Communication control device and communication control method

Abstract

This communication control device includes: a prediction unit configured to calculate a predicted value of throughput for each of a plurality of edge servers on the basis of a correspondence relationship between a terminal and the edge server, which is set as a connection destination, the correspondence relationship being in a corresponding relay unit among a plurality of relay units that are associated with one or more mutually different terminals and that transmit data from the associated terminal to the edge server; and a control unit configured to change the connection destinations of relay units of which the connection destination is an edge server having a predicted value that exceeds an upper limit value to an edge server of which the predicted value is less than the upper limit value. The communication control device is thereby able to suppress a deterioration in the quality of communication with the edge servers.

Description

Communication control device and communication control method

[0001] The present invention relates to a communication control device and a communication control method.

[0002] Conventionally, due to the congestion or failure situation of the edge server, a technique for routing user access to a secondary edge server instead of the primary edge server that is normally used has been studied (for example, Patent Document 1).

[0003] Also, in MEC (Multi-access Edge Computing), a technique has been proposed to directly monitor the communication congestion situation and switch the edge server to which the terminal 50 is connected according to the congestion situation.

[0004] Japanese Patent Application Laid-Open No. 2019-41266

[0005] 3GPP TS 23.503 version 17.10.0 Release 17, "5G; Policy and charging control framework for the 5G System (5GS); Stage 2", [online], Internet <https: / / www.etsi.org / deliver / etsi_ts / 123500_123599 / 123503 / 17.10.00_60 / ts_123503v171000p.pdf>

[0006] However, in the prior art, since switching is performed according to the communication status of the edge server, it may be a post-coping measure for switching after the communication with the edge server becomes congested, or since switching is performed based on the instantaneous throughput of the edge server, server switching may occur frequently due to a temporary increase or decrease in throughput. As a result, there is a problem that the communication quality of the service may deteriorate.

[0007] The present invention has been made in view of the above points, and an object thereof is to suppress a decrease in the quality of communication with an edge server.

[0008] To solve the above problem, the communication control device includes a prediction unit configured to calculate a predicted throughput value for each edge server based on the correspondence between the terminals and the edge servers in a plurality of relay units, each of which is associated with one or more different terminals and transmits data from the associated terminals to an edge server among a plurality of edge servers that is set as the connection destination, and a control unit configured to change the connection destination of the relay unit to an edge server whose predicted value exceeds an upper limit, if the predicted value exceeds an upper limit.

[0009] This makes it possible to suppress the degradation of communication quality with edge servers.

[0010] This figure shows an example configuration of the MEC (Multi-access Edge Computing) system in the first embodiment. This figure illustrates the correspondence between terminal 50, AN40, VTM21, and edge server 60. This figure shows an example hardware configuration of the NW controller 10 in the first embodiment. This figure shows an example functional configuration of the NW controller 10 in the first embodiment. This flowchart illustrates an example of the processing procedure for predicting throughput for each edge server 60 in the first embodiment. This flowchart illustrates an example of the processing procedure for switching the destination edge server 60 in the first embodiment. This flowchart illustrates an example of the processing procedure for predicting throughput for each edge server 60 in the second embodiment.

[0011] Embodiments of the present invention will be described below with reference to the drawings. Figure 1 is a diagram showing an example configuration of a MEC (Multi-access Edge Computing) system in a first embodiment. In Figure 1, the MEC system includes a plurality of edge servers 60 located at different locations (such as edge server 60a at location A and edge server 60b at location B), a plurality of GW devices 20 such as GW devices 20-1 and 20-2, a plurality of AN40 such as AN40-1 and 40-2, a plurality of terminals 50, an access line authentication server 30, and an NW controller 10.

[0012] An edge server 60 is one or more computers (physical or virtual) that process data transmitted from a terminal 50. For example, since low-latency processing is required for data transmitted from terminal 50, the edge server 60 processes this data. Each edge server 60 is capable of performing the same processing. A "location" is a concept corresponding to the geographical location where the edge server 60 is deployed. The locations of the edge servers 60 are located closer to the terminal 50 than the data centers where the cloud servers are deployed. For example, if in Japan there are two data centers where the cloud servers are deployed, one in Tokyo and one in Osaka, the edge servers 60 are expected to be installed at base stations, prefectural data center locations, or regional data center locations.

[0013] Terminal 50 is a communication terminal that transmits data to be processed by the edge server 60. For example, XR devices and IoT devices are examples of terminal 50. One terminal 50 can connect to multiple types of access lines. Terminal 50 communicates with the edge server 60 by switching the access network (AN40) to which it is connected each time, such as using WLAN40 indoors, a dedicated mobile network in smart factories, and a public mobile network outdoors.

[0014] AN40 is a network (access line) used to accommodate terminals 50 and transmit data from terminals 50 to edge servers 60. Examples of AN40 include public mobile networks, dedicated mobile networks, and fixed networks. Terminals 50 connect wirelessly to AN40 via base stations owned by AN40. AN40 includes RAN40 (Radio Access Network) and the core network. The core network is the central part of a network owned by a telecommunications carrier, connecting switches and connecting to the Internet and the communication networks of other telecommunications carriers.

[0015] Each AN 40 connects to the GW device 20. Each GW device 20 connects to each edge server 60 via the network. The GW device 20 is one or more computers on which a VTM 21 (Virtual Tunneling Machine) operates, and in response to instructions from the NW controller 10, it creates, deletes, switches the connection destination of the VTM 21, etc.

[0016] VTM21 is a virtual gateway that corresponds to one AN40 and simultaneously to one edge server 60. It generates a tunnel (closed network connection) for transferring data from terminals 50 connected to the AN40 that are associated with VTM21 to the edge server 60 that corresponds to VTM21, and transfers the data through this tunnel. This tunnel may be an encrypted tunnel like a VPN (Virtual Private Network) or an unencrypted tunnel like a GRE (Generic Routing Encapsulation).

[0017] The access line authentication server 30 connects to the core network of each AN 40 via the network and also connects to the NW controller 10. The access line authentication server 30 is one or more computers in MEC (Multi-Access Edge Computing) that processes authentication requests made by terminals 50 from various access lines in addition to mobile communication lines. Specifically, the access line authentication server 30 authenticates closed network connections from terminals 50 housed in each AN 40 to the edge server 60, and only allows communication to the edge server 60 for authenticated terminals 50. Therefore, the information of terminals 50 housed in each AN 40 is known to the access line authentication server 30. For example, the access line authentication server 30 receives terminal-specific information such as username, password, and SIM information from the terminal 50, compares it with terminal-specific information stored in the access line authentication server 30 beforehand, and determines that authentication was successful if the terminal-specific information from the terminal 50 matches the terminal-specific information stored in the access line authentication server 30; otherwise, it determines that authentication failed. Specifically, this is intended for APN settings for LTE and 5G.

[0018] The NW controller 10 is one or more computers connected to each GW device 20 via the network. The NW controller 10 performs resource operations (creation, deletion) on the VTM 21 and issues instructions to the VTM 21 to create a tunnel with the edge server 60.

[0019] Figure 2 is a diagram illustrating the correspondence between terminal 50, AN40, VTM21, and edge server 60.

[0020] One AN40 can accommodate one or more terminals 50. Figure 3 shows an example in which two terminals 50 are accommodated in AN40-1 and AN40-2, respectively.

[0021] One or more VTMs 21 can be connected to a single AN 40, and each VTM 21 can connect to a maximum of one AN 40. Furthermore, one VTM 21 can correspond to one or more terminals 50 housed in the AN 40 to which it is connected. Also, one VTM 21 can simultaneously connect to one edge server 60. Therefore, when distributing the connections of multiple terminals 50 housed in a single AN 40 to multiple edge servers 60, a number of VTMs 21 equal to the number of edge servers 60 to which the connections are made will be generated for that AN 40.

[0022] In Figure 2, VTM21-1a is connected to AN40-1, and VTM21-2a and VVTM21-2b are connected to AN40-2. VTM21-1a corresponds to two terminals 50 housed in AN40-1 and provides a private network connection to these terminals 50 with the edge server 60a. VTM21-2a corresponds to one of the two terminals 50 housed in AN40-2 and provides a private network connection to this terminal 50 with the edge server 60a. VTM21-2a corresponds to the other of the two terminals 50 housed in AN40-2 and provides a private network connection to this terminal 50 with the edge server 60b.

[0023] Furthermore, a VTM21 that supports multiple terminals 50 generates a tunnel for each terminal 50. That is, a tunnel between the VTM21 and the edge server 60 is generated for each terminal 50. Therefore, in Figure 2, a VTM21-1a that supports two terminals 50 generates a tunnel between each terminal 50 and the edge server 60a, and uses each terminal 50 to establish a closed network connection between the terminal 50 and the edge server 60a.

[0024] The correspondence between terminal 50 and edge server 60 can be changed by configuring the VTM21. Specifically, each VTM21 is configured with the IP address of the terminal 50 it corresponds to and information (identification name) of the edge server 60 to which the VTM21 is connected. The VTM21 sends packets related to the configured IP address of terminal 50 to the edge server 60a configured as the connection destination. Therefore, by changing the edge server 60 to which the VTM21 is connected, the degree of congestion of each edge server 60 can be changed. Congestion of an edge server 60 refers to bandwidth constraints in the part connecting the site where the edge server 60 is installed to the outside. For example, in Figure 2, if edge server 60a is congested, changing the connection destination of VTM21-2a from edge server 60a to edge server 60b will reduce the degree of congestion of edge server 60a. As a result, the degree of congestion of each edge server 60 can be made more uniform.

[0025] In Figure 2, an example is shown where one AN 40 is associated with each connected edge server 60, but multiple VTM 21s may be associated with the same edge server 60 for a single AN 40. For example, by associating a VTM 21 with each group of terminals 50, the connected edge server 60 can be dynamically changed on a per-group basis of terminals 50. An example of a group of terminals 50 is multiple terminals 50 belonging to the same company.

[0026] Figure 3 shows an example of the hardware configuration of the NW controller 10 in the first embodiment. The NW controller 10 in Figure 3 includes a drive device 100, an auxiliary storage device 102, a memory device 103, a processor 104, and an interface device 105, etc., which are all interconnected by bus B.

[0027] The program that enables processing by the NW controller 10 is provided on a recording medium 101 such as a CD-ROM. When the recording medium 101 containing the program is set in the drive device 100, the program is installed from the recording medium 101 to the auxiliary storage device 102 via the drive device 100. However, the program does not necessarily have to be installed from the recording medium 101; it may also be downloaded from another computer via the network. The auxiliary storage device 102 stores the installed program as well as necessary files and data.

[0028] The memory device 103 reads and stores a program from the auxiliary storage device 102 when a program startup command is received. The processor 104 is either a CPU or a GPU (Graphics Processing Unit), or both a CPU and a GPU, and executes functions related to the NW controller 10 according to the program stored in the memory device 103. The interface device 105 is used as an interface for connecting to the network.

[0029] Figure 4 is a diagram showing an example of the functional configuration of the NW controller 10 in the first embodiment. As shown in Figure 4, the NW controller 10 has a prediction unit 11 and a control unit 12. Each of these units is realized by processing that one or more programs installed in the NW controller 10 cause the processor 104 to execute.

[0030] The prediction unit 11 predicts the throughput for each edge server 60 based on the correspondence between terminals 50 and edge servers 60 in multiple VTMs 21. If there is an edge server 60 whose predicted value exceeds the upper limit, the prediction unit 11 instructs the control unit 12 to switch the connection destination of the VTM 21 that is connected to that edge server 60.

[0031] The control unit 12 performs tasks such as generating VTM21, setting VTM21, deleting VTM21, and switching the connection destination of VTM21. For example, in response to an instruction from the prediction unit 11 to switch the connection destination, the control unit 12 executes tunnel switching control (switching process of the connection destination edge server 60) for the VTM21 related to the instruction. At this time, the control unit 12 sums the predicted throughput values ​​of all terminals 50 that the VTM21 to be switched to corresponds to, and selects an edge server 60 as the switching destination that does not exceed the upper limit of throughput when the sum is added to the current predicted value. Note that the upper limit of throughput may differ for each site (each edge server 60).

[0032] The following describes the processing procedure performed by the NW controller 10 in the first embodiment. Figure 5 is a flowchart illustrating an example of the processing procedure for predicting throughput for each edge server 60 in the first embodiment. The processing in Figure 5 is performed periodically. The execution interval may be set externally, such as once every hour or once every 10 minutes.

[0033] In step S101, the prediction unit 11 obtains connection terminal information for each VTM21 corresponding to each AN40 from the access line authentication server 30. Connection terminal information for each VTM21 refers to information that includes identification information (e.g., IP address) of the terminal 50 (connection terminal) corresponding to the VTM21 for each VTM21.

[0034] Next, the prediction unit 11 obtains connection destination edge server information from each VTM 21, and also obtains the maximum value from the throughput history of each terminal 50 related to the connection terminal information obtained for the VTM 21 as the predicted throughput value for each terminal 50 (S102). Connection destination edge server information is, for example, the identification name of the edge server 60 that the VTM 21 will connect to.

[0035] Furthermore, connected terminal information and destination edge server information can be obtained using the authentication information when terminal 50 connects to AN40. In this case, the NW controller 10 does not need to access each location.

[0036] Next, the prediction unit 11 calculates a predicted throughput value for each edge server 60 by summing the predicted throughput values ​​(maximum values ​​from past history) of the connected terminals of the VTM 21 that connect to the edge server 60 (S103).

[0037] Figure 6 is a flowchart illustrating an example of the processing procedure for switching the destination edge server 60 in the first embodiment. The processing procedure in Figure 6 is executed following the completion of the processing procedure in Figure 5.

[0038] In step S201, the prediction unit 11 compares the predicted throughput value calculated by the processing procedure in Figure 5 with the upper throughput limit set for each edge server 60.

[0039] Next, the prediction unit 11 determines whether or not there are any edge servers 60 whose predicted value exceeds the upper limit (S202). If there are no such edge servers 60 (No in S203), the processing procedure in Figure 6 ends.

[0040] If there is one or more matching edge servers 60 (hereinafter referred to as "switching source candidates") (Yes in S202), the prediction unit 11 instructs the control unit 12 to switch the connection destination of the VTM 21 to the switching source candidate. The control unit 12 executes steps S203 and later in response to the instructions from the prediction unit 11.

[0041] In step S203, the control unit 12 selects one switching source candidate as the processing target (hereinafter referred to as the "target server"), starting with the edge server 60 with the relatively largest difference between the upper limit and the predicted value (upper limit - predicted value). The larger the difference between the upper limit and the predicted value, the higher the likelihood of congestion occurring. Therefore, the intention is to reduce throughput starting with the edge server 60 that is most likely to experience congestion.

[0042] Next, the control unit 12 determines whether there is an edge server 60 of a neighboring site of the site to which the target server belongs (hereinafter referred to as "neighboring server") whose predicted value is less than the upper limit value (S204). The neighboring site of a certain site refers to, for example, a site whose geographical distance from a certain site is below a predetermined value. A list of neighboring sites of each site may be stored in the NW controller 10 in advance. In this case, the control unit 12 may identify the interest rate server of the target server based on the list.

[0043] If there is no corresponding neighboring server (No in S204), the process proceeds to step S213. If there is one or more corresponding neighboring servers (Yes in S204), the control unit 12 selects, as the switching target server, the edge server 60 with the largest (upper limit value - predicted value) among the corresponding neighboring servers (that is, the edge server 60 predicted to have the most margin) (S205).

[0044] Next, the control unit 12 determines whether there is a VTM 21 that uses the target server as a connection destination and has not yet been selected as a processing target in steps S207 and subsequent steps (hereinafter referred to as "unselected VTM 21") (S206). If all of the VTMs 21 that use the target server as a connection destination have already been selected as processing targets (No in S206), the process proceeds to step S213.

[0045] If there is an unselected VTM 21 (Yes in S206), the control unit 12 selects, as the target VTM 21, the VTM 21 with the largest predicted throughput among the unselected VTMs 21 (S207). The predicted throughput of the VTM 21 is the sum of the predicted throughputs of the connection terminals of the VTM 21.

[0046] Next, the control unit 12 determines whether the predicted throughput of the switching target server is below the upper limit value when the connection destination of the target VTM 21 is switched to the switching target server (S208). Specifically, the control unit 12 determines whether the predicted value of the switching target server + the predicted value of the target VTM 21 (hereinafter referred to as "predicted value after switching") is below the upper limit value of the switching target server.

[0047] If the predicted value after switching exceeds the upper limit value of the target server (No in S208), the process proceeds to step S213. If the predicted value after switching is less than or equal to the upper limit value of the target server (Yes in S208), the control unit 12 adds the set of {target VTM21 (identification information), target server (identification name)} to the switching list (S209). The switching list is data in a list format that stores the set of the VTM21 whose connection destination is to be switched and the edge server 60 that is the switching destination of the VTM21 as one element, and its initial value is empty.

[0048] Subsequently, the control unit 12 changes the predicted value of the throughput of the target server to the predicted value after switching (S210). This is to prevent the predicted value of the throughput of the target server from being underestimated in the next loop, as the processes after step S203 are loop processes that are repeatedly executed.

[0049] Subsequently, the control unit 12 subtracts the predicted value of the throughput of the target VTM21 from the predicted value of the throughput of the target server (S211). That is, the predicted value of the target server is updated to the value after the connection destination of the target VTM21 is switched.

[0050] Subsequently, the control unit 12 determines whether the predicted value of the target server is less than or equal to the upper limit value of the target server (S212). If the predicted value of the target server exceeds the upper limit value (No in S212), the processes after step S206 are repeated.

[0051] If the predicted value for the target server is less than or equal to the upper limit (Yes in S212), the control unit 12 changes the connection destination of some or all of the VTM 21s that connect to the target server according to the switching list (S213). Specifically, the control unit 12 takes elements one by one from the switching list and sets the connection destination of the VTM 21 included in the taken element to the edge server 60 included in that element. As a result, the connection destination of the VTM 21 switches to the edge server 60. This means that data from the terminal 50 corresponding to the VTM 21 will be transferred to the edge server 60. Note that the switching list becomes empty after step S213 is executed. Note that if the switching list is empty, the connection destination of the VTM 21s that connect to the target server will not be changed.

[0052] Next, the control unit 12 determines whether steps S203 and later have been executed for all switching source candidates (S214). If there are any switching source candidates that have not been processed (No in S224), steps S203 and later are repeated. If steps S203 and later have been executed for all switching source candidates (Yes in S214), the processing procedure in Figure 6 is completed.

[0053] As described above, according to the first embodiment, by switching the edge server 60 to which each VTM 21 is connected using the predicted throughput value of each edge server 60, it becomes possible to prevent congestion at each edge server 60 before communication congestion occurs. Here, since the terminals connected to the VTM 21 corresponding to each AN 40 obtained from the access line authentication server 30 may include terminals 50 that have not yet started communication, it becomes possible to predict the throughput when such terminals 50 have started communication.

[0054] Furthermore, in this embodiment, it is possible to anticipate the most congested case when all terminals 50 counted as connected terminals are communicating, thus suppressing frequent server switching compared to a method that switches based on real-time changing throughput.

[0055] As described above, according to this embodiment, it is possible to suppress the deterioration of the quality of communication with the edge server 60.

[0056] Next, a second embodiment will be described. The differences between the second embodiment and the first embodiment will be described. Points not specifically mentioned in the second embodiment may be the same as in the first embodiment.

[0057] In the second embodiment, the throughput prediction process for each edge server 60 differs from that in the first embodiment. Specifically, in the second embodiment, the processing procedure in Figure 5 is replaced by the processing procedure in Figure 7.

[0058] Figure 7 is a flowchart illustrating an example of the processing procedure for predicting throughput for each edge server 60 in the second embodiment.

[0059] In step S301, the prediction unit 11 obtains the number of connected terminals for each VTM21 corresponding to each AN40 from the access line authentication server 30. The number of connected terminals for each VTM21 refers to the number of terminals 50 (connected terminals) corresponding to each VTM21.

[0060] Next, the prediction unit 11 obtains information about the destination edge server from each VTM 21 (S302).

[0061] Next, the prediction unit 11 calculates the total number of connected terminals of the VTM 21 that connect to the edge server 60 as the number of connected terminals for that edge server 60 (S303). The number of connected terminals calculated for a given edge server 60 is the total number of terminals 50 that have the potential to communicate with that edge server 60.

[0062] Next, the prediction unit 11 sets a predetermined fixed value for the predicted throughput of each terminal 50 for each edge server 60, and calculates the predicted throughput of the edge server 60 by multiplying the number of connected terminals calculated for that edge server 60 by this fixed value (S304).

[0063] In step S301, an example was shown in which the actual number of connected terminals is obtained. However, for example, the prediction unit 11 may linearly predict the number of connected terminals (for example, the predicted number of connected terminals after a certain period) based on the trend of the number of connected terminals obtained periodically in the past, taking such trends into account. A machine learning model may be used for such predictions.

[0064] As described above, the second embodiment can also obtain the same effects as the first embodiment.

[0065] In this embodiment, VTM21 is an example of a relay unit. NW controller 10 is an example of a communication control device.

[0066] Although embodiments of the present invention have been described in detail above, the present invention is not limited to these specific embodiments, and various modifications and changes are possible within the scope of the gist of the present invention as described in the claims.

[0067] 10 Network Controller 11 Prediction Unit 12 Control Unit 20 Gateway Device 21 VTM 30 Access Line Authentication Server 40 AN 50 Terminal 60 Edge Server 100 Drive Device 101 Recording Medium 102 Auxiliary Storage Device 103 Memory Device 104 Processor 105 Interface Device B Bus

Claims

1. A communication control device comprising: a prediction unit configured to calculate a predicted throughput value for each edge server based on the correspondence between terminals and edge servers in a plurality of relay units, each of which is associated with one or more different terminals, and which transmits data from the associated terminals to an edge server among a plurality of edge servers that is set as the connection destination; and a control unit configured to change the connection destination of a relay unit that is set as the connection destination to an edge server whose predicted value exceeds an upper limit to an edge server whose predicted value is less than an upper limit.

2. The communication control device according to claim 1, characterized in that the prediction unit is configured to calculate a predicted throughput value for each edge server based on the throughput history of each terminal and the corresponding relationship.

3. The communication control device according to claim 1, characterized in that the prediction unit is configured to determine the number of connected terminals for each edge server based on the respective correspondence relationships, and to calculate a predicted throughput value for each edge server by multiplying the number of connected terminals for each edge server by a fixed value.

4. A communication control method characterized in that a computer executes the following steps: a prediction step, which calculates a predicted throughput value for each edge server based on the correspondence between terminals and edge servers in a plurality of relay units, each of which is associated with one or more different terminals, and which transmits data from the associated terminals to an edge server among a plurality of edge servers that is set as the connection destination; and a control step, which changes the connection destination of a relay unit that is set as the connection destination to an edge server whose predicted throughput value exceeds an upper limit to an edge server whose predicted throughput value is less than an upper limit.

Images