KR100685747B1 - System and method for processing sensor data stream based on xml - Google Patents

Abstract

The present invention relates to an eXtended Markup Language (XML) -based sensor data stream processing system and method, and defines an event (that is, whether or not a specific situation occurs) configured with various conditions for data streams coming from various types of sensors. It is possible to easily build various ubiquitous service environments by defining services that should be applied when specific events occur and enabling them to be executed.

In addition, the XML-based sensor data stream processing system and method of the present invention is configured to process sensor data based on an XML model having excellent expressive power and flexibility, so that interworking with various sensors providing various types of sensor data is difficult. It is easy. In addition, by processing a large amount of sensor data stream coming from the interlocked sensors in real time to provide the real-time information required by various applications in the ubiquitous environment, the service provider can concentrate on the development of services.

XML, Sensor Data, Stream Processing, Events, Ubiquitous

Description

WML-based sensor data stream processing system and method {SYSTEM AND METHOD FOR PROCESSING SENSOR DATA STREAM BASED ON XML}

1 is a block diagram of a sensor data stream processing system according to the present invention;

FIG. 2 is a flowchart illustrating an operation for automatically calling an associated service by processing a sensor data stream coming from a sensor in an XML-based sensor data stream processing system according to the present invention;

3 is a diagram illustrating a PML core sensor data model, which is a format of sensor data processed by a sensor data stream processing system according to the present invention, in UML;

4 is a diagram illustrating an XML element definition in FIG. 3;

5 is a diagram illustrating the definition of AnyXMLContentType, which is a data type of an XML element of FIG. 4;

6 is a diagram showing an example of sensor data format registration;

7 is a view showing a sensor data format registration process of FIG. 6;

8 is a diagram illustrating information provided by a user when registering a context-service.

<Description of Symbols for Major Parts of Drawings>

10: sensor

20: Sensor Data Processor (SDPS)

20a: sensor manager

20b: Sensor Data Manager

20c: context-service manager

20d: sensor data access manager

20e: SDPS Interoperation Manager

30: Data Stream Manager (DSMS)

30a: Data Stream Source Manager

30b: CQ Manager

30c: Performance Monitor

30d: data stream access manager

40: sensor data storage manager

40a: XML / Relation Converter

40b: MMDBMS

The present invention relates to an XML-based sensor data stream processing system that automatically calls a linked service after converting and processing data streams from various sensors such as GPS, temperature, and bio signals into XML (eXtended Markup Language) format. More specifically, the present invention relates to an XML-based sensor data stream processing system and method that can effectively process sensor data in the PML Core format, which is a data standard format proposed by EPCglobal.

Sensor data stream processing technology is essential for processing various kinds of high-capacity, real-time sensor data that is constantly generated in ubiquitous computing environment. Sensor data stream processing technology is a technology that processes, processes and provides various sensor data such as RFID, GPS, temperature, and bio signals to application services, and is a core technology of ubiquitous computing. In particular, RFID-based sensor data stream processing technology has recently been actively studied in industry and academia. Industrial companies have developed or are developing various middleware for RFID-based sensor data processing, and academia is conducting research on Data Stream Management System (DSMS), which is the basis of sensor data stream processing.

Examples of RFID-based sensor data processing middleware developed by companies include SUN's Savant, CapTech's TagsWare, GlobeRanger's iMotion, ConnecTerra's RFTagWare, Manhattan Associates 'RFID Middleware, and OATSystems' EPC-IS Edge Server. Most of these middlewares aim to provide a foundation for integrating RFID systems to reduce the significant amount of custom development required for RFID projects, and are designed to work with sensors or readers, filtering or smoothing sensor data processing. It provides a logging function for sensor data and supports the real-time processing of sensor data and interworking with application services. For reference, see their web sites for an overview of these middleware. The detailed technical contents of these are not disclosed in detail in the nature of the company, and there are few related patents. An apparatus, method and program for tracking and inventory management in the field of SCM have been proposed.

Since the sensor data is generated in a continuous and unlimited amount of data, continuous query processing technology for a large amount of stream data of a data stream management system (DSMS) is essential to process the data. Research in the data stream management system, which is mainly conducted in academia, is based on the Stanford University STEAM (Standford stream data Manager) project, which is designed to define continuous queries for stream data through CQL (Continuous Query Language), which greatly expands SQL. In the continuous query processing based on PostgreSQL, an object-relational DBMS, Berkeley University's TelegraphCQ project seeks to improve performance through sharing and adaptability, and allows users to create graphs via the “Box-and-Arrows” interface. There is an Aurora project in collaboration between Brandeis University and MIT University, which has formed a network of operators to enter queries and process continuous queries. Others include NiagaraCQ for XML document processing, OpenCQ for monitoring changes on the Internet, and YFilter as a filtering system that connects documents expressed in XML to XPath expressions of interest.

Meanwhile, EPCglobal, which leads the de facto standard of RFID technology, proposes a reference architecture called EPC network and provides related standards to make it easier and more convenient to develop application services for processing sensor data streams. EPCglobal defines EPC and Physical Markup Language (PML) as a standard for the format and semantics of the data exchanged in the EPC network architecture. The EPC is a recognition system for universally recognizing physical objects. The EPC includes an EPC identifier field for uniquely identifying an individual object and a filter value field for efficiently reading an EPC tag. PML is an application of XML. It is a standard for expressing and exchanging information about objects on the EPC Network. PML Core is a standard format for exchanging data obtained from sensors such as RFID readers, which are described in detail in PML Core specification 1.0. have.

Since middleware developed by companies is based on relational models, most of them are efficient to process fixed-type data with simple data types, but are not suitable for processing various types of sensor data with complex data types. In addition, since the developed middleware is based on middleware that is biased in a specific field such as SCM, ERP, etc., the application service developer has a function to define an event according to the needs of the user and the application field and to link the service according to the situation. It lacks a lot of time and effort in developing application services.

Therefore, the present invention has been proposed to solve the above problems, and an object of the present invention is a problem of the conventional sensor data stream processing system, which is difficult to model and process sensor data having various types and complex formats. In order to improve the performance, we provide an XML-based sensor data stream processing system and method that automatically invokes the associated service after processing data streams from various sensors based on XML with rich modeling ability and flexibility.

Hereinafter, an XML-based sensor data stream processing system and method of the present invention will be described in detail with reference to the accompanying drawings.

1 is a configuration diagram of an XML-based sensor data stream processing system according to the present invention, wherein the sensor data stream processing system includes a sensor 10, a sensor data processor (SDPS) 20, and a data stream manager (DSMS) 30. And a sensor data storage manager 40 or the like.

The sensor data processor 20 provides a sensor data stream processing function from the sensor 10, and includes a sensor manager 20a that manages registration and connection information of a sensor, a sensor data source, a sensor data browsing, and a sensor data log. Sensor data manager 20b for managing the data, context-service information management, context-service manager 20c for service message generation, remote service calls, and sensor data access manager 20d for setting sensor data access rights And an SDPS interworking manager 20e which is in charge of interworking with other sensor data processors 50 and 52.

The data stream manager 30 provides a continuous query processing function for a large amount of stream data, and the data stream source manager 30a in charge of managing schema and input queues for the data stream coming from the sensor data processor 20. , CQ manager 30b for continuous query parsing, execution planning, scheduling and execution, performance monitor 30c for measuring continuous query executor performance and providing scheduling information, user access rights and authentication to data streams And a data stream access manager 30d for managing the information.

The sensor data storage manager 40 is in charge of storing and managing data that needs to be permanently stored in order to support a historical query among sensor data transmitted from the sensor 10, and stores data or queries in XML format. It consists of an XML / Relation converter 40a and a MMDBMS (Main Memory Database Management System) 40b that converts relational data or SQL and vice versa.

2 is a flowchart illustrating an operation of automatically invoking an associated service by processing a sensor data stream coming from a sensor in an XML-based sensor data stream processing system according to the present invention.

The operation of the sensor data stream processing system goes through two main processes. That is, there are an environment building step (S100) and a sensor data stream processing step (S200) for preparing the sensor data stream processing.

The environment building step (S100) consists of a sensor registration step (S110), a sensor data source registration step (S120), and a context-service registration step (S130). Each step is briefly described as follows.

◆ Sensor registration step (S110):

Register the sensor by inputting the sensor information such as sensor name, sensor model, number of connected antennas, connected antenna list, connection method, address, and connection port.

◆ Sensor data source registration step (S120):

The sensor data source refers to an input data source that groups the associated sensors and delivers them to the data stream manager 30. The sensor data source corresponds in one-to-one correspondence with the input source of the data stream manager 30, and generates and registers a source by inputting a source identifier, a sensor list, a source schema, an input queue size, an input queue type, and the like. The details of the sensor data source registration will be described below with reference to FIGS. 6 and 7.

◆ context-service registration step (S130):

In context-services, the context refers to a situation that must be satisfied in order for the service to be performed, and the service refers to an action that is performed when a particular situation is satisfied. Context-service registration registers {context information, service information, mapping information, execution interval} as input. Detailed information required for registration will be described below with reference to FIG. 8.

In the process of registering a context-service, the context-service manager 20c calls the CQ manager 30b of the DSMS 30 as a context (i.e., a continuous query), and the CQ manager 30b calls the continuous This is done by analyzing the query and creating and returning an output queue to return the processing result of the relevant processing structure and context.

When the registration step is completed, it is possible to receive the data stream from the sensor 10, process and then call the service according to the result.

On the other hand, the sensor data stream processing step (S200) is composed of a sensor data input step (S210), a sensor data processing step (S220), a service interworking step (S230), briefly described for each step as follows.

◆ Sensor data input step (S210):

The data is received from the sensor 10 and placed in the input queue of the corresponding source. At this time, the XML document or the document is parsed according to the type of the input queue, converted into a tree structure, and put into a tree form.

◆ Sensor data processing (continuous query processing) step (S220):

The CQ manager 30b reads data from the input queues for each successive query, processes the successive query, and puts the query result into the corresponding result queue.

◆ Service Interlocking Steps (S230):

After mapping data to the service bound to the result queue by using the registered mapping information, call the service referring to the execution interval.

The sensor data stream processing system according to the present invention converts sensor data of various formats from interlocked sensors into PML Core schema format, which is a sensor data exchange standard format proposed by EPCglobal. The PML core schema will be described with reference to FIGS. 3, 4, and 5 as follows.

3 is a PML Core sensor data model expressed in UML.

The sensor element is the main element of the PML core and is composed of an ID element and one or more observation elements.

The observation element is an element that includes data measured by a specific sensor and must include a DateTime element, and may include an ID element, a Command element, and the like. In addition, the data element may include several data elements or tag elements.

The data element should be used as an element for representing the measurement data measured by the sensor, otherwise the tag element should be used to represent the measurement data. In addition, data elements can be used to represent structured data by representing unstructured data in simple data blobs, such as strings or hexbinaries, or by inserting additional XML tags that exist in other namespaces. Can be. In addition, the data element is configured by selecting one of three elements, that is, a text element, a binary element, and an XML element.

XML element is used to express sensor data with XML element beyond PML Core specification. The 'any' element in the schema definition statement (XSD) allows you to represent elements not specified by the PML Core schema. As such, the reason for providing openness locally to the PML Core schema is to allow users to model the sensor data in various ways as desired.

4 is a definition of an XML element in the PML Core schema, and FIG. 5 is a definition of AnyXMLContentType which is a data type of an XML element in the PML Core schema.

Except for XML elements whose data type is AnyXMLContentType, the PML Core schema is already formatted and fixed. Therefore, only the XML element part whose format is not determined can be received and inserted into the corresponding part of PML Core schema.

FIG. 6 shows an example of sensor data source registration of FIG. 2.

An XML schema definition statement with the source identifier i, the list of sensors to register, S3 and S4, the source schema (that is, the schema of the XML element of AnyXMLContentType) as hatched, the input queue size is qs, the input queue This is the situation in the sensor data stream processing system, assuming the type accepts data in the form of an XML tree.

FIG. 7 shows a sensor data source registration process in the situation of FIG. 6.

The sensor data manager of the sensor data processor creates a new source item in the sensor data source information management table and registers the related information (S120a), and inserts the input source schema into the XML data portion of the sensor data schema template for a new source. The entire schema (ie, SrcSchema1 of FIG. 7) is completed (S120b), and the data stream source manager of the data stream manager is called while giving the schema (S120c).

The data stream source manager of the data stream manager uses the source schema (SrcSchema1) information received as input to create a new source in the DSMS and registers the related information in the internal source information management table (S120d).

In operation S120e, an input queue IQi having a queue size of qs and a type of input data as a tree is generated in the data stream manager as a device to receive data continuously (S120e).

FIG. 8 shows information necessary for registering a context-service of FIG. 2.

In order to register a context-service, a user must provide {context information, service information, context-service mapping information, service execution interval}, and the like as an input. Each information is described as follows.

◆ Context Information:

The context information is expressed using a continuous query provided by the data stream manager. This continuous query is an extension of the XML query language that allows you to retrieve streaming data in the form of XML from a sensor.

◆ Service Information:

A service is an action provided to a user, such as a web service and a Java messaging service (JMS).

◆ Context-to-service mapping information:

The mapping information indicates information on how to accept a predetermined value (static information) and a result of a situation check query (dynamic information) when defining a context-service as an input to a service.

In the case of static information, it is expressed as a pair of argument names and values, and dynamic information is expressed as a pair of paths to argument names and values. The path to the value representing the dynamic information uses the full path to the element in the XML document that represents the query return value.

◆ Service intervals:

The interval for invoking the service, expressed in days: hours: minutes: seconds. If not specified, the service is called as soon as it is satisfied.

As described above, the XML-based sensor data stream processing system and method according to the present invention provides the following advantages over existing systems.

First, sensor data streams with various XML formats are processed in real time. By providing a means of registering data types for user-defined types so that any type of XML data can come, new sensors can be easily interlocked without modifying internal programs.

Second, specific events based on sensed data can be defined using extended XQuery. Thus, multiple continuous XQuery processing is supported for continuous incoming sensor data streams.

Third, when a defined event occurs, web services, JMS, etc. can be linked.

Claims (12)

In the XML-based sensor data stream processing system that processes services by processing various XML sensor data streams, A sensor data processor comprising a sensor manager, a sensor data manager, a context-service manager, a sensor data access manager, and an SDPS interworking manager to provide a sensor data stream processing function; A data stream manager comprising a data stream source manager, a CQ manager, a performance monitor, and an access manager to provide continuous query processing for a large amount of stream data; And Sensor data storage manager, consisting of XML / Relation converter and MMDBMS, responsible for storage management of data that needs to be stored permanently to support historical queries XML-based sensor data stream processing system comprising a. The method of claim 1, wherein in the sensor data processor, The sensor manager manages registration and connection information of the sensor, the sensor data manager manages sensor data sources, sensor data browsing, and sensor data logs, and the context-service manager manages context-service information and generates service messages. And a remote service call, the sensor data access manager sets sensor data access rights, and the SDPS interworking manager is in charge of interworking with other sensor data processors. system. The method of claim 1, wherein in the data stream manager, The data stream source manager is responsible for schema and input queue management for data streams coming from the sensor data processor, the CQ manager is responsible for continuous query parsing, execution planning, scheduling, and execution, and the performance monitor is continuous. And measuring scheduling performance of the query executor to provide scheduling information, wherein the data stream access manager manages a user's access authority and authentication information on the data stream. The method of claim 1, wherein in the sensor data storage manager, And the XML / relation converter and the MMDBMS convert XML data or queries into relational data or SQL, or vice versa. The method according to any one of claims 1 to 4, The sensor data stream processing system converts sensor data of various formats from interlocked sensors into PML core schema format, which is a sensor data exchange standard format proposed by EPCglobal, and processes the converted sensor data. In the XML-based sensor data stream processing system according to any one of claims 1 to 4, the XML-based sensor data stream processing method for processing the sensor data stream coming from the sensor to automatically call the associated service In An environment building step for preparing the sensor data stream processing comprising a sensor registration step, a sensor data source registration step, and a context-service registration step; And The sensor data stream processing step comprising a sensor data input step, a sensor data processing step, and a service interworking step XML-based sensor data stream processing method comprising a. The method of claim 6, In the sensor registration step, the sensor is registered by inputting information about the sensor to a sensor name, a sensor model, the number of connected antennas, a connected antenna list, a connection method, an address, a connection port, In the step of registering a sensor data source, a sensor data source is generated by inputting a source identifier, a sensor list, a source schema, an input queue size, an input queue type, and the like, which are input data sources that are grouped with the connected sensors and delivered to the data stream manager. To register, The context-service registration step of the XML-based sensor data stream processing method, characterized in that for registering the context information, service information, mapping information, execution interval. The method of claim 7, wherein When registering the sensor data source, the XML-based sensor data stream processing method comprising receiving and registering only an XML data portion having an AnyXMLContentType as an input instead of the entire sensor data schema. 8. The process of claim 7, wherein the processing for context-service registration is: The context-service manager calls the CQ manager of the DSMS with the context as input, and the CQ manager analyzes the context and generates and returns an output queue for returning the related processing structure and the processing result of the context. How to process XML-based sensor data streams. The method according to claim 7 or 9, XML-based sensor characterized in that the context of the event, the service information, the context-service mapping information, the service execution interval while defining the context of the event and the service to associate the service according to the situation of the input sensor data when registering the context-service How to handle data streams. The method according to claim 7 or 9, The context information is expressed using a continuous query that extends an XML query provided by the data stream manager. The service information is a web service and a Java messaging service provided to the user, The context-service mapping information represents information on how to accept a predetermined value (static information) and a result of a situation check query (dynamic information) when defining a context-service as an input to a service. The service execution interval is expressed as days: hours: minutes: seconds in a cycle of invoking the service, and if not specified, the service processing method for an XML-based sensor data stream, characterized in that the service is called immediately when a satisfactory situation occurs. The method of claim 6, In the sensor data input step, the sensor data is parsed into a tree structure by parsing an XML document or a document according to the type of the input queue with respect to an input queue of a corresponding source, and converting the sensor data into a tree structure. In the sensor data processing step, the CQ processor reads data from the input queues for each continuous query, processes the continuous query, and puts the query result into the corresponding result queue. In the service interworking step, the XML-based sensor data stream processing method, characterized in that the service is called with reference to the execution interval after mapping the data to the service bound to the result queue using the registered mapping information.

Images