JP6872919B2 - Sensor network management method and sensor network management system - Google Patents

Description

The present invention relates to a sensor network management method and a sensor network management system, and more specifically, to a technique capable of flexibly providing sensing data of a physical sensor in a sensor network in a form according to application requirements.

IoT (Internet of Things), which creates new value by combining information on various things that exist all over the world on a network, has become widespread.

With IoT, sensors with sensing and communication functions are installed in various places and industrial equipment, and a large amount of sensor data obtained from the sensors is managed, and the user can use these sensor data according to the purpose. It is to be used. In order to enable the collection, management, and seamless use of such sensor data, the mainstream is to network sensors installed in various locations and centrally manage sensor data on a cloud platform. It has become to.

In such a situation, there are cases where the owner of the sensor and the user of the sensor data described above are different. In this case, since the data is traded as a resource by both parties, it is necessary to establish a sensor network mechanism for appropriately distributing the sensor data according to the purpose of the user.

Against this background, for example, information on the sensor-side metadata acquisition means for acquiring sensor-side metadata, which is information on the sensor that outputs sensing data, and information on applications that provide services using the sensing data. By matching the app-side metadata acquisition means for acquiring the app-side metadata with the sensor-side metadata and the afri-side metadata, the sensor capable of providing sensing data satisfying the requirements of the application is extracted. It is characterized in that it has a matching means to be used, and an instruction means for transmitting a data flow control command specifying the sensor extracted by the matching means and the friction to the sensor management device that manages the sensor. A data flow control command generator (see Patent Document 1) and the like have been proposed.

WO2014 / 041826

However, in the prior art, when providing sensing data, only the requirements of the physical sensor are matched with the requirements of the application. In other words, the sensing data of the physical sensor that matches the requirements is provided to the application as it is. However, depending on the specifications of the physical sensor included in the sensor network and its variations, there may be many cases where the content of the sensing data is not the attribute desired by the application.

In that case, even if the requirements as in the conventional technique described above are matched, the physical sensor that provides appropriate sensing data to the application cannot be specified, and the distribution of the sensing data may be hindered. This makes it difficult to flexibly examine alternative sensors when a failure occurs in a physical sensor or the like, and leads to a malfunction of an application that uses sensing data.

Therefore, an object of the present invention is to provide a technique capable of flexibly providing sensing data of a physical sensor in a sensor network in a form according to application requirements.

The sensor network management method of the present invention that solves the above problems includes metadata including the operating rate of the physical sensor included in the sensor network and the accuracy of the output data, metadata regarding the accuracy of the output data of the predetermined data extraction means, and the above. Metadata that includes the mobility and the accuracy of the sensing data, which is the percentage of time that the physical sensor provides normal sensing data, which is required by the sensing data of the physical sensor or the application that uses the predetermined data based on the sensing data. An information processing system equipped with a storage device to hold
The physical sensor and the data extraction means that satisfy the accuracy of the mobility and the sensing data required by the application based on the accuracy of the output data of each of the data extraction means and the accuracy of the operating rate and the output data of each of the physical sensors. It is characterized in that a matching process for specifying a sensing means, which is a combination with the above, is executed.

Further, the sensor network management system of the present invention includes metadata including the operating rate of the physical sensor included in the sensor network and the accuracy of the output data, metadata regarding the accuracy of the output data of the predetermined data extraction means, and the physical sensor. Storage that holds metadata including mobility and accuracy of sensing data, which is the percentage of time that the physical sensor provides normal sensing data, which is required by the sensing data or an application that uses predetermined data based on the sensing data. A physical sensor that satisfies the mobility and sensing data accuracy required by the application based on the accuracy of the output data of the device and each of the data extraction means, and the operating rate and output data accuracy of each of the physical sensors. A computing device that executes a matching process for specifying a sensing means that is a combination with a data extraction means is provided.

According to the present invention, it is possible to flexibly provide the sensing data of a physical sensor in a sensor network in a form according to the requirements of an application.

It is a figure which shows the configuration example of the sensor network management system in this embodiment. It is a figure which shows the hardware configuration example of the management server in this embodiment. This is an example of the application attribute table of this embodiment. This is an example of the physical sensor attribute table of this embodiment. This is an example of the extraction processing management table of this embodiment. This is an example of a combination candidate table of this embodiment. This is an example of the processing chain table of this embodiment. This is an example of the processing chain table of this embodiment. This is an example of the processing chain table of this embodiment. This is an example of the processing chain table of this embodiment. It is a figure which shows the flow example 1 of the sensor network management method of this embodiment. It is a figure which shows the flow example 2 of the sensor network management method of this embodiment. This is an example of the selection policy table of this embodiment. It is a figure which shows the flow example 3 of the sensor network management method of this embodiment. It is a figure which shows the flow example 4 of the sensor network management method of this embodiment. It is a figure which shows the flow example 5 of the sensor network management method of this embodiment. It is a figure which shows the flow example 6 of the sensor network management method of this embodiment. It is a figure which shows the flow example 7 of the sensor network management method of this embodiment. It is a figure which shows the screen example 1 in this embodiment. It is a figure which shows the screen example 2 in this embodiment. It is a figure which shows the screen example 3 in this embodiment.

--- System configuration ---
Hereinafter, embodiments of the present invention will be described in detail with reference to the drawings. FIG. 1 is a diagram showing a configuration example of the sensor network management system 1 of the present embodiment.

The sensor network management system 1 shown in FIG. 1 is a computer system that can flexibly provide the sensing data of the physical sensor 20 in the sensor network 5 in a form according to the requirements of the application 30. Further, the sensor network management system 1 controls the sensing means in response to the request from the application 30 that uses the sensing data, and returns the requested data.

Such a sensor network management system 1 is composed of a sensor network 5 composed of a physical sensor 20, an IoT hub server 300, a data providing server 200, and an application 30, a management server 100, and a monitoring server 400. Of these, the management server 100 and the data providing server 200 are the minimum configurations of the sensor network management system 1.

The physical sensor 20 included in the sensor network 5 described above is connected to the IoT hub server 300 through the network 11. The physical sensor 20 is a device having a predetermined sensing function and a communication function. The physical sensor 20 periodically transmits the sensing data obtained by sensing by the sensing function to the IoT hub server 300 via the network 11 by the communication function.

On the other hand, the IoT hub server 300 is a management device for a plurality of physical sensors 20 included in the sensor network 5. The IoT hub server 300 receives the sensing data sent from each of the physical sensors 20 and provides the sensing data to the data providing server 200. Further, the IoT hub server 300 provides the management server 100 with metadata such as the type, operating state, and position information of the sensing data of each of the physical sensors 20 via the output processing unit 310.

Further, the above-mentioned IoT hub server 300 monitors the data provision state of each of the physical sensors 20 connected via the network 11. Therefore, the IoT hub server 300 transmits a predetermined alert to the management server 100 with respect to the physical sensor 20 indicating the abnormality.

On the other hand, the data providing server 200 receives the sensing data from the above-mentioned IoT hub server 300, executes unprocessed or arbitrary extraction processing (210, 211, 212) on the sensing data, and outputs the output data which is the processing result. , It is provided to the corresponding application via the endpoint process (213, 214) created for each application.
The application 30 is software that uses the output data (the sensing data as it is or the data obtained by extracting the sensing data) from the data providing server 200 described above. The application 30 transmits a data request to the above-mentioned data providing server 200 at an arbitrary timing.

The monitoring server 400 and the management server 100 monitor and control the sensor network 5. The monitoring server 400 and the management server 100 are connected to the sensor network 5 via the network 10.

Of these, the monitoring server 400 is a server device that monitors the IoT hub server 300, the data providing server 200, and the network 10 and notifies the management server 100 when a predetermined abnormality is detected in any of them. Such a function may be provided in the management server 100 (in that case, the monitoring server 400 becomes unnecessary).

On the other hand, the management server 100 executes a matching process between the sensing means and the application 30. Further, the management server 100 builds a sensing means matching the application 30 on the data providing server 200, and performs basic control for providing the output data by the sensing means to the application 30.

The sensing means in the present embodiment is appropriate data processing such as selection, extraction, calculation, etc. for one or a plurality of physical sensors 20 and the sensing data output from the physical sensors 20 (in the present embodiment). It is composed of one or a plurality of data extraction means, which is a program for performing (collectively referred to as "data extraction").

When a plurality of data extraction means are included in such a sensing means, the usage relationship of the output data between the data extraction means is such that the output data of one data extraction means is used as an input by the other data extraction means. I can imagine. Of course, the usage relationship between such data extraction means is not limited to one-to-one. For example, the output data of one data extraction means (or the sensing data of one physical sensor 20) may be used by a plurality of data extraction means, or the output data of a plurality of data extraction means (or the plurality of physical sensors 20) may be used. (Sensing data) can be assumed to be used by one data extraction means.

The management server 100 switches the sensing means according to a predetermined event, such as an abnormality occurrence of each physical sensor 20 in the sensor network 5 or an instruction reception by an operator of the sensor network management system 1, which may be included in the sensing means.

The management server 100 described above corresponds to the function of the sensor network management system 1 of the present embodiment, which is the sensor information acquisition processing unit 110, the input / output processing unit 111, the matching processing unit 112, the infrastructure control processing unit 113, and the alert reception processing. Each function unit includes a unit 114, a reselection processing unit 115, and an undeploy processing unit 116.

Further, as data necessary for processing, the management server 100 includes an application attribute table 125, a physical sensor attribute table 126, an extraction processing management table 127, a combination candidate table 128, a processing chain table 129, and a selection policy table 130, as appropriate. It is held in a storage device.

Further, the management server 100 is a data extraction means constructed and executed by the above-mentioned data providing server 200, and the source code of each extraction means (in FIG. 1, the extraction process A source code 141, the synthesis process B source code 142, Is stored in the source code repository 140. The source code repository 140 described above does not necessarily have to be configured on the management server 100 if it can be accessed from the management server 100 and the data providing server 200, and is held on any device on the network 10. Just do it. Details of these data will be described later.

Among the functional units included in the management server 100, the sensor information acquisition processing unit 110 acquires metadata related to the physical sensor 20 from the IoT hub server 300 and registers it in the physical sensor attribute table 126.

Further, the input / output processing unit 111 acquires the requirements of the application 30 in advance from the user of the application 30 that uses the sensing data, and registers the requirements in the application attribute table 125, which is a table of metadata related to the requirements of the application 30. Is.

Further, the matching processing unit 112 can provide data satisfying the requirements of the application 30 specified to the above-mentioned user based on the physical sensor attribute table 126, the application attribute table 125, and the extraction processing management table 127. It finds a combination of the physical sensor 20 and the data extraction means.

The matching processing unit 112 includes a combination extraction processing unit 1121 and a combination selection processing unit 1122. Of these, the combination extraction processing unit 1121 refers to the extraction processing management table 127 and the above-mentioned application attribute table 125, and can output the data of the data type required by the requirements of the application 30. The physical sensor 20 and the data extraction means. In other words, a combination candidate table 128 which is a list of processing chains (hereinafter, processing chains) which is a sensing means is created.

On the other hand, the combination selection processing unit 1122 refers to the above-mentioned application attribute table 125 and the selection policy table 130 in which the priority for specifying an appropriate one from the plurality of processing chains is defined, and the above-mentioned combination. From the candidate table 128, the processing chain that meets the requirements of the application 30 is selected to be prioritized. The processing chain table 129 is a table generated by the combination extraction processing unit 1121, and is a table that manages the combination of the physical sensor 20, the data extraction means (hereinafter, extraction processing), and the endpoint processing.

Further, the infrastructure control processing unit 113 is called after the matching processing unit 112 described above, builds a processing chain selected by the matching processing unit 112 on the data providing server 200, and makes it available to the corresponding application 30. It is a thing.

The above-mentioned infrastructure control processing unit 113 is composed of a construction connection processing unit 1131 and an endpoint creation processing unit 1132. Of these, the construction connection processing unit 1131 performs the extraction processing included in the processing chain selected by the matching processing unit 112 on the data providing server 200 based on the source code stored in the source code repository 140. Make it move. Further, the construction connection processing unit 1131 connects the extraction processes to make the data requested by the application 30 in a state in which it can be generated. On the other hand, the endpoint creation processing unit 1132 creates an endpoint for making the data that can be generated by the construction connection processing unit 1131 accessible from the corresponding application 30.

Further, when the alert reception processing unit 114 receives a predetermined alert from the monitoring server 400 and the IoT hub server 300, the alert reception processing unit 114 calls the reselection processing unit 115 so that the data provision to the application 30 is continued.

When the alert reception processing unit 114 receives the information regarding the abnormality of the physical sensor 20 or the like acquired by the IoT hub server 300, the application 30 using the physical sensor 20 (hereinafter, referred to as an abnormal physical sensor) in which the abnormality occurs. Identify the existence of. When the alert reception processing unit 114 identifies the existence of the application 30 here, the alert reception processing unit 114 executes the reselection processing unit 115, and the processing chain constructed for the application 30 is a processing chain that does not include the abnormal physical sensor 20. Switch to.

Further, as described above, the reselection processing unit 115 is called when switching the processing chain, removes the abnormal physical sensor 20, and then reselects the processing chain that satisfies the requirements of the application 30.

After that, the reselection processing unit 115 calls the undeploying processing unit 116 for deleting the processing chain that has been constructed and operated for the application 30 from the data providing server 200. As a result, the unnecessary processing chain is deleted from the data providing server 200.
--- Hardware configuration ---

Further, the hardware configuration of the management server 100 constituting the sensor network management system 1 will be described. FIG. 2 is a diagram showing a hardware configuration example of the management server 100 of the present embodiment.

The management server 100 is held in a storage device 101 composed of an appropriate non-volatile storage element such as an SSD (Solid State Drive) or a hard disk drive, a memory 103 composed of a volatile storage element such as a RAM, and a storage device 101. Communication that is connected to a computing device 104 such as a CPU that executes various determinations, calculations, and control processes while executing the program 102 by reading it into a memory 103 to perform overall control of the device itself, and is responsible for communication processing with other devices. The device 105 is provided at least.

In the storage device 101, in addition to the program 102 for implementing the functions required as the device constituting the sensor network management system of the present embodiment, the application attribute table 125, the physical sensor attribute table 126, and the extraction processing management table At least 127, combination candidate table 128, processing chain table 129, selection policy table 130, and source code repository 140 are stored. These data structures and the like will be described later.
--- Data example ---

FIG. 3 shows an example of the application attribute table 125 in this embodiment. The application attribute table 125 is a table that manages metadata related to the requirements of the application 30 that uses the sensing data itself of the physical sensor 20 and the output data obtained by the extraction process for the sensing data.

The application attribute table 125 of the present embodiment includes an "app ID" column 201 for storing the identification information of the application 30, a "data type" column 202 for storing the type information of the sensing data required by the application 30, and the corresponding The "accuracy" column 203 for storing information on the accuracy of sensing data required by the application 30, the "cost" column 204 for storing information on the upper limit of the sensing data required by the application 30, and the application 30. The "mobility rate" column that stores information on the upper limit of the ratio between the time when data was normally provided and the time when data was expected to be provided for the sensing data required by the application 30 and the sensing required by the application 30 A "data acquisition interval" column 206 that stores information about the time interval in which the data is provided, and a data source (mainly the physical sensor 20), extraction process, and end that can provide the sensing data required by the application 30. It is composed of a processing chain ID column 207 that stores identification information of a processing chain that is a combination of point processing, and an endpoint ID column 208 that stores identification information of an endpoint of the processing chain.

Here, the value stored in the above-mentioned "data type" column 202 is an identifier of the type of sensing data by the physical sensor 20 such as temperature, voice, image, and position information.

In the application attribute table 125 illustrated in FIG. 3, for example, the application requirement of row 210 is that the accuracy "error radius 10 m or less" is obtained for the sensing data of the data type "Type- # 1" with respect to the application 30 having the application ID "App- # 1". And also
Under the condition of the cost "0.5 yen / minute" as the upper limit of the data usage fee, the data is requested to be provided at the data acquisition interval "1 minute" at the mobility rate "any". For the processing chain ID and the endpoint ID, "-" indicating no value is set until a processing chain that meets the requirements is specified.

FIG. 4 is an example of the physical sensor attribute table 126 in this embodiment. The illustrated physical sensor attribute table 126 is a table that manages the metadata of the physical sensor 20 connected to the above-mentioned IoT hub server 300.

The physical sensor attribute table 126 of the present embodiment shows a "device ID" column 301 for storing the identifier of each physical sensor 20, an "installation position" column 302 for storing information on the installation location of the physical sensor 20, and the physical sensor 20. "Data type" column 303 for storing the type information of the sensing data by the physical sensor 20, "Data acquisition interval" column 304 for storing the time interval value at which the physical sensor 20 transmits the sensing data to the IoT hub server 300, sensing by the physical sensor 20 "Accuracy" column 305-307 for storing data accuracy information, "Cost" column 308 for storing sensing data usage unit price information, "Operating rate" column 308 for storing operating rate information of the physical sensor 20, the physical It is composed of an "operating status" column 310, which stores a value of the operating status of the sensor 20.

The above-mentioned "data type" column 303 stores an identifier of the type of sensing data by the physical sensor 20, such as temperature, voice, image, and position information. Further, in the "accuracy" columns 305 to 307, the types of accuracy corresponding to the types of sensing data ("accuracy 1" to "accuracy N") by the physical sensor 20 are stored.

In such a physical sensor attribute table 126, for example, the physical sensor 20 represented by row 312 holds the device ID “Dev- # 2”, and its installation position is the latitude / longitude information “35.399057, 139.53258”. The data type of the sensing data is "Type- # 3", the data acquisition interval is "10 seconds", the accuracy of the sensing data is "error radius 5 m or less" with respect to the distance, and "resolution 200 dpi" with respect to the fineness of the image. The usage cost of the sensing data is "0.05 yen / minute", the operating rate is "100%", and the operating status is "ON".

FIG. 5 is an example of the extraction processing management table 127 in the present embodiment. The extraction process management table 127 is a table that manages the metadata of the program that extracts the sensing data (“data extraction” described above) stored in the source code repository 140. The extraction process management table 127 of the present embodiment includes an "extraction process program ID" column 401 that stores an identifier that uniquely identifies the extraction process, a "process name" column 402 that stores the type information of the extraction process, and the extraction process. "Input data type number" column 403 for storing the number of input data types required for the input data, "input data type" columns 404 to 407 for storing the input data type information, and output data type information in the extraction process are stored. It is composed of an "output data type" column 408 and an "accuracy" column 409 for storing the accuracy information of the output data.

In the extraction process management table 127 of FIG. 5, for example, the extraction process represented by line 416 holds the extraction process program ID “extraction process- # F”, the type of which is the process name “CC synthesis”, and the input data type to be handled. The number of is "2", the type of the input data is input 1 "Type- # 7" and the input 2 "Type- # 9", and the output data type is "Type- # 1". The accuracy is "an error of half 1 m or less".

FIG. 6 is an example of the combination candidate table 128 in this embodiment. The combination candidate table 128 is an output of the combination extraction processing unit 112 in the management server 100, and is a table that manages a list of sensing means capable of providing the data requested by the application 30.

The combination candidate table 128 of the present embodiment includes a "processing chain ID" column 501 that stores an identifier of a processing chain, which is a sensing means in which a data source, an extraction process, and an endpoint are combined in a series, and the process. The "accuracy" column 502 that stores the accuracy information of the data output from the chain, the "cost" column 503 that stores the charge information of the data output from the processing chain, and the operation rate information of the processing chain are stored. It is composed of an "operation rate" column 504 and a "data acquisition interval" column 505 that stores information on the data output interval of the processing chain.

In the combination candidate table 128 of FIG. 6, for example, the processing chain of row 505 is a processing chain whose processing chain ID is identified by “processing chain ID # 1” (processing chain table 129 of FIG. 7A), and the output data thereof. The accuracy is "error radius 5 m or less", the cost is "0.07 yen / minute", and the operating rate is "90%".

The above-mentioned accuracy "error radius of 5 m or less" is the sensor "Dev- # 1" included in the processing chain table 129 of FIG. 7A and the extraction processing "extraction processing- # A", "extraction processing- # B", "extraction processing- # B". Of the "extraction process- # C", the value "with an error radius of 5 m or less", which is the least accurate value, is specified by the management server 100. In addition, the cost "0.07 yen / minute" is similarly determined by the management server 100 with the cost of the sensor "Dev- # 1" and the extraction processes "extraction process- # A", "extraction process- # B", and ". Each cost of "Type- # 2", "Type- # 3", and "Type- # 7" used for each input of "Extraction process- # C" is specified in the cost column 308 of the physical sensor attribute table 126. However, it is the total value of these. Further, the operating rate "90%" means that the management server 100 determines the operating rate of the sensor "Dev- # 1", the extraction process "extraction process- # A", the "extraction process- # B", and the "extraction process- # C". The operating rates of "Type- # 2", "Type- # 3", and "Type- # 7" used for each input are specified in the operating rate column 309 of the physical sensor attribute table 126, and these are specified. Is the value multiplied by.
The above-mentioned various metadata may be acquired from another information processing system.

7A to 7D are examples of the processing chain table 129 in this embodiment. The processing chain table 129 is a table that defines a series of physical sensors 20, extraction processing, and endpoints that constitute the above-mentioned processing chain, and the data input / output relationship between each of them.

In the processing chain table 129, the physical sensor 20 constituting the processing chain or the "input" field 601 for storing the information of the extraction processing, and the sensing data of the physical sensor 20 in the input field 601 or the output data of the extraction processing are output. It is composed of an "output" column 604 that stores information on the previous extraction process or endpoint process.

For example, in the processing chain table 129 of FIG. 7A, row 603 is a processing chain starting from the physical sensor 20 “Dev- # 1”, and the output destination of the sensing data of the physical sensor 20 “Dev- # 1” is It indicates that it is "extraction process- # A". Similarly, line 604 indicates that the sensing data output from "extraction process- # A" is input to "extraction process- # B" at the output destination. Similarly, line 605 indicates that the sensing data output from "extraction process- # B" is input to "extraction process- # C" at the output destination. Similarly, line 606 indicates that the sensing data output from "extraction process- # C" is input to the output destination "endpoint process A".

As described above, the processing chain table 129 of FIG. 7A has a series of flow of "Dev- # 1 ⇒ extraction process- # A ⇒ extraction process- # B ⇒ extraction process- # C ⇒ endpoint process A" as a whole. It represents a processing chain in which data is processed.
--- Flow example 1 ---

Hereinafter, the actual procedure of the sensor network management method in the present embodiment will be described with reference to the drawings. Various operations corresponding to the sensor network management method described below are realized by a program that is read and executed by the management server 100 that mainly constitutes the sensor network management system 1 into a memory or the like. Then, this program is composed of the code for performing various operations described below.

FIG. 8 is a diagram showing a flow example 1 of the sensor network management method in the present embodiment, and specifically, is an example of a flowchart in the matching processing unit 112 of the management server 100.

In this case, the matching processing unit 112 of the management server 100 selects a combination of an application that requires sensing data and a sensing means that can provide sensing data that satisfies the requirements of the above-mentioned application. Further, the matching processing unit 112 is executed when the requirements of the application are input from the user of the above-mentioned application through the input / output processing unit 121.

Here, the matching processing unit 112 stores three types of tables, an application attribute table 125, a physical sensor attribute table 126, and an extraction processing management table 127, which are generated by a user or the like via the input / output processing unit 111. The combination extraction processing unit 1121 is called with these as inputs, and the combination candidate table 128 is created (701). The flow (FIG. 9) in the combination extraction processing unit 1121 will be described later.

Next, the management server 100 calls the combination selection processing unit 1122 using the combination candidate table 128 generated in step 701 as an input, and selects one row of the combination candidate table 128 as the combination candidate sensing means. (702). Details of this step 702 will be described later based on FIG.

Subsequently, the management server 100 sets the processing chain ID (example: “processing chain ID # 3”) of the combination candidate selected in step 702 described above in the processing chain ID column 207 in the record of the corresponding application in the application attribute table 125. Register at (703).
The above flow determines the combination of the application and the sensing means that meets the requirements.
--- Flow example 2 ---

FIG. 9 is a diagram showing a flow example 2 of the sensor network management method in the present embodiment, and specifically, is a flow example of processing in the combination extraction processing unit 1121 in the above-mentioned flow example 1.

This flow corresponds to a process in which the combination extraction processing unit 1121 lists a combination of an application that requires sensing data and a sensing means that can provide sensing data that satisfies the requirements of the above-mentioned application. The combination extraction processing unit 1121 handles the endpoint processing as one of the extraction processes.

In this case, the combination extraction processing unit 1121 sets the "desired data type" (801).
). This process is recursively called multiple times. When this process is first executed, the data type specified in the data type of the application attribute table 125 is treated as an input to the endpoint process, and the above data type is set as the desired data type. In the second and subsequent calls, the input to the specified extraction process is set as the desired data type.

For example, the combination extraction processing unit 1121 in the step 801 sets the desired data type to "Type- # 1" with reference to the data type column 202 of the row 210 in the application attribute table 125.

Subsequently, the combination extraction processing unit 1121 refers to the data type column of the physical sensor attribute table 126, and lists the physical sensors that output the desired data type (802).

If one or more physical sensors 20 are found during this process (802: YES), the combination extraction processing unit 1121 registers the physical sensors 20 in the candidate list of this input (803). On the other hand, when the physical sensor 20 is not found in the above processing (802: NO), the combination extraction processing unit 112 skips step 803.

In the case of the above example, when the combination extraction processing unit 1121 refers to the physical sensor attribute table 126 in this step 802, since there is no physical sensor 20 whose output data type matches "Type- # 1", the step 803 will be skipped.

Subsequently, the combination extraction processing unit 1121 refers to the extraction processing management table 127, and executes the processing up to the following step 808 for each row representing the extraction processing for outputting the data of the desired data type (804).

In the case of the above example, among the processes included in the extraction process management table 127, the output matches "Type- # 1" only in the four processes of line 413, line 415, line 416, and line 418. Therefore, the combination extraction processing unit 1121 repeatedly executes the processing up to step 808 for each of the four rows.

Of these, in step 805, the combination extraction processing unit 1121 refers to the searched list, and if the extraction processing specified in step 804 has already completed the search, the process proceeds to step 808. In the case of the above example, since the searched list is empty, the process proceeds to step 806. Further, the combination extraction processing unit 1121 is recursively called in step 806 for each of the four processes of line 413, line 415, line 416, and line 418.

On the other hand, when the search has not been completed for the extraction process specified in step 804, the combination extraction processing unit 1121 adds the extraction process to the searched list in step 806. Further, the combination extraction processing unit 1121 is recursively called for the extraction processing. When the above processing is completed, the combination extraction processing unit 1121 deletes the extraction processing from the searched list.
Further, in step 807, the combination extraction processing unit 1121 adds the call result to the candidate list of this input.

Further, after the completion of the above routines 804 to 808, in step 809, the combination extraction processing unit 1121 determines whether or not there are candidates for all the inputs. As a result of this determination, if there is no candidate in the input (809: NO), the combination extraction processing unit 1121 returns no candidate in step 810.

On the other hand, if there are candidates for all the inputs as a result of the above determination (809: YES), the combination extraction processing unit 1121 creates a processing chain table 129 for the combination of the candidate list of each input in step 811. And register it in the result list.

Next, in step 812, the combination extraction processing unit 1121 determines whether or not the extraction processing is for the endpoint processing. As a result of this determination, when the present extraction process is for the endpoint process (812: YES), the combination extraction processing unit 1121 records the contents of the result list as the combination candidate table 128 in step 813. Further, the combination extraction processing unit 1121 obtains and registers the values of the columns such as the accuracy and cost of the combination candidate table 128 from the processing chain table 129 of each entry.

Subsequently, in step 813, the combination extraction processing unit 1121 assigns a processing chain ID to each row of the result list, and refers to the physical sensor attribute table 126 and the extraction processing management table 127 for each row. Register for accuracy, cost, and operating rate.

In the accuracy column, for the sensors and extraction processing included in the processing chain table 129, the lowest accuracy value is entered with reference to the accuracy column 307 of the physical sensor attribute table 126 and the accuracy column 409 of the extraction processing management table 127. sign up.
Further, in the cost column, for the sensors included in the processing chain table 129, the sum of the charges is registered with reference to the cost column 308 of the physical sensor attribute table 126.
Further, in the operating rate column, the product of the operating rates is registered for the sensors included in the processing chain table 129 with reference to the operating rate column 309 of the physical sensor attribute table 126.

On the other hand, as a result of the above determination, when the present extraction process is not for the endpoint process (812: NO), the combination extraction process unit 1121 returns the result list in step 814 and ends the flow.

Here, a further specific example of the above-mentioned flow will be shown. Here, the operation of the combination extraction processing unit 1121 called for "extraction processing- # C" in line 413 of the extraction processing management table 127 as the processing target in the above-mentioned step 804 will be described. For convenience, this combination extraction process is referred to as extraction process 413.

In this case, the combination extraction processing unit 1121 sets the desired data type to "Type- # 7" in step 801. Further, the combination extraction processing unit 1121 determines in step 802 that there is no sensor that outputs the data of "Type- # 7".

Further, in step 804, the combination extraction processing unit 1121 refers to the extraction processing management table 127 again, and specifies that the extraction processing for outputting the data of "Type- # 7" is only the row 412.

Subsequently, in step 805, the combination extraction processing unit 1121 identifies that only line 413 is described in the searched list, and line 412 is not included.

Further, in step 806, the management server 100 calls the combination extraction processing unit 1121 again for the above-mentioned line 412. For convenience, the combination extraction processing unit called here is referred to as extraction processing 412. In the extraction process 412, the combination extraction processing unit 1121 is similarly called for the line 411. For convenience, the combination extraction processing unit called here is referred to as extraction processing 411. In this extraction process 411, the desired data type is set to "Type- # 2" in step 801.

In step 802, it can be seen that the sensor that outputs the data of "Type- # 2" has line 311. Therefore, in step 803, the combination extraction processing unit 1121 registers the line 311 in the input candidate list of the line 411.

Further, in step 804, the combination extraction processing unit 1121 refers to the extraction processing management table 127 again, and finds that there is no process for outputting the data of "Type- # 2", and therefore proceeds to step 809.

Further, the combination extraction processing unit 1121 knows that the input candidate for the line 411 is registered in the input candidate list in the input candidate list in the step 809, and therefore proceeds to the step 811.

Further, since the number of input data types for the row 411 is 1, the combination extraction processing unit 1121 creates one processing chain table 129 for which the contents of the combination of the candidate list are empty in step 811. Further, when the combination extraction processing unit 1121 inputs the data of the data type "Type- # 2" into the input field 601 of the "sensor Dev- # 1" that outputs the data of the data type "Type- # 2", the combination extraction processing unit 1121 inputs the data of the data type "Type- # 2". A line 603 is generated in which the "extraction process- # A" for outputting the data of the data type "Type- # 3" is held in the output column 602.
Further, the combination extraction processing unit 1121 adds row 603 to the processing chain table 129 and registers it in the result list.

Further, the combination extraction processing unit 1121 proceeds to step 814 because the input to the main extraction processing is line 411 and is not an endpoint in step 812.

Further, in step 814, the combination extraction processing unit 1121 returns the result list in which the line 605 is registered, ends the extraction processing 411, and returns to step 806 of the extraction processing 412. Subsequently, the extraction process 412 is deleted from the searched list.

Further, in step 807, the combination extraction processing unit 1121 adds the processing chain table 129 holding only the row 603 which is the result of the extraction processing 411 obtained in step 806 to the input candidate list for the processing 412.

Further, in step 808, the combination extraction processing unit 1121 ends the loop starting from step 804 in the extraction processing 412. In step 809, only the result of the extraction process 411 is stored in the input candidate list for the row 412.

Subsequently, in step 811 the combination extraction processing unit 1121 inputs "extraction processing- # A" for outputting data of the data type "Type- # 3" into the input field 601 and data of the data type "Type- # 3". Is input, a line 604 that holds the "extraction process- # B" that outputs the data of the data type "Type- # 7" in the output field 602 is generated, so the generated line 607 is added to the result list. To do. After that, the combination extraction processing unit 1121 returns the result list to which the line 604 is added and returns to step 806 of the extraction processing 413, similarly to the extraction processing 411. In step 806 and subsequent steps of the extraction process 413, the same process as the extraction process 412 and 411 is performed, and "extraction process- # B" for outputting the data of the data type "Type- # 7" is entered in the input field 601 and the data type "Type- # 7" is input. -When the data of "# 3" is input, the line 605 that holds the "extraction process- # C" that outputs the data of the data type "Type- # 7" in the output field 602 is generated, so the line 605 is added. The result list is returned, and the process returns to step 806. By the processing up to this point, the processing chain table 129 excluding row 606 is generated.

In step 806, the combination extraction processing unit 1121 has the combination extraction processing unit 1121 (extraction processing 415) targeting the row 415 and the combination extraction processing unit 1121 (extraction processing) targeting the row 416 in the same manner as the extraction processing 413. 416), the combination extraction processing unit 1121 (extraction processing 418) for the line 418 is executed.

Processing chain table 129 (FIG. 7B) excluding row 615 as a result of extraction processing 415, processing chain table 129 (FIG. 7C) excluding row 622 as a result of extraction processing 416, and row 626 excluding row 626 as a result of extraction processing 418. The processing chain table 129 (FIG. 7D) is generated.

In step 807 of each extraction process, each result is registered in the input candidate list. In step 809, the input candidate list for endpoint processing is the processing chain table 129 excluding row 606, the processing chain table 129 excluding row 615, the processing chain table 129 excluding row 622, and the processing excluding row 626. There are four tables in the chain table 129.

In step 811, since the input to the endpoint process is only one type of data of "Type- # 1", the input combination is only four elements of the input candidate list. The combination extraction processing unit 1121 adds information regarding input to the endpoint to the combinations in the input candidate list.

The combination extraction processing unit 1121 inputs "extraction processing- # C" for outputting data of the data type "Type- # 1" to the processing chain table 129 excluding the row 609 in the input field 601 and the data type "Type- # 1". -Add line 606 to hold "endpoint processing A" with the data of "# 1" in the output field 602 to complete the processing chain table 129.

The combination extraction processing unit 1121 also performs the same processing on the processing chain table 129 excluding row 615, the processing chain table 129 excluding row 622, and the processing chain table 129 excluding row 626, and the respective results are obtained. As a result, each processing chain table 129 in FIGS. 7A to 7C is completed.

Further, as a result of step 811, four processing chain tables 129 in FIGS. 7A to 7D are registered in the result list. In step 812, the combination extraction processing unit 1121 proceeds to step 813 because each of the above-mentioned extraction processes is a combination extraction process with the endpoint as an output.

Further, in step 813, the combination extraction processing unit 1121 assigns a processing chain ID to each row of the result list, and refers to the physical sensor attribute table 126 and the extraction processing management table 127 for each row to obtain accuracy. , Cost, and occupancy rate.

Row 505 of the combination candidate table 128 is generated as follows. The combination extraction processing unit 1121 assigns "processing chain ID # 1" as the processing chain ID to the processing chain table 129 (FIG. 7A). Further, the combination extraction processing unit 1121 registers "error radius 5 m or less", which is the lowest accuracy value among the sensors and extraction processes included in the processing chain table 129, in the accuracy column. Further, the combination extraction processing unit 1121 refers to the cost column of the physical sensor attribute table 126 as the sum of the charges described in the cost column 308 of rows 314 and 315, which is "0.07 yen / minute". Is registered. Further, the combination extraction processing unit 1121 registers "90%", which is the product of the operating rates of the sensors included in the processing chain table 129, in the operating rate column. As a result of the above-mentioned extraction process, the combination candidate table 128 is generated.

Here, an example of the selection policy table 130 is shown in FIG. This selection policy table 130 is used in the combination selection processing unit 1122 to select a priority candidate from a plurality of combination candidates.

The selection policy table 130 in the present embodiment has a "priority" column 901 for storing the value of the priority relationship between a plurality of indicators (each row of the table) and an "index" column for storing the indicators for selecting candidates. It is composed of 902 and a "direction" column 903 that stores information indicating how the index has a high priority.

The index column 902 described above is newly constructed on, for example, the cost for providing the sensing means, the accuracy of the sensing data provided by the sensing means, and the data providing server 200 and the data providing server 200 for providing the sensing means. It represents numerical values such as the number of extraction processes required, the amount of resources such as a CPU required by the data providing server 200, and the operation cost of the system operation manager.

For example, row 906 of the selection policy table 130 holds "2" as the priority. Therefore, it is referred only when the candidate to be output is not determined by the index of priority "1". Further, since line 906 holds "number of extraction processes" as an index and "minimization" as a direction, the number of extraction processes that need to newly construct a sensing means on the data providing server 200 is reduced. Treat candidates as higher priority candidates.
--- Flow example 3 ---

FIG. 11 is a diagram showing a flow example 3 of the sensor network management method in the present embodiment, and specifically, is an example of a flowchart in the combination selection processing unit 1122.

In this case, the process by the combination selection processing unit 1122 is a process of selecting sensing means candidates from the combination candidate table 128 generated by the combination extraction processing unit 1121 described above. The combination selection processing unit 1122 is called by the matching processing unit 112 following the combination extraction processing unit 1121.

First, in step 1001, the combination selection processing unit 1122 in this flow repeats each process up to step 1004 below for each row included in the combination candidate table 128.

Of these, in step 1002, the combination selection processing unit 1122 specifies the combination candidates represented by the rows specified in step 1001 by the data acquisition interval 203, accuracy 204, cost 205, and operating rate 206 of the application attribute table 125. Determine if it meets the requirements of your application.

As a result of the above determination, when the combination candidate satisfies the requirements of the application (1002: YES), the combination selection processing unit 1122 adds the combination candidate to the final candidate list in step 1003.

On the other hand, if the combination candidate does not meet the requirements of the application (1002: NO), the combination selection processing unit 1122 skips step 1003.
Subsequently, the combination selection processing unit 1122 ends the iterative processing starting from step 1001 in step 1004.

Then, in step 1005, the combination selection processing unit 1122 determines whether the final candidate list is empty. As a result of this determination, if the final candidate list is empty (1005: YES), the combination selection processing unit 1122 notifies the user in step 1006 that there is no candidate satisfying the requirements of the application, and ends this processing. ..

On the other hand, if the final candidate list is not empty as a result of the determination in step 1005 (1005: NO), the combination selection processing unit 1122 refers to the priority of the selection policy table 130 in Tep 1007, and the desired candidate is the upper part of the table. Sort each row of the combination candidate table 128 so that it appears in.
Finally, the combination selection processing unit 1122 outputs the candidate at the top of the table in step 1008.

A more specific example accompanying the above flow will be described. In this case, the combination selection processing unit 1122 executes the processing up to step 1004 for each row of the combination candidate table 128 in step 1001. Further, the combination selection processing unit 1122 determines in step 1002 whether or not the items in each row satisfy the requirements of the application. In the following, in order to simplify the explanation, a case where all the items of the combination candidate form 128 satisfy the conditions will be described as an example.

Since all the items of rows 505, 506, and 507 of the combination candidate table 128 illustrated in FIG. 6 satisfy the conditions, the combination selection processing unit 1122 adds to the final candidate list in step 1003. In addition, line 508 has "0.5 yen / minute" registered in the cost column 503, and does not satisfy the cost "0.1 yen / minute or less" specified in the cost column 204 of the application attribute table 125. The combination selection processing unit 1122 skips step 1003 and proceeds to step 1004.
Subsequently, in step 1004, the combination selection processing unit 1122 ends the iterative processing starting from step 1002.

Further, the combination selection processing unit 1122 proceeds to step 1007 because the final candidate list is not empty in step 1005. Further, in step 1007, the combination selection processing unit 1122 sorts each row with reference to the selection policy table 130.

Since the index with the highest priority is cost and direction is minimized from the row 905 of the selection policy table 130, the combination selection processing unit 1122 is among the rows 505, 706, and 707 registered in the final candidate table. Then, the row 507 whose cost is "0.02 yen / minute" is sorted so as to come to the top of the final candidate table.

Further, in step 1008, the combination selection processing unit 1122 outputs the row 507 at the top of the final candidate table as a candidate. As a result, line 507 is the result of the selection process.
--- Flow example 4 ---

FIG. 12 is a diagram showing a flow example 4 of the sensor network management method in the present embodiment, and specifically, is an example of a flowchart in the board control processing unit 113.

In this case, the processing in the infrastructure control processing unit 113 is a processing in which the sensing means specified by the matching processing unit 112 is constructed and connected on the data providing server 200, and the output data by the sensing means can be used by the application. .. Such a basic control processing unit 113 is called by the management server 100 following the matching processing unit 112.

In this case, the infrastructure control processing unit 113 executes the construction connection processing unit 1131 in step 1101, and performs the extraction processing specified from the input combination candidate table 128 and the construction connection processing of the endpoint.

Further, in step 1102, the construction connection processing unit 1131 constructs an endpoint for accessing data from the application on the data providing server 200.
Further, in step 1103, the construction connection processing unit 1131 sets the endpoint ID of the endpoint in the endpoint ID column 208 of the row representing the application currently attracting attention in the application attribute table 125, and performs processing. finish.
As an operation example of the construction connection processing unit 1131 in the above flow, a case where row 507 of the combination candidate table 128 is input will be described.

In this case, in step 1101, the management server 100 calls the construction connection processing unit 1131 following the matching processing unit 112. A detailed operation example in step 1101 by the construction connection processing unit 1131 will be described with reference to FIG.

Subsequently, in step 1102, the construction connection processing unit 1131 enables the application to access the sensing means constructed and connected on the data providing server 200 in step 1101. When the application is accessed via the web, a unique URL is set in the sensing means, and a process of returning the sensing data when a request comes to the URL is constructed on the data providing server 200.

Subsequently, in step 1103, the construction connection processing unit 1131 sets a unique identifier for the endpoint constructed in step 1102 described above, and registers it in the endpoint ID column 208 of the application attribute table 125. When a unique URL is set for the endpoint in step 1102, the URL can be regarded as the endpoint ID. As described above, it becomes possible to access desired sensing data from the application.
--- Flow example 5 ---
FIG. 13 is a diagram showing a flow example 5 of the sensor network management method in the present embodiment, and specifically, is an example of a flowchart of the construction connection processing unit 1131.

The process in the construction connection processing unit 1131 is a process for constructing and connecting the sensing means specified by the matching processing unit 112 on the data providing server 200. The construction connection processing unit 1131 is called from the management server 100 as the first processing of the infrastructure control processing unit 113 following the matching processing unit 112.

In this case, in step 1201, the construction connection processing unit 1131 repeatedly executes the processing up to step 1204 for each extraction processing appearing in the input field and the output field of the processing chain table 129.

Of these, in step 1202, the construction connection processing unit 1131 determines whether or not the extraction processing specified in step 1201 has already been constructed on the data providing server 200. As a result of this determination, if the construction has been completed (1202: YES), the construction connection processing unit 1131 skips step 1203.

On the other hand, if the result of the above determination is that the construction has not been completed (1202: NO), the construction connection processing unit 1131 refers to the extraction processing program ID of the extraction processing program to be constructed in the extraction processing management table 127 in step 1203. Then, using this as a key, the relevant source code is specified in the source code repository 140, and the relevant source code is compiled on the data providing server 200, and appropriate implementation processing is performed. Note that step 1204 represents the end of the iterative process starting from step 1201.
Next, in step 1205, the construction connection processing unit 1131 displays the processing chain table 129.
The process up to step 1207 is repeatedly executed for each line of.

Of these, in step 1206, the construction connection processing unit 1131 sets the output of the data source or extraction processing specified in the input field to the input of the extraction processing or endpoint processing specified in the output field. Note that step 1207 represents the end of the iterative process starting from step 1205.

A specific operation example of the construction connection processing unit 1131 in the above flow will be described. Here, a case where row 507 of the combination candidate table 128 is input as a combination candidate as a result of the matching processing unit will be described.

In this case, the construction connection processing unit 1131 can trace from the processing chain ID “processing chain # 3” registered in the processing chain ID column 501 of line 507 in step 1201, the processing chain table 129 of FIG. 7C. With reference to each of the "extraction process- # B", "extraction process- # F", and "extraction process- # G" appearing in the processing chain, the processes of steps 1202 and 1203 are performed.

In the process executed for the "extraction process- # B", since the "extraction process- # B" has not been constructed in step 1202, the process proceeds to step 1203. Further, in step 1203, the construction connection processing unit 1131 implements the source code capable of executing the "extraction process- # B" stored in the source code repository 140 on the management server 100 on the data providing server 200. Build the corresponding process.

As a result of the iterative processing of steps 1201 to 1204 described above, the construction connection processing unit 1131 performs "extraction processing- # G" and "extraction processing- # F" in the same manner as the processing executed for "extraction processing- # B". Also, a process capable of executing each process is constructed on the data providing server 200.
In this case, the construction connection processing unit 1131 performs the processing of step 1206 for each row of the processing chain table 129 (FIG. 7C) in step 1205.

In step 1206, the construction connection processing unit 1131 has the output of the sensor "Dev- # 2" as the input of "extraction processing- # B" and the output of "extraction processing- # B" as "extraction processing- # F". For input, the output of the sensor "Dev- # 3" is the input of "Extraction process- # G", the output of "Extraction process- # G" is the input of "Extraction process- # F", and "Extraction process- #" Each setting is made so that the output of "F" becomes the input of "endpoint processing A". As a result, the sensing means represented by line 507 can be constructed and connected on the data providing server.
--- Flow example 6 ---

FIG. 14 is a diagram showing a flow example 6 of the sensor network management method in the present embodiment, and specifically, is an example of a flowchart of the reselection processing unit 115 in the management server 100.

The process of the reselection processing unit 115 in the present embodiment is a process for reselecting the sensing means once selected through the matching processing unit 112. The reselection processing unit 115 is called from the management server 100 when the provision of sensing data to the application is interrupted due to, for example, an abnormality in the physical sensor 20.
In this case, the reselection processing unit 115 identifies the application 30 using the physical sensor 20 having an abnormality in step 1301.
The reselection processing unit 115 executes each process up to step 1307 below for the application 30 specified here.

When the application 30 is specified in step 1301 described above, the reselection processing unit 115 saves the processing chain ID of the application attribute table 125 of the application 30 in step 1302.
Next, in step 1303, the reselection processing unit 115 creates a combination candidate table 128 excluding the combinations using the physical sensor 20 having an abnormality.

Subsequently, in step 1304, the reselection processing unit 115 executes the combination selection process using the combination candidate table 128 created in step 1303 described above, and reselects the sensing means.

Further, in step 1305, the reselection processing unit 115 constructs and connects the sensing means selected in step 1304 described above on the data providing server 200, and uses the endpoint used by the application 30 to create a new sensing means. Ensure that data is provided by sensing means.

Subsequently, in step 1306, the reselection processing unit 115 performs extraction processing that is no longer necessary by switching to a new sensing means on the data providing server 200 based on the processing chain ID saved in step 1302 described above. Remove from. Step 1307 represents the end of the iterative process starting from step 1301.

As an operation example of the reselection processing unit 115 in the above flow, the monitoring server 400 detects that the communication from the physical sensor "Dev- # 3" is interrupted, and notifies the management server 100 of this fact. A processing example will be described.

In this case, in step 1301, the reselection processing unit 115 traces the processing chain from the information registered in the processing chain ID column 207 of the application attribute table 125, and uses the application 30 in which the abnormal physical sensor 20 is used. Identify.

If "processing chain # 1" is registered in the processing chain ID column 207 of the application represented by row 210 of the application attribute table 125 (unregistered state in the example of FIG. 3), the reselection processing unit 115 The processing chain table 129 corresponding to the processing chain ID is specified as the processing chain table of FIG. 7C.

Processing chain of FIG. 7C There are two sensors included in Table 129, “Dev- # 2” and “Dev- # 3”, and the application 30 represented by line 210 uses “Dev- # 3” in which an abnormality has occurred. You can see that it is doing.

In step 1302, the reselection processing unit 115 temporarily saves the processing chain ID in the application attribute table 125. Further, in step 1303, the reselection processing unit 115 executes the combination extraction processing unit 1121 by inputting a table excluding the row 313 representing the physical sensor “Dev- # 3” as an input.
As a result of this, the reselection processing unit 115 outputs a table obtained by removing row 507 from the combination candidate table 128.

Subsequently, in step 1304, the reselection processing unit 115 executes the combination selection process by inputting the table obtained by removing the row 507 from the combination candidate table 128 as an input. At the time of step 1005 of the combination selection process, rows 506 and rows 507 of the combination candidate table 128 are registered in the final candidate list. In step 1007 of the combination selection process, the combination candidate table 128 is sorted with reference to the selection policy table 130 and the priority. Lines 506 and 507 both have a cost of "0.05 yen / minute", which is a key to be reprioritized. However, from the viewpoint of the number of extraction processes, which is the second priority key, since "extraction process- # B" has already been constructed on the data providing server 200, the extraction process that must be newly constructed must be constructed. In line 506, there are two numbers, "extraction process- # A" and "extraction process- # C", and in line 507, "extraction process- # A", "extraction process- # D", and "extraction process- # E". Since there are three, rows 506 are sorted so that they are at the top.
Therefore, the combination candidate output in step 1304 is row 506 of the combination candidate table 128.

In step 1305, the reselection processing unit 115 constructs the processing chain specified in row 506 of the combination candidate table 128 on the data providing server 200, and maintains a state in which the output data of the processing chain can be accessed from the application.

In step 1306, the reselection processing unit 115 refers to the endpoint ID column 208 of row 210 in the application attribute table 125 to identify the endpoint processing 136 that is providing data to the application 30. The reselection processing unit 115 reconnects the endpoints by setting the input to the endpoint processing 136 to the output of "extraction processing- # C" which is the end of the processing chain constructed in step 1305.

Further, the reselection processing unit 115 updates the processing chain ID column 207 of line 210 to “processing chain ID # 1”. Step 1307 represents the end of the iterative process starting with step 1301.

In step 1308, the reselection processing unit 115 deletes the extraction processing that is no longer used from the data providing server 200 based on the processing chain ID saved in step 1308. As a result of the processing in the reselection processing unit 115, the data is provided to the application 30 by the means represented by "processing chain # 1".
--- Flow example 7 ---

FIG. 15 is a diagram showing a flow example 7 of the sensor network management method in the present embodiment, and specifically, is an example of a flowchart of the undeployment processing unit 116.

The process of the undeployment processing unit 116 in the present embodiment is a process for erasing the extraction process that is no longer used by the application 30 from the data providing server 200, and when the application 30 is terminated or the sensing means is switched. Is called when As an operation example here, a case where the user ends the use of the application 30 ("App App- # 1" in FIG. 1) will be described.

First, in step 1401, the undeploy processing unit 116 traces the processing chain table 602 from the processing chain ID column 207 of the row 210 representing the application 30 in the application attribute table 125, and from the application 30, “extraction processing- # A”, “ It is specified that "extraction process- # B" and "extraction process- # C" are used, and each extraction process is deleted from the data providing server 200.

Next, in step 1402, the undeploy processing unit 116 identifies the endpoint processing 136 from the endpoint ID column 208 of the row 210 of the application attribute table 125, and deletes it from the data providing server 200.

As a result of the undeployment processing unit 116, the above-mentioned "App App- # 1"
The extraction process that became unnecessary due to the end of is deleted from the data providing server 200.
--- Example of interface ---

The sensor network management system 1 in the present embodiment can provide a user interface to a predetermined user or administrator. FIG. 16 is an example of a screen 1500 provided by the management server 100 so that the user can input the data requirements required by the application 30.

On this screen 1500, the user operates a predetermined terminal and specifies a file in which the requirements of the application 30 are described in the file selection box 1501. After that, the user presses the upload button 1502. The terminal described above receives the press of the upload button 1502 and transmits the specified file to the management server 100.

Subsequently, FIG. 17 shows an example of the screen 1601 provided by the management server 100 for allowing the user of the application 30 to select a sensing means that satisfies the requirements of the application 30.

In the flow of FIG. 11, an example has been described in which the combination selection processing unit 1122 automatically selects a combination determined according to the priority shown in the selection policy table 130. However, the user may manually select a candidate for the sensing means from the combination candidate table 128. In that case, the user of the application 30 is made to operate the combination selection screen 1601 to select a desired sensing means.

FIG. 17 shows an example of the combination candidate confirmation screen 1601 displayed on a predetermined terminal of the user after the processing of the combination extraction processing unit 1121 in step 702 in FIG. 8 is completed. The user operates the terminal and searches for the sensing means to be used in the application 30 from the combination candidate table 128 displayed on the screen 1601. In the combination candidate table 128, one row represents one combination.

In the screen 1601 of FIG. 17, the information "processing chain # 1" is displayed in the hyperlink format in the column of the processing chain ID column 1602 in the combination candidate 1608. By displaying this link destination, the user can display the processing chain table 129. After confirming the contents of each combination, the user presses the build button displayed in the selection column 1606 of the selected candidate row. In response to this pressing, the terminal notifies the management server 100 of the sensing means of the combination desired by the user. The management server 100 builds the sensing means on the data providing server 200.

Further, FIG. 18 shows an example of the management screen 1701 for the operator of the sensor network management system 1 in a certain implementation. The management screen 1701 for the operator is a screen for the operator to switch the sensing means, and can manually perform the process of selecting a candidate represented by steps 1007 and 1008 in the combination selection process. It is a screen to do.

The operator management screen 1701 selects combination candidate display units 1704 and 1705 and combination candidate detail display units 1706 and 1707 for displaying combination candidates included in the combination candidate table 128 for a certain application 30. It is composed of a switching button 1708 for executing switching to the sensing means.

As an operation example here, the provider of the physical sensor 20 informs the operator of the sensor network management system 1 that the physical sensor 20 having the device ID "Dev- # 3" is urgently stopped. Upon arrival, the operator of the sensor network management system 1 uses the sensing means using the physical sensor 20 having "Dev- # 3" and other sensing means not using the physical sensor 20 having "Dev- # 3". The case of switching to is described.

In this case, the operator of the sensor network management system 1 operates a predetermined terminal, traces the processing chain from the processing chain ID in the application attribute table 125, and uses the physical sensor 20 having "Dev- # 3". Identify the endpoint of the sensing instrument.

Here, as the sensing means using the physical sensor "Dev- # 3", the combination display unit 1704 shows the sensing means corresponding to the row 507 of the combination candidate table 128, and the detailed display unit 1706 shows the details of the sensing means. It is displayed.

The operator of the sensor network management system 1 operates a terminal, confirms other combination candidates displayed on the management screen 1701 for the operator, and selects a sensing means to be switched to. As another sensing means candidate for the "endpoint A", a diagram showing the sensing means corresponding to row 505 of the combination candidate table 128 is displayed on the combination display unit 1705, and the details of the sensing means are displayed on the detail display unit 1707. .. At the bottom of the screen, another sensing means candidate for "Endpoint A" is continuously displayed.

When the operator of the sensor network management system 1 presses the switching button 1708, step 1305 and subsequent steps of the reselection processing unit 115 are executed with row 505 of the combination candidate table 128 specified.

As described above, the sensing data and the data generated by applying the extraction process to the data can be matched with the requirements of the application 30. Further, when the sensing means being used becomes unavailable, the data can be continuously provided to the application 30 by switching to the alternative means. Further, the sensing means can be switched in order to reduce the cost of providing data and the management cost of the physical sensor 20 and the extraction process.

The data to be searched for by the application 30 is not limited to sensing data, but also data that is open to any user on an appropriate network such as the Internet or private data that is permitted to be accessed by the application executor. Any data existing on the network and accessible from the application 30, or a combination thereof may be targeted.

Further, in the present embodiment, the source code managed in the source code repository 140 of the management server 100 is targeted as the extraction process, but this may be registered in an executable form. Further, a process provided in the form of SaaS (Software as a Service) or the like on another server reachable by the network may be an extraction process.

Although the best mode for carrying out the present invention has been specifically described above, the present invention is not limited to this, and various modifications can be made without departing from the gist thereof.

According to the present embodiment, not only the sensing data directly provided from the physical sensor but also the data generated by applying a predetermined process to the sensing data can be matched with the requirements of the application. it can. In addition, when the physical sensor in use becomes unavailable due to the application, it is possible to switch to an alternative physical sensor, thereby maintaining the continuity of data provision to the application. further,
It is possible to suppress the number of physical sensors and the number of processes applied to the sensor data, which are required when providing data to each of a plurality of applications, and it is possible to reduce the cost for data provision and management.
That is, it is possible to flexibly provide the sensing data of the physical sensor in the sensor network in a form according to the requirements of the application.

The description herein reveals at least the following: That is, in the sensor network management method of the present embodiment, the information processing system may further execute a process of constructing a predetermined data extraction means included in the sensing means specified by the matching process.

According to this, for example, a processing chain selected by matching processing can be constructed on a data providing server and implemented in a state where it can be used by the corresponding application.

Further, in the sensor network management method of the present embodiment, the endpoint for enabling the information processing system to access the processing result by the predetermined data extraction means in the sensing means specified by the matching process from the predetermined application. The process of constructing may be further executed.

According to this, the data finally output by the sensing means can be used by the application, such as the acquisition destination of the data, that is, the endpoint (eg, the destination of the command requesting the output data, the location address of the output data, etc. ) Can be generated and used by the application.

Further, in the sensor network management method of the present embodiment, the information processing system further executes a process of connecting a predetermined physical sensor, a predetermined data extraction means, and a predetermined endpoint process in the sensing means specified by the matching process. You may do.

According to this, for example, the extraction process included in the process chain selected by the matching process can be made to run on the data providing server based on the source code stored in the source code repository. It is possible to connect these extraction processes so that the data requested by the relevant application can be generated.

Further, in the sensor network management method of the present embodiment, the information processing system executes a process of switching the sensing means of the application to another sensing means at a predetermined timing including at least a timing when a predetermined defect occurs. , May be.

According to this, even if a defect occurs in a physical sensor, for example, it is possible to newly identify a sensing means that satisfies the requirements of the application and provide it to the application.

Further, in the sensor network management method of the present embodiment, when an event occurs in which the information processing system does not use the data extraction means included in the predetermined sensing means, the constructed data extraction means included in the sensing means is used. The process of undeploying may be further executed.
According to this, it is possible to remove the data extraction means whose usage opportunity has disappeared from the sensing means and improve the utilization efficiency of the release.

Further, in the sensor network management method of the present embodiment, the information processing system
When an event occurs in which the endpoint process included in the predetermined sensing means is no longer used, a process of deleting the constructed endpoint process included in the sensing means may be further executed.
According to this, it is possible to delete unnecessary endpoint processing and improve the efficiency of the entire system.

Further, in the sensor network management method of the present embodiment, the information processing system has the metadata of the physical sensor, the metadata of the predetermined data extraction means, the sensing data of the physical sensor or the predetermined data based on the sensing data, and , The sensing data of the physical sensor or the metadata of the application that uses the predetermined data based on the sensing data, may be retained including the predetermined data acquired from another information processing system.

According to this, the sensor data of the physical sensor can be provided more flexibly and efficiently in a form according to the requirements of the application without limiting the range of the sensor network.

Further, in the sensor network management method of the present embodiment, the information processing system satisfies all the items of the predetermined requirements indicated by the metadata of the application among the plurality of sensing means obtained as the matching result in the matching process. Alternatively, a sensing means that satisfies some of the predetermined requirements indicated by the metadata of the application may be specified.

According to this, it is possible to identify and provide a sensing means capable of acquiring the optimum data for the application. In addition, for example, even if the requirements of the application are diverse, it is possible to flexibly respond to the situation where the essential ones are limited, and to specify and provide the sensing means within the range not limited more than necessary. You can.

Further, in the sensor network management method of the present embodiment, when the information processing system cannot specify a sensing means that satisfies a predetermined requirement indicated by the metadata of the application in the matching process, the requirement is satisfied for the application. The process of notifying that the sensing means could not be specified may be further executed.
According to this, the application can recognize the situation where appropriate data cannot be obtained in the sensor network in the first place, and can avoid unnecessary measures.

Further, in the sensor network management method of the present embodiment, the information processing system may further execute a process of accepting selection by a predetermined user with respect to a plurality of sensing means obtained as matching results in the matching process.
According to this, it is possible not only to specify an appropriate sensing means on the sensor network management system side but also to accept selection from the user.

Further, in the sensor network management system of the present embodiment, when the application stops using the data provided by the sensing means, or when the physical sensor included in the predetermined sensing means cannot be used. In that case, the deletion process of the data extraction means in the sensing means already constructed for the application may be further executed.

Further, in the sensor network management system of the present embodiment, when the data extraction means included in the predetermined sensing means becomes unavailable, the arithmetic unit deploys another constructed data extraction means included in the sensing means. It may be that the process of deploying is further executed.
According to this, it is possible to efficiently manage the data extraction means that configure the sensing means and connect each other's outputs as inputs according to the necessity of use.

Further, in the sensor network management system of the present embodiment, the arithmetic unit may further execute a process of constructing a predetermined data extraction means included in the sensing means specified by the matching process.

Further, in the sensor network management system of the present embodiment, the arithmetic unit constructs an endpoint for making the processing result by the predetermined data extraction means in the sensing means specified by the matching process accessible from the predetermined application. It may be said that the processing to be performed is further executed.

Further, in the sensor network management system of the present embodiment, the arithmetic unit further executes a process of connecting a predetermined physical sensor, a predetermined data extraction means, and a predetermined endpoint process in the sensing means specified by the matching process. It may be a thing.

1 Sensor network management system 5 Sensor network 10, 11 Network 20 Sensor 30 Application 100 Management server 101 Storage device 102 Program 103 Memory 104 Computing device 105 Communication device 110 Sensor information acquisition processing unit 111 Input / output processing unit 112 Matching processing unit 1121 Combination extraction Processing unit 1122 Combination selection processing unit 113 Infrastructure control processing unit 1131 Construction connection processing unit 1132 Endpoint creation processing unit 114 Alert reception processing unit 115 Reselection processing unit 116 Undeploy processing unit 125 App attribute table 126 Physical sensor attribute table 127 Extraction processing Management table 128 Combination candidate table 129 Processing chain table 130 Selection policy table 140 Source code repository 200 Data providing server 300 IoT hub server 310 Output processing unit 400 Monitoring server

Claims (15)

Metadata including the operating rate of the physical sensor included in the sensor network and the accuracy of the output data, the metadata regarding the accuracy of the output data of the predetermined data extraction means, and the sensing data of the physical sensor or the predetermined data based on the sensing data. An information processing system equipped with a storage device that holds metadata including mobility and the accuracy of sensing data, which is the percentage of time that the physical sensor provides normal sensing data, which is required by the application to be used.
The physical sensor and the data extraction means that satisfy the accuracy of the mobility and the sensing data required by the application based on the accuracy of the output data of each of the data extraction means and the accuracy of the operating rate and the output data of each of the physical sensors. Performs a matching process that identifies the sensing means in combination with
A sensor network management method characterized by this. The information processing system
Further executing the process of constructing the predetermined data extraction means included in the sensing means specified by the matching process.
The sensor network management method according to claim 1. The information processing system
Further, a process of constructing an endpoint for making the processing result of the predetermined data extraction means in the sensing means specified by the matching process accessible from the predetermined application is executed.
The sensor network management method according to claim 2. The information processing system
Further executing a process of connecting a predetermined physical sensor, a predetermined data extraction means, and a predetermined endpoint process in the sensing means specified by the matching process.
The sensor network management method according to claim 3, wherein the sensor network management method is characterized. The information processing system
A process of switching the sensing means of the application to another sensing means is executed at a predetermined timing including at least a timing at which a predetermined defect occurs.
The sensor network management method according to claim 4. The information processing system
When an event occurs in which the data extraction means included in the predetermined sensing means is no longer used, a process of undeploying the constructed data extraction means included in the sensing means is further executed.
The sensor network management method according to claim 4, wherein the sensor network is managed. The information processing system
When an event occurs in which the endpoint processing included in the predetermined sensing means is no longer used, a process of further deleting the constructed endpoint process included in the sensing means is further executed.
The sensor network management method according to claim 4, wherein the sensor network is managed. The information processing system
The metadata of the physical sensor, the metadata of the predetermined data extraction means, the sensing data of the physical sensor or the predetermined data based on the sensing data, and the sensing data of the physical sensor or the predetermined data based on the sensing data are used. Holds predetermined data acquired from other information processing systems as at least one of the application metadata.
The sensor network management method according to claim 1. The information processing system
Of the plurality of sensing means obtained as the matching result in the matching process, the sensing means that satisfies all the predetermined requirements indicated by the application metadata, or some items among the predetermined requirements indicated by the application metadata are satisfied. Identify the sensing means,
The sensor network management method according to claim 1. The information processing system
In the matching process, if the sensing means satisfying the predetermined requirement indicated by the metadata of the application cannot be specified, the process of notifying the application that the sensing means satisfying the requirement cannot be specified is further executed.
The sensor network management method according to claim 1. The information processing system
In the matching process, a process of accepting selection by a predetermined user is further executed for the plurality of sensing means obtained as the matching result.
The sensor network management method according to claim 1. Metadata including the operating rate of the physical sensor included in the sensor network and the accuracy of the output data, the metadata regarding the accuracy of the output data of the predetermined data extraction means, and the sensing data of the physical sensor or the predetermined data based on the sensing data. A storage device that holds metadata that includes the mobility and the accuracy of the sensing data, which is the percentage of time that the physical sensor provides normal sensing data, as required by the application to be used.
The physical sensor and the data extraction means that satisfy the accuracy of the mobility and the sensing data required by the application based on the accuracy of the output data of each of the data extraction means and the accuracy of the operating rate and the output data of each of the physical sensors. An arithmetic device that executes matching processing to specify the sensing means, which is a combination of
Sensor network management system with. The arithmetic unit
The process of constructing a predetermined data extraction means included in the sensing means specified by the matching process is further executed.
The sensor network management system according to claim 12. The arithmetic unit
The process of constructing an endpoint for making the processing result of the predetermined data extraction means in the sensing means specified by the matching process accessible from the predetermined application is further executed.
The sensor network management system according to claim 13. The arithmetic unit
Further, a process of connecting a predetermined physical sensor, a predetermined data extraction means, and a predetermined endpoint process in the sensing means specified by the matching process is executed.
The sensor network management system according to claim 14.

Images