TWI684154B - Service-oriented architecture (SOA) for walk-through command reorganization of real-time business intelligence systems - Google Patents

Abstract

A service-oriented architecture (SOA) for walk-through command reorganization of an instant business intelligence system, including a walk-through command reorganization module consisting of a command reorganization and analysis module, a first service definition file, a second service definition file and a The inheritance module is composed of the data source obtained by the reorganized instruction of the walk-through instruction reorganization module in real time, and then the data source obtains the required data from multiple operating systems or data warehouses, and is service-oriented. Obtain the analysis environment and calculation of multi-dimensional data with data warehousing. After the user transforms the demand, even if different dimensions, different operating systems or multiple operating systems and data warehousing or homogeneous or heterogeneous databases can be obtained immediately after the transformation The immediacy of the data.

Description

Service-oriented architecture (SOA) for walk-through command reorganization of real-time business intelligence systems

The technical field of a business intelligence system of the present invention, in particular, it provides a service-oriented architecture (SOA) walk-through instruction reorganization design implemented in a real-time business intelligence system, service-oriented, when user needs change instruction reorganization And the analysis module provides the flexibility of strain, so as to obtain the analysis environment and calculation of multi-dimensional data with data warehousing. After the user changes the requirements, even if the different dimensions, different operating systems or multiple operating systems and data warehousing are homogeneous or heterogeneous The database can also immediately obtain the real-time nature of the converted data, and there is no need to write an independent specific application program interface (API).

According to, see the first picture, the first picture is a system model of the current BIS standard, its process can be roughly divided into three stages:

(1) Data aggregation: use the ETL (Extract-Transform-Load) tool (see the second figure) to extract, convert and load the source database data source 10 into the Operational data store database; then after finishing, The accumulated data is stored 12 in the data warehouse (Data Warehouse) database.

(2) Data analysis: use the ETL tool (see the second figure) to extract the data of the data warehouse 24 and store it in the data supermarket (Data Mart) database constructed based on the analysis (in some cases, the Data Mart can be omitted and directly Retrieve 21, convert 22, and load 23 from the trading system 20 Data Warehouse to obtain analysis data). Then use OLAP (Online Analytical Processing) or data mining (Data Mining) technology for data analysis.

(3) Data display 13: The results of data analysis are presented to users by means of reporting tools or Web Portal.

And the problem facing enterprises today, can we achieve the dream of enterprise BI by building a good BIS? Before BIS's voice was loud and loud, and enterprises invested a lot of resources such as cost, time, and manpower, the benefits and expectations of users from BIS were useful for Information Quality (IQ) and system usefulness. There is often a big gap in the use of information. In the end, 90% of the employees in the company still used Excel to analyze the data, which caused the effectiveness of BI to be ineffective. The present invention can be roughly summarized into the following two points on the user's perception level of poor information quality:

(1) Data warehousing plays a core role in BIS and is the basis for all subsequent information analysis. However, with the rapid changes in the company’s operating environment, the data in the data warehousing cannot provide decision makers with timely and stable information in a fiercely competitive environment. Information, so even a good BIS cannot meet the needs of enterprises. On the other hand, the development of data storage must be an extremely large and Complex investment is not only risky, but also requires a large amount of corporate resources. If there is no clear business case (business case), it may be a long-term "work in progress" or "Project black hole", so many companies are very hesitant based on investment costs and risk factors; Gartner Group's 2004 survey from Intelligent Enterprise showed that 55% of enterprise users questioned and believed that The information provided by their data warehouse did not hit the target. Another disadvantage is that the development is quite time-consuming. If the traditional system development method usually takes several years, if the demand will continue to increase during the development process, the prototype method will be used to make a prototype first, and then it will be developed repeatedly. The time spent is not conspicuous; in the end, companies have invested a lot of cost and time, and there is a significant gap between the benefits obtained and expectations.

(2) Many companies will mistakenly believe that the introduction of BIS can smoothly operate to obtain due analysis data. However, companies did not expect that the data in the data warehouse will also need to be adjusted in real time as the analysis needs change and the type of business changes. On the other hand, IT staff often spend a lot of time to integrate a large number of different heterogeneous data sources. In particular, data integration plays an important role in information quality (IQ), especially the quality of information content (Information content quality) ultimately affects the use of information; the effectiveness of BIS lies in its ability to provide timely business information, so BIS The success or failure of the import depends on the available information; when relying on the IT staff for a longer time or the IT staff’s lack of skills (the lack of IT skills) , Which means that the longer the decision maker waits for analysis, the resulting analysis report can never meet the decision maker’s on demand expectations, which means that it is difficult to reflect the changes in the current operating environment. , It doesn't make much sense to do it.

In 2006, it was proposed that BI 2.0 gives different thinking, but it advocates immediacy. However, there are many gaps and obstacles to achieve the concept of BI 2.0. Judging from the current status of BI applications in most enterprises, if it is compared with BI 2.0’s main demands one by one, it will It is not difficult to find that the gap between them can be described as very large. Among them, what are the current BI obstacles? The following explains:

(1) The purpose of traditional BIS is to convert data into information, and then convert the information into wisdom. If an enterprise only imports BIS for the purpose of using BI, it is easy to fall into the process of evaluating tools and stick to the comparison of who has good scalability, The blind spot of whose calculation speed is fast, in the end, BIS can only improve the accuracy, timeliness and integration of information acquisition at best, and may not be able to reach the needs of business operations and management. There is a significant gap between the two, and the benefits are difficult to assess; and the IT integration of data is time-consuming and laborious. Finally, BIS is reduced to a report generator, and the application value of such things, the benefits will naturally not be too high, and it is far from the BI 2.0 appeal concept.

(2) Looking at the BI 2.0 argument, many scholars admit that it is really not easy to realize. Especially in real-time decision-making, because the premise is that the detailed data in the enterprise's complex information system must be sorted and calculated, and the data transmission also wastes some time, sometimes even "add 1 column" Simple requirements can take several working days. Therefore, the demand for immediacy is undoubtedly a key obstacle faced by policy makers, because once the time of data integration is prolonged, it means that business decisions are based on non-immediate data, and it is difficult to reflect the current Changes in the operating environment, not to mention the decision-makers' needs for analysis will change at any time in order to respond to the rapid changes in the market. When the current consolidated data (data warehouse) cannot meet their needs, they must rely on IT staff to invest again A large amount of time (if the IT staff's professional literacy is insufficient, the timing of their decision-making will undoubtedly be postponed), such a process does not make much sense for immediate decision-making needs.

The analytical data produced by the BI 2.0 environment must be able to be embedded in different operating processes. In the past, BIS was based on a specific application program interface (API), which was developed by IT personnel through development procedures. Fei Zhouzhang retrieved the analysis results. As for what new approach to take, any component in the workflow can easily obtain the BIS analysis results, the premise is to get rid of the independent API developed by the IT staff using the program, and change to service-oriented Architecture (SOA) approach.

The design of the present invention provides an organization (enterprise) with benefits in management and decision-making. Based on the enterprise transaction information system (operational IS), a smart "walking instruction reorganization" design with service-oriented architecture is introduced, which overcomes the organization (enterprise)'s The analysis environment and calculation of multi-dimensional data with data storage (drilldown, rollup, slice or pivoting etc.)" to realize the demands of real-time decision-making, and has been actually imported into the enterprise's internal information system (operational IS), which verifies the real-time interactive relationship between the user and the source of information needs, BIS, and the specific reaction in the information Quality assessment and the degree of system usefulness, thereby improving the use of information and greatly improving the success of BIS; it is worth mentioning that the SOA architecture designed by the present invention becomes a new service through recombination between services Innovative design architecture makes it easy to integrate heterogeneous data.

The design framework proposed by the present invention can immediately verify the aforementioned key factors for the success of the BIS from the intra-enterprise transaction information system (operational IS), that is, data integration, information quality and use of information, etc., without It takes a long time to invest a lot of resources such as cost, time, and manpower before discovering that there is a large gap between the expected benefits; in addition, the data in the data warehouse has been cleaned before being provided to the user (cleaned ), decoded, reorganized and reordered, however, through the use of raw transactional data, broken and ugly data can only be revealed together with good data during the analysis process, and corporate decisions The person will have the opportunity to evaluate what is the damaged data and why, so as to have the opportunity to solve it. Therefore, from the perspective of the benefits of IT development BIS, it can be expected that the value of such a design framework can improve the correlation between the understanding of the problem and the user's key factors in cognition, greatly reducing the investment in real-time BIS Resources and changes to meet the needs of subsequent users, providing users with the analysis data they need in real time 2. Improve the efficiency of BIS information usage, and allow IT staff to save time for data integration, while focusing on other business processes of the enterprise to maximize IT benefits.

Therefore, under this design architecture, the system has a high degree of expansion flexibility and implements the core spirit of SOA emphasized abstraction, composability, and reusability to implement BI 2.0 real-time analysis services To help companies achieve:

(1) Increase the profitability of enterprises or enhance the competitiveness of enterprises.

(2) Provide variable service types.

(3) Reduce the cost and time of enterprise introduction.

(4) Reduce the time to develop services to meet immediate needs.

(5) Integrate enterprise network service technical resources.

(6) Reduce overall risks and accidents.

Regarding the technology, means and functions adopted by the present invention, a preferred embodiment is described in detail with reference to the drawings. It is believed that the above-mentioned objects, structures and features of the present invention can be gained a deep and specific understanding .

[Knowledge]

10‧‧‧Source

11‧‧‧Extraction, conversion and loading

12‧‧‧Data storage

13‧‧‧Information

20‧‧‧ Trading System

21‧‧‧ Capture

22‧‧‧Conversion

23‧‧‧ Load

24‧‧‧ data storage

[this invention]

30‧‧‧ Trading system

31‧‧‧ data storage

40‧‧‧Service Oriented Architecture (SOA)

41‧‧‧ Walking Reorganization Module (ReSQL)

411‧‧‧Command reorganization and analysis module

412‧‧‧First service definition file (eistable.def)

413‧‧‧Second service definition file (valueSQL.def)

414‧‧‧Inheritance module

42‧‧‧ Dataset (DS)

43‧‧‧grammar conversion module

44‧‧‧Multi-value data processing module

45‧‧‧Dimensional Integration Module

461‧‧‧User interface output module

462‧‧‧User interface control module

47‧‧‧filter

48‧‧‧ homogeneous or heterogeneous database

49‧‧‧Cache

The first diagram is a conventional business intelligence (BI) system architecture.

The second graph is a conventional data extraction transformation loading (ETL) architecture.

The third diagram is a service-oriented architecture (SOA) system diagram of the present invention.

The fourth figure is a system diagram of the ReSQL reorganization module (ReSQL) architecture of the present invention.

The fifth figure is the result of the implementation of Example 1 of the present invention: 94 quarterly summary data (not yet explored).

The sixth figure is the result of the first example of the present invention: the aggregated data down to the departmental dimension.

The seventh figure is the result of the implementation of the first example of the present invention: down to the summary data of the department dimension.

The eighth figure is the execution result of Example 1 of the present invention: digging down to the aggregated data of the dimension of the physician.

One of the ninth figures is the execution result of the first example of the present invention: the detailed information of the doctor's visit declaration.

The second figure 9 is the execution result of the first example of the present invention: quarterly data summary.

The third of the ninth figure is the execution result of the first example of the present invention: Drill-down drills down the first quarter data to the month.

The fourth of the ninth figure is the execution result of the first example of the present invention: Drill-down to February department summary data.

The fifth of the ninth figure is the execution result of the first example of the present invention: Drill-down will summarize medical data in February.

The sixth of the ninth figure is the execution result of the first example of the present invention: Drill-down drills down the surgical summary data in February to daily.

The tenth figure is the execution result of Example 2 of the present invention: inherit the definition of Example 1 and change the dimension attribute.

The eleventh figure is the result of the implementation of the second example of the present invention: the aggregated data down to the dimension of the payment category.

The twelfth figure is the result of the implementation of the second example of the present invention: down to the aggregated data of the classification dimension of the hospitalized case.

Figure 13 is the execution result of Example 2 of the present invention: pivot analysis (pivot). Figure 14 is the execution result of Example 3 of the present invention: inheriting the definition of Example 1 and generating a new definition in eistable.def.

The fifteenth figure is the result of the implementation of the third example of the present invention: digging down to the summary data of the doctor dimension.

The sixteenth figure is the result of the implementation of the third example of the present invention: the doctor summarizes the data annually.

One of the seventeenth figures is the fourth example of the present invention: multi-data source integration architecture (numerical integration).

Figure 17 bis is a schematic diagram of multiple data source integration architecture (classified integration) of the present invention.

Figure 18 is the implementation result of Example 4 of the present invention: integration of multiple data sources.

Figure 19 is the execution result of Example 4 of the present invention: integration of multiple data sources (the dive dimension is the department).

Figure 20 is the execution result of Example 4 of the present invention; integration of multiple data sources (drill down to the details of physicians in the HR database).

One of the twenty-first figures is the result of the implementation of Example 4 of the present invention: integration of multiple data sources (compared with the same period last year), shown in general dimensions.

Figure 21 is the implementation result of Example 4 of the present invention: integration of multiple data sources (compared with the same period last year), displayed in the period dimension.

The third figure of the twenty-first figure is the result of the fourth example of the present invention: the integration of multiple data sources (compared with the same period last year), shown in the following dimension.

The present invention provides a service-oriented architecture (SOA) designer of walk-through instruction reorganization of an instant business intelligence system.

In order for your reviewing committee to have a better understanding and understanding of the purpose, features and efficacy of the present invention, the detailed description of the implementation mode and the drawings is as follows: referring to the third figure, the present invention provides an instant business intelligence system The service-oriented architecture (SOA) of walk-through command reorganization includes: a service-oriented architecture (SOA) 40, a user end and an information technology (IT) staff, and the service-oriented architecture (SOA) 40 Including a walk-around instruction reorganization module (ReSQL) 41, the walk-around instruction reorganization module (ReSQL) 41 includes a command reorganization and analysis module 411, a first service definition file (eistable.def) 412, a first Two service definition files (valueSQL.def) 413 and an inheritance module 414, the first service definition file (eistable.def) 412 and the second service definition file (valueSQL.def) 413 are composed of the information technology (IT) Simple definition of personnel, data is taken from operating system or data is taken from multiple operating systems or data is taken from multiple operating systems and data warehousing It is based on the data source obtained by the reorganized instruction of the walking instruction reorganization module (ReSQL) 41, and then obtains the required data from the data source. The service-oriented architecture (SOA) 40 further includes a syntax conversion module 43. Through the grammar conversion module 43, the command reorganization and analysis module 411 and the homogeneous or heterogeneous database 48 of each operating system or data warehouse can be communicated to make all data automatically accessible, and the inheritance module 414 The requirements of the user end can be modified between each service through inheritance and recombined between the services to form a new service. The command reorganization and analysis module 411 is responsible for setting the definition of the service and the parameters entered by the user to The operation mode of the reorganization instruction of the walking instruction reorganization completes real-time demand service analysis, and is service-oriented. When the user's demand changes, the instruction reorganization and analysis module 411 provides flexibility to adapt to obtain data storage Multi-dimensional data analysis environment and calculation, after the user changes the demand, even if the different dimensions, different operating systems or multiple operating systems and data storage or homogeneous or heterogeneous database 48, you can immediately get the real-time data after conversion It does not need to write an independent specific application programming interface (API).

The service-oriented architecture (SOA) of the walk-through instruction reorganization of the real-time business intelligence system, wherein the instruction reorganization and analysis module 411 is constructed with a complete retrieval instruction, and the instruction reorganization and analysis module 411 further has the reorganized Command optimization to shorten the time of data retrieval.

The service-oriented architecture (SOA) of the walk-through command reorganization of the real-time business intelligence system described above, in which the method for the user to access data includes: Drill-down, roll-up, pivot analysis or slice.

The service-oriented architecture (SOA) of the walk-through instruction reorganization of the real-time business intelligence system, wherein the service-oriented architecture (SOA) 40 further includes a multi-value data processing module 44 which is connected to the The walking instruction reorganization module (ReSQL) 41, the grammar conversion module 43, and the homogeneous or heterogeneous database 48 use the multiple-value data processing module 44 to perform multiple data calculation and integration.

The service-oriented architecture (SOA) of the walk-through instruction reorganization of the real-time business intelligence system, wherein the service-oriented architecture (SOA) 40 further includes a data set (DS, dataset) 42 and a cache memory (cache) 49 , The data set 42 connects the walking command reorganization module (ReSQL) 41, the multi-value data processing module 44, the homogeneous or heterogeneous database 48 and the user terminal, and then proceeds to the data set 42 after the command reorganization, the data The set 42 only retrieves data from the homogeneous or heterogeneous database 48 corresponding to the instruction when the cache 49 cannot find the data.

The service-oriented architecture (SOA) of the walk-through instruction reorganization of the real-time business intelligence system described above, wherein the service-oriented architecture (SOA) 40 further includes a one-dimensional integration module 45, which is connected to the data set (DS) , dataset) 42 and the user end, the dimension integration module 45 may integrate two or more different dimension data into one data, and then display the integrated data on the user end.

The service-oriented architecture (SOA) of the walk-through command reorganization of the real-time business intelligence system, wherein the service-oriented architecture (SOA) 40 further includes a filter 47 that connects the dimension integration module 45 and the use On the user side, the filter 47 filters out the excess part of the data integrated by the dimension integration module 45 according to the conditions set by the user side, and then displays the filtered data on the user side.

The service-oriented architecture (SOA) of the walk-through command reorganization of the real-time business intelligence system, wherein the service-oriented architecture (SOA) 40 further includes a user interface control module 462, and the user interface control module 462 is connected to the On the user side, the IT staff side, the walking command reorganization module (ReSQL) 41, the data set 42 and the filter 47, the user interface control module 462 accepts the user side control input and inputs The signal is sent to the relevant IT staff, the walking instruction reorganization module (ReSQL) 41, the data set 42 and the filter 47, or receives the IT staff, the walking instruction reorganization module (ReSQL) 41, The data set 42 and the filter 47 signals control the output to the user terminal.

The service-oriented architecture (SOA) of the walk-through instruction reorganization of the real-time business intelligence system, wherein the service-oriented architecture (SOA) 40 further includes a user interface output module 461, and the user interface output module 461 is connected to the The filter 47, the user interface control module 462 and the user terminal, the user interface output module 461 can output data to the user terminal display.

A Service-Oriented Architecture (SOA) is composed of different components behind. These components can be placed in different processes according to different purposes to complete on-demand services. The structure of its design can be summarized into three levels:

(1) The BIS designed by the present invention is built on the operational IS of the enterprise (such as online transaction system, ERP, MES, CRM...), which saves the time of data conversion and integration into the data warehouse and provides the most real-time analysis data ( The development of data warehousing is often an extremely large investment in money, personnel and time).

(2) Based on changes in the external environment and increased competitive pressures, companies must rely on real-time data to make real-time analysis and make decisions, and on the other hand, IT personnel must support and satisfy the decision-makers to adjust and analyze at any time in the shortest time. Requirements; moreover, because the data source for analysis is not from the aggregated data warehouse, but from the "status" data at the time of the online transaction or the data needs to integrate different data such as homogeneous or heterogeneous databases Source, so in order to obtain the "analysis environment and calculation of multi-dimensional data with data warehousing (drilldown, rollup, slice or pivot, etc.) in real time", the present invention designs a service-oriented architecture (SOA) walking instruction reorganization module (ReSQL), IT staff provides real-time analysis services within a short period of time through simple definition settings according to the analysis needs of decision makers, such as multi-dimensional dip, roll-up, slicing, annual comparison, integration and comparison of multiple data sources, annual Advanced dimensional analysis of trends and detailed data...; This walking command reorganization module (ReSQL) dynamically reorganizes the optimized commands and analyzes them correctly (parse) data source, providing real-time data to achieve real-time decision analysis.

(3) Since the data warehouse is the "historical data" after consolidation, it can predict future trends and results and quickly provide decision makers with analysis data, or the data warehouse has been built within the enterprise, based on the company's movement under different elastic demand factors. The Recommanded Reorganization Module (ReSQL) can synchronously support data sources from data warehousing. On the other hand, when the data comes from the data warehouse, the decision-maker needs to view more detailed real-time data (raw data or some sort of aggregated data) during the analysis process, but these data do not exist in the data In warehousing, based on this, the Service Oriented Architecture (SOA) support can be connected to the transaction system in real time to present complete details while browsing the data from the data warehousing.

In a Service Oriented Architecture (SOA), software resources are packaged as services. After modularizing this service, it provides standard business functions. In order to support shared business processes, services can be viewed as building blocks to communicate with other services to support shared business processes; services are a small part of the functional side in a service-oriented architecture (SOA) The copy is accompanied by some important features, including services are autonomous, services hide underlying logic, services are loosely coupled, and complete service definitions (Services are well defined). Service Oriented Architecture (SOA) provides a flexible framework for modular applications into services. In addition, Service Oriented Architecture (SOA) allows enterprises to create, deploy and Integrate multiple services. So in this way, Service Oriented Architecture (SOA) can provide the needs of enterprise users for flexibility and agility, and define services that can be integrated and reused as key blocks to promote continuous and changing business needs.

The purpose of Service Oriented Architecture (SOA) is to build a flexible and reusable integrated interface for enterprises. The combined elements usually include software components, services and processes. When the decision maker is faced with a new demand, the process is responsible for defining the processing steps required by the decision maker; the service includes all program elements of a specific step, while the software element is responsible for the program that performs the task. The following separately illustrates how the design architecture of the present invention is based on a service-oriented architecture (SOA) to implement a BI 2.0 real-time analysis environment.

The third picture is the service-oriented architecture (SOA) BIS architecture, in which the walk-through instruction reorganization module (ReSQL) plays the main core role of the real-time analysis environment in the service-oriented architecture (SOA) and is responsible for constructing the user requirements (service) definition The required multi-dimensional data environment and processing flow to provide decision-makers with analysis services in real time; the undoubted command reorganization and analysis module needs to optimize the reorganized commands in addition to constructing complete retrieval commands In order to shorten the time of data retrieval; during the decision-making, the decision-makers according to their analysis mode (or the service required by the decision-maker), such as drilldown, rollup or pivot, etc., walk-through calculations, service-oriented architecture (SOA) method through walk-through command reorganization module (ReSQL) and communication link design, dynamically load data into data sets (DS, dataset) to achieve real-time response to user needs and changes; The Removable Command Reorganization Module (ReSQL) can quickly solve the aforementioned problems faced when analyzing data by integrating different data sources (homogeneous or heterogeneous databases).

The data loading (DS, dataset) in the service-oriented architecture (SOA) uses the server's cache key-value design concept, and the key obtained after the instruction reorganization is not in the cache memory (cache) before going to operation. The system database retrieves data.

Referring to the fourth diagram, the fourth diagram illustrates the design architecture of the operation of the walking instruction reassembly module (ReSQL). The present invention designs a first service definition file (eistable.def) corresponding to the transaction system 30 and a second service definition file (valueSQL.def) corresponding to the data warehouse 31 under the walk-through instruction reorganization module (ReSQL) as the entrance of the service process ( Or service interface), mainly responsible for the IT staff to define the data environment and process steps required by the demand (service) through the interface and behavior attributes according to the user's needs, as description, access, transmission, and understanding of each item The elements of the service; the present invention proposes a design framework for service innovation that includes different definitions (services) that can be modified by the inheritance module to modify the needs of decision makers and recombined between services to become a new service. The Removable Command Reorganization Module (ReSQL) is responsible for the definition of the service and the setting of user parameters, and completes the real-time demand (service) analysis in the design mode of "Removable Command Reorganization"; and the service-oriented concept, when When user needs change, the command reorganization and profiling modules provide resilience. From the fourth picture, we can see that the walking instruction reassembly module (ReSQL) is the entire service-oriented architecture (SOA) At the core, it provides a walk-through analysis mode to achieve an immediate analysis environment.

The following introduces the present invention into the actual hospital internal operational IS development supervisor information system (EIS) as an example to illustrate the structure and execution results of command reorganization and analysis of module design under the service-oriented architecture (SOA):

(1) Behavior attribute setting of the first service definition file (eistable.def)-multi-dimensional data cube

The following example illustrates how IT personnel define user analysis needs (services) in the first service definition file. Table 1 shows the attributes, set values and meanings of the services in the first service definition file, while "Hospitalization Date", "Department", "Department", ... etc. are the expressions corresponding to the dimension attribute name and measurement attribute, respectively . When the service definition is set up, the command reorganization and analysis module will obtain the definition content through the function of the first service definition file (National Science Fan One), combined with the user parameter settings, reorganize the data operation and retrieve the command to load the data into the data set (DS), as shown in the fifth figure; then provide multi-dimensional data analysis environment according to user analysis needs, when performing the following drilldown or rollup operations, the command reorganization and analysis module can Identify the current dimension and dynamically provide real-time analysis data, making the hierarchical landscape easier. Please refer to Figures 6 to 8 for execution results; even in different dimensions, you can view more detailed data or carry out For pivot analysis, please refer to the ninth figure for execution results. It is known from Table 1 that the IT staff only needs to understand the internal database structure of the enterprise and simple SQL commands to simply write the first service definition file to quickly construct the service-oriented architecture (SOA) of the present invention An analysis environment with multi-dimensional data for data storage is provided to provide decision-makers with drilldown, rollup and other analytical operations.

As mentioned earlier, data loading in the data set (DS) uses the server's cache key-value design concept. The key obtained after command reorganization is not in the cache memory to retrieve data from the operating system database. In order to speed up the reading time, the first service definition file will be loaded into the cache memory (when the first service definition file is modified, the relevant key-values in the cache memory will be automatically removed).

(2) Service definition inherits module design

The invention incorporates the object-oriented inheritance feature, allowing the child definition to modify the decision maker's needs by inheriting the parent definition (the child definition field setting can override or extend the parent definition setting), and its inheritance methods are: Two (1) No new definition is generated in the first service definition file after inheritance (2) A new definition is generated in the first service definition file after inheritance. The following two methods of inheritance are described: After inheritance, no new definition is generated in the first service definition file: if the data requirements of the decision-maker's probe are adjusted, the IT staff obtains the definition content of example 1 through the eistable function and redefines it The settings of drilldown are modified to the dimensions required by decision makers, as shown in Table 2. The tenth to twelfth figures illustrate the change of the dip dimension after inheritance, but the data source maintains the setting of Example 1. It is also worth mentioning that if the dip dimension does not include the original analysis requirement dimension, if the decision maker needs to see the department The summary data of the branch can be entered into the details. The instruction reorganization and analysis module will reassemble the retrieval instructions to the operating system database to retrieve the detailed data and provide decision makers with pivot analysis (dimension analysis). Please refer to the tenth Three pictures.

Inherit the definition of the original example 1, and create a new definition in the first service definition file (may be allowed to be inherited again): If the dimension attribute is to be modified into a department and a physician, please refer to Table 3 and Figures 14 to 15 The description of the figure, in which "ref=National Science Fanyi" in Table 3 will inherit all the definitions of National Science Fanyi and "xcols=Ministry" specifies the first-level dimension attribute as the ministry, and "drilldown=Ministry," Physician" overrides the definition of the parent definition to redefine the dimension Sex. In addition, the command reorganization and analysis module supports the user side to provide analysis services in different dimensions (please refer to the annual analysis data in Figure 16).

(3) Second service definition file (valueSQL.def)-integration of multiple data sources: There are often different transaction systems operating in the environment in which companies operate. These systems use different database management systems (DBMS) due to different construction times Even the attribute naming methods (data naming), data representation (data representation), data encoding (data encoding) or data content (data content), and data units (data scaling) on different systems are inconsistent. A BI solution must be seamlessly integrated with the overall information environment of the enterprise. It is very difficult to quickly integrate multiple data sources into a single real-time analysis environment to provide real-time analysis data for decision makers without the establishment of a data warehouse or additional data storage (physical data repository) Tasks; on the other hand, IT staff often spend a lot of time to complete the integration work. The service-oriented architecture (SOA) introduced by the invention becomes a new service innovative design architecture through the recombination of services, which makes the integration of heterogeneous systems easy, not only quickly solves the problem of heterogeneous data integration that enterprises often face, but also meets The dynamic process requirements of enterprises support the innovation of business models. Table 4 is an example of hospitalization declaration The database (EIS1) and the personnel database (HR), EIS1 stores the hospital's daily patient inpatient transaction information, each transaction records which hospitalized patient is visited by which doctor and the required medical expenses (MedFee) , Part of the patient's burden (PortFee) and application fee (AppFee). The HR database stores the information of each doctor's personnel information and classification. In order to analyze the data related to medical expenses, the current BIS must be structured in a multi-dimensional model data storage environment, as shown in Table 5. The data warehouse C (Data Mart C) uses ETL (Extraction, Transformation, and Load) to obtain information from the EIS1 and HR databases, respectively, in order to reach decisions across all relevant medical analysis dimensions and based on existing and available operational data Process. The doctor dimension integrates the doctor data table from EIS1 and HR database. Once the multi-dimensional data warehouse is built, the OLTP tool is usually used to allow users to analyze the data according to different dimensions. On the contrary, if the data warehouse is not built, it is assumed that the data of the doctors of grade A (leading physician) in the past three years in August will be aggregated into a single real-time analysis environment to provide decision-makers with real-time analysis data. Highly complex and challenging tasks.

The present invention uses the following examples to illustrate how to have a multi-dimensional data analysis environment and operations for data storage without quickly building a data storage, and quickly achieve integration of multiple data sources. Assume that the data source in the preceding example is the EIS1 database (hospital declaration database), where “medical expenses” refers to the total expenses (including the consultation fee) that have to be reported to the Health Insurance Bureau due to the medical behavior between the doctor and the patient , Medical expenses, inspection fees and equipment usage fees, etc.), based on the supervisor’s analysis, the “medical expenses” should include all the resources of the department’s intervention (including the nurses, administrators or some doctors who have not seen the department) Resources), and the supervisor wants to know the total number of physicians in the department in order to analyze whether there is a gap between income and cost, and further make certain decisions; however, to obtain the total number of physicians in the department cannot be obtained from the EIS1 single database, it must also be obtained from Complete information can only be obtained from another database. The following will explain how to combine services to form a new service to achieve the integration of multiple data: Table 6 is the second service definition file designed by the present invention to provide a service mechanism for processing multiple data integrations, where "specialist" refers to the name of the data source of the service after the content of the data source is set, that is, all the doctors in the department in the foregoing ( Suppose that includes the number of doctors who have been seen and not seen in the month. The data source is from the HR database; in the "specialist.setting", it means the attribute setting of the data source. For example, the data source in this example is The HR database and the detailed data to be drilled down to the doctor are specified in the <RawSelect></RawSelect> tag (if ID_DOCTOR_NAME is the doctor’s name); once the IT staff completes the settings, they can combine the information in the first service definition file The integration of the defined services into a single data set (DS) completes the integration of different data sources, as shown in Table 7. The data definition of the first service definition file of this example 4 inherits the definition of example 3, and it can be integrated into the data set (DS) by adding "specialist" ("+") to vcols Calculation of other measurement attributes (such as the average number of visits); then the user in different dimensions, the command reorganization and analysis module reorganizes the retrieval commands of two different data sources according to the current dimension to achieve the purpose of numerical integration , Please refer to the integrated architecture description in Figure 17. In addition, the data on the classification of the doctors mentioned above comes from the HR database in Table 4. Assuming that the data of the doctors of Grade A in August in the past three years have been aggregated into a single OLAP analysis environment, please refer to the table 8. The definition of the integration definition of the classification in Table 9, please refer to Figure 17 bis for the integrated architecture diagram; even the comparative data of different years can be quickly integrated into the data set (DS), please refer to the 19th To the results of the twenty-first figure execution.

In addition, how to effectively solve the aforementioned inconsistencies such as attribute naming, data representation, etc. that are faced by the data integration mentioned above. The invention takes the inconsistent attribute naming as an example to explain the solution method. If the attribute name or data representation of the key (dimension) between the multiple data sources in Figure 18 is inconsistent; for example, suppose the attribute name of the data source key defined in the first service definition file is xyz, xyz1,... The attribute name of the data source key defined in the second service definition file is abc, abc1.... The design method of the present invention defines the mapping of the attribute in the second service definition file to achieve agreement. Please refer to the settings in Table 10; if there is a conflict in the data representation (data type), you can use the functions provided by the database management system (such as convert, substring, ...) or custom functions (such as: to_char,: substr ,...), as shown in Table 10, xyz=: to_char(abc), converted to a consistent data type.

Another example illustrates the combination of the second service definition file and the first service definition file service to solve the problem of integrating multiple data of different types. Assuming that the decision maker needs a certain analysis data, this data needs to be aggregated from two different databases (possibly summed up, averaged other calculations) according to different drill-down dimensions. The names are not necessarily consistent), as long as similar to Table 6 define two data sources in the second service definition file (assuming DS1 and DS2), then add DS1 and DS2 to vcols in the first service definition file (as shown in Table 7) "Specialists"), and then use compute (such as "average number of patients" in Table 7) to consolidate and achieve integration.

Looking at the design purpose of the second service definition file above, in addition to the rapid integration of data sources from different databases, in some cases, more complicated calculations are required, as long as they are defined in the second service definition file and then combined with the first service definition file. For the purpose of integration, please refer to Table 11 and we will discuss this innovative design architecture in more depth, such as the "specialist" defined in Table 6, in fact, it represents the design concept of an object, and this object Can be used in any service definition process in the first service definition file (as long as the service process and the object can be linked), through the link of the message to achieve the purpose of integration and object reuse, also echoes the previously described service room Can be combined into a new service.

According to the above description, the combination of the second service definition file and the first service definition file service definition can quickly complete different service needs of users, and solve more complicated problems and quickly achieve integration of multiple data sources. It is also worth mentioning that the architecture is based on the Removable Command Reorganization Module (ReSQL), which can easily expand the definition of different service processes and complete different service functions. It also shows the Reactive Command Reorganization Module (ReSQL) Strict architecture design, complete integration and high cohesion.

(3) Data warehouse (Cube) and online trading system data synchronization integration

As mentioned above, the Data warehouse and Cube have a certain degree of support or the enterprise has built a data warehouse. Based on the enterprise's different elastic demand factors, the service-oriented architecture (SOA) designed by the present invention is also suitable for data. On the basis of warehousing, it can further integrate data storage (or cube) and transaction system data at the same time. The integration method is shown in Table 12. Specify which database system the source is from in the initial settings of the data source , Such as CUBE1=summary 1 (assuming CUBE1’s data source is Data warehouse environment); EIS1=real-time one (assuming EIS1’s data source is the hospital declaration transaction system), and the data in the transaction system can be specified from EIS1, as shown in the table Twelve Cube1.RawDataDB=EIS1 and EIS1.IsCubeDB=false. Modification example 1 of the present invention shows how to integrate data warehouse (Cube) and transaction system data in the first service definition file; please refer to Table 13 "table.cube=eiscubeview(DTLFBC)" means that the data source of the dimension attribute is from The data table "DTLFBC" in CUBE1 is obtained, and the real-time transaction data source is from EIS1. The data source definitions of the above-mentioned example 2 to example 4 are all inherited from the definition of example 1. Once the definition of example 1 is modified, all the inherited examples are modified together. Since the current data collected by CUBE1 is consistent with the result of the above example, the data source is EIS1, so there is no need to provide an integrated result screen here.

(4) Integrate definition files of other service functions

Based on the above description, the walking instruction reorganization module (ReSQL) designed by the present invention truly implements the core spirit of the service-oriented architecture (SOA), that is, emphasizes abstraction, composability, and reusability (reusability) to achieve BI 2.0 real-time analysis services.

40‧‧‧Service Oriented Architecture (SOA)

41‧‧‧ Walking Reorganization Module (ReSQL)

411‧‧‧Command reorganization and analysis module

412‧‧‧First service definition file (eistable.def)

413‧‧‧Second service definition file (valueSQL.def)

414‧‧‧Inheritance module

42‧‧‧ Dataset (DS)

43‧‧‧grammar conversion module

44‧‧‧Multi-value data processing module

45‧‧‧Dimensional Integration Module

461‧‧‧User interface output module

462‧‧‧User interface control module

47‧‧‧filter

48‧‧‧ homogeneous or heterogeneous database

49‧‧‧Cache

Claims (4)

A service-oriented architecture (SOA) for walk-through instruction reorganization of real-time business intelligence systems, including: a service-oriented architecture (SOA) (40) connected to a user terminal and an information technology (IT) personnel-side telecommunications ; It is characterized in that the service-oriented architecture (SOA) (40) includes a walking instruction reorganization module (ReSQL) (41), and the walking instruction reorganization module (ReSQL) (41) includes a command reorganization and analysis The module (411), a first service definition file (eistable.def) (412), a second service definition file (valueSQL.def) (413) and an inheritance module (414) are composed of mutual telecommunication links. A service definition file (eistable.def) (412) and the second service definition file (valueSQL.def) (413) are simply defined by the information technology (IT) staff, with a homogeneous or heterogeneous database (48) telecommunications link The walking instruction reorganization module (ReSQL) (41), the data is taken from the homogenous or heterogeneous database (48), which is based on the moving instruction reorganization module (ReSQL) (41) dynamically and automatically reorganization optimization Instruction, and then go to the homogeneous or heterogeneous database (48) to obtain the required data, the service-oriented architecture (SOA) (40) further includes a syntax conversion module (43) telecommunications link to the homogeneous or heterogeneous database (48) and the Removable instruction reorganization module (ReSQL) (41), through which the syntax conversion module (43) can grammatically convert the instruction reorganization and analysis module (411) and the homogeneous or heterogeneous database (48), so that all Automatic data acquisition can be achieved due to grammatical conversion. The inheritance module (414) can modify the needs of the user side through inheritance between the services of the walk-around command reorganization module (ReSQL) (41) The reorganization between the services through the Removable Command Reorganization Module (ReSQL) (41) is sought as a new service. The Reorganization and Analysis Module (411) is responsible for the ReSQL Reorganization Module (ReSQL) ( 41) The definition of the service and the parameter settings input by the user end complete the real-time demand service analysis in the order reorganization operation mode of the walk-through order reorganization. The service-oriented architecture (SOA) of the walk-through instruction reorganization of the real-time business intelligence system as described in item 1 of the patent scope, wherein the instruction reorganization and profiling module (411) is constructed with a complete retrieval instruction, the instruction reorganization and The profiling module (411) also has reorganized instruction optimization. The service-oriented architecture (SOA) of the walk-through instruction reorganization of the real-time business intelligence system as described in item 1 of the patent scope, in which the user side accesses data, including drill-down and roll-up -up), pivot analysis (Pivot) or slice (Slice). A service-oriented architecture (SOA) for walk-through instruction reorganization of real-time business intelligence systems, including: a service-oriented architecture (SOA) (40) connected to a user terminal and an information technology (IT) personnel-side telecommunications ; It is characterized in that the service-oriented architecture (SOA) (40) includes a walking instruction reorganization module (ReSQL) (41), and the walking instruction reorganization module (ReSQL) (41) includes a command reorganization and analysis The module (411), a first service definition file (eistable.def) (412), a second service definition file (valueSQL.def) (413) and an inheritance module (414) are composed of mutual telecommunication links. A service definition file (eistable.def) (412) and the second service definition file (valueSQL.def) (413) are simply defined by the information technology (IT) staff, and the mobile or heterogeneous database (48) telecommunications link the walking command Reorganization module (ReSQL) (41), the data is taken from the homogenous or heterogeneous database (48), which is a real-time reorganization module (ReSQL) (41) _ dynamically and automatically reorganizes the optimized instruction according to the walking instruction. Then go to the homogeneous or heterogeneous database (48) to obtain the required data. The service-oriented architecture (SOA) (40) further includes a syntax conversion module (43) to connect the homogeneous or heterogeneous database (48) and the walk Instruction Reorganization Module (ReSQL) (41), through which the grammar conversion module (43) can grammatically convert the instruction reorganization and analysis module (411) and the homogeneous or heterogeneous database (48), so that all data The automatic acquisition can be achieved due to the grammatical conversion. The inheritance module (414) can modify the needs of the user terminal and inherit the walkthrough through inheritance between each service of the walkable instruction reorganization module (ReSQL) (41). The reorganization of the instruction reorganization module (ReSQL) (41) into a new service, the instruction reorganization and analysis module (411) is responsible for the definition and service of the walk-through instruction reorganization module (ReSQL) (41) The parameter settings entered by the user terminal complete the real-time demand service analysis by the command reorganization operation mode of the walk-through command reorganization; wherein the service-oriented architecture (SOA) (40) further includes a multi-value data processing module (44). The value data processing module (44) telecommunications connects the walking instruction reorganization module (ReSQL) (41), the grammar conversion module (43) and the homogeneous or heterogeneous database (48), using the multiple value data processing module Group (44) performs multiple data calculation and integration; The service-oriented architecture (SOA) (40) further includes a data set (DS, dataset) (42) and a cache (cache) (49), the data set (42) telecommunications link to the walk-through command reorganization The module (ReSQL) (41), the multi-value data processing module (44), the homogeneous or heterogeneous database (48) and the user terminal, after the instruction reorganization, the data set (42), the data set ( 42) When the cache (49) cannot find data, it will only retrieve data from the homogeneous or heterogeneous database (48) corresponding to the command; the service-oriented architecture (SOA) (40) is more It includes a one-dimensional integration module (45). The one-dimensional integration module (45) connects the data set (42) (DS, dataset) with the user terminal through telecommunications. The one-dimensional integration module (45) has integrated measurement Analyze the data scattered in different homogeneous or heterogeneous databases (48), and then display the integrated data on the user-side display; the service-oriented architecture (SOA) (40) further includes a filter (47) , The filter (47) telecommunications connects the dimension integration module (45) and the user terminal, and the filter (47) filters out the dimension integration module (45) after integration according to the conditions set by the user terminal The redundant part of the data is displayed on the user-side display as filtered data; wherein the service-oriented architecture (SOA) (40) further includes a user interface control module (462), the user interface control module (462) Telecommunications link the user terminal, the IT personnel terminal, the walking command reorganization module (ReSQL) (41), the data set (42) and the filter (47), the user interface control module (462) Accept the control input from the user terminal and send the input signal to the IT staff , The walking instruction reorganization module (ReSQL) (41), the data set (42) and the filter (47), or receiving the IT staff, the walking instruction reorganization module (ReSQL) (41), The data set (42) and the filter (47) signals control the output to the user end; wherein the service-oriented architecture (SOA) (40) further includes a user interface output module (461), the user The interface output module (461) is electrically connected to the filter (47), the user interface control module (462) and the user terminal, and the user interface output module (461) can output data to the user terminal monitor.

Images