CN101303646A - A Modeling Method for Telecommunications Domain Based on Executable Meta-Language - Google Patents

Abstract

The telecommunications field modeling method based on executable metalanguage of the present invention belongs to the technical field of telecommunications network management, and specifically relates to a modeling method of network equipment, network performance, and network failure in the telecommunications field and a verification method of model correctness. The core of the modeling process is the executable meta-language xKL suitable for the telecommunications field; it guides the meta-modeling process, describes and dynamically instantiates the domain model, verifies and executes the instantiated model, and makes the generated model executable. The present invention has the following advantages: good scalability and adaptability; good reusability; executable object model; domain model can establish a good mapping relationship between system requirements and underlying implementation; can shorten product delivery Time to market, improve, improve product quality, and increase customer satisfaction.

Description

A Modeling Method for Telecommunications Domain Based on Executable Meta-Language

technical field

The invention belongs to the technical field of telecommunication network management, and in particular relates to a modeling method of network equipment, network performance and network failure in the telecommunication field and a verification method of model correctness.

Background technique

With the rapid development of the telecommunication industry, the competition among various telecommunication operators is becoming increasingly fierce, and the rapidly developing user group, diversified business and business competition environment put forward higher and newer requirements for the service quality of telecommunication enterprises. Facing market challenges, telecom operators need to improve management efficiency and network management level. Therefore, it is imperative to build a cross-platform integrated network management system. This platform can intuitively, comprehensively and quickly analyze and judge the operation quality of the operator's telecommunication service network from the height of the entire network, and perform fault discovery, location and monitoring. Exclude, analyze business and customer impact. In the telecommunications field, network equipment is diversified and network services are diversified. The idea of domain modeling can fundamentally solve the problems brought about by the continuous development of new services and new equipment in the current telecommunications network management to the construction of network management systems.

Currently, the popular object-oriented modeling tool is Rational Rose developed by Rational Corporation. There are Beida Jade Bird object-oriented modeling tools in China. These modeling tools are based on a common modeling language - Unified Modeling Language. The Unified Modeling Language is an object-oriented modeling language standard initiated by the Object Management Organization, and has now become the standard for object-oriented modeling. The unified modeling language describes the static structure information of the system through classes, objects, relationships, etc., and describes the dynamic behavior of the system through sequence diagrams and state transition diagrams. This abstract description of the software system facilitates the exchange of ideas between developers, and also provides convenience for document processing. Unified Modeling Language is a very successful modeling language, which has made great contributions to the standardization of software process and the improvement of the efficiency of software development. However, problems have also been encountered in the process of practical application. The most important problem is that the unified modeling language is only applied in the analysis and design stages of the software process at present. The models established by using this modeling language are not executable, and the analysts and designers can only manually check the correctness of the models. It brings hidden dangers to the design of the model and the implementation of the code.

Invention content:

The technical problem to be solved by the present invention is to provide a modeling method in the telecommunication field based on the executable meta-language, and solve the problem that the field model in the telecommunication field cannot be verified after being established.

The model is the formal representation of the function, behavior and structure of the system constructed by the present invention. By expressing the domain model as an executable model, the correctness of the model can be tested at the domain level without considering platform-related issues. In this way, the domain The model will never become obsolete due to technological changes, so it can be continuously developed and reused.

A method for modeling the telecommunications domain based on executable meta-language, the method includes the following steps:

a) Based on the extension of the Unified Modeling Language based on the Meta Object Facility (MOF), an executable metamodel language suitable for the telecommunications field—xKL language is designed, and the modeling tools and verification tools based on the xKL language are designed and constructed.

b) Collect relevant information in the telecommunications field, integrate them into a domain model and describe it as a class diagram structure with visual components; use xKL to fill the constraints and static semantic parts in the class diagram to generate a complete graphical domain model.

c) Mapping the graphical domain model to the telecommunications domain model completely based on xKL text.

d) Starting from the telecom domain model, according to the problems to be solved, select the classes in the domain model for instantiation, generate memory dynamic operable objects based on xKL, and instantiate the relationship between classes in the domain model as xKL memory objects Relationships among objects form an object model.

e) Execute the xKL-based constraints in the domain model, verify whether the objects in the object model meet the due constraints, if not, dynamically modify the model according to the corresponding constraints, and return the modified results.

f) Send an instruction to a specific object in the model to make it execute the corresponding method, and check the effect of the method execution.

The xKL language is composed of a static syntax meta-model and a dynamic behavior meta-model, as shown in FIG. 1 . The static grammatical metamodel extends the EMOF metamodel and defines the structure required to represent the model; the dynamic behavioral metamodel extends the OCL peripherally to form the xOCL language, and xOCL defines the operation of the model.

As can be seen from Figure 1, a more specific description of the composition of the xKL language is:

1. Static syntax metamodel. EMOF is the basic core of MOF2.0, and the static grammatical metamodel forms the grammatical structure part of the xKL language by extending the EMOF metamodel. In order to meet the requirements of platform independence, it only defines the structure required by a model, and does not give any specification of model behavior.

2. Dynamic Behavior Metamodel. Behavioral meta-model makes peripheral extensions to OCL and forms xOCL language. The xOCL specification is completely based on the core definitions of UML and MOF, which makes it applicable to MOF and UML. Since EMOF is non-executable, some behavioral elements need to be added, and the behavioral metamodel is formed by adding some operations on the model in the OCL language. The structural model and the behavioral model are connected by writing behavioral statements in the operating body of the class.

The system architecture of modeling tools and verification tools based on xKL language is shown in Figure 2. In Fig. 2, the bottom layer is the Java virtual machine JVM part, which provides basic instruction support for the virtual machine of the executable metamodel. Part of the core semantics and basic input and output of the executable meta-model are implemented in Java, while the rest of the extended parts can be realized by the executable meta-model itself. The executable meta-model virtual machine is implemented based on JVM, and the executable virtual machine provides a virtual interpretation and execution environment for the model of the executable meta-kernel language xKL. xKL provides the basic executable model concept on the EMOF virtual machine, and other executable meta-models are implemented based on xKL. xOCL provides a constraint language, and other modeling languages for the executable meta-model and its description provide a well-formed relational constraint or model query function of abstract syntax concepts.

Domain information mainly comes from existing standards, existing systems, domain experts, requirements manuals, etc.

The domain model includes concepts abstracted from the telecommunications domain, namely classes and relationships between classes. The relationships between classes mainly include inheritance, association, and aggregation. The inheritance relationship is the relationship between a general class and a special class, and the parent class can describe all common attributes and operations among all subclasses. Association is a connection relationship between classes, so that one association end knows the properties and methods of another association end, which can be divided into one-way and two-way associations. Aggregation is a special case of a unidirectional relationship, meaning that one end of the association owns the other end.

The constraints in the class diagram include two types: constraints in the class and constraints in the method. The former constraint is described as a member in the class, and the latter is described as the precondition and postcondition in the method.

The mapping of the graphical domain model to the telecommunications domain model based entirely on xKL text is a static mapping from graphic grammar to text grammar, mainly including: mapping of class elements in graphic grammar to classes in xKL; aggregation Mapped to the attribute in xKL; association is mapped to the reference in xKL; inheritance is mapped to the inherits relationship in xKL; other grammar filling elements (constraints and methods, etc.) are directly mapped to the xKL grammar.

The relationship between the classes can be instantiated as a relationship between xKL memory objects, that is, the original association relationship is instantiated as a reference relationship between xKL memory objects, and the original aggregation relationship is instantiated as Attribute relationships between xKL memory objects.

After forming the object model, a verification process should be carried out to test the correctness of the model at the domain level and express the domain model as an executable model.

The said verification is to verify whether the objects in the object model satisfy the due constraints, including syntax mapping, constraint checking and method execution.

The meta-model and object model established by using the above architecture are verifiable, which ensures that the correctness of the model can be verified at the PIM level when user requirements change. Model executable in the modeling process is mainly manifested in the following aspects:

1. Domain model verifiability and correctness

There are three main steps to establish a domain model: create a domain framework model based on graph grammar; use xKL grammar to fill the content of graph elements; finally, map the class diagram mixed with graph elements and xKL grammar to a domain model completely based on xKL grammar. First, the verifiability of the domain model can be fully transformed into the grammar verification of the xKL meta-language. Obviously, the grammar of the xKL meta-language is verifiable. One-to-one correspondence and correctness, since the xKL metalanguage is completely based on MOF, and MOF provides corresponding graph-grammar examples for metadata, and the graph elements used in the present invention also refer to MOF specification documents.

2. Dynamic Executability of the Model

The final representation of the model is an object graph, which logically describes the objects in the model and the relationships between them. Whether the model is dynamically executable is mainly manifested in: the constraints satisfied by the objects in the model can be dynamically verified; the information of the object itself can be dynamically modified as needed; different objects can cooperate with each other to simulate the specific workflow of the model. To realize the above actions in the model, these actions must be translated into executable instructions to run on the machine. It can be seen from the above that the object graph of the model is generated by xKL, which provides conditions for the feasibility of the model's executable mapping to xKL, and the actions performed by the corresponding model can be mapped to the action semantics of the xKL language. In summary, according to xKL is executable, it is not difficult to get the dynamic executable of the model.

In summary, the core of the whole modeling process is the executable meta-language xKL. It guides the meta-modeling process, describes and dynamically instantiates the domain model, verifies and executes the instantiated model, and makes the generated model executable.

Implementing this method to establish a model in the telecommunications domain and the software for implementing this method have the following advantages:

1. Has good scalability and adaptability. According to different situations in the communication field, a new domain model can be regenerated by modifying the model and the relationship between the models to form a new domain application.

2. Has good reusability. A complete domain model can be a combination of multiple small domain models, which is low-coupling. Each small domain model has independent and complete logic, which can be reused to different degrees and different granularities, and can also be reused to a certain extent. Modifications migrated to similar domains.

3. Executability of the object model. The created object model is instantiated in the memory, and the execution of the method can be simulated, and the simulation of the program function can be realized through the call between the methods.

4. The domain model can establish a good mapping relationship between system requirements and underlying implementation, which not only meets the needs of requirements analysts and developers, but also improves the traceability between models and the reproducibility of modeling results usability.

5. It can shorten the time for products to be put on the market, improve and improve product quality, and increase customer satisfaction.

6. Compared with the background technology, the present invention designs a kind of executable meta-model language suitable for the field of telecommunication---xKL language; Adopt the two-step modeling method of meta-modeling and modeling; Through the verified process, prove The implementability of modeling methods in the field of telecommunications.

Description of drawings

Fig. 1 is a diagram of the formation process of the executable metalanguage xKL of the present invention.

Fig. 2 is a system architecture diagram of the xKL language-based modeling tool and verification tool of the present invention.

FIG. 3 is a structural diagram of a specific network domain meta-model of the present invention.

Fig. 4 is a specific network domain object diagram of the present invention.

Detailed ways

Example 1 Establishing a meta-model

First, domain experts and modelers abstract domain knowledge and establish domain metamodels through domain tools. Including the establishment of domain concepts and domain relationships and domain constraints, as well as icons presented by model elements, views of the internal structure of complex model elements, and so on.

Figure 3 is a specific network domain meta-model.

The meta-model shows that a network topology diagram includes one or more meta-model elements: router (Router), switch, host (Host). In addition, the relationship between each element also needs to be included as a meta-modeling element. Including association relationship and aggregation relationship. A router and a switch may contain one or more ports (Port), which are represented by an aggregation relationship, and links between a router and a switch, and a switch and a server are represented by an association relationship. At the same time, the elements in the meta-model need to be constrained. If the router needs to be the core device of the entire network, a router with dual CPU engines is required, and the router is required to be equipped with a fast Ethernet port. So the constraints on the router can be written as

Router.CPUNum==2

Router.fastEthNum==1

Similarly, the constraint on the switch is that there are two Fast Ethernet ports,

Switch.fastEthNum==2

Embodiment 2 builds a model

When modeling, corresponding model objects can be established according to various types of elements and their relationships established by the domain meta-model.

As shown in Figure 4, a network model is established, that is, the network core uses 2 CoreBuilder9000s, which are interconnected through 2 Gigabit Ethernet links, which improves reliability and adds 2 bandwidth between devices. Each CoreBuilder 3500 is connected to different CoreBuilder 9000 through 2 optical fiber fast Ethernet links, and all servers are inserted with 2 network cards, which are respectively connected to different core switches, thus forming a network backbone with high reliability and high bandwidth . When you want to establish a link between routers, switches and routers, the constraint validator of the domain model tool will perform verification to ensure that this modeling behavior that does not conform to the domain meta-model will not occur. This ensures that the established model follows the domain knowledge.

Embodiment 3 verifies the correctness of the model

(1) Syntax mapping

Carry out grammatical mapping on the network domain model diagram, and map it into a logic diagram represented by xKL. This part of the function is realized by a converter from graphic elements to text elements. The converted code is as follows:

    abstract class Device{

        //Represents the aggregation relationship between Device and Port

        attribute ports: Port[1..*]

      attribute deviceType: String

}

class Router extends Device{

attribute cpuNum:Integer

property fastEthNum:Integer

Getter is do

` ` ` ` ` ` ` ` this.ports.each{e|

If(e instanceof FastEthernet)

result:=result+1

}

end

//Association from Router to Switch, and specify your participating role name as

routers

reference switches:Switch[1..*]#routers

//Constraints satisfied by the number of CPUs in Router

constraint cpuNumConstraint is do

this.cpuNum==2

end

//Constraints satisfied by the number of Fast Ethernet in Router

constraint fastEthNumConstraint is do

this.fastEthNum==1

end

//Set the number of CPUs in Router

operation setCPUNum(num:Integer):void

// Precondition: The set quantity should satisfy the non-negative invisible constraint

Pre notPositiveInput is do

num＞0

end

is

do

`` this.cpuNum=num

end

// Get the number of CPUs in the Router

operation getCPUNum():Integer

//Post-condition: The obtained quantity should satisfy the non-negative invisible constraint

Post notPositiveOutput is do

num＞0

end

is

do

      result = num

end

.........

}

class Switch extends Device{

  attribute fastEthNum:Integer

//Switch-to-Router association, and specify your participating role name

switch

reference routers:Router[1..*]#switches

  constraint fastEthNumConstraint is do

** this.fastEthNumConstraint==1

end

.........

}

class Sever extends Device{

  attribute fastEthNum:Integer

.........

}

class Port{

.........

}

Next, the xKL syntax checker automatically checks the generated xKL code, thereby replacing the syntax check on the graphical class diagram. The program automatically generates the result of syntax checking and displays it to the meta-modeller in the form of a text report. The content of the report includes the number of created classes, the number of relationships, the number of constraints, etc., and their list. If an error occurs in the grammar check, suggestions for modification will be given in the report, and the error will be automatically located on the corresponding graphic element. For example, in the Router::setCPUNum() method body, "num>0" is changed to "nun>0", and the system prompts "undefined variable "nun"!" and displays the graph element Sever in red.

(2) Constraint check

When building a model, create corresponding objects according to the classes in the domain model, for example, create a HWRouter object coreRouterL, create a CiscoRouter object coreRouterR, and so on. After the model is created, the constraints satisfied by coreRouterL and coreRouterR can be checked in real time, and the action of verifying the constraints is translated into the following code:

//Set the number of CPUs for coreRouterL

      coreRouterL.setCPUNum(2)

...

Do

// Obtain the number of Fast Ethernet ports of coreRouterL

coreRouterL.fastEthNum.getter

// Check all constraint members of coreRouterL

coreRouterL. checkConstraints

// Capture the exception information generated by verification

Rescue(err:ConstraintException)

// Handle exceptions generated by verification

stdio.writeln(err.toString)

stdio.write(err.message)

end

Similarly, constraints satisfied by coreRouterR and other devices can be handled in the same way. According to the captured abnormal information, the model diagram is corrected accordingly, and even the domain model diagram is modified to achieve the goal. If it is necessary to upgrade the network hardware facilities, for example, the number of Fast Ethernet ports of coreRouterR is required to reach 3, the model diagram of this field needs to be revised so that the number of Fast Ethernet ports of Router meets the quantity constraint 3; and then coreRouterR in the model is corrected The number of Ethernet ports actually connected; then verify the constraints of the model.

(3) Method execution

The model diagram shows a memory object diagram, and the method of the corresponding object in the model can be called directly through the elements on the model diagram. For example, coreRouterR has the method getCPUNum(), and its call can be activated through the mouse event.

Claims (7)

1. A method for modeling the telecommunications domain based on an executable metalanguage, the method comprising the following steps: a) Extending the Unified Modeling Language based on the Meta Object Facility (MOF) and designing an executable metamodel suitable for the telecommunications domain Language——xKL language, and designed and built modeling tools and verification tools based on xKL language; b) Collect relevant information in the telecommunications field, integrate them into a domain model and describe it as a class diagram structure with visual components; use xKL to fill the constraints and static semantic parts in the class diagram to generate a complete graphical domain model; c) Mapping the graphical domain model to the telecommunications domain model completely based on xKL text; d) Starting from the telecom domain model, according to the problems to be solved, select the classes in the domain model for instantiation, generate memory dynamic operable objects based on xKL, and instantiate the relationship between classes in the domain model as xKL memory objects The relationship between them forms an object model; e) Execute the xKL-based constraints in the domain model, verify whether the objects in the object model meet the due constraints, if not, dynamically modify the model according to the corresponding constraints, and return the modified results; f) Send an instruction to a specific object in the model to make it execute the corresponding method, and check the effect of the method execution. 2. According to the method for modeling the telecommunications field based on executable meta-language according to claim 1, it is characterized in that the xKL language is composed of a static syntax meta-model and a dynamic behavior meta-model; the static syntax meta-model is an extension of the EMOF The meta-model defines the structure required to represent the model; the dynamic behavior meta-model is a peripheral extension of OCL to form the xOCL language, and xOCL defines the operation of the model. 3. The method for modeling the telecommunications domain based on executable meta-language according to claim 1, wherein the domain model includes concepts abstracted from the telecommunications domain, that is, classes and relationships between classes; The relationship between classes mainly includes inheritance, association, and aggregation; among them, the inheritance relationship is the relationship between a general class and a special class, and the parent class can describe all common attributes and operations between all subclasses; the association relationship is the connection between classes Relationship, which enables one association end to know the properties and methods of another association end, which is divided into one-way and two-way associations; aggregation relationship means that one association end owns another association end. 4. According to the executable meta-language-based modeling method in the telecommunications field according to claim 1, it is characterized in that the constraints in the class diagram include constraints in the class and constraints in the method, and the constraints in the class describe For members of a class, constraints in a method are described as preconditions and postconditions in the method. 5. The telecommunications domain modeling method based on executable meta-language according to claim 1, characterized in that the mapping of the graphical domain model to the telecommunications domain model based entirely on xKL text is a graphic method Static mapping to text grammar mainly includes: class elements in graph grammar are mapped to class in xKL; aggregation is mapped to attribute in xKL; association is mapped to reference in xKL; inheritance is mapped to inherits relationship in xKL; constraints and methods directly primitively map into the xKL grammar. 6. According to the executable meta-language-based modeling method in the telecommunications field according to claim 1, it is characterized in that the relationship between the classes is instantiated as the relationship between xKL memory objects, that is, the original Some association relationships are instantiated as reference relationships between xKL memory objects, and the original aggregation relationships are instantiated as attribute relationships between xKL memory objects. 7. The executable meta-language-based modeling method in the telecommunications field according to claim 1, wherein the verification is to verify whether the objects in the object model satisfy the due constraints, that is, syntax mapping, constraint Inspection and method execution.

Images