reality) or VR (Virtual Reality) technology is applied as a user terminal UE has a short required delay and is therefore preferably arranged in the MEC host 110m. Also, an application for delivering map information to a navigation device mounted in an automobile as a user terminal UE does not have a very short required delay and does not have a large amount of processing, so it may be arranged in any host 110. Also, an application corresponding to a surveillance camera as a user terminal UE has a certain margin in required delay but a large amount of processing, so if there is a margin in processing capacity, it may be arranged in the MEC host 110m, or may be arranged in the central cloud host 110c.
[0018] In a mobile communications network, when user equipment UE moves between base station areas or when user equipment UE is concentrated in some base station areas due to an event or the like, the initial application placement may not be able to maintain sufficient service quality. For example, in FIG. 1, if an event is held in area 2 and congestion occurs in area 2 due to a large number of user equipment UE, the load may be concentrated on the MEC host 110m corresponding to area 2, making it difficult to maintain a specified service quality. In such a case, an attempt is made to maintain the specified service quality by relocating applications.
[0019] In this embodiment, application reallocation refers to changing the allocation of an application for realizing a service to be provided to a certain user equipment UE. In other words, application reallocation refers to changing the allocation of a certain user equipment UE that is assigned to an application executed on a certain host 110 to an application executed on another host 110.
[0020] In Fig. 1, for example, in a situation where an application for realizing a service to be provided to a certain user equipment UE is located in a MEC host 110m (area 2), user equipment UE may be concentrated in a base station area (area 2) close to the MEC host 110m (area 2), causing the load on the MEC host 110m (area 2) to increase and making it impossible to maintain sufficient service quality. In such a case, in order to maintain a predetermined service quality, an attempt is made to reduce the load on the MEC host 110m (area 2) by relocating the application to the MEC host 110m (area 2).
[0021] The control device 1 transmits and receives information through communication with a mobile communication network, a MEC host 110m, a central cloud host 110c, etc. The control device 1 determines a host 110 that executes an application for realizing a service to be provided to a user in the distributed cloud system 100. The control device 1 controls new deployment and re-deployment of applications in the distributed cloud system 100.
[0022] In the relocation of applications according to this embodiment, the cost associated with the relocation of applications is reduced by giving priority to the relocation of applications that do not satisfy the service request delay condition.
[0023] In this embodiment, an application that does not satisfy the service request delay condition is called an over-threshold application. An over-threshold application is an application in which the ratio (E2E delay rate) of the E2E service delay time (E2E service delay time) of the application to a request delay threshold based on the end-to-end (E2E) service request delay time required for the application exceeds a predetermined value (for example, 1). The E2E service delay time of an application is represented by the sum of the data transfer delay time in the network section between the user equipment UE and the application on the host 110 and the processing delay time on the host 110. The request delay threshold may be the E2E service request delay time, or may be a value determined using the E2E service request delay time as an index.
[0024] Fig. 2 is a block diagram showing an example of the configuration of a control device according to this embodiment. In Fig. 2, the control device 1 includes a user equipment (UE) request acceptance unit 11, a database (DB) 13, a distributed cloud information collection unit 14, a service quality monitoring unit 15, a placement calculation unit 16, a placement control unit 17, and an external collaboration unit 18. In Fig. 2, solid arrows indicate the flow of control traffic, and dashed arrows indicate the flow of service traffic.
[0025] The functions of the control device 1 are realized by the control device 1 including computer hardware such as a CPU (Central Processing Unit) and memory, and the CPU executing a computer program stored in the memory. The control device 1 may be configured using a general-purpose computer device, or may be configured as a dedicated hardware device. The control device 1 may be realized by a single computer, or may be realized by distributing the functions of the control device 1 among multiple computers.
[0026] The UE request reception unit 11 receives an application start request from a user equipment UE present in each base station area of a mobile communication network. The application start request includes application identification information (AppID) for identifying an application for implementing a service desired by the user equipment UE (user) and terminal identification information (terminal ID) of the user equipment UE.
[0027] When the UE request reception unit 11 receives an application start request from the user equipment UE, it registers the application start request in DB13 (UE information). Furthermore, the UE request reception unit 11 acquires radio information from the user equipment UE that has made the application start request, and stores the acquired radio information in DB13 (UE information) in association with the terminal ID of the user equipment UE. The radio information is measurement values such as reception strength and radio parameters related to radio communication between the user equipment UE that has made the application start request and the base station gNB. By registering the application start request in DB13 (UE information), a request for new placement calculation of the application is made to the placement calculation unit 16 via DB13.
[0028] When the UE request acceptance unit 11 receives an application placement update notification from the placement control unit 17 via the DB 13, it refers to the application (App) placement information in the DB 13 and checks for updates (additions and changes) to the information on new placement or relocation of the application. When the information on new placement or relocation of the application is updated, the UE request acceptance unit 11 notifies the relevant user equipment UE of information on the host 110 (the communication destination host of the user equipment UE) on which the new placement or relocation is indicated in the updated information. For example, the UE request acceptance unit 11 notifies the user equipment UE that originated the application startup request of information on the host 110 on which the application is newly placed. As a result, the user equipment UE that originated the application startup request accesses the host 110 indicated in the notified information and uses the application corresponding to the application startup request. For example, the UE request acceptance unit 11 notifies the user equipment UE that originated the application of information on the host 110 on which the application is relocated. As a result, the user equipment UE corresponding to that application can continue to receive the service of that application by accessing the relocation destination host 110 indicated in the notified information and using that application.
[0029] The DB 13 stores various information. Examples of information stored in the DB 13 include a service catalog, a network topology, host information, link information, user terminal (UE) information, and App placement information. The service catalog is information indicating service request delay conditions (such as E2E service request delay time) for each application, required computational resources (requested resources), and an overview of the service. The network topology is information indicating the connection form of the network in the distributed cloud system 100. The host information is information on each host 110, such as information on computational resources. Examples of information on computational resources include the maximum resource amount available for each host 110 and the amount of resources currently being used (resource consumption amount). The link information is information such as the maximum bandwidth available for each link shown in the network topology and the bandwidth currently being used (bandwidth consumption). The UE information is information notified from each user terminal UE, such as wireless information such as the terminal ID and reception strength of each user terminal UE, and an application startup request. The App placement information is information indicating the placement or relocation of an application.
[0030] The distributed cloud information collector 14 periodically collects host information of each host 110 in the distributed cloud system 100 and link information of each link. The distributed cloud information collector 14 stores the collected host information and link information in the DB 13. The host information and link information are collected from each MEC host 110m included in the MEC system 4, each central cloud host 110c included in the central cloud system 5, etc.
[0031] The service quality monitoring unit 15 monitors the E2E service delay time of each application arranged in each host 110. In monitoring the E2E service delay time, the service quality monitoring unit 15 estimates the current E2E service delay time for each application arranged in each host 110, and determines whether or not the service request delay condition is satisfied based on the estimation result.
[0032] Specifically, the service quality monitor 15 estimates the current E2E service delay time based on the UE information, host information, link information, and app deployment information stored in the DB 13. The E2E service delay time is the time (delay time) required for an end-to-end (E2E) service. The E2E service delay time is the sum of the time (processing delay time) required for application processing and the time (data transfer delay time) required for data transfer in the network section.
[0033] The service quality monitor 15 compares the current E2E service delay time estimation result with a threshold based on the service request delay condition (E2E service request delay time). If the comparison result shows that the current E2E service delay time estimation result exceeds the threshold, the service quality monitor 15 requests the placement calculation unit 16 to perform a relocation calculation, specifying the host 110 where the application exceeding the threshold (over-threshold application) is currently placed as the relocation source. The request for the relocation calculation is sent from the service quality monitor 15 to the placement calculation unit 16 via the DB 13.
[0034] The placement calculation unit 16 performs calculation of new placement of applications (new placement calculation) and calculation of relocation of applications (relocation calculation) in the distributed cloud system 100. In the new placement calculation, the placement calculation unit 16 determines the host 110 in which to newly place the application indicated in the application start request (start request application) based on various information in the DB 13 such as UE information of the user terminal UE that has requested the application start in response to a request for new placement calculation of the application by the application start request. In this new placement calculation, the host 110 in which the new placement is to be placed is determined based on the request delay of the start request application and the cost required for the new placement of the application. The placement calculation unit 16 stores the result of the new placement calculation, "the combination of the user terminal UE that has requested the application start and the host 110 in which the new placement is to be placed," in the App placement information in the DB 13. Note that, if the placement calculation unit 16 cannot find a host 110 that satisfies the service request delay condition or the requested resource of the start request application, it performs the relocation calculation and then tries to perform the new placement calculation of the start request application again.
[0035] The placement calculation unit 16 performs a placement calculation when it cannot find a new placement destination host 110 for the start-requested application or when it receives a request for placement calculation from the service quality monitoring unit 15. The method of placement calculation will be described later in detail. The placement calculation unit 16 stores the result of the placement calculation, "a combination of the application to be placed, the host 110 from which the application is placed, and the host 110 to which the application is placed," in the App placement information in the DB13.
[0036] The placement control unit 17 executes control of new placement and re-placement of applications in the distributed cloud system 100 (application placement control) based on the app placement information stored in the DB 13 by the placement calculation unit 16. As a result of executing this application placement control, new placement and re-placement of applications based on the app placement information are realized. After completing the application placement control, the placement control unit 17 sends an application placement update notification to the UE request acceptance unit 11 via the DB 13.
[0037] The external collaboration unit 18 performs processing in collaboration with a control system (mobile network control system) 3 of the mobile communication network and a control system (central cloud control system) 6 of the central cloud system 5. For example, the external collaboration unit 18 acquires information that can be used for new deployment or re-deployment of applications from each system 3, 6. For example, the external collaboration unit 18 may acquire wireless quality information of the mobile communication network and location information of the user terminal UE from the mobile network control system 3. In addition, the external collaboration unit 18 notifies the mobile network control system 3 of a setting change related to the MEC host 110m.
[0038] (Relocation calculation method) The relocation calculation method according to this embodiment will be described. The relocation calculation unit 16 performs relocation calculation when it cannot find a new relocation destination host 110 for the start-requested application or when it receives a relocation calculation request from the service quality monitoring unit 15.
[0039] When the host 110 for the new deployment destination of the start-requested application cannot be found, the relocation source host is selected according to a predetermined decision rule as a host 110 suitable as a new deployment destination of the start-requested application. For example, when the E2E service request delay time of the start-requested application is short, the MEC host 110m closest to the user equipment UE that originated the application startup request is determined as the relocation source host. On the other hand, the relocation source host is specified in the request for relocation calculation from the service quality monitor unit 15.
[0040] The rearrangement calculation method according to this embodiment will be described with reference to FIGS. FIG. 3 is an explanatory diagram of an outline of the relocation calculation method according to the present embodiment. In the example of FIG. 3, the MEC host A is a relocation source host. The MEC host A has an over-threshold application such as "App a1" and a non-over-threshold application such as "App a2" that is not an over-threshold application. In the relocation calculation method according to the present embodiment, first, a process S1 is executed to preferentially relocate the over-threshold application. In the process S1, a list of candidates (relocation candidate hosts) of the host 110 to which the over-threshold application is to be relocated is created, and the relocation destination host 110 (relocation destination host) is determined from the relocation candidate hosts in the list. In the example of FIG. 3, the MEC host C is determined by the process S1 as the relocation destination host to which the over-threshold application "App a1" of the relocation source MEC host A is to be relocated. Next, when an over-threshold application occurs in the relocation destination MEC host C to which the over-threshold application "App a1" is to be relocated, the process S2 is executed. In process S2, a predetermined relocation calculation is executed to relocate an application from the MEC host C to another host 110, with the application to be relocated to the relocation destination MEC host C being treated as a relocation candidate application.
[0041] 4 is a flowchart showing an example of the procedure of the rearrangement calculation method according to this embodiment. Hereinafter, an example of the procedure of the rearrangement calculation method according to this embodiment will be described with reference to FIG.
[0042] When the host 110 for the new placement of the start-requested application cannot be found, or when a request for relocation calculation is made from the service quality monitor 15, the relocation calculation process (referred to as a "host-determined App relocation process" in this embodiment for the sake of convenience) S1000 shown in Fig. 4 is executed. In the host-determined App relocation process S1000, the relocation source host h is specified as input information.
[0043] In the host-judged App relocation process S1000, steps S1-1 to S1-8 are first executed as process S1.
[0044] (Step S1-1) The placement calculation unit 16 creates an over-threshold App list tap_list_h of the relocation source host h. The over-threshold App list tap_list_h is a list of over-threshold applications in the relocation source host h, and is a list in descending order of the E2E delay rate “(E2E service delay time)÷(request delay threshold)”.
[0045] Next, the process of Loop 1 is executed. In the process of Loop 1, over-threshold applications are extracted in order from the top of the over-threshold App list tap_list_h, and this process is repeated until all over-threshold applications in the over-threshold App list tap_list_h are extracted.
[0046] (Step S1-2) The placement calculation unit 16 extracts one over-threshold application eap_h from the top of the over-threshold App list tap_list_h.
[0047] (Step S1-3) The allocation calculation unit 16 creates a list eap_h_list of candidates (relocation candidate hosts) for the host 110 to which the over-threshold application eap_h is to be relocated. The relocation candidate hosts may include the relocation source host h.
[0048] Examples of conditions for relocation candidate hosts are shown below.
[0049] (Example 1 of conditions for relocation candidate hosts) When the over-threshold application eap_h is relocated from the relocation source host h to the relocation candidate host, the maximum E2E delay rate of the application on each host 110 is equal to or less than the maximum value when the relocation is not performed. In other words, even if the over-threshold application eap_h is relocated from the relocation source host h to the relocation candidate host, the maximum E2E delay rate of the application on each host 110 does not worsen.
[0050] (Example 2 of relocation candidate host conditions) The upper limit of the resource capacity of each host 110 and each link is not exceeded.
[0051] (Step S1-4) The placement calculation unit 16 calculates an expected relocation cost for each relocation candidate host in the list eap_h_list. The expected relocation cost of a certain relocation candidate host is an estimated value of the relocation cost when the over-threshold application eap_h is relocated to the relocation candidate host.
[0052] Here, a method for calculating the expected reallocation cost will be described. The expected relocation cost is calculated by the following formula (1): The expected relocation cost is calculated for the relocation candidate host h'.
[0053] Expected relocation cost = "Relocation cost C_1st in process S1" + "Relocation cost C_2st in process S2" (1)
[0054] The relocation cost C_1st is the cost required to relocate the over-threshold application eap_h. Note that if the relocation candidate host h' is the relocation source host h, the over-threshold application eap_h is not relocated, so the relocation cost C_1st is 0.
[0055] The relocation cost C_1st is calculated using the following formula (2).
[0056] Relocation cost C_1st = w_1 × Cres(eap_h,h,h') + w_2 × Crel(eap_h,h,h') (2)
[0057] w_1 and w_2 are preset weighting coefficients, and "w_1+w_2=1". Cres(eap_h,h,h') is an estimate of the total fluctuation in resource consumption cost in the relocation source host h, the relocation candidate host h', and each link when the over-threshold application eap_h is relocated from the relocation source host h to the relocation candidate host h'. Cres(eap_h,h,h') is calculated from the information (required resources) in the service catalog of the over-threshold application eap_h, the host information (resource unit price) of each host h, h', and the link information (bit unit price) of each link.
[0058] Crel(eap_h,h,h') is the cost required to execute the relocation of the over-threshold application eap_h from the relocation source host h to the relocation candidate host h', and is an estimate of the cost incurred in transferring information expanded in memory by the over-threshold application eap_h between hosts h and h'. Crel(eap_h,h,h') is calculated from the information (required resources) in the service catalog of the over-threshold application eap_h and the link information (bit price) of each link, etc.
[0059] The relocation cost C_2st is calculated using the following equation (3).
[0060] Relocation cost C_2st=w_1×Cres_ev(h',h'')+w_2×Crel_ev(h',h'')×n_ev(h') (3)
[0061] Cres_ev(h',h'') is the expected value of the total change in resource consumption cost at relocation candidate host h' and host h'' when an application on relocation candidate host h' is relocated from relocation candidate host h' to another host h''. Cres_ev(h',h'') is calculated from information in the service catalog (required resources) of various applications on relocation candidate host h' and host information (resource unit price) of each host h', h'', etc.
[0062] For example, in the case where three types of applications "App a1", "App a2", and "App a3" are allocated to a relocation candidate host h', with N1 instances of "App a1", N2 instances of "App a2", and N3 instances of "App a3", if the resource consumption cost of an application type "App aX" on host x is C'res(aX,x), Cres_ev(h',h'') is calculated by the following formula (4): "N_all=N1+N2+N3".
[0063] Cres_ev(h',h'')=(N1÷N_all)×(C'res(a1,h'')-C'res(a1,h')) +(N2÷N_all)×(C'res(a2,h'')-C'res(a2,h')) +(N3÷N_all)×(C'res(a3,h'')-C'res(a3,h')) (4)
[0064] Crel_ev(h',h'') is the expected cost required to execute the relocation of one application on relocation candidate host h' from relocation candidate host h' to another host h''. Crel_ev(h',h'') is calculated from information in the service catalog of various applications on relocation candidate host h' (required memory amount) and the number of applications of each type of application, etc.
[0065] For example, in the case where three types of applications "App a1", "App a2", and "App a3" are allocated to a relocation candidate host h', with N1 instances of "App a1", N2 instances of "App a2", and N3 instances of "App a3", and the memory requirements of each application "App a1", "App a2", and "App a3" are ma1 for "App a1", ma2 for "App a2", and ma3 for "App a3", respectively, Crel_ev(h',h'') is calculated by the following formula (5): "N_all=N1+N2+N3".
[0066] Crel_ev(h',h'')=(N1÷N_all)×ma1+(N2÷N_all)×ma2+(N3÷N_all)×ma3 ···(5)
[0067] n_ev(h') is the expected number of applications that need to be relocated from relocation candidate host h' to eliminate applications that exceed the threshold on relocation candidate host h'. n_ev(h') is calculated from the expected request processing volume Pev(h') of the applications on relocation candidate host h' and the maximum E2E delay rate at relocation candidate host h'. When one application is relocated from relocation candidate host h' to another host, it is expected that the CPU load of relocation candidate host h' will decrease by the expected request processing volume Pev(h'), and therefore the maximum E2E delay rate at relocation candidate host h' is expected to decrease.
[0068] n_ev(h') is found as a solution to the problem of how many applications need to be relocated from relocation candidate host h' to another host in order to resolve the applications exceeding the threshold on relocation candidate host h'. The expected required processing volume Pev(h') is calculated from information (required processing volume) in the service catalog of various applications on relocation candidate host h', the number of applications for each type of application, etc.
[0069] For example, in the case where three types of applications "App a1", "App a2", and "App a3" are allocated to a relocation candidate host h', with N1 instances of "App a1", N2 instances of "App a2", and N3 instances of "App a3", and the required processing volume of each application "App a1", "App a2", and "App a3" is pa1 for "App a1", pa2 for "App a2", and pa3 for "App a3", respectively, the expected required processing volume Pev(h') is calculated by the following formula (6): "N_all=N1+N2+N3".
[0070] Pev(h')=(N1÷N_all)×pa1+(N2÷N_all)×pa2+(N3÷N_all)×pa3 ···(6)
[0071] Note that instead of n_ev(h'), "α×n_ev(h')" may be used, which is n_ev(h') multiplied by coefficient α. Coefficient α is a coefficient set for each host 110, and represents the ease of relocating an application from each host 110. The larger the value of coefficient α, the easier it is to relocate an application.
[0072] For example, the coefficient α may be a wireless mismatch rate that indicates the proportion of user equipment UE that is present in a base station area other than the base station area corresponding to the host 110, among the user equipment UE accommodated in the host 110. In Fig. 1, for example, the greater the proportion (wireless mismatch rate) of user equipment UE that is present in other areas 1 and 3, among the user equipment UE of an application accommodated in the MEC host 110m in area 2, the more applications there are candidates for relocation from the MEC host 110m in area 2, and the relocation is expected to eliminate the mismatch between the area where the user equipment UE is present and the accommodated MEC host 110m, thereby improving the E2E delay rate.
[0073] For example, the coefficient α may be the ratio of stateless applications among the applications on the host 110. Since stateless applications do not manage their state, it is expected that the greater the ratio of stateless applications on the host 110, the less the impact on application services on that host 110 due to relocation will be.
[0074] For example, the coefficient α may be a value proportional to "NU_all÷NUm." NU_all is the total number of user equipments UE of the applications accommodated in the host 110. NUm is the number of user equipments UE that are present in a base station area other than the base station area corresponding to the host 110, among the user equipments UE of the applications accommodated in the host 110.
[0075] For example, the coefficient α may be a value proportional to “NA_all / NSA.” NA_all is the total number of applications on the host 110. NSA is the number of stateless applications on the host 110.
[0076] (Step S1-5) The placement calculation unit 16 judges whether the relocation candidate host with the smallest expected relocation cost among the expected relocation costs of each relocation candidate host calculated in step S1-4 is the relocation source host h. If the result of this judgment is that the relocation candidate host with the smallest expected relocation cost is the relocation source host h (step S1-5, YES), the process returns to step S1-2 and extracts the next highest over-threshold application eap_h from the over-threshold App list tap_list_h. On the other hand, if the relocation candidate host with the smallest expected relocation cost is not the relocation source host h (step S1-5, NO), the process proceeds to step S1-6.
[0077] (Step S1-6) The placement calculation unit 16 determines the relocation candidate host with the smallest expected relocation cost as the host (relocation destination host h') to which the over-threshold application eap_h (application to be relocated) is to be relocated.
[0078] (Step S1-7) The placement calculation unit 16 updates the over-threshold App list tap_list_h. In this update, if there is an over-threshold application that is resolved by the relocation (combination of the over-threshold application eap_h (application to be relocated) and the relocation destination host h′) determined in step S1-6, the placement calculation unit 16 deletes the resolved over-threshold application from the over-threshold App list tap_list_h.
[0079] (Step S1-8) The placement calculation unit 16 determines whether or not an over-threshold application will occur in the relocation destination host h' after the relocation determined in step S1-6 (combination of the over-threshold application eap_h (application to be relocated) and the relocation destination host h'). If the result of this determination is that an over-threshold application will occur in the relocation destination host h' (step S1-8, YES), the process proceeds to step S2. On the other hand, if an over-threshold application will not occur in the relocation destination host h', the process returns to step S1-2 and extracts the next highest over-threshold application eap_h from the over-threshold App list tap_list_h.
[0080] (Step S2) A new host-determined App relocation process S1000 starts to be executed as a recursively called process. The relocation destination host h' is specified as the new relocation source host h' as input information to this new host-determined App relocation process S1000. Therefore, in this new host-determined App relocation process S1000, the relocation destination host h' in the relocation (combination of the over-threshold application eap_h (application to be relocated) and the relocation destination host h') determined in step S1-6 is used as the new relocation source host h' for the relocation calculation.
[0081] Next, when the Loop1 process is completed, the process proceeds to step S10.
[0082] (Step S10) The placement calculation unit 16 checks whether or not there are any over-threshold applications remaining in the over-threshold App list tap_list_h. If the result of this check is that there are no over-threshold applications remaining (step S10, YES), the host-determined App relocation process S1000 is terminated. On the other hand, if there are over-threshold applications remaining (step S10, NO), the process proceeds to step S20.
[0083] (Step S20) The placement calculation unit 16 executes a predetermined placement calculation for the relocation source host h. For example, the application relocation technique described in Non-Patent Document 1 may be used as this predetermined placement calculation. In the application relocation technique described in Non-Patent Document 1, a combination of a relocation target application to be relocated from the relocation source host h to another host 110 and a relocation destination host is selected in consideration of the fluctuation in E2E delay of each application accompanying the relocation and the cost of the relocation. The candidate combination of the relocation target application and the relocation destination host is a combination that satisfies each constraint condition (request delay of each process, resource capacity upper limit of each host and link) after the relocation. After step S20, the host determination type App relocation process S1000 is terminated.
[0084] The predetermined relocation calculation executed for the relocation source host h in step S20 is not limited to the application relocation technique described in Non-Patent Document 1. The predetermined relocation calculation may be any calculation that determines a relocation (a combination of an application to be relocated and a relocation destination host) required to eliminate an application exceeding a threshold from the relocation source host h. Furthermore, it is preferable that the cost associated with the relocation can be kept low. Also, it is preferable that the overall service quality in the entire distributed cloud system 100 is taken into consideration.
[0085] The relocation calculation process called in step S2 may be a relocation calculation process other than the host-determined App relocation process S1000. For example, a relocation calculation process using the application relocation technology described in Non-Patent Document 1 may be called in step S2.
[0086] Fig. 5 is an explanatory diagram showing an example of relocation according to this embodiment. In Fig. 5, first, it is determined by process S1 (host-determined App relocation process S1000) that the over-threshold application "App a1" is relocated from the relocation source MEC host A to the relocation destination MEC host B. Since the over-threshold application "App b1" occurs in the relocation destination MEC host B after this relocation, process S2 (host-determined App relocation process S1000) is executed with the MEC host B as the relocation source host. By this process S2, the host 110 to which the over-threshold application "App b1" is relocated from the MEC host B is determined.
[0087] Fig. 6 is an explanatory diagram showing an example of relocation according to this embodiment. Fig. 6 shows the result of the relocation calculation in step S20. In Fig. 6, the over-threshold application "App a1" remains in the relocation source MEC host A. Therefore, in step S20, the relocation target application selected according to a predetermined selection rule from all applications of the MEC host A and the relocation destination host to which the relocation target application is to be relocated are determined.
[0088] This concludes the explanation of the rearrangement calculation method according to this embodiment.
[0089] Next, an example of a control method according to this embodiment will be described with reference to FIGS.
[0090] First, an example of a control method for new application deployment will be described with reference to Fig. 7. Fig. 7 is a flowchart showing an example of the procedure of a control method for new application deployment according to this embodiment.
[0091] The control device 1 repeats the loop process while incrementing the time t1 by 1 from the initial value "t1=0" until the time t1 reaches a predetermined time.
[0092] (Step S101) When a new arrangement request for an application (activation requested application) occurs, the control device 1 shifts the process to step S102.
[0093] (Step S102) If the control device 1 finds a host 110 capable of newly allocating the start-requested application as a result of the application new allocation calculation, the process proceeds to step S106, and if not, the process proceeds to step S103.
[0094] (Step S103) The control device 1 determines the host 110 (provisional destination host) in which the new application (activation requested application) is to be provisionally deployed.
[0095] (Step S104) The control device 1 designates the temporary destination host as the relocation source host and executes the host-determined App relocation process S1000.
[0096] (Step S105) If the control device 1 executes the host-determined App relocation process S1000 and the start-requested application can be newly allocated to the temporary allocation destination host, the control device 1 proceeds to step S109, otherwise the control device 1 proceeds to step S110.
[0097] (Step S106) If the control device 1 proceeds to step S109 when there is no over-threshold application among the applications already deployed on the host 110 on which the launch-requested application discovered in step S102 can be newly deployed, or proceeds to step S107 when there is no over-threshold application among the applications already deployed on the host 110,
[0098] (Step S107) The control device 1 specifies the host 110 on which the start-requested application found in step S102 can be newly allocated as the re-allocation source host, and executes the host-determined App re-allocation process S1000.
[0099] (Step S108) If, as a result of executing the host-determined App relocation process S1000, the control device 1 proceeds to step S109 if the threshold-exceeding application is improved in the host 110 on which the launch-requested application discovered in step S102 can be newly located, or proceeds to step S110 if not.
[0100] (Step S109) The control device 1 executes new placement of the start-requested application or re-placement of the existing application, and then proceeds to step S111.
[0101] (Step S110) The control device 1 cancels the rearrangement of the existing application and determines that the new arrangement of the start-requested application has failed, and then proceeds to step S111.
[0102] (Step S111) The control device 1 increments the time t1 by 1. After that, the process returns to step S101.
[0103] Next, an embodiment of a control method for relocating an existing application will be described with reference to Fig. 8. Fig. 8 is a flowchart showing an example of the procedure of a control method for relocating an existing application according to this embodiment.
[0104] The control device 1 repeats the loop process while incrementing the time t2 by 1 from the initial value "t2=0" until the time t2 reaches a predetermined time.
[0105] (Step S201) The control device 1 monitors the E2E delay rate of an existing application.
[0106] (Step S202) When the E2E delay rate of the existing application exceeds a predetermined threshold, the control device 1 shifts the process to step S203.
[0107] (Step S203) The control device 1 specifies the host 110, which has an existing application (exceeding-threshold application) whose E2E delay rate exceeds a predetermined threshold, as the relocation source host, and executes the host-determined App relocation process S1000.
[0108] (Step S204) If the execution result of the host-judged App relocation process S1000 indicates that the over-threshold application is improved in the relocation source host, the control device 1 proceeds to step S205, otherwise proceeds to step S206.
[0109] (Step S205) The control device 1 executes the relocation of the existing applications obtained as a result of the execution of the host-determined App relocation process S1000. After that, the process proceeds to step S207.
[0110] (Step S206) The control device 1 stops the rearrangement of the existing applications, and then proceeds to step S207.
[0111] (Step S207) The control device 1 increments the time t2 by 1. After that, the process returns to step S201.
[0112] According to this embodiment, in the distributed cloud system 100, the effect is that it is possible to reduce the costs associated with application relocation by considering the preferential relocation of applications that do not satisfy the service request delay condition (applications exceeding the threshold).
[0113] According to this embodiment, when determining a destination host to which an over-threshold application is to be relocated, the relocation cost C_1st (first relocation cost) when relocating the over-threshold application and the relocation cost C_2st (second relocation cost) when relocating the application from the destination host to which the over-threshold application is to be relocated are taken into consideration, thereby achieving the effect of reducing the overall relocation cost.
[0114] According to this embodiment, a relocation destination host is selected from among the relocation candidate hosts based on the expected number of applications to be relocated from the relocation candidate host that is required to eliminate applications exceeding the threshold at the relocation candidate host, thereby achieving the effect of reducing the number of applications to be relocated.
[0115] This will, for example, make it possible to improve the overall service quality in wireless networks, thereby contributing to Goal 9 of the United Nations-led Sustainable Development Goals (SDGs), which is to "Build resilient infrastructure, promote sustainable industrialization and foster innovation."
[0116] Although an embodiment of the present invention has been described in detail above with reference to the drawings, the specific configuration is not limited to this embodiment, and design changes and the like that do not deviate from the gist of the present invention are also included.
[0117] In addition, a computer program for implementing the functions of each of the above-mentioned devices may be recorded on a computer-readable recording medium, and the program recorded on the recording medium may be read into a computer system and executed. Note that the "computer system" referred to here may include hardware such as an OS and peripheral devices. In addition, the term "computer-readable recording medium" refers to writable non-volatile memory such as a flexible disk, an optical magnetic disk, a ROM, or a flash memory, a portable medium such as a DVD (Digital Versatile Disc), or a storage device such as a hard disk built into a computer system.
[0118] Furthermore, the term "computer-readable recording medium" includes devices that retain a program for a certain period of time, such as volatile memory (e.g., DRAM (Dynamic Random Access Memory)) within a computer system that serves as a server or client when a program is transmitted via a network such as the Internet or a communication line such as a telephone line. The program may be transmitted from a computer system in which the program is stored in a storage device or the like to another computer system via a transmission medium, or by a transmission wave in the transmission medium. Here, the "transmission medium" that transmits the program refers to a medium that has a function of transmitting information, such as a network (communication network) such as the Internet or a communication line (communication line) such as a telephone line. The program may be for implementing some of the functions described above, or may be a so-called differential file (differential program) that can implement the functions described above in combination with a program already recorded in the computer system. [Explanation of symbols]
[0119] 1...control device, 11...user equipment (UE) request reception unit, 13...database (DB), 14...distributed cloud information collection unit, 15...service quality monitoring unit, 16...placement calculation unit, 17...placement control unit, 18...external linkage unit, 100...distributed cloud system, 110m...MEC host, 110c...central cloud host, UE...user equipment, gNB...base station, MBN...mobile backhaul network, MNW...wireless access network, INW...Internet
Claims
1. A control device that determines an information processing device that executes an application for realizing a service to be provided to a user in an information processing system configured with a plurality of information processing devices, creating a list of relocation candidate information processing devices that are candidates for a destination to which a relocation target application, which is an application that does not satisfy a service request delay condition that is a condition related to an end-to-end service delay time required by a user, is to be first relocated; Calculate an expected relocation cost consisting of a first relocation cost required for a first relocation of the application to be relocated for each of the relocation candidate information processing devices in the list and a second relocation cost required for a second relocation of the application on the relocation candidate information processing device to another information processing device, for one or more second relocations; determining the relocation candidate information processing device having the smallest expected relocation cost as a relocation destination information processing device for the application to be relocated; a placement calculation unit that, when an application that does not satisfy a service request delay condition occurs in the relocation destination information processing device to which the relocation target application is first relocated, recursively repeats a predetermined relocation calculation for secondly relocating an application from the relocation destination information processing device to another information processing device as a relocation candidate application, until no application that does not satisfy the service request delay condition occurs in the relocation destination information processing device or until a predetermined number of repetitions is reached; a placement control unit that executes application placement control based on the relocation content determined by the placement calculation unit; A control device comprising:
2. The relocation candidate information processing device is an information processing device in which a maximum value of an end-to-end delay rate in an application on each of the information processing devices does not deteriorate when the relocation target application is first relocated. The control device according to claim 1 .
3. The first relocation cost is A fluctuation amount of resource consumption cost when the application to be relocated is relocated for the first time; and A required cost for executing the first relocation of the application to be relocated is calculated based on the required cost. The control device according to claim 1 .
4. The second relocation cost is an expected value of a fluctuation amount of resource consumption cost when the application to be allocated to the relocation candidate information processing device is secondly relocated to another information processing device; an expected value of a required cost in executing a second relocation of the application to be relocated to the relocation candidate information processing device to another information processing device; an expected value of the number of applications to be secondly relocated from the relocation candidate information processing device to another information processing device in order to eliminate applications that do not satisfy a service request delay condition occurring in the relocation candidate information processing device, The control device according to claim 1 or 3.
5. 1. A control method for determining an information processing device that executes an application for realizing a service to be provided to a user in an information processing system including a plurality of information processing devices, comprising: The control device, creating a list of relocation candidate information processing devices that are candidates for a destination to which a relocation target application, which is an application that does not satisfy a service request delay condition that is a condition related to an end-to-end service delay time required by a user, is to be first relocated; Calculate an expected relocation cost consisting of a first relocation cost required for a first relocation of the application to be relocated for each of the relocation candidate information processing devices in the list and a second relocation cost required for a second relocation of the application on the relocation candidate information processing device to another information processing device, for one or more second relocations; determining the relocation candidate information processing device having the smallest expected relocation cost as a relocation destination information processing device for the application to be relocated; a placement calculation step of recursively repeating a predetermined placement calculation for secondly placing an application from the relocation destination information processing device to another information processing device as a relocation candidate application when an application not satisfying a service request delay condition occurs in the relocation destination information processing device to which the relocation target application is first relocated, until no application not satisfying the service request delay condition occurs in the relocation destination information processing device or until a predetermined upper limit of the number of repetitions is reached; a placement control step in which the control device executes application placement control based on the relocation content determined in the placement calculation step; A control method comprising:
6. In an information processing system including a plurality of information processing devices, a computer of a control device that determines an information processing device that executes an application for realizing a service to be provided to a user includes: creating a list of relocation candidate information processing devices that are candidates for a destination to which a relocation target application, which is an application that does not satisfy a service request delay condition that is a condition related to an end-to-end service delay time required by a user, is to be first relocated; Calculate an expected relocation cost consisting of a first relocation cost required for a first relocation of the application to be relocated for each of the relocation candidate information processing devices in the list and a second relocation cost required for a second relocation of the application on the relocation candidate information processing device to another information processing device, for one or more second relocations; determining the relocation candidate information processing device having the smallest expected relocation cost as a relocation destination information processing device for the application to be relocated; a placement calculation step of recursively repeating a predetermined placement calculation for secondly placing an application from the relocation destination information processing device to another information processing device as a relocation candidate application when an application not satisfying a service request delay condition occurs in the relocation destination information processing device to which the relocation target application is first relocated, until no application not satisfying the service request delay condition occurs in the relocation destination information processing device or until a predetermined upper limit of the number of repetitions is reached; a placement control step of executing application placement control based on the relocation content determined in the placement calculation step; A computer program for executing the above.
Citation Information
Patent Citations
Virtual machine rearranging apparatus
JP2016057851A
Profile-based SLA guarantees under workload migration in distributed cloud
JP2018198068A
Method and apparatus for providing edge computing services
US20210243264A1
Load distribution device and method
WO2014192132A1
