JP7048402B2 - Data store system and data store management method - Google Patents

Description

The present invention generally relates to data stores.

Regarding the data store, for example, it is conceivable to use a virtualized IT system. Patent Document 1 describes "a fiber capable of setting a virtual SAN by connecting one or more computers capable of setting a virtual computer, one or more storage devices capable of setting virtual storage, and the computer and the storage. In a system including the computer, the storage device, and a management computer connected to the fiber channel switch via a network in which a virtual network can be set, the management computer can be used as virtual resource configuration management information. , Has physical resource configuration management information, searches for virtual resources and physical resource operations at a specified time, and dynamically outputs the search results as a checklist. "

Japanese Unexamined Patent Publication No. 2011-081579

In recent years, in the ever-increasing social infrastructure (water and sewage, gas, electricity, etc.), we believe that the replacement of aging pipes and accidents are also increasing. In the future, IoT (Internet of things) systems and IoT services will be required to curb the increase in TCO (Total Cost of Ownership), including the operation and maintenance of this social infrastructure. In particular, for IoT services that are shared by multiple local governments and lifeline companies, we believe that expectations for added value will increase, such as feedback of information that is valuable to users, depending on the usage status of the service. .. Municipalities and lifeline companies are examples of users.

Against this background, sensors (sensor devices) installed in water pipes under the jurisdiction of each municipality and measuring their status are arbitrarily grouped and connected to the sensor communication network wirelessly or by wire. Each of these groups is further connected to the public network (Internet).

A sensor data set from a sensor (eg, a data set containing a measurement date and time and a measurement value) is an example of a data store system, a data lake service system (eg, a flexible and scalable storage system) via the Internet. It is stored in the included system).

For exclusive use IoT systems rather than shared use, data lake service storage systems scale out and scale in to prevent wasting resources and service stagnation as data volumes and throughput increase and decrease. Is possible. However, in response to a sudden increase in the amount of data due to a wide-area disaster, etc., the peak of the amount of data is estimated in advance based on empirical knowledge, and operations are carried out on a scale with a margin. In addition, for known batch processing, schedules are set and executed in advance so that IT resources (for example, computer resources and storage resources) do not run short.

In the future, in order to reduce the increase in TCO of social infrastructure, it is expected that the use of IoT services that share IoT systems will be required.

In realizing such a service, the operation manager who manages the data lake service system as an IoT system can manage multiple users who share the system, so that the characteristics of each user (for example, customer) can be understood. It gets lower. As one method to solve this problem, a method of analyzing a time-series sensor data set for each user can be considered. However, since the data format of the sensor data set may differ depending on the user, it is necessary to manage the data format for each user and analyze the sensor data set according to the data format. It is considered that the burden is heavy.

Such issues may also apply to other data store systems that are shared by multiple users.

Each time the data store system satisfies a predetermined condition for the measured value of the sensor for each of the plurality of sensors existing in the plurality of segments related to the plurality of regions corresponding to the plurality of users, the sensor data of the sensor. It is designed to receive write requests for sets. When the data store system receives a write request for the sensor data set of any sensor, it corresponds to the user corresponding to the region where the segment in which the sensor is located is involved among the multiple volumes corresponding to each of the multiple users. Issue a write request for the sensor dataset to the volume. Each of the plurality of volumes is based on a group of scale-out storage nodes which are one or more storage nodes. A plurality of storage nodes corresponding to a plurality of volumes are a plurality of storage nodes existing in the same storage system. In each of the plurality of volumes, each sensor data set stored in the volume is associated with the geographical location of the segment in which the sensor data set resides and the time associated with the sensor data set. The data store system monitors the write frequency of each of the plurality of volumes for each segment constituting the region regarding the user corresponding to the volume.

According to the present invention, even an operation administrator who is not exclusive to a single user and does not understand the characteristics of each user can generate a specific event such as an abnormality based on the change in write frequency and the position of the segment in which the change occurs. It is possible to identify the area where it is located.

FIG. 1 is a configuration diagram of an entire system including a data lake service system according to an embodiment of the present invention. FIG. 2 is a diagram showing an example of a data structure of a data lake system. FIG. 3 is a diagram showing an example of the correspondence between the sensor installation position and the area. FIG. 4 is a diagram showing the configuration of the management table in the management DB. FIG. 5 is a diagram showing a configuration of a management table in the management DB. FIG. 6 is an internal configuration diagram of the sensor. FIG. 7 is an internal configuration diagram of the redirect server. FIG. 8 is an internal configuration diagram of the application server. FIG. 9 is an internal configuration diagram of the user terminal. FIG. 10 is a diagram showing an example of a setup screen. FIG. 11 is an internal configuration diagram of the management node. FIG. 12 is an internal configuration diagram of the storage node. FIG. 13 is an operation flow diagram of the sensor transfer program. FIG. 14 is an operation flow diagram of the user terminal program. FIG. 15 is an operation flow diagram of the user service program. FIG. 16 is an operation flow diagram of the monitoring program. FIG. 17 is an operation flow diagram of the file access program.

In the following description, the "interface unit" may be one or more interfaces. The one or more interfaces may be one or more communication interface devices of the same type (for example, one or more NICs (Network Interface Cards)) or two or more different types of communication interface devices (for example, NIC and HBA). (Host Bus Adapter)).

Further, in the following description, the "memory unit" is one or more memories, and may be typically a main storage device. At least one memory in the memory unit may be a volatile memory or a non-volatile memory.

Further, in the following description, the "PDEV unit" is one or more PDEVs, and typically may be an auxiliary storage device. "PDEV" means a physical storage DEVice, typically a non-volatile storage device such as an HDD (Hard Disk Drive) or SSD (Solid State Drive). That is, the PDEV unit is an example of the storage device unit.

Further, in the following description, the "storage unit" is at least one of the memory unit and the PDEV unit (typically, at least the memory unit).

Further, in the following description, the "processor unit" is one or more processors. The at least one processor is typically a microprocessor such as a CPU (Central Processing Unit), but may be another type of processor such as a GPU (Graphics Processing Unit). At least one processor may be single-core or multi-core. The at least one processor may be a processor in a broad sense such as a hardware circuit (for example, FPGA (Field-Programmable Gate Array) or ASIC (Application Specific Integrated Circuit)) that performs a part or all of the processing.

Further, in the following description, information that can be output with respect to an input may be described by an expression such as "xxx table", but this kind of information may be data of any structure and may be input. It may be a learning model such as a neural network that produces an output for. Therefore, the "xxx table" can be referred to as "xxx information". Further, in the following description, the configuration of each table is an example, and one table may be divided into two or more tables, or all or a part of the two or more tables is one table. May be.

Further, in the following description, the process may be described with "program" as the subject, but the program is executed by the processor unit to appropriately perform the specified process in the storage unit and / or the interface unit, etc. The subject of the process may be a processor unit (or a device such as a controller having the processor unit). The program may be installed from the program source into a device such as a calculator. The program source may be, for example, a program distribution server or a computer-readable (eg, non-temporary) recording medium. Further, in the following description, two or more programs may be realized as one program, or one program may be realized as two or more programs.

Further, in the following description, the function may be described by the expression of "kkk unit" (excluding the interface unit, the storage unit, and the processor unit), but the function is executed by one or more computer programs by the processor unit. It may be realized by the above, or it may be realized by one or more hardware circuits. The description of each function is an example, and a plurality of functions may be combined into one function, or one function may be divided into a plurality of functions.

Further, in the following description, the "storage system" is a system including a plurality of storage nodes. From the plurality of storage nodes, a plurality of storage node groups that are the basis of a plurality of volumes corresponding to a plurality of users may be constructed. The relationship between the volume and the storage node group may be 1: 1, 1: many, many: 1, many: many. The storage node group may be typically referred to as a scale-out storage system or a distributed storage system. The "storage node" is a device as a member of the storage system, and may be a general-purpose computer or a dedicated computer (for example, a storage device such as a so-called disk array device having a plurality of PDEVs). The storage node group may have a redundant configuration group. The redundant configuration may be a configuration with a plurality of storage nodes such as Erasure Coding, RAIN (Redundant Array of Independent Nodes) and inter-node mirroring, or one or more RAIDs (Redundant Array of Independent) as at least a part of the PDEV part. (or Inexpensive) Disks) It may be configured by a single computer (for example, a storage node) such as a group.

Further, in the following description, a "data set" is a logical mass of electronic data seen from a program such as an application program, and is, for example, a record, a file, a key-value pair, or a taple. But it's okay.

Further, in the following description, the common part of the reference codes may be used when the same type of elements are not distinguished, and the reference code may be used when the same type of elements are described separately. .. For example, when the storage nodes are described without any distinction, they are described as "storage node 14", and when the individual nodes are described separately, they are described as "storage node 14a" and "storage node 14b". May be described.

Hereinafter, an embodiment of the present invention will be described with reference to the drawings. The "segment" referred to in the following description may be, for example, a geometric segment.

Hereinafter, details will be described using the configuration of each element and the operation flow.

In the following description, the operation and specifications of the storage node group (for example, the distributed file storage system) (for example, the data access program 1410, the node expansion program 1411, and the node reduction program 1412 described later) that realize the volume 103 will be described. , For example, a known technique such as Ceph may be followed.

FIG. 1 is a configuration diagram of an entire system including a data lake service system according to an embodiment of the present invention.

A plurality of sensors 201 are arranged in a plurality of areas composed of a plurality of segments.

For each user, a sensor data set (for example, a data set including the measured value of the sensor 201 and the measurement date and time) from each of one or more sensors 201 located in the area related to the user is collected, analyzed, and visualized. The user application system 101 in which the processing of the above is executed is provided. At least a part of the user application system 101 may be common to a plurality of users.

A data lake service system 102 is provided that stores the sensor data set from the user application system 101 and provides the requested data from the user application system 101. The data lake system 102 includes a management node 13 and one or more storage nodes 14. At least the management node 13 of the data lake system 102 is an example of a data store system.

Between each sensor 201 and the user application system 101, there is a sensor network 2 that mediates a sensor data set. Further, there is a data lake network 1 that mediates data communication between the user application system 101 and the data lake service system 102. The sensor network 2 may be either a wireless network (WiFi (registered trademark), LPWA (Low Power Wide Area), Bluetooth (registered trademark), etc.) or a wired network. The data lake network 1 may be any communication network that can provide data communication, such as the Internet, LAN (Local Area Network), and WAN (Wide Area Network).

The schematic geography 100 in the figure is a notation that simplifies the explanation that the plurality of sensors 201 are set in a plurality of zones including a plurality of actual regions. As an example, the boundaries of the Japanese continent are also shown. For example, it can be seen that a certain sensor 201 is installed in the northwestern part of Kyushu.

The user application system 101 includes a redirect server 9, an application server 10, and a user terminal 11.

The redirect server 9 collects sensor data sets from each sensor 201 via the sensor network 2, and transmits (stores) the collected sensor data sets to the volume 103 of the data lake service system 102.

The application server 10 acquires a sensor data set held by the volume 103 of the data lake service system 102 via the data lake network 1 and executes a process of analyzing or visualizing the sensor data set.

The user terminal 11 displays the result of the visualization process executed by the application server 10 and performs a setting operation of the data lake service system 102.

As described above, the data lake service system 102 has one or more storage nodes 14 for holding data and a management node 13 for managing them. The storage node 14 has a PDEV unit 15. The management node 13 has a management DB (database) 4 that holds a plurality of management tables described later.

A storage node group (one or more storage nodes 14) as a distributed file storage system is used as a volume 103 from the user application system 101. Further, when the scale-out is performed, as shown by the arrow 105, some new storage nodes 14d are added as members to provide the volume 103A while maintaining the retained data, and as a result, the volume is added. 103A (volume based on storage nodes 14a to 14c) is expanded to volume 103B. The expanded volume 103B is used from the user application system 101. On the contrary, as shown by the arrow 106, when the scale-in is performed, some storage nodes 14d are reduced from the members providing the volume 103B while maintaining the retained data, resulting in the result. As a result, the volume 103B is reduced to the volume 103A. The reduced volume 103A is used from the user application system 101. However, since the data set is multiplexed by the distributed file storage system, the number of storage nodes 14 is generally larger than the number of multiplexes.

The overall configuration of the system of this embodiment has been described above.

FIG. 2 shows an example of the data configuration of the volume 103 included in the data lake service system 102.

This data structure is grasped by the data lake access programs 913 and 1011 (FIGS. 7 and 8) that access the data lake service system 102.

The sensor data set held by the volume 103 is a file in this embodiment. In the volume 103, the data structure is a tree structure (hierarchical structure) including nodes such as folders (directories) and files and edges between the nodes, like a namespace. The file is stored in the folder at the measurement date and time and the measurement position associated with the file among the hierarchical folders.

According to the example of FIG. 2, one or more folders “Daytime” 213 classified by the measurement date and time belong to the folder “root” 214 at the head of the volume 103. Each folder "Daytime" 213 belongs to one or more folders "Segment" 212 classified by measurement position. One or more files 211 are stored in the folder "Segment" 212. Specifically, according to the data structure of FIG. 2, the storage destination folder of a certain file 211 is a plurality of folders "Segment" belonging to the folder "Daytime" 213 to which the measurement date and time associated with the certain file 211 belongs. "212, the folder" Segment "212 corresponding to the segment (district) 202 to which the measurement position associated with the certain file 211 belongs.

The folder "Segment" 212 may be defined by, for example, information obtained by a calculation formula from the latitude and longitude of the measurement position. It is also possible to use information corresponding to a place name that identifies a region.

The hierarchical relationship between the folder "Daytime" 211 and the folder "Segment" 212 may be reversed, but if the file (sensor data set) is not continuously retained due to the problem of sensor life, etc., the file In the process of specifying the folder, if the folder date and time is higher, the narrowing down is performed first, and unnecessary selection processing can be prevented. If there is no performance requirement, either the folder "Daytime" 211 or the folder "Segment" 212 may be higher.

FIG. 2 also shows an outline of the flow of an example of the processing performed in this embodiment. The data set from the sensor 201 is uploaded (transmitted) as a file 211 to the volume 103 for the user corresponding to the area where the sensor 201 is located (S1) via the redirect server 9. .. The upload destination of the file 211 is the folder “Segment” 212 corresponding to the measurement date and time and the measurement position of the sensor 201. When the write frequency for the segment 202 corresponding to the upload destination folder "Segment" 212 satisfies the notification condition for the user corresponding to the area including the segment 202, the management node 13 assigns the segment 202 to the upload. Detect (S2). If necessary, the management node 13 sends a notification message to the application server 10 of the user corresponding to the area including the detected segment 202 (S3), and scales the volume 103 of the upload destination. Is requested to the storage system (for example, at least one storage node 14 which is the basis of the volume 103) (S4). When the application server 10 receives the cancellation request during the display operation from the user terminal 11, the application server 10 transmits the above-mentioned scaling (scale-out or scale-in) cancellation request to the management node 13. In response to the cancellation request, the management node 13 requests the storage system to cancel the scaling requested in S4.

The method of placing the file (sensor data set) on the volume 103 has been described above.

FIG. 3 shows an example of the correspondence between the installation position of the sensor 201, the segment 202, and the area 303 represented by the place name.

In this embodiment, a large number of sensors 201 used by a plurality of users are widely installed. Further, each sensor 201 may be the property of a certain user, or may be a sensor provided by a third party across the region. When a user uses a sensor data set for a certain area 303, it is necessary to access the sensor data set narrowed down to the sensor 201 belonging to the area 303.

According to FIG. 3, the region 303 and the segment 202 are schematically associated with each other based on the two-dimensional plane defined from the latitude and longitude. For example, for a rectangular range 302 (range of a broken line frame) including a region 303, one or more segments 202 overlapping the rectangular range 302 (for example, forming the rectangular range 302 (or region 303)) are defined. .. Therefore, the segment 202 related to the region 303 is all or part of the segment 202 that overlaps the rectangular range 302 (for example, the segment 202 in which the area of the overlapping portion with the region 303 is equal to or larger than a predetermined ratio of the area of the segment 202). .. The correspondence between the segment 202 and the region 303 may be 1: 1, 1: many, many: 1 or many: many. Since the segment 202 corresponds to the above-mentioned folder “Segment” 212, the sensor data set of the sensor 201 installed in the segment 202 can be taken out from the folder “Segment” 212 corresponding to the segment 202.

Further, whether or not the position (measurement position) of the sensor 201 on the two-dimensional position coordinate system represented by latitude and longitude is included in the area 303 represented as a broken line designated by a plurality of position coordinates (vertices). As a method of determining whether or not, a known method such as a method according to the selection function of OpenGL can be adopted.

4 and 5 show the configuration of the management table included in the management DB 4.

The regional property management table 401 holds information about regional properties. For example, the regional property management table 401 has columns "regional ID", "regional boundary adjustment" and "regional area". That is, the area property management table 401 stores the area ID for each area, and also holds the information indicating the area property such as the area boundary adjustment and the area area.

The column "region ID" holds a unique region ID corresponding to the place name. For example, as the area ID, the identification number used in the map data published by the Geospatial Information Authority of Japan may be used.

The column "Regional Boundary Adjustment" holds a list of the number of vertices and coordinates when the boundary line is expressed as a broken line as the boundary line information that defines the boundary line of the region. The boundary line information may include coordinate information and the like used in the map data published by the Geospatial Information Authority of Japan.

The column "regional area" holds the area of the region. The area of this area may be obtained based on the coordinates of the vertices of the area by using, for example, an equation such as Equation 1.

The area-related management table 402 of FIG. 4 holds information about the area other than the information belonging to the area property. For example, the region-related management table 402 has columns "region ID", "region name", and "user ID". That is, the area-related management table 402 stores the area ID for each area and holds information such as the area name and the user ID.

The column "region ID" holds the above-mentioned region ID. The column "region name" holds information indicating the place name of the region. The column "user ID" holds the user ID of the user who handles the area.

The user management table 501 of FIG. 5 holds information about the user. For example, the user management table 501 has columns "user ID", "user name" and "notification condition". That is, the user management table 501 holds information such as a user ID, a user name, and notification conditions for each user (for example, a customer who uses the data lake service).

The column "user ID" holds the user ID. The column "user name" holds the name of the user. The column "Notification condition" holds the notification condition entered by the user. Details of the notification conditions will be described later.

The statistical management table 502 holds the write frequency statistical information for each segment. For example, the statistical management table 502 has columns "segment ID", "folder path", "long-term write frequency", and "short-term write frequency". That is, the statistical management table 502 stores the area ID for each segment, and also holds information such as the area name and the user ID.

The column "segment ID" holds the ID of the segment. The segment ID may be determined based on the measurement position (for example, latitude and longitude) of the sensor by using an equation such as Equation 2. As a result, the range of the installation position (measurement position) of the sensor that sent the sensor data set can be calculated from the segment ID.

The column "folder path" holds the path name of the path to the corresponding folder "Segment".

The column "Long-term write frequency" is the value of the long-term write frequency (for example, the moving average per week when writes to the corresponding folder "Segment" occur, which is the write frequency over a relatively long period, per day. Average write frequency). The column "Long-term write frequency" is a short-term write frequency that is a write frequency in a relatively short period (for example, an average write frequency such as a moving average value per day when a write to the corresponding folder "Segment" occurs). To hold. Each of the "relatively long period" and "relatively short period" is a period in which one is longer or shorter than the other.

The management table of the management DB 4 has been described above.

FIG. 6 shows the internal configuration of the sensor 201.

Sensor 201 is a device that includes one or more sensor modules 806. For example, the sensor 201 has a microcontroller 802, a sensor network interface 803, a memory 804, a real-time clock module 805, one or more sensor modules 806, and a data path 801 that mediates data communication between these elements.

The sensor network interface 803 is an example of an interface unit, and is, for example, a NIC (Network Interface Card) module for transmitting and receiving data via an Ethernet protocol (Ethernet is a registered trademark) or LPWA. It can be connected to the sensor network 2 via the sensor network interface 803.

The memory 804 is an example of a storage unit (for example, a memory unit). The memory 804 is, for example, a sensor measurement program 810 and a sensor transfer program as one or more programs (one or more programs executed by a microcontroller 802 (an example of a processor unit)) for the operation of the sensor 201. Stores 811 and. Further, the memory 804 has a sensor buffer 812 as a data holding area. Further, the memory 804 stores the sensor position information 813. The sensor position information 813 is position information (for example, latitude and mild) indicating the installation position of the sensor 201. During the installation work of the sensor 201, the sensor position information 813 indicating the installation position of the sensor 201 is recorded. Sensor position information 813 is associated with the sensor data set from sensor 201. Further, when the sensor 201 has a position detection function such as a GPS (Global Positioning System) device, information indicating the position detected by the function may be recorded as the sensor position information 813. The explanation of each program will be described later.

The real-time clock module 805 may be a clock module that outputs the current date and time.

Each of the one or more sensor modules 806 may be a sensor module adopted according to the application. For example, the sensor module 806a may be a module for measuring the vibration value, and the sensor module 806b may be a module for measuring the acceleration.

FIG. 7 shows the internal configuration of the redirect server 9.

The redirect server 9 has an MPU 902, a memory 903, a network interface 904, and a data path 901 that mediates data communication between these elements.

The MPU 902 is an example of a processor unit and is a microprocessor. The memory 903 is an example of a storage unit (for example, a memory unit). The memory 903 stores a data conversion program 911, a data recording program 912, and a data lake access program 913 as one or more programs (one or more programs executed by the MPU 902) for the operation of the redirect server 9. .. Further, the memory 903 has a detection data buffer 914 as a data holding area. The explanation of each program will be described later.

The network interface 904 is an example of an interface unit, and is, for example, a NIC module that can handle an Ethernet protocol. As a result, the sensor network 2 and the data lake network 1 can be connected.

FIG. 8 shows the internal configuration of the application server 10.

The application server 10 has an MPU 1002, a memory 1003, a network interface 1004, and a data path 1001 that mediates data communication between these elements.

The MPU 1002 is an example of a processor unit and is a microprocessor. The memory 1003 is an example of a storage unit (for example, a memory unit). The memory 1003 stores the sensor program 1010, the data lake access program 1011 and the user program 1012 as one or more programs (one or more programs executed by the MPU 1002) for the operation of the application server 10. Further, the memory 1003 has a monitoring data buffer 1013 as a data holding area. The explanation of each program will be described later.

The network interface 1004 is an example of an interface unit, and is a NIC module that can handle an Ethernet protocol. As a result, it is possible to connect to the data lake network 1.

FIG. 9 shows the internal configuration of the user terminal 11.

The user terminal 11 has an MPU 1102, a user interface 1103, a memory 1104, a network interface 1105, and a data path 1101 that mediates data communication between these elements.

The MPU 1102 is an example of a processor unit and is a microprocessor.

The user interface 1103 is an interface module for transmitting and receiving display data (display information) to the display device 12 and input data from the input device 1106, respectively. The input device 1106 may be a keyboard or a pointing device. A touch panel display device in which the display device 12 and the input device 1106 are integrated may be adopted.

The memory 1104 is an example of a storage unit (for example, a memory unit). The memory 1104 stores the user terminal program 1111 as one or more programs (one or more programs executed in the MPU 1102) for the operation of the user terminal 11. The explanation of each program will be described later.

The network interface 1105 is an example of an interface unit, and is a NIC module that can handle an Ethernet protocol. As a result, it is possible to connect to the data lake network 1.

FIG. 10 shows an example of a setup screen displayed on the display device 12 of the user terminal 11.

The setup screen 1201 is a user interface screen (for example, GUI (Graphical User Interface)) that displays a notification from the management node 13 in the data lake service system 102. The setup screen 1201 displays a user name 1202, a user ID 1203, a notification condition UI (User Interface) 1204, a notification message 1206, a button "Cancel" 1207, a button "OK" 1208, and a button "Exit" 1209. The display of the setup screen 1201 follows the display data received from the management node 13 via the application server 10, and is controlled by the user program 1012 (FIG. 8) and the user terminal program 1111 (FIG. 9). The display data includes data for displaying the notification.

In the example of FIG. 10, the notification condition UI 1204 and the notification message 1206 are displayed for each area where a specific event such as an abnormality has been detected. The notification message 1206 is a message indicating the content of the notification issued when the notification condition is satisfied. The notification condition UI 1204 is a UI that displays the current notification conditions for the area related to the notification. When a predetermined operation is performed with respect to the UI 1204, a UI (for example, a pull-down menu) 1205 for specifying notification conditions is displayed. The notification conditions include, for example, "high frequency" and "low frequency". The button "Cancel" 1207 is a button for requesting cancellation of the operation indicated in the content of the notification. The operation includes scaling (scale-out or scale-in) with respect to the volume 103 corresponding to the user. If the notification message 1206 does not include a notification regarding a cancelable operation, the button "Cancel" 1207 is hidden (or invalid).

The button "OK" 1208 is a button for transmitting the specified notification condition to the data lake service system 102. The button "Exit" 1209 is a button for terminating the user terminal program 1111.

FIG. 11 shows the internal configuration of the management node 13.

The management node 13 has an MPU 1302, a memory 1303, a network interface 1304, and a data path 1301 that mediates data communication between these elements.

The MPU 1302 is an example of a processor unit and is a microprocessor.

The memory 1303 is an example of a storage unit (for example, a memory unit). The memory 1303 includes a user service program 1310, a monitoring program 1311, a file access program 1312, and a scaling program 1313 as one or more programs (one or more programs executed by the MPU 1302) for the operation of the management node 13. To store. By executing the user service program 1310, the monitoring program 1311, the file access program 1312, and the scaling program 1313 in the MPU 1302, functions such as a user service unit, a monitoring unit, a file access unit, and a scaling unit are realized. The explanation of each program will be described later.

The management DB 4 is stored in a memory 1303 (or a PDEV unit (not shown)). The details of the management table included in the management DB 4 have already been described with reference to FIGS. 4 and 5.

The network interface 1304 is an example of an interface unit, and is a NIC module that can handle an Ethernet protocol. As a result, it is possible to connect to the data lake network 1.

FIG. 12 shows the internal configuration of the storage node 14.

The storage node 14 has a MPU 1402, a memory 1403, a PDEV unit 15, a network interface 1405, and a data path 1401 that mediates data communication between these elements.

The MPU 1402 is an example of a processor unit and is a microprocessor.

The memory 1403 (an example of the memory unit) and the PDEV unit 15 are examples of the storage unit. The memory 1403 stores the data access program 1410, the node expansion program 1411, and the node reduction program 1412 as one or more programs (one or more programs executed by the MPU 1402) for the operation of the storage node 14. .. The explanation of each program will be described later. The PDEV unit 15 is one or more PDEVs for holding a sensor data set (file).

The network interface 1405 is an example of an interface unit, and is a NIC module that can handle an Ethernet protocol. As a result, it is possible to connect to the data lake network 1.

The system configuration and the internal configuration of each device have been described above. The operation of each program will be described below.

First, the operation of each program in the sensor 201 will be described.

The sensor measurement program 810 acquires the measured value from one or more sensor modules 806, acquires the measurement date and time (current date and time) from the real-time clock module 805, and stores the sensor data set including the measured value and the measurement date and time in the sensor buffer 812. Save to.

FIG. 13 shows the operation flow of the sensor transfer program 811.

In S1601, the sensor transfer program 811 calculates the amount of change in the amount of sensor data (total amount of the sensor data set) stored in the sensor buffer 812.

In S1602, the sensor transfer program 811 determines that the calculated change amount satisfies the predetermined condition. If the determination result of S1602 is true (S1602: YES), S1604 is executed, and if not (S1602: NO), S1603 is executed. The "default condition" may be, for example, a change amount of a default value or more.

In S1603, the sensor transfer program 811 acquires the current date and time from the real-time clock module 805, and determines whether the acquired current date and time is the default periodic report date and time. If the determination result of S1603 is true (S1603: YES), S1604 is executed, and if not (S1603: NO), the operation ends.

In S1604, the sensor transfer program 811 sends the sensor data set held by the sensor buffer 812 (for example, the measured value, the measurement date and time, and the measurement position (position indicated by the sensor position information 813) to the redirect server 9 via the sensor network interface 803. ) Is included in the data set).

Next, the operation of each program on the redirect server 9 will be described.

The data conversion program 911 receives the sensor data set from the sensor 201 via the sensor network interface 1105. The data conversion program 911 performs a default detection calculation from the received sensor data set, and as a sensor data set including the measured value (for example, the value after the detection calculation), the detection date and time (for example, the measurement date and time), the measurement position, and the like. The detection result of is saved in the detection data buffer 914. Note that this detection calculation may be a calculation that converts the measured values in the data set sent from the sensor 201 into the user's specifications. For example, as the detection calculation, conversion to spectrum data by Fourier transform calculation, conversion to state value by comparison of numerical value, calculation to correlate measured values of a plurality of sensors, and the like may be adopted.

The data recording program 912 transfers the detection result (sensor data set including the measurement date and time, the measurement position, the value after the detection calculation, etc.) held by the detection data buffer 914 to the data lake service system 102 via the data lake access program 913. Record on the volume 103 of.

The data lake access program 913 stores the sensor data set received from the data recording program 912 in the volume 103 with a default path structure via the file access program 1312 of the management node 13. This default path structure has already been described in FIG.

Next, the operation of each program of the application server 10 will be described.

The sensor program 1010 is a sensor data set (file) as a measurement result of a default area with a measurement date and time in the range from the current date and time to the past predetermined time from the volume 103 of the data lake service system 102 via the data lake access program 1011. ) Is acquired, and the acquired sensor data set is saved in the monitoring data buffer 1013.

The data lake access program 1011 is based on the measurement date and time and the measurement position in the sensor data set received from the sensor program 1010, and is in the folder corresponding to the applicable sensor data set (for example, the measurement date and time and the measurement position) from the volume 103. Data set) is acquired, and the acquired sensor data set is stored in the monitoring data buffer 1013.

The user program 1012 returns at least a part of the sensor data set held by the monitoring data buffer 1013 to the user terminal 11 if there is a connection request from the user terminal 11.

Next, the operation of each program of the user terminal 11 will be described.

FIG. 14 shows an operation flow of the user terminal program 1111.

In S2301, the user terminal program 1111 acquires a connection destination by an input operation from the operator by the input device 1106, and transmits a connection request to the acquired connection destination. Here, there are two connection destination candidates in this embodiment, the application server 10 and the management node 13, and the connection with the management node 13 will be mainly described.

In S2302, the user terminal program 1111 displays the display data sent from the connection destination (management node 13) on the display device 12.

In S2303, the user terminal program 1111 transmits the content of the input operation of the operator by the input device 1106 to the connection destination (management node 13).

In S2304, the user terminal program 1111 determines whether or not the button "Exit" 1209 is selected. If selected (S2304: YES), the process ends, otherwise (S2304: NO), the process returns to S2302.

Next, each program of the management node 13 will be described.

FIG. 15 shows the operation flow of the user service program 1310.

In S2401, the user service program 1310 transmits the display data of the setup screen 1201 if there is a connection request from the user terminal 11.

In S2402, the user service program 1310 determines whether the operation content from the user terminal 11 is completed. If it is finished (S2402: YES), the process is finished, otherwise (S2402: NO), the process proceeds to S2403.

In S2403, the user service program 1310 determines whether the operation content from the user terminal 11 is the selection (pressing) of the button "Cancel" 1207. If so (S2403: YES), the process proceeds to S2406, otherwise (S2403: NO), the process proceeds to S2404.

In S2406, the user service program 1310 sends a scaling (scale-out or scale-in) cancellation request executed on the management node 13 to the management node 13. The cancellation request corresponds to a request to stop scaling and restore the number of storage nodes 14 that are the basis of the volume 103.

In S2404, the user service program 1310 receives the record data of the user management table 501 from the user terminal 11, and registers the record according to the record data in the user management table 501 in the management DB 4.

In S2405, the user service program 1310 transmits the updated display data for changing the display contents of the setup screen 1201 according to the registered record to the user terminal 11.

FIG. 16 shows the operation flow of the monitoring program 1311.

In S2501, the monitoring program 1311 has a long-term access target segment (segment corresponding to the access destination folder) according to the access target file path (file path as an argument from the file access program 1312). It is determined whether the difference (absolute value) between the write frequency ( _FL ) and the short-term write frequency ( _FS ) is equal to or more than the default value (Th ₁ ). The long-term write frequency and the short-term write frequency for the segment are specified from the statistical management table 502. If it is equal to or more than the default value (S2501: YES), the process proceeds to S2502, and if not (S2501: NO), the process ends.

In S2502, the monitoring program 1311 extracts the segment ID corresponding to the access destination segment from the statistical management table 502 based on the above file path, and obtains the regional ID corresponding to the segment ID from the regional property management table 401. Find the latitude and longitude range of the target area. The monitoring program 1311 obtains the regional area of the region having the above latitude and longitude ranges from the regional property management table 401. The monitoring program 1311 specifies the user ID corresponding to the above-mentioned area ID and the notification condition from the area-related management table 402 and the user management table 501.

In S2503, the monitoring program 1311 generates display data including a notification (notification message 1206) according to the notification condition specified in S2502.

In S2504, the monitoring program 1311 determines whether the short-term write frequency exceeds the long-term write frequency. If so (S2504: YES), the process proceeds to S2505, otherwise (S2504: NO), the process proceeds to S2507.

In S2505, the monitoring program 1311 determines whether the notification condition specified in S2502 is "high frequency". If so (S2505: YES), the process proceeds to S2506, otherwise (S2505: NO), the process ends.

In _S2506 , the monitoring program 1311 is based on the area area specified in _S2502 , the two write frequencies (FL and HS), and the current number of storage nodes that are the basis of the volume 103 corresponding to the user specified in S2502. According to the default scale evaluation formula having and as an argument, the number of storage nodes to be added (or the number of storage nodes after expansion) is calculated, and a scale-out request according to the calculated number of storage nodes is transmitted to the management node 13.

In S2507, the monitoring program 1311 determines whether the notification condition specified in S2502 is "low frequency". If so (S2507: YES), the process proceeds to S2508, otherwise (S2507: NO), the process ends.

In S2508, the monitoring program 1311 is based on the area area specified in _S2502 , the two write frequencies (FL and _HS ), and the current number of storage nodes that are the basis of the volume 103 corresponding to the user specified in S2502. According to the default scale evaluation formula with and as an argument, the number of storage nodes to be reduced (or the number of storage nodes after reduction) is calculated, and a scale-in request according to the calculated number of storage nodes is sent to the management node 13. do.

As the above-mentioned default scale evaluation formula, for example, the following number 3 can be adopted. This makes it possible to estimate the scale-out and scale-in scales according to the area of the area corresponding to the user.

FIG. 17 shows the operation flow of the file access program 1312.

In S2601, the file access program 1312 determines whether the request from the data lake network 1 is a write request. If so (S2601: YES), the process proceeds to S2602, otherwise (S2601: NO), the process proceeds to S2605.

In S2602, the file access program 1312 transmits a file write request according to the file path specified from the request to the volume 103 based on the write request.

In S2603, the file access program 1312 updates the long-term write frequency and the short-term write frequency of the statistical management table 502 (two write frequencies corresponding to folders (segments) according to the file path in S2602). The write frequency (write request reception frequency) from the data lake network 1 increases when a specific event such as an abnormality occurs in the sensor 201, and the long-term write frequency and the short-term write frequency are managed for each segment. ..

In S2604, the file access program 1312 calls the monitoring program 1311 with the file path in S2602 and the two write frequencies after the update in S2603 as arguments.

In S2605, the file access program 1312 determines whether the request from the data lake network 1 is a read request. If so (S2605: YES), the process proceeds to S2606, otherwise (S2605: NO), the process ends. The operation flow shown in FIG. 17 takes two examples of a write request and a read request as requests from the data lake network 1, but if there are requests other than those requests, after S2605: NO, , The request may be processed.

In S2606, the file access program 1312 sends a file read request according to the file path specified from the request to the volume 103 based on the read request, and sends the file read in response to the file read request to the volume 103. , Send to the requester.

In the case of scale-out, the scaling program 1313 sends a request to join the volume 103 to the unused storage node 14. In the case of scale-in, the scaling program 1313 transfers the data held by the storage node 14 to be withdrawn from the volume 103 to another storage node 14 in use, and sends a withdrawal request to the storage node 14 to be withdrawn.

Next, the operation of each program of the storage node 14 will be described.

If the request from the management node 13 is a data write (file write request), the data access program 1410 writes the data (data constituting the file) from the management node to the PDEV unit 15 (or cache memory). If the request from the management node 13 is a data read (file read request), the data access program 1410 reads the data from the PDEV unit 15 (or cache memory) and returns it to the management node 13.

In the node expansion program 1411, the storage node 14 to be expanded is set to be in use and the permission for data transmission / reception operation from the management node 13 is set.

The node reduction program 1412 sets the storage node 14 to be reduced to a free state and prohibits the data transmission / reception operation from the management node 13.

The outline of the system configuration and operation according to this embodiment has been described above.

The system according to this embodiment can be applied to social infrastructure management and the IT system that realizes it. In particular, it can be applied to the shared use of IT systems that provide data lakes that hold monitoring data (sensor data) for social infrastructure.

Although one embodiment has been described above, these are examples for explaining the present invention, and the scope of the present invention is not limited to this embodiment. The present invention can also be practiced in various other forms. For example, any storage node 14 may also function as the management node 13. Further, for example, the setup screen 1201 is a screen that serves both as a notification condition setting and a notification (for example, a screen that enables the notification condition to be set at the time of notification). The screen for notification may be different.

The above description can be summarized as follows, for example. The following summary may include matters not included in the above description.

The management node 13 has a data set access unit (for example, a file access unit) and a monitoring unit.

The data set access unit receives each time a predetermined condition is satisfied with respect to the measured value of the sensor 201 for each of the plurality of sensors 201 existing in the plurality of segments 202 related to the plurality of regions 303 corresponding to the plurality of users. A write request for the sensor data set of the sensor 201 is received. The “predetermined condition” referred to here may be, for example, the amount of increase in the sensor data set per unit time, or the amount of fluctuation in the measured value per unit time. Each sensor 201 (or a repeater such as the redirect server 9) issues a write request for the sensor data set to the management node 13 when a specific event such as an abnormality occurs in the measurement of the sensor 201. There is. When the data set access unit receives a write request for the sensor data set of any of the sensors 201, the data set access unit enters the area 303 in which the segment 202 in which the sensor 201 exists is involved among the plurality of volumes 103 corresponding to the plurality of users. Issue a write request for the sensor data set to the volume 103 corresponding to the corresponding user. Each of the plurality of volumes 103 is based on a scale-out type storage node group which is one or more storage nodes 14 existing in the same storage system. In each of the plurality of volumes 103, each sensor data set (for example, a file) stored in the volume 103 is associated with the measurement position of the sensor 201 corresponding to the sensor data set and the measurement date and time for the sensor data set. (For example, the sensor data set is stored in a folder (an example of a storage area) corresponding to the measurement position and the measurement date and time).

The monitoring unit monitors the write frequency of each of the plurality of volumes 103 for each segment 202 constituting the area 303 for the user corresponding to the volume.

This makes it possible to identify the area 303 where a specific event such as an abnormality has occurred due to a change in the write frequency of the sensor data set, without analyzing the sensor data set itself.

The management DB 4 may include a management table showing the correspondence between the area 303 and the segment 202.

The management node 13 further includes a user service unit that issues a condition applicable notification to the user corresponding to the area 303 in which the target segment 202, which is the segment 202 satisfying the condition regarding the write frequency, is involved. The "condition applicable notification" is a notification that the condition relating to the write frequency is satisfied, and is, for example, the notification message 1206.

When the system is shared, it may be difficult to specify which user is the notification destination even if the conditions for writing the sensor data set are satisfied. For example, assuming that a large number of sensors installed and operated by a third party independent of the user are used for the purpose of reducing TCO, it is possible to identify which user needs notification about which sensor. It is not considered easy. According to the above-mentioned user service unit, a user who has jurisdiction over the area 303 where a specific event such as an abnormality has occurred is specified, and the user is notified.

The management node 13 further includes a scaling unit. The scaling unit includes the property of the area 303 (for example, the value related to the size of the area such as the area area and the area population) related to the target segment (the segment satisfying the condition regarding the write frequency) 202 and the area 303 related to the target segment 202. Based on the number of storage nodes 14 that are the basis of the target volume 103, which is the volume 103 for the user corresponding to the above, the scale of the storage node that is the basis of the target volume 103 (for example, the number of storage nodes) is determined. , Issue a scaling execution request to the storage system according to the determined scale.

Thereby, the operation manager of the system shared by a plurality of users can set the scale of the volume 103 to the scale corresponding to the region 303 without having to know the region 303.

When the scale of the target volume 103 is determined, the user service unit issues a scaling notification to the user corresponding to the target volume 103. Issuing a scaling notification means, for example, that the above-mentioned notification message 1206 includes a notification indicating that scaling has been performed, and that the button "Cancel" 1207 is enabled for the notification message 1206. It may be. When the scaling unit accepts the cancellation in response to the scaling notification, the scaling unit issues a request for cancellation of the scaling according to the determined scale to the storage system. In response to the cancellation request, the storage system reverts the configuration of the storage node group on which the volume 103 is based.

As a result, in a normal case where scale-out is performed due to a wide-area disaster (for example, when a wide-area disaster occurs, the frequency of writing sensor data sets for many sensors 201 increases, and scale-out is performed to increase the storage capacity accordingly. It becomes possible to cancel the scaling (for example, the scaling according to the write frequency increased or decreased due to the malfunction due to the error of the sensor 201) in a case different from the above case). In the above embodiment, at least for scale-out, scaling is started without receiving a scaling instruction from the user so that even if the writing frequency of the sensor data set suddenly increases due to a wide area disaster or the like, it can be dealt with. However, even if such scaling occurs due to a malfunction of the sensor 201 or the like, it can be canceled.

For each of the plurality of volumes 103, there is a short-term write frequency and a long-term write frequency for each segment 202 constituting the region 303 for the user corresponding to the volume. The above-mentioned condition regarding the write frequency is that the difference (for example, an absolute value) between the short-term write frequency and the long-term write frequency is greater than or equal to the default value. The scaling according to the determined scale is as follows.
-If the short-term write frequency is higher than the long-term write frequency, and the user corresponding to the target volume 103 wants high-frequency notification (when the notification condition is "high frequency"), scale out.
-If the short-term write frequency is lower than the long-term write frequency, and the user corresponding to the target volume 103 wants low-frequency notification (when the notification condition is "low frequency"), scale-in.

As a result, each user is notified of the desired notification frequency according to the relationship between the short-term write frequency and the long-term write frequency when the difference between the short-term write frequency and the long-term write frequency is greater than or equal to the default value. It can be expected that notifications can be made as often as possible.

Even if there is no desired notification condition for each user, depending on the relationship between the short-term write frequency and the long-term write frequency when the difference between the short-term write frequency and the long-term write frequency is greater than or equal to the default value. , Notification to the user is possible. Further, regardless of the relationship between the short-term write frequency and the long-term write frequency, the user may be notified when the difference between the short-term write frequency and the long-term write frequency is greater than or equal to the default value.

102: Data lake service system

Claims (9)

For each of a plurality of sensors existing in a plurality of segments related to a plurality of regions corresponding to a plurality of users, a write request for a sensor data set of the sensor is requested each time a predetermined condition is satisfied with respect to the measured value of the sensor. When a write request for a sensor data set of any of the sensors is received, it corresponds to the area where the segment in which the sensor is located is involved among the plurality of volumes corresponding to the plurality of users. A dataset access unit that issues a write request for the sensor dataset to the volume corresponding to the user,
Each of the plurality of volumes is based on a scale-out type storage node group which is one or more storage nodes existing in the same storage system.
In each of the plurality of volumes, each sensor data set stored in the volume is associated with the measurement position and the measurement date and time of the sensor corresponding to the sensor data set for the sensor data set.
A data store system including a monitoring unit that monitors the write frequency of each of the plurality of volumes for each segment constituting the area related to the user corresponding to the volume. User service department that issues condition applicable notifications to users corresponding to the area related to the target segment, which is a segment that meets the conditions related to write frequency.
The data store system according to claim 1. Storage that is the basis of the target volume based on the properties of the area that the target segment is involved in and the number of storage nodes that are the basis of the target volume that is the volume for the user corresponding to the area that the target segment is related to. A scaling unit that determines the scale of a node and issues a scaling execution request to the storage system according to the determined scale.
The data store system according to claim 2. When the scale of the target volume is determined, the user service unit issues a scaling notification to the user corresponding to the target volume.
When the scaling unit receives the cancellation in response to the scaling notification, the scaling unit issues a request for cancellation of the scaling according to the determined scale to the storage system.
The data store system according to claim 3. For each of the plurality of volumes, there is a short-term write frequency and a long-term write frequency for each segment constituting the region related to the user corresponding to the volume.
The condition is that the difference between the short-term write frequency and the long-term write frequency is equal to or greater than a predetermined value.
Scaling according to the determined scale is
When the short-term write frequency is higher than the long-term write frequency and the user corresponding to the target volume desires high-frequency notification, it is scale-out.
When the short-term write frequency is lower than the long-term write frequency and the user corresponding to the target volume desires low-frequency notification, the scale-in is performed.
The data store system according to claim 4. For each of the plurality of volumes, there is a short-term write frequency and a long-term write frequency for each segment related to the region related to the user corresponding to the volume.
Under the above conditions, the difference between the short-term write frequency and the long-term write frequency is equal to or greater than the default value.
Scaling according to the determined scale is
If the short-term write frequency is higher than the long-term write frequency, it is a scale-out.
If the short-term write frequency is lower than the long-term write frequency, it is a scale-in.
The data store system according to claim 3. For each of the plurality of volumes, there is a short-term write frequency and a long-term write frequency for each segment related to the region related to the user corresponding to the volume.
Under the above conditions, the difference between the short-term write frequency and the long-term write frequency is greater than or equal to the default value.
The data store system according to claim 2. For each of a plurality of sensors existing in a plurality of segments related to a plurality of regions corresponding to a plurality of users, a write request for a sensor data set of the sensor is requested each time a predetermined condition is satisfied with respect to the measured value of the sensor. When a write request for a sensor data set of any of the sensors is received, it corresponds to the area where the segment in which the sensor is located is involved among the plurality of volumes corresponding to the plurality of users. Issue a write request for the sensor data set to the volume corresponding to the user,
Each of the plurality of volumes is based on a scale-out type storage node group which is one or more storage nodes existing in the same storage system.
In each of the plurality of volumes, each sensor data set stored in the volume is associated with the measurement position and the measurement date and time of the sensor corresponding to the sensor data set for the sensor data set.
For each of the plurality of volumes, the write frequency is monitored for each segment constituting the region for the user corresponding to the volume.
Data store management method. For each of a plurality of sensors existing in a plurality of segments related to a plurality of regions corresponding to a plurality of users, a write request for a sensor data set of the sensor is requested each time a predetermined condition is satisfied with respect to the measured value of the sensor. When a write request for a sensor data set of any of the sensors is received, it corresponds to the area where the segment in which the sensor is located is involved among the plurality of volumes corresponding to the plurality of users. Issue a write request for the sensor data set to the volume corresponding to the user,
Each of the plurality of volumes is based on a scale-out type storage node group which is one or more storage nodes existing in the same storage system.
In each of the plurality of volumes, each sensor data set stored in the volume is associated with the measurement position and the measurement date and time of the sensor corresponding to the sensor data set for the sensor data set.
For each of the plurality of volumes, the write frequency is monitored for each segment constituting the region for the user corresponding to the volume.
A computer program that lets a calculator do things.

Images