JP7809032B2 - Proxy device, proxy method and program - Google Patents

Description

This disclosure relates to a proxy device, a proxy method, and a program.

5GC (5G Core) adopts a service-based architecture (SBA) (see, for example, Non-Patent Documents 1-2).
Techniques are known for selecting an instance of a Network Function (NF) within an SBA.

3GPP (registered trademark) TS 23.501 "System architecture for the 5G System (5GS)" Tulja Vamshi Kiran Buyakar et al. , “Prototyping and Load Balancing the Service Based Architecture of 5G Core using NFV”, 2019 IEEE Conference on Network Softwarization (Netsoft), p. 228-232 Liang Bao et al. , “Performance Modeling and Workflow Scheduling of Microservice-Based Applications in Clouds”, IEEE TRANSACTIONS ON PARALLEL AND DISTRIBUTED SYSTEMS, VOL. 30, NO. 9, SEPTEMBER 2019

However, conventionally, the selection of an instance of an NF (NF instance) is performed by a producer based only on the load information of the NF.
Note that an NF instance must be selected from a set of multiple NF instances called an NF set, and the NF instances in an NF set may be configured in a geographically redundant manner.
Furthermore, since a proxy device called a Service Communication Proxy (SCP) is placed between the NF instances, it is conceivable that the processing delay between the NF instances will increase.

The present disclosure has been made in consideration of these circumstances, and aims to provide a proxy device, proxy method, and program that can improve efficiency by selecting an NF instance while taking network delay into consideration.

As one configuration example, the apparatus includes a request acquisition unit that acquires a request from an NF instance of a client, and an NF instance selection unit that selects an NF instance of a server based on a processing delay including a network delay in accordance with the request acquired by the request acquisition unit, wherein the NF instance selection unit acquires, as the processing delay, a processing delay of the entire process in which the request is transmitted from the NF instance of the client to its own device, the request is transferred from its own device to the NF instance of the server, and a response is transmitted from the NF instance of the server to the NF instance of the client via its own device, and a proxy device that selects an NF instance of the server that can process the entire process within the timeout value of the request based on a processing delay, and the NF instance selection unit calculates an estimate of the transfer timeout value by subtracting the time to send the request from the NF instance of the client from the time to transfer the request from the device itself, and subtracting the result from the timeout value of the request, and selects an NF instance of the server that produces a smaller result than the calculated estimated value by adding the network delay time for the request to be transferred from the device itself to the NF instance of the server, the processing time at the server, and the network delay time for the response to be sent from the NF instance of the server to the device itself .
As one configuration example, the apparatus includes a request acquisition unit that acquires a request from an NF instance of a client, and an NF instance selection unit that selects an NF instance of a server based on a processing delay including a network delay in accordance with the request acquired by the request acquisition unit, wherein the NF instance selection unit acquires, as the processing delay, a processing delay of the entire process in which the request is transmitted from the NF instance of the client to its own device, the request is transferred from its own device to the NF instance of the server, and a response is transmitted from the NF instance of the server to the NF instance of the client via its own device, and selects the entire process based on the acquired processing delay of the entire process. a proxy device that selects an NF instance of the server that can process the request within a timeout value, and the NF instance selection unit calculates an estimate of a transfer timeout value by subtracting from the timeout value of the request the sum of the network delay time when the request is transferred from the NF instance of the client to the device itself, the processing time at the device itself, and the network delay time when the response is sent from the device itself to the NF instance of the client, and the result is compared with the calculated estimated value to select the NF instance of the server that produces a smaller result of adding the network delay time when the request is transferred from the device itself to the NF instance of the server, the processing time at the server, and the network delay time when the response is sent from the NF instance of the server itself to the device itself.

The proxy device, proxy method, and program disclosed herein can improve efficiency by selecting an NF instance taking network delay into consideration.

FIG. 1 is a diagram illustrating a schematic configuration example of 5GC architecture. FIG. 10 is a diagram illustrating an example of a communication model for NF/NF service interaction. FIG. 10 is a diagram illustrating an example of a communication model for NF/NF service interaction. FIG. 10 is a diagram illustrating an example of a communication model for NF/NF service interaction. FIG. 10 is a diagram illustrating an example of a communication model for NF/NF service interaction. FIG. 1 is a diagram illustrating a schematic example of an overall configuration using an SCP. FIG. 2 is a diagram illustrating a schematic functional block configuration of a proxy device according to an embodiment. FIG. 2 is a diagram illustrating a schematic configuration of a first model according to the embodiment. FIG. 10 is a diagram illustrating a schematic configuration of a second model according to the embodiment. FIG. 2 is a diagram illustrating a schematic configuration of an assumed model according to the embodiment. FIG. 10 is a diagram illustrating an example of an instance processing delay retention table according to the embodiment. FIG. 10 is a diagram illustrating an example of a processing delay holding table of the proxy agent according to the embodiment. FIG. 10 is a diagram illustrating an example of an instance registration request according to the embodiment.

The following describes the embodiments with reference to the drawings.

[5GC Architecture]
FIG. 1 is a diagram illustrating a schematic configuration example of 5GC architecture 1 (for details, see Non-Patent Document 1).
Figure 1 shows NSSF (Network Slice Selection Function) 11, NEF (Network Exposure Function) 12, and NRF (Network Repository). Function) 13, PCF (Policy Control Function) 14, UDM (Unified Data Management) 15, AF (Application Function) 16, EASDF (Edge Application Server Discovery Function) 17, NSSAAF (Network Slice-specific and SNPN Authentication and Authorization Function) 18, AUSF (Authentication Server Function) 19, AMF (Access and Mobility) Management Function) 20, SMF (Session Management Function) 21, SCP 22, NSACF (Network Slice Admission Control) Function) 23, UE (User Equipment) 24, (R)AN ((Radio) Access Network) 25, UPF (User Equipment) 25, A Plane Function 26 and a Data Network (DN) 27 are shown.

In 5G, SBA is adopted as the control plane (C-Plane) configuration. In SBA, the C-Plane is divided into NFs based on its functional units. Each NF is loosely coupled, and communication between NFs is achieved using APIs (Application Programming Interfaces).

2A-2D, an example of a communication model between NFs is shown.
2A and 2B are examples of the Direct method, and FIGS. 2C and 2D are examples of the Indirect method.
2A, 2B, 2C, and 2D are diagrams each showing an example of a communication model for NF/NF service interaction.

In the example of Figure 2A, consumer 111 sends a service request to producer 112, producer 112 sends a service response to consumer 111, and consumer 111 sends a subsequent request to producer 112.

In the example of FIG. 2B, the consumer 131 sends a discovery to the network repository function (NRF) 133, and the NRF 133 sends NF profiles (NF profile(s)) to the consumer 131.
Here, the NF profile can hold the load status of the target NF instance, and based on this information, it is possible to perform load balancing and selection of the NF instance of the producer 132.
The consumer 131 selects an NF set or an NF instance, which in turn selects a producer 132. Communication between the consumer 131 and the selected producer 132 is generally similar to that in FIG. 2A.

In the example of FIG. 2C, the consumer 151 sends a discovery to the NRF 153, and the NRF 153 sends an NF profile to the consumer 151.
The consumer 151 selects an NF set or an NF instance, which in turn selects a producer 152.
The consumer 151 sends a service request to the SCP 154, and the SCP 154 sends the service request to the producer 152. At this time, the SCP 154 may obtain parameters from the NRF 153.
The producer 152 sends a response to the SCP 154 , and the SCP 154 sends the response to the consumer 151 .
Consumer 151 sends subsequent requests to SCP 154 , which in turn sends subsequent requests to producer 152 .

2D , the consumer 171 sends a service request and parameters to the SCP 174, and the SCP 174 sends the service request to the producer 172. At this time, the SCP 174 uses the parameters to send a discovery to the NRF 173 and receives an NF profile from the NRF 173. Then, the SCP 174 selects an NF instance.
The producer 172 sends the response to the SCP 174 , which sends the response to the consumer 171 .
Consumer 171 sends subsequent requests to SCP 174 , which in turn sends subsequent requests to producer 172 .

Here, the service framework was reviewed with the aim of improving cooperation between NFs and enabling flexible scale-out, and as a result, the following functions have been added from Rel16 onwards.
First, as shown in Figures 2C and 2D, an Indirect method is added, and the SCP performs NF discovery and signal routing processing, thereby simplifying the NF consumer.
Second, NF Set/NF Service Set has been added. All instances in an NF Set maintain the same context, so if one instance becomes unavailable, other instances can take over.
NFs in the same NF set may be deployed in different data centers (DCs) or different geographic locations to increase service availability.

FIG. 3 is a diagram showing a schematic example of the overall configuration using an SCP.
In the example of FIG. 3, in the region 231 (Region A), an NF producer 271 is formed in a data center 251 (DC1), and an NF producer 272 is formed in a data center 252 (DC2).
In the region 232 (Region B), an NF producer 281 is formed in the data center 261 (DC1), and an NF producer 282 is formed in the data center 262 (DC2).
The SCP 211 has access to each of the producers 271, 272, 281, and 282.

Currently, communication between NF instances is generally performed in the form shown in Fig. 2B. However, in the future, it is expected that communication between NF instances will generally be performed in the forms shown in Fig. 2C and Fig. 2D.
As the installation method for each NF progresses, and in order to achieve high availability, it is expected that a method will be used in which NFs are installed in different DCs in different locations, as shown in Figure 3. For this reason, it is expected that the variance in the (round-trip) transmission time expected for each NF when viewed from a certain SCP will become so large that it cannot be ignored. In addition, in the future, it is expected that NF instances will become more stateless in order to flexibly expand SBA.

Two models will be exemplified as models according to the embodiment.
In the first model, the proxy device queries the repository instance for information about a server instance, and the proxy device selects the server instance.
In the second model, a client instance queries a repository instance for information about server instances, and the client instance selects the server instance.

In this embodiment, the proxy device and/or repository instance stores a table that holds processing delay values and network delay values for each instance.
In this embodiment, the header customization function of http is used to add the following header options: That is, the header customization function of http2 is used to add a specified optional header (parameter), and when using other protocols, specified parameters are also added.
Here, the added parameters include, for example, send time (transmission time), load information (instance processing delay time), and a request timeout value. The send time (transmission time) is the request transmission time when a request is sent, and is the response transmission time when a response is sent. Note that for such information, a header option described in 3GPP (registered trademark) 29.500 may be used.

FIG. 4 is a diagram showing a schematic functional block configuration of the proxy device 311 according to the embodiment.
The proxy device 311 is an example of an SCP.
The proxy device 311 includes a communication unit 331 , a storage unit 332 , and a control unit 333 .
The control unit 333 includes a request acquisition unit 351 , a query unit 352 , a delay calculation unit 353 , and an NF instance selection unit 354 .
The delay calculation unit 353 includes a delay acquisition unit 371 .

The communication unit 331 communicates with an external device. Note that the communication may be, for example, wired communication or wireless communication.
The storage unit 332 stores various types of information.
The control unit 333 executes various controls and processes.

In this embodiment, the proxy device 311 is configured using a computer.
The control unit 333 has a processor such as a CPU (Central Processing Unit), and executes a predetermined program using the processor to perform various controls and processes.
The program may be stored in the storage unit 332, for example.

The request acquisition unit 351 acquires a request for a service.
The inquiry unit 352 makes an inquiry about the NF based on the request.
The delay acquisition unit 371 acquires information about delays in the processing in response to the request. In this embodiment, the delays include network delays.
The delay calculation unit 353 calculates a delay (for example, a delay time) based on the acquired information about the delay.
The NF instance selection unit 354 selects an NF instance based on the calculated delay.
The function of the proxy device 311 may be configured as an instance, for example.

(Example of the first model)
FIG. 5 is a diagram showing a schematic configuration of a first model (first model configuration 401) according to the embodiment.
In the first model, the proxy device 311A inquires about server instance information from the repository instance 413, and the proxy device 311A selects a server instance.

FIG. 5 shows a client instance 411, a proxy device 311A, a server instance 412, and a repository instance 413.
Here, the first model is applicable to the configuration of FIG. 2D, where the client instance 411 corresponds to the consumer, the proxy device 311A corresponds to the SCP, the server instance 412 corresponds to the producer, and the repository instance 413 corresponds to the NRF.

The proxy device 311A is an example of the proxy device 311 shown in FIG. 4, and will be described using the same reference numerals as the respective functional units shown in FIG.
The client instance 411 sends a service request to the proxy device 311A. The request may include parameters and the like.
In the proxy device 311 A, the communication unit 331 communicates with the client instance 411 , and the request acquisition unit 351 acquires the request received from the client instance 411 .
In the proxy device 311A, the communication unit 331 communicates with the repository instance 413, and the query unit 352 queries the repository instance 413 for a server instance based on the request acquired by the request acquisition unit 351.

Here, in the proxy device 311A, information related to the processing time is acquired by the delay acquisition unit 371. In the proxy device 311A, the delay calculation unit 353 calculates the delay related to the server instance based on the information acquired by the delay acquisition unit 371, and acquires information related to the delay (e.g., the processing time).
In the proxy device 311 A, the communication unit 331 communicates with the server instance 412 based on the calculation result by the delay calculation unit 353 , and transfers the request to the server instance 412 .
Then, the server instance 412 sends a response to the client instance 411 via the proxy device 311A.

In the first model, the proxy device 311A queries the repository instance 413 for a server instance, so the processing delay of the proxy device 311A includes the processing time of the repository instance 413 and the network delay time between the repository instance 413 and the proxy device 311A.

(Example of the second model)
FIG. 6 is a diagram showing a schematic configuration of a second model (second model configuration 501) according to the embodiment.
In the second model, the client instance 511 queries the repository instance 513 for server instances, and the client instance 511 selects a server instance (e.g., one server instance or a set of multiple server instances).

FIG. 6 shows a client instance 511, a proxy device 311B, a server instance 512, and a repository instance 513.
Here, the second model is applicable to the configuration of FIG. 2C, where the client instance 511 corresponds to the consumer, the proxy device 311B corresponds to the SCP, the server instance 512 corresponds to the producer, and the repository instance 513 corresponds to the NRF.

The proxy device 311B is an example of the proxy device 311 shown in Fig. 4, and will be described using the same reference numerals as the respective functional units shown in Fig. 4. Note that in the second model, the function of the query unit 352 shown in Fig. 4 does not necessarily have to be provided in the proxy device 311B.
The client instance 511 queries the repository instance 513 for the server instance.
Here, this inquiry may be made in advance, for example, separately from the timing of the subsequent processing.

The client instance 511 sends a request for service to the proxy device 311B, which may include the results of a query of a server instance.
In the proxy device 311B, the communication unit 331 communicates with the client instance 511, and the request acquisition unit 351 acquires the request received from the client instance 511.

Here, in the proxy device 311B, information related to the processing time is acquired by the delay acquisition unit 371. In the proxy device 311B, the delay calculation unit 353 calculates the delay related to the server instance based on the information acquired by the delay acquisition unit 371, and acquires information related to the delay (e.g., the processing time).
In the proxy device 311 B, the communication unit 331 communicates with the server instance 512 based on the calculation result by the delay calculation unit 353 , and transfers the request to the server instance 512 .
Then, the server instance 512 sends a response to the client instance 511 via the proxy device 311B.
Here, in the case of the second model, for example, a query to the server instance is executed in advance, so there is no need to consider delays in query processing to the repository instance 513 .

FIG. 7 is a diagram showing a schematic configuration of an assumed model (an assumed model configuration 601) according to the embodiment.
In the example of FIG. 7, a plurality of n client instances 611-1 to 611-n, a proxy device 311C, a plurality of m server instances 612-1 to 612-m, and a repository instance 613 are shown.

The proxy device 311C includes a proxy agent 711 (proxy agent a), a proxy agent 712 (proxy agent b), a proxy control plane 713, and a proxy storage unit 714.
The proxy device 311C is an example of the proxy device 311 shown in Fig. 4. The proxy agents 711 and 712 and the proxy control plane 713 are configured, for example, by the functions of the communication unit 331 and the control unit 333 shown in Fig. 4, and the proxy storage unit 714 is configured, for example, by the functions of the storage unit 332 shown in Fig. 4.
In the example of FIG. 7, communication between proxy agents is performed inside the proxy device 311C, so communication between proxy agents is also taken into consideration.

The proxy agent 711 has the function of querying the repository instance 613 for server instances.
The proxy agent 712 has the function of accessing the server instances 612-1 to 612-m.
In the example of FIG. 7, for convenience of explanation, the functions of the proxy agent 711 and the proxy agent 712 are distinguished from each other, but the distinction is not necessarily limited to this example.

The total request processing delay RD _total (t) is expressed by (Equation 1).
In this embodiment, the processing time required for communication between two instances, etc. is described as being the same for the processing time from one to the other and the processing time from the other to the one, but as another example, these may be determined as values that can be different from each other.

The definition of formula (1) is as follows.
The request processing delay time is represented by RD(t).
The network delay time between instances is represented as ND(t).
The network delay time between a client (such as client instance 611-1) and proxy device 311C is represented as ND _client-Proxy (t).
The network delay time between the proxy device 311C and a server (eg, server instance 612-1) is represented as ND _Proxy-server (t).
The processing time in the proxy device 311C is represented by RD _proxy (t).
The processing time at the server is represented by RD _server (t).

The definition of formula (2) is as follows.
The processing time in the proxy agent 711 of the proxy device 311C is represented by RD _{proxy_agent_a} (t).
The processing time in the proxy agent 712 of the proxy device 311C is represented by RD _{proxy_agent_b} (t).
The processing time (delay time due to communication) between the proxy agent 711 and the proxy agent 712 is represented by ND _{proxy_agent_ab} (t).

Here, Request send time represents the time when each instance sends a request. In this embodiment, the client instance stamps the time when it sends a request in the send time parameter in the http2 header.
The Request received time indicates the time when each instance or the proxy device 311C receives the request. Each instance or the proxy device 311 acquires the request reception time.
The Request transferred time indicates the time when the proxy device 311 transfers the request. The proxy device 311C stamps this time in the send time parameter in the header of http2. The proxy device 311C can also update this value.

Response send time indicates the time when the server or proxy agent sends the response. The server or proxy agent stamps the time when the response is sent in the send time parameter in the http2 header.
Response received time indicates the time when the proxy agent receives the response. The proxy device 311 acquires the time when the response is received.

Request received time _{proxy_agent_a} indicates the time when the proxy agent 711 (proxy agent a) received the request.
Request transferred time _{proxy_agent_a} indicates the time when the proxy agent 711 (proxy agent a) transfers the request.

Request send time _{proxy_agent_a} indicates the time when the proxy agent 711 (proxy agent a) sends a request.
Request received time _{proxy_agent_b} indicates the time when the proxy agent 712 (proxy agent b) received the request.

Response send time _server indicates the time when the server (server instance) sends the response.
Response received time _{proxy_agent_b} indicates the time when the proxy agent 712 (proxy agent b) received the response.

Request send time _Client indicates the time when the client (client instance) sends a request.

Request transferred time _{proxy_agent_b} indicates the time when the proxy agent 712 (proxy agent b) transfers the request.

A method for selecting an instance that takes into consideration a timeout value will be described.
The selection of the instance may be determined based on the timeout value of the request.
In the proxy device 311, when a request is received from a client instance, the control unit 333 may execute the following process: The proxy agent 712 on the server side determines whether to transfer the request.

First, the proxy device 311 checks the timeout value in the request.

Second, the proxy device 311 performs a calculation of an estimated transferred timeout value.
The Estimation transferred timeout value is the timeout value minus the time elapsed from when a request transferred from a client instance was generated until it reached the predetermined processing. The timeout value is set by the client instance.

Here, two examples (first calculation method and second calculation method) will be shown as methods for calculating the Estimation transferred timeout value.
The first calculation method is shown in equation (8).
The second calculation method is shown in equation (9).
Regarding the second calculation method, since it is possible to take into account the processing time between each instance and the processing time of each instance, it is also possible to take into account the time consumed in the future between the client and the proxy.

Request send time _{proxy_agent_b} indicates the time when the proxy agent 712 (proxy agent b) sends the request.

Third, the proxy device 311 compares the Estimation transferred timeout value with the processing delay time and network delay time stored in the proxy device, selects a server instance whose required processing time is smaller than the Estimation transferred timeout value, and transfers the request to the selected server instance.
The condition for selecting a server instance is expressed by, for example, expression (10).

Fourth, the proxy device 311 overwrites the Estimation transferred timeout value with the request timeout value when the request is transferred.

If the proxy device 311 detects multiple candidate server instances (target server instances) based on the calculation result of equation (10), the proxy device 311 may, for example, control the request to be forwarded to the server instance with the largest delay value, in order to prevent the request from being forwarded only to the server instance with the smallest delay value.
Furthermore, if the proxy device 311 is unable to find the target server instance, it returns an error to the client instance side.
Here, the candidate server instances are server instances that do not time out.

The processing delay of each instance and the network delay between each instance will be explained.
In the proxy device 311, the proxy agents 711 and 712 store (hold) information on processing delays of subordinate instances and information on processing delays between proxy agents with which they communicate, as a table.
As a method of obtaining the processing delay of a server instance, for example, a proxy server (proxy device 311) can detect (determine) the processing delay of a server instance by referring to the processing delay time of the instance in relation to the load information in the response header.

FIG. 8 is a diagram showing an example of the instance processing delay retention table 2011 according to the embodiment.
In the proxy device 311, for example, the storage unit 332 may store a processing delay retention table 2011. The values in the processing delay retention table 2011 may be set to be fixed, or may be variably updated, for example.

In the processing delay holding table 2011, ND _{instance-Proxy} (t) and RD _instance (t) are shown for each instance identification information (ID).
In the example of FIG. 8, instance IDs are shown as instance_A, instance_B, and instance_C.
Here, the network delay between the proxy device 311 and each instance is represented as ND _{instance-Proxy} (t).
The processing time for each instance is represented by RD _instance (t).
Examples of instances include a client instance, a server instance, and a repository instance.

FIG. 9 is a diagram showing an example of the processing delay holding table 2021 of the proxy agent according to the embodiment.
In the proxy device 311, for example, the storage unit 332 may store a processing delay retention table 2021. The values in the processing delay retention table 2021 may be set to be fixed, or may be variably updated, for example.

The processing delay holding table 2021 shows ND _{Proxy_agent} (t) and RD _{proxy_agent} (t) for each identification information (ID) of a proxy agent instance.
In the example of FIG. 9, Proxy agent b, Proxy agent c, and Proxy agent d are shown as instance IDs.
Here, the network delay between the proxy agent 711 (proxy agent a) and each proxy agent (between proxy agents) is represented by ND _{Proxy_agent} (t).
The processing time in each instance (each proxy agent) is represented by RD _{proxy_agent} (t).

In addition, processing time information such as the contents of the processing delay retention table 2011 shown in FIG. 8 and the contents of the processing delay retention table 2021 shown in FIG. 9 may be managed (stored), for example, in the repository instance, or may be managed (stored) in both the proxy device 311 and the repository instance.

The processing delay of each instance and the update of the network delay between each instance will be explained.
In the proxy device 311, for example, updating of table values may be performed only when a request has ended normally, because if a value when a request has ended abnormally is used, an incorrect value may be stored.
The value of each processing delay time may be set or updated, for example, based on only one request, or may be set or updated using a value (such as an average value) calculated based on the results of multiple requests.

The initial values of the processing delay of each instance and the network delay between each instance will be explained.
Any value may be specified for the processing delay of each instance.
Also, for example, a prediction model of microservices may be used (see, for example, Non-Patent Document 3).

When an NF instance is generated (created), its information is registered with the NRF via the SCP, and the network delay between each instance can be calculated using the request made at that time.

FIG. 10 is a diagram illustrating an example of an instance registration request according to the embodiment.
In the example of FIG. 10, a generated instance (generated instance 821), a proxy device 311D, and a server instance 822 are shown.
Here, the proxy device 311D is an example of the proxy device 311 shown in FIG.

In the example of FIG. 10, the created instance 821 transmits a request for instance registration (instance registration request) to the proxy device 311D.
The proxy device 311D may detect (determine) a network delay based on the time when the instance registration request is sent from the created instance 821 (Request send time) and the time when the instance registration request is received (Request received time).

Account for network delays between proxy agents.
Any value may be specified for the processing delay of each instance.
The proxy device 311D may set the network delay by referring to values obtained by executing communication between proxy agents, for example. The communication may be, for example, ICMP (Internet Control Message Protocol) communication.

This section describes a process when a server instance needs to transfer a request to another instance when processing the request.
First, the server instance checks the request timeout value if serialization is required.
Second, the server instance performs a calculation of the Estimation transferred timeout value, for example, using equation (8) or equation (9).

Third, the server instance overwrites the Estimation transferred timeout value with the request timeout value at the time of request transfer, and transfers the request to the proxy server (proxy device 311).
Fourth, the proxy server (proxy device 311) executes processing for selecting an instance based on the received request, taking into account the timeout value.
By repeating the above, it is possible to update the timeout value and prevent inconsistencies such as data updates in the processing between each instance.

Next, a process for transferring data to another instance when parallel processing is required will be described.
A case where synchronous processing is required for request processing while executing parallel processing is explained.
Here, synchronous processing is, for example, processing in which a request is sent to a certain instance, a response to the request is received, and then a request is sent to another instance.

First, the server instance checks the timeout value in the request.
Second, the server instance performs a calculation of the Estimation transferred timeout value, for example, using equation (8) or equation (9).

Third, the server instance overwrites the Estimation transferred timeout value with the request timeout value at the time of request transfer, and transfers the request to the proxy server (proxy device 311). At this time, the server instance stores the transfer time and the Estimation transferred timeout value.
Fourth, after the transfer, the server instance stores the time when the response is received (response time).
Fifth, the server instance calculates the request processing delay time based on the stored response time and transfer time, and determines the timeout value for the next request by subtracting this value from the stored Estimation transferred timeout value.

When performing parallel processing, it is sometimes possible to execute request processing asynchronously.
In such a case, a process similar to the serial process is executed.

As described above, the proxy device 311 according to this embodiment can improve efficiency by selecting an NF instance taking network delay into consideration.
The proxy device 311 according to this embodiment can take into account the network delay between NF instances and can also take into account the processing delay of the SCP when selecting an NF instance. As a result, the proxy device 311 according to this embodiment can perform load balancing based on not only the processing delay of the instance but also the network delay between the instances, and can select an instance to which a request is to be sent, taking into account the processing delay of the proxy that selects the instance.
The proxy device 311 according to this embodiment can select an NF instance taking into consideration communication processing delays between a plurality of agents.

In the proxy device 311 according to this embodiment, for example, by grasping the delay value of the entire SBA, it is possible to process a request by using the most suitable NF instance for each request of the NF instance.
In the proxy device 311 according to this embodiment, for example, for high availability, an appropriate NF instance can be selected even when each NF instance is installed in a different location or a different data center (DC).

In the proxy device 311 according to this embodiment, the timeout value can be dynamically updated.
The proxy device 311 according to this embodiment can store the processing delay value and the network delay value of each instance.
The proxy device 311 or each instance according to this embodiment can transmit a request while retaining its own request transmission time and timeout value.
In the proxy device 311 according to this embodiment, the timeout value and the request transfer time can be updated when a request is transferred.

The proxy device 311 according to this embodiment can select an instance by comparing the timeout value with each processing delay.
Each instance can retain the processing delay of each instance in the response and transfer it to the proxy device 311 .
The proxy device 311 according to this embodiment can autonomously calculate the processing delay time when, for example, a new instance is added.
In the proxy device 311 according to this embodiment, for example, it is possible to reduce the timeout and reduce the occurrence of error termination.

The processing method performed by the proxy device 311 in this embodiment can also be implemented as a proxy method.

This will enable, for example, improvements to the overall quality of services 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."

A program for implementing the functions of any of the components of any of the devices described above may be recorded on a computer-readable recording medium, and the program may be loaded and executed by a computer system. Note that the term "computer system" as used herein includes hardware such as an operating system or peripheral devices. Furthermore, "computer-readable recording medium" refers to portable media such as flexible disks, optical magnetic disks, ROMs, and CD (Compact Disc)-ROMs (Read Only Memory), as well as storage devices such as hard disks built into computer systems. Furthermore, "computer-readable recording medium" also includes devices that retain a program for a certain period of time, such as volatile memory within a computer system that acts as a server or client when a program is transmitted over a network such as the Internet or over a communication line such as a telephone line. Such volatile memory may be, for example, RAM (Random Access Memory). The recording medium may also be, for example, a non-transitory recording medium.

The above program may be transmitted from a computer system storing the program in a storage device or the like to another computer system via a transmission medium or by transmission waves in the transmission medium. Here, the "transmission medium" that transmits the program refers to a medium that has the function of transmitting information, such as a network such as the Internet or a communication line such as a telephone line.
The above program may also be one that realizes part of the above-mentioned functions. Furthermore, the above program may be a so-called differential file that can realize the above-mentioned functions in combination with a program already recorded in a computer system. A differential file may also be called a differential program.

Furthermore, the functions of any of the components in any of the devices described above may be implemented by a processor. For example, each process in the embodiments may be implemented by a processor that operates based on information such as a program, and a computer-readable recording medium that stores information such as the program. Here, the functions of each unit of the processor may be implemented by separate hardware, or the functions of each unit may be implemented by integrated hardware. For example, the processor may include hardware, and this hardware may include at least one of a circuit for processing digital signals and a circuit for processing analog signals. For example, the processor may be configured using one or more circuit devices mounted on a circuit board, or one or both of one or more circuit elements. An IC (Integrated Circuit) or the like may be used as the circuit device, and a resistor or capacitor may be used as the circuit element.

Here, the processor may be, for example, a CPU. However, the processor is not limited to a CPU, and various types of processors may be used, such as a GPU (Graphics Processing Unit) or a DSP (Digital Signal Processor). The processor may also be a hardware circuit, such as an ASIC (Application Specific Integrated Circuit). The processor may also be configured, for example, by multiple CPUs, or by a hardware circuit made up of multiple ASICs. The processor may also be configured, for example, by a combination of multiple CPUs and a hardware circuit made up of multiple ASICs. The processor may also include, for example, one or more amplifier circuits or filter circuits that process analog signals.

The above describes in detail an embodiment of this disclosure with reference to the drawings, but the specific configuration is not limited to this embodiment and includes designs that do not deviate from the gist of this disclosure.

1...5GC architecture, 311, 311A-311D...proxy device, 331...communication unit, 332...storage unit, 333...control unit, 351...request acquisition unit, 352...query unit, 353...delay calculation unit, 354...NF instance selection unit, 371...delay acquisition unit, 401...first model configuration, 411, 511, 611-1 to 611-n...client instances, 412, 512, 612-1 to 612-m, 822...server instance, 413, 513, 613...repository instance, 501...second model configuration, 601...assumed model configuration, 711, 712...proxy agent, 713...proxy control plane, 714...proxy storage unit, 821...generated instance, 2011, 2021...processing delay retention table

Claims (8)

a request acquisition unit that acquires a request from an NF instance of a client;
an NF instance selection unit that selects an NF instance of a server based on a processing delay including a network delay in response to the request acquired by the request acquisition unit;
Equipped with
the NF instance selection unit acquires, as the processing delay, a processing delay of the entire process in which the request is transmitted from the NF instance of the client to its own device, the request is transferred from its own device to the NF instance of the server, and a response is transmitted from the NF instance of the server to the NF instance of the client via its own device, based on the acquired processing delay of the entire process, and selects an NF instance of the server that can process the entire process within a timeout value of the request ;
the NF instance selection unit calculates an estimated value of a transfer timeout value by subtracting the time when the request is transmitted from the NF instance of the client from the time when the request is transferred from its own device, and subtracting the result from the timeout value of the request, and selects the NF instance of the server for which the sum of the network delay time when the request is transferred from its own device to the NF instance of the server, the processing time at the server, and the network delay time when the response is transmitted from the NF instance of the server to its own device is smaller than the calculated estimated value.
Proxy device. a request acquisition unit that acquires a request from an NF instance of a client;
an NF instance selection unit that selects an NF instance of a server based on a processing delay including a network delay in response to the request acquired by the request acquisition unit;
Equipped with
the NF instance selection unit acquires, as the processing delay, a processing delay of the entire process in which the request is transmitted from the NF instance of the client to its own device, the request is transferred from its own device to the NF instance of the server, and a response is transmitted from the NF instance of the server to the NF instance of the client via its own device, based on the acquired processing delay of the entire process, and selects an NF instance of the server that can process the entire process within a timeout value of the request ;
the NF instance selection unit calculates an estimated value of a transfer timeout value by subtracting from a timeout value of the request the sum of a network delay time when the request is transmitted from the NF instance of the client to the own device, a processing time in the own device, and a network delay time when the response is transmitted from the own device to the NF instance of the client, and comparing the sum with the calculated estimated value, the NF instance selection unit selects the NF instance of the server for which the sum of a network delay time when the request is transmitted from the own device to the NF instance of the server, a processing time in the server, and a network delay time when the response is transmitted from the NF instance of the server to the own device is smaller.
Proxy device. a query unit that queries a repository instance for an NF instance of the server;
3. The proxy device according to claim 1. It has multiple agents
The processing delay of the entire process includes a communication processing delay between the agents.
3. The proxy device according to claim 1 . The request acquisition unit acquires a request from the client's NF instance,
an NF instance selection unit selects an NF instance of a server based on a processing delay including a network delay in response to the request acquired by the request acquisition unit;
the NF instance selection unit acquires, as the processing delay, a processing delay of the entire process in which the request is transmitted from the NF instance of the client to its own device, the request is transferred from its own device to the NF instance of the server, and a response is transmitted from the NF instance of the server to the NF instance of the client via its own device, based on the acquired processing delay of the entire process, and selects an NF instance of the server that can process the entire process within a timeout value of the request ;
the NF instance selection unit calculates an estimated value of a transfer timeout value by subtracting the time when the request is transmitted from the NF instance of the client from the time when the request is transferred from its own device, and subtracting the result from the timeout value of the request, and selects the NF instance of the server for which the sum of the network delay time when the request is transferred from its own device to the NF instance of the server, the processing time at the server, and the network delay time when the response is transmitted from the NF instance of the server to its own device is smaller than the calculated estimated value.
Proxy method. The request acquisition unit acquires a request from the client's NF instance,
an NF instance selection unit selects an NF instance of a server based on a processing delay including a network delay in response to the request acquired by the request acquisition unit;
the NF instance selection unit acquires, as the processing delay, a processing delay of the entire process in which the request is transmitted from the NF instance of the client to its own device, the request is transferred from its own device to the NF instance of the server, and a response is transmitted from the NF instance of the server to the NF instance of the client via its own device, based on the acquired processing delay of the entire process, and selects an NF instance of the server that can process the entire process within a timeout value of the request ;
the NF instance selection unit calculates an estimated value of a transfer timeout value by subtracting from a timeout value of the request the sum of a network delay time when the request is transmitted from the NF instance of the client to the own device, a processing time in the own device, and a network delay time when the response is transmitted from the own device to the NF instance of the client, and comparing the sum with the calculated estimated value, the NF instance selection unit selects the NF instance of the server for which the sum of a network delay time when the request is transmitted from the own device to the NF instance of the server, a processing time in the server, and a network delay time when the response is transmitted from the NF instance of the server to the own device is smaller.
Proxy method. On the computer that configures the proxy device,
A request acquisition function that acquires requests from the client's NF instance;
an NF instance selection function that selects an NF instance of a server based on a processing delay including a network delay in response to the acquired request;
A program for realizing the above,
the NF instance selection function acquires, as the processing delay, a processing delay of the entire process in which the request is transmitted from the NF instance of the client to its own device, the request is transferred from its own device to the NF instance of the server, and a response is transmitted from the NF instance of the server to the NF instance of the client via its own device, and selects an NF instance of the server that can process the entire process within a timeout value of the request based on the acquired processing delay of the entire process ;
The NF instance selection function calculates an estimated value of a transfer timeout value by subtracting the time when the request is transmitted from the NF instance of the client from the time when the request is transferred from the own device, and subtracting the result from the timeout value of the request, and selects the NF instance of the server for which the sum of the network delay time when the request is transferred from the own device to the NF instance of the server, the processing time at the server, and the network delay time when the response is transmitted from the NF instance of the server to the own device is smaller than the calculated estimated value.
program. On the computer that configures the proxy device,
A request acquisition function that acquires requests from the client's NF instance;
an NF instance selection function that selects an NF instance of a server based on a processing delay including a network delay in response to the acquired request;
A program for realizing the above,
the NF instance selection function acquires, as the processing delay, a processing delay of the entire process in which the request is transmitted from the NF instance of the client to its own device, the request is transferred from its own device to the NF instance of the server, and a response is transmitted from the NF instance of the server to the NF instance of the client via its own device, and selects an NF instance of the server that can process the entire process within a timeout value of the request based on the acquired processing delay of the entire process ;
The NF instance selection function calculates an estimated value of a transfer timeout value by subtracting from a timeout value of the request the sum of a network delay time when the request is transmitted from the NF instance of the client to the own device, a processing time in the own device, and a network delay time when the response is transmitted from the own device to the NF instance of the client, and comparing the sum with the calculated estimated value, the NF instance of the server is selected, which produces a smaller sum of a network delay time when the request is transmitted from the own device to the NF instance of the server, a processing time in the server, and a network delay time when the response is transmitted from the NF instance of the server to the own device.
program.