JP7266970B2 - Application servers and how business applications are executed by servers - Google Patents

Description

The present invention relates to an application server and a method of executing a business application by a server that improves performance problems when modifying an application and migrating to a new platform.

As a performance issue when retrieving data from a columnar database, if data is transferred to the main memory in units of columns, the data of the same column will be placed at consecutive addresses, resulting in a single row. In spite of access, discrete addresses must be accessed multiple times, resulting in an increase in the load on the CPU.

In order to solve this problem, Patent Document 1 discloses that "a memory device, a storage device, and a database processing engine are provided, and the database processing engine converts column type pages stored in the storage device into row type pages. By having a module that stores information in a memory device, the memory device manages data in row-type pages, so efficient access to the database can be achieved.”

WO2014/010038

When a database (hereinafter referred to as "DB") is switched to a new DB, performance problems may occur if the structure of the originally used business application is applied as is due to differences in physical data arrangement.

Therefore, in the present invention, when modifying an application caused by a platform change and migrating to a new platform, the DB can be switched without modifying the business program as much as possible to solve the performance problem. With the goal.

In order to solve the above problems, an application server according to the present invention includes a communication unit that transmits and receives data to and from a columnar database, and an auxiliary storage unit that stores a business program that performs business processing and an analysis program that analyzes the business program. a main memory having a temporary storage area for temporarily storing data ; The past data used at the time of execution is read as input data, the read input data is analyzed, the first syntax is extracted from the business program, and the extracted first syntax is used in the conditional expression. Create a second syntax that analyzes the values of the data to be input to the variables in the business program and the input data from each of the business program and the input data, and acquires the analyzed data values from the columnar database all at once. It is characterized in that it is registered in a temporary storage area of a storage unit, and corrected by converting the first syntax of the business program into the second syntax.

According to the present invention, programmers can implement application migration without being conscious of the characteristics of the DB after migration, and solve performance problems.
Problems, configurations, and effects other than those described above will be clarified by the following description of the embodiments.

1 is a block diagram showing a configuration of an AP server and columnar DBs as an example of a system configuration according to the present invention; FIG. 4 is a diagram showing a data structure of table information stored in a columnar DB; FIG. It is a figure which shows the processing procedure (flowchart) based on this invention. FIG. 4 is a diagram showing a processing procedure (flowchart) of program analysis processing and conversion processing according to the present invention; FIG. 4 is a diagram showing a processing procedure (flowchart) of execution processing and analysis processing according to the present invention; FIG. 10 is a diagram showing a processing procedure (flowchart) when the memory capacity is insufficient as a result of checking the free size of the cache area in step 503; FIG. 10 is a diagram showing a processing procedure (flowchart) when there is no target data in the cache in step 510; FIG. 10 is a diagram showing an SQL statement to be analyzed in program analysis processing; FIG. 10 is a diagram showing a procedure for analyzing and specifying values in a program that apply to conditional expressions of SQL statements; FIG. 10 is a diagram showing a procedure for analyzing input data and specifying a value that fits a conditional expression of an SQL statement; FIG. 10 is an image diagram of an SQL sentence as a result of analysis; FIG. 10 is a diagram showing data information stored in a cache device when data is acquired from a columnar DB by executing an SQL statement; FIG. 10 is a diagram showing an SQL statement obtained from a columnar DB when target data does not exist in a cache device; FIG. 4 is a diagram showing data information stored in a cache device when data is acquired from a columnar DB; It is a figure which shows the merged SQL sentence. FIG. 10 is a diagram showing log output processed in step 509;

EMBODIMENT OF THE INVENTION Below, an Example is described in detail using drawing as a form for implementing this invention.

FIG. 1 is a block diagram showing a configuration of an application server (hereinafter referred to as "AP server") 100 and columnar DB 200 as an example of a system configuration according to the present invention.

The AP server 100 is a device that processes and executes applications, and includes a memory 101 that is a main storage unit, a CPU 105 that is a processing device, a storage 106 that is a data storage area, and a communication device 109 that communicates with the columnar DB 200. Prepare.

The memory 101 has a cache device 102 which is an area for managing data. The cache device 102 is a cache area 103 which is an area for storing data and a common processing component for storing data in the cache area 103. It has part 104 .
The storage 106 stores programs such as the business program 107 and analysis program 108 as well as required data.

The business program 107 is a program that performs business processing.
On the other hand, the analysis program 108 analyzes the business program 107 to specify the data used by the business program 107, and acquires a certain range of data including the data in a batch from the columnar DB 200 as a common component. Register with 104. Also, the analysis program 108 modifies the business program 107 so that it uses the data registered by the business program 107 .
The business program 107 and the analysis program 108 are developed in the memory 101 when executed by the CPU 105 .

Note that the business program 107 according to this embodiment uses, as an example, a program for analyzing whether weather and time affect the number of visitors to a store. This business program 107 is merely an example, and it goes without saying that other business programs may be targeted.

The columnar DB 200, which is a device for managing and storing a large amount of data, is a device for managing and storing a large amount of data. 204 and a communication device 203 for communicating with the AP server 100 .
FIG. 2 is a diagram showing the data configuration of table information stored in the columnar DB 200. As shown in FIG. This table information is saved in the storage 202 . Here, the table name shown in FIG. 2 is assumed to be "visitor information table".

FIG. 3 is a diagram showing a processing procedure (flowchart) according to the present invention. A processing mode based on this flow chart will be described below. Here, although the entity that executes the processing is the CPU 201 of the AP server 100, it is collectively referred to as the AP server 100. FIG.

<Step 301>
The AP server 100 executes program analysis processing for the business program 107 .

<Step 302>
The AP server 100 executes conversion processing that reflects the analysis result of step 301 on the business program 107 .

<Step 303>
The AP server 100 executes the business program 107 reflecting the processing result of step 301 .

<Step 304>
The AP server 100 accumulates the execution log of the SQL statement of the business program 107 executed in step 303 and performs analysis processing.

Next, details of each step described above will be described.
FIG. 4 is a diagram showing a processing procedure (flowchart) of the program analysis processing in step 301 and the conversion processing in step 302 shown in FIG. A processing mode based on this flow chart is shown below.

<Step 401>
The AP server 100 reads the business program 107 and input data.
In this embodiment, the input data is data for the past three times used when the business program 107 was executed, such as the data shown in FIG. Here, the input data is data for the past three times, but the numerical value "3" is an example and is not limited to this.

<Step 402>
The AP server 100 analyzes the data read in step 401 and extracts SQL sentences.
In this embodiment, as shown in FIG. 8, when the variable 1 is true and the variable 2 is true from the visitor information table, the SQL statement for extracting the data matching the conditions is extracted.

<Step 403>
The AP server 100 analyzes the value of the data to be input to the variable used in the conditional expression of the SQL statement extracted in step 402 from the business program 107 .
In this embodiment, the analysis program 108 analyzes the business program 107, and from the conditional expression of the SQL statement shown in FIG. 8, it is found that the variable 2 is "sunny" as shown in FIG.

<Step 404>
The AP server 100 analyzes the value of the data to be input to the variable used in the conditional expression of the SQL statement from the input data.
In this embodiment, the analysis program 108 analyzes the business program 107, and from the conditional expression of the SQL statement shown in FIG. 8, it is found that the variable 2 is the variable inputdata as shown in FIG. Further analysis reveals that the variable inputdata stores the value obtained by reading the input data, and the analysis program 108 analyzes the input data, revealing that the variable inputdata is "cloudy". do.

<Step 405>
The AP server 100 creates an SQL statement for collectively obtaining a certain range of data including the input data specified in steps 403 and 404 and registers it in the common component 104 . Also, the data acquired by this SQL statement is stored in the cache area 103 .
In this embodiment, as shown in FIG. 11, an SQL statement is created by combining the SQL statement when the variable 2 is "sunny" and the SQL statement when it is "cloudy", and is registered in the common component 104. will do.

<Step 406>
The AP server 100 converts and corrects the SQL statement of the business program so that the SQL statement extracted in step 402 is obtained from the common component 104 of the cache device 103 .

<Step 407>
The business program 107 and the common component 104 are deployed after being built.

FIG. 5 is a diagram showing a processing procedure (flowchart) of the execution processing of step 303 and the analysis processing of step 304 shown in FIG. A processing mode based on this flow chart is shown below.

<Step 501>
The AP server 100 executes business programs 107 . By executing this program, the cache device 102 is notified of the record to be acquired. The AP server 100 uses the common component 104 of the cache device 102 to acquire records from the columnar DB 200 .

<Step 502>
The AP server 100 uses the cache device 102 to check the free size of the cache area 103 .

<Step 503>
The AP server 100 uses the cache device 102 to check whether or not the acquired record can be stored in the cache area 103, that is, whether or not the checked free size (memory capacity) is sufficient for this storage. If it can be stored (YES), go to step 504 . If it cannot be stored (NO), go to step 505 .

<Step 504>
The AP server 100 uses the cache device 102 to store the acquired record in the cache area 103 . FIG. 12 is a diagram showing the data structure of data acquired from the columnar DB 200 and stored in the cache area 103 by executing the SQL statement shown in FIG.
In this embodiment, data having information of "sunny" and "cloudy" are collectively acquired from the table information of the columnar DB 200 shown in FIG. 2, and stored in the cache area 103 as shown in FIG.

<Step 505>
The AP server 100 executes the processing procedure (flowchart) shown in FIG. 6 so that the acquired record can be stored in the cache area 103 . After completing the processing procedure (flow chart) shown in FIG. The processing procedure (flowchart) shown in FIG. 6 will be described later.

<Step 506>
The AP server 100 executes business processing according to the business program 107 .

<Step 507>
The AP server 100 requests the data to be acquired from the cache device 102 according to the business program 107 . The AP server 100 accesses the cache area 103 using the cache device 102 and the common component 104 .

<Step 508>
The AP server 100 uses the cache device 102 to check whether or not the data to be acquired exists in the cache area 103 . If there is data to be acquired (YES), the process proceeds to step 509 . If there is no data to be acquired (NO), the process proceeds to step 510 .

<Step 509>
The AP server 100 uses the common component 104 to log and store the program of the request source, the date and time when the request was received, and the executed SQL statement as an execution history. FIG. 16 is a diagram showing the log output processed in step 509. As shown in FIG. The storage destination of this log output is the storage 106, but may be a cache area if there is room.
In this embodiment, the execution history of the SQL statement accessed by the business program at 12:00 on March 1, 2017 and the SQL statement accessed at 12:00 on April 1, 2017 is saved as a log for the visitor information table. be.

<Step 510>
The AP server 100 executes the processing procedure (flowchart) shown in FIG. 7 so that the data to be acquired can be stored in the cache device 102 . After completing the processing procedure (flow chart) shown in FIG.

<Step 511>
The acquired data is delivered to the business program, and the AP server 100 completes the business processing when the business program receives the data.

<Step 512>
The AP server 100 analyzes the execution log output in step 509 , merges the executed SQL statements, and registers the merged SQL statements in the common component 104 .
FIG. 15 is a diagram showing merged SQL statements. In this embodiment, the analysis program 108 acquires the data having the information of "sunny" and "cloudy", which is found and created by analyzing the business program 107 and the input data, and the SQL statement for acquiring the data having the information of "sunny" and "cloudy", and the step 601 described later. Merge with the SQL statement that acquires the newly created "rainy". As a result, an SQL statement for acquiring “sunny”, “cloudy” and “rainy” is created and registered in the common component 104 .

FIG. 6 is a diagram showing a processing procedure (flow chart) when the memory capacity is insufficient (NO) as a result of checking the free size of the cache area in step 503 of FIG. A processing mode based on this flow chart is shown below.

<Step 601>
The AP server 100 creates an alternative SQL statement and registers it in the common component 104 so that a certain amount of data including the data to be acquired can be collectively acquired from the columnar DB 200 using the cache device 102 .
FIG. 13 is a diagram showing an SQL statement for acquiring target data from the columnar DB 200 when target data does not exist in the cache device 102 . In this embodiment, since there is no data having the information of “rainy” in the cache device 102 , an alternative SQL statement that can acquire “rainy” is created and registered in the common component 104 .

<Step 602>
The AP server 100 uses the common component 104 to execute the alternative SQL statement and acquires the data to be acquired from the columnar DB 200 .
In this embodiment, an alternative SQL statement (FIG. 13) capable of acquiring "rainy" is executed, and as shown in FIG. Add information.

<Step 603>
The AP server 100 uses the cache device 102 to check the free size of the cache area 103 .

<Step 604>
The AP server 100 uses the cache device 102 to check whether or not the acquired record can be stored in the cache area 103, that is, whether or not the checked free size is sufficient for this storage. If it can be stored (YES), the process flow shown in FIG. 6 is terminated and the process returns to step 504 in FIG. If it cannot be stored (NO), go to step 605 .

<Step 605>
The AP server 100 uses the cache device 102 to check whether or not data that can be evicted from the cache area 103 exists among the data stored in the cache area 103 . If there is data that can be evicted (YES), go to step 606 . If there is no data that can be evicted (NO), go to step 607 .

<Step 606>
The AP server 100 uses the cache device 102 to move data stored in the cache area 103 that is least referenced to the storage 106 as a file. Then, return to step 504 of FIG.

<Step 607>
The AP server 100 uses the cache device 102 to create an SQL statement with a narrower range than the range obtained in the previous step 601 so as to include the data to be acquired from the columnar DB 200, and registers it in the common component 104. to get the desired data. Then, return to step 504 of FIG.

FIG. 7 is a diagram showing a processing procedure (flow chart) when there is no target data in the cache in step 510 of FIG. A processing mode based on this flow chart is shown below.

<Step 701>
The AP server 100 uses the cache device 102 to check whether or not the data to be acquired exists in the storage 106 of the AP server 100 . If there is data to be acquired (YES), the process proceeds to step 702 . If there is no data to be acquired (NO), the process proceeds to step 704 .

<Step 702>
The AP server 100 uses the cache device 102 to move file data from the storage 106 to the cache area 103 .

<Step 703>
The AP server 100 uses the common component 104 to acquire data to be acquired from the cache area 103 . Then, return to step 509 of FIG.

<Step 704>
The AP server 100 uses the cache device 102 to execute the flowchart shown in FIG. 6 so that the record to be acquired can be stored in the cache area 103 . After completing the processing flow shown in FIG. 6, the process proceeds to step 705 .

<Step 705>
The AP server 100 stores the acquired record in the cache area 103 using the cache device 102, and proceeds to step 703 after storing.

In this way, the AP server 100 analyzes the business program 107 and the input data, identifies the data used in business by the business program 107, creates an SQL statement for batch acquisition, and stores it in the cache device 102 as the common component 104. do. By executing the common component SQL statement in advance and storing the acquisition target data in the cache area 103, the number of accesses to the columnar DB 200 can be reduced and the processing time can be shortened.

The configuration diagram of the table information of the columnar DB 200 shown in FIG. 2, the image of the SQL statement shown in FIG. 8, and the image of the business program and input data shown in FIGS. do not have.
Also, in this embodiment, the DB is a columnar DB, but a structured DB or a relational DB may also be targeted.

In addition, although the AP server 100 and the columnar DB 200 are described as separate devices in this embodiment, for example, they operate as separate virtual servers in a virtual environment running on a single physical computer. may be one physical server or one virtual server having each function.
Furthermore, although the input data is analyzed in this embodiment, only the business program may be analyzed.

100... AP server, 101... memory, 102... cache device,
103... cache area, 104... common component, 105, 201... CPU,
106... Storage, 107... Business program, 108... Analysis program,
109... Communication device, 200... Columnar DB, 202... Storage, 203... Communication device,
204 memory

Claims (11)

a communication unit that transmits and receives data to and from the columnar database;
an auxiliary storage unit that stores a business program that performs business processing and an analysis program that analyzes the business program;
a main storage unit having a temporary storage area for temporarily storing data;
A control unit that executes the business program or the analysis program,
The control unit executes the analysis program ,
reading the business program and past data used when the business program was executed as input data;
analyzing the read input data to extract a first syntax from the business program;
analyzing values of data to be input to variables used in the extracted conditional expression of the first syntax from each of the business program and the input data;
creating a second syntax for collectively acquiring the values of the analyzed data from the columnar database;
registering the created second syntax in the temporary storage area of the main storage unit;
modifying the first syntax of the business program by converting it into the second syntax
An application server characterized by: The application server according to claim 1,
providing an area for registering a common component, which is a processing component for storing data, in the temporary storage area;
The application server, wherein the control unit registers the second syntax in the common component. The application server according to claim 2,
When executing the business program , the control unit checks whether or not the data acquired from the columnar database using the common component can be stored in the temporary storage area based on the free size of the temporary storage area. An application server characterized by: An application server according to claim 3,
When the data obtained from the columnar database cannot be stored in the free size of the temporary storage area, the control unit checks whether there is data that can be ejected from the temporary storage area, and if there is data that can be ejected, the data is stored. An application server that stores the obtained data in the temporary storage area by moving it to the auxiliary storage unit. The application server according to claim 4,
The application server, wherein the data that can be evicted from the temporary storage area is the data with the oldest reference in the temporary storage area. An application server according to claim 3 ,
When the data obtained from the columnar database cannot be stored in the free size of the temporary storage area, the control unit checks whether there is data that can be ejected from the temporary storage area, and if there is no data that can be ejected, the temporary storage. An application server characterized by creating an alternative syntax that can collectively acquire a data range that can be stored in a free size remaining in an area, registering it in the common component, and executing it. A method for executing a business application by a server using a columnar database,
Using an analysis program,
a first step of reading, as input data, a business program for performing business processing and past data used when the business program was executed;
a second step of analyzing the read input data and extracting a first syntax from the business program;
a third step of analyzing values of data to be input to variables used in the extracted conditional expression of the first syntax from each of the business program and the input data;
a fourth step of creating a second syntax for collectively acquiring the values of the analyzed data from the columnar database, and registering the second syntax in a temporary storage area of the main storage unit of the server;
a fifth step of converting the first syntax of the business program into the second syntax and correcting it;
execution method of a business application having A method for executing a business application according to claim 7,
Registering the second syntax in the temporary storage area in the fourth step includes registering in a common component provided as a processing component for storing data in the temporary storage area.
A business application execution method characterized by: A business application execution method according to claim 8 ,
a fifth step of checking whether or not the data obtained from the columnar database using the common component can be stored in the temporary storage area during execution of the business program, based on the free size of the temporary storage area;
execution method of a business application having A business application execution method according to claim 9,
If the data obtained from the columnar database cannot be stored in the free size of the temporary storage area, the presence or absence of data that can be evicted from the temporary storage area is checked. a sixth step of moving old referenced data to another storage unit of the server and storing the acquired data in the temporary storage area;
execution method of a business application having A business application execution method according to claim 9 ,
If the free size of the temporary storage area cannot store the data obtained from the columnar database, the presence or absence of data that can be evicted from the temporary storage area is checked. A seventh step of creating an alternative syntax capable of collectively acquiring the range of data that can be stored, registering it in the common component, and executing it
execution method of a business application having

Images