CN111274142A

CN111274142A - Software communication system architecture conformance test modeling method based on extended finite-state machine

Info

Publication number: CN111274142A
Application number: CN202010063310.3A
Authority: CN
Inventors: 唐麒; 伍旭东; 魏急波; 连进; 赵海涛
Original assignee: National University of Defense Technology
Current assignee: National University of Defense Technology
Priority date: 2020-01-20
Filing date: 2020-01-20
Publication date: 2020-06-12
Anticipated expiration: 2040-01-20
Also published as: CN111274142B

Abstract

The invention provides a software communication system architecture conformance test modeling method based on an extended finite-state machine, which solves the problem that a complete test sequence is difficult to construct only by analyzing the dependency relationship among tests in the prior art. The model of the invention can visually display the operation process of the tested component, analyze the dependency relationship among the required tests, guide the construction of the test sequence and provide effective support for the test automation. The method can visually display the operation flow of the software radio system in the test process, improves the readability of the test program, makes up for the domestic vacancy of testing and researching the conformity of the software communication system architecture, and lays a good foundation for the subsequent research of test automation.

Description

Software communication system architecture conformance test modeling method based on extended finite-state machine

Technical Field

The invention relates to a modeling method based on Software Defined Radio (SDR) system conformance test, in particular to a Software Communication Architecture (SCA) conformance test modeling method based on an extended finite state machine.

Background

SDR technology is a radio communication technology which is very popular at present, and has been emerging from the last 90 th century, and it has been made by organically combining software, hardware and radio technology to form a system with strong flexibility and openness. The SDR system relies on a standardized hardware platform, controls bottom hardware by using upper software, thereby realizing various waveforms and releasing radio technology from hardware-based radio station technology.

In order to establish a standardized software framework which is independent of specific hardware and mainly applied to the development and implementation of waveforms, the SCA specification, which is a standard specification set issued by the Joint Tactical Network Center (JTNC) on the software radio architecture, is mandatory for the adoption of the software radio of the United states army. The specification is a design specification at the top of the system, which relates to the software, hardware architecture, and interfaces of upper applications of the SDR system. The SCA architecture is divided into 6 layers: the system comprises a bus driver and board level hardware driver layer, a network and serial interface service layer, a POSIX operating system interface layer, a CORBA middleware layer, a Core Framework (CF) layer and an application layer.

The CF is an important component of the SCA, which contains a set of application programming interfaces, components, and configuration files. The configuration file is mainly written by using extensible description language (XML) and is mainly used for describing the structure and the function of the component; the interface is defined by an Interface Definition Language (IDL); the component is a carrier of the interface and the configuration files, is associated with one or more configuration files, and has corresponding functions by realizing the corresponding interface. All components need to implement the basic interfaces defined by the CF so that the CF can access and control in a unified manner.

The SCA specification is a top-level framework design specification that only specifies the composition of the core framework and its functions to be performed, and does not set any requirements on specific implementation details. The software radio system runs on various hardware platforms, and there may be many differences in software products developed by various units, so the software radio system must be subjected to a compliance test to verify whether it meets the SCA standard. The method aims to ensure the universality and cross-platform performance of waveform software.

The conformance test of the SCA core framework plays a crucial role in guaranteeing the quality of a software radio system, the research on the SCA core framework test automation is relatively lacked at home and abroad at present, and a plurality of complex conditions exist in the actual test process, and the main problems faced at the present stage are as follows:

(1) however, since some tests may affect or even fail to perform subsequent tests, it is difficult to construct a complete test sequence by analyzing the dependency between tests.

(2) The traditional test method is to manually appoint a test sequence, the method has low test efficiency, is easy to cause omission, cannot well ensure the test coverage, and can cause test result errors due to mutual influence among tests.

(3) In order to ensure that an SDR system can operate properly, some requirements require multiple tests under a variety of different conditions. How to implement automatic construction of a test sequence for loop testing of some functions or the whole system is a current challenge.

(4) The lack of models can intuitively present the whole test process, and when a system has a plurality of errors in the test process, the reason of the errors is difficult to analyze and correct.

Disclosure of Invention

Based on the problems faced by the SCA core frame conformance test at the present stage, the invention provides a method for constructing a test model based on the SCA core frame conformance test by combining the principle of Extended Finite State Machine (EFSM).

The EFSM model is a common test model in software testing, and is widely applied to modeling of testing of communication protocols, software embedded systems and the like. EFSM consists of a six-membered group, denoted M = < S, S0, V, O, I, T >. Wherein S represents a state set of the EFSM model; s0 denotes the initial state of EFSM; v is a variable set in the system; i and O represent the input and output set of the system; t is the set of transitions between states in the system.

In the migration set T, each migration T may be composed of a five-tuple, T = < source (T), target (T), condition (T), event (T), action (T) > represents, source (T) and target (T) represent the source state and target state of the migration T, respectively; condition (t) is a precondition triggered by migration of t, and when the precondition is not satisfied, migration cannot be triggered, so the precondition is often called guard (guard) in the literature; event (t), action (t) represent the t-triggered event and the executed operation, respectively. event (t), condition (t) and action (t) together form a label of t, and the expression is event [ guard ]/action. It shows that in the case of the system in the state source (t), event (t) is triggered and the current condition satisfies condition (t), the system will transfer from the state source (t) to the state target (t) and execute action (t).

The method for constructing the EFSM model based on the SCA conformance test comprises the following steps:

step one, determining an initial state S0 of a test, and constructing a state set S and a transition set T. In the SCA conformance test, each required test can be abstracted into a small EFSM model, the state existing in the required test is an initial state si, an execution action state sj and a test passing state (a test end state) sz after the required test passes; there is a migration ta where si points to sj, a migration tb where sj points to sz, and a migration tc where sj points to si. The process of testing can be described as the system first starts from si, reaches the action execution state sj via ta, the system executes the test at sj, if the test condition is satisfied and the test passes, the state from sj reaches the end state via tb, and executes the corresponding operation, if the system fails or the test does not satisfy the condition, the system returns to si via tc and throws the exception. Analyzing the test process of each requirement, and adding the state and the transition existing in the test process of each requirement into S and T respectively;

and step two, constructing a variable set V. The variable set V represents a set of data that needs to be manipulated during the test, usually as a decision condition for state transition. In the SCA conformance test, there are variables such as the number of connections, the number of loads, boolean type input values, and the like. Adding the variables existing in the test process of each requirement into a variable set V;

analyzing the dependency relationship among the demand tests according to the SCA specification, and constructing a test dependency tree diagram, wherein the demand test of each node corresponds to an EFSM model;

fourthly, constructing an EFSM model for SCA conformance test according to the test dependence tree model and the defined state set S, the migration set T and the variable set V:

(1) each requirement test corresponds to different EFSM models, the same state exists in the models, the models are connected by taking the same state as a node, and the repeated state existing in S is removed;

(2) specifying an initial state s0 to point to a migration of the execution state of the root node in the tree graph;

(3) in the dependency tree model, testing of a child node needs to be done with the testing of its parent node passed. Thus, for each node's requirement test, a transition is constructed in which the test end state points to the state of the execution action of the requirement test for its child node.

(4) For each node's requirement test, a migration is constructed that points to the execution action state of the requirement test for its siblings from the test end state.

According to the steps, an EFSM model based on the SCA core framework conformance test can be constructed, in the model, the test of specific requirements is included in the migration of the model, the precondition of the test is judged on the migration, and a feasible migration sequence set (path) can be used for guiding the construction of test cases;

with the increase of the test requirement, the number of nodes and the number of transitions in the model will increase, and the scale of the model will increase continuously, which is a research and exploration of the automation technology. Therefore, the model needs to be reduced to a certain degree, so that the model is more suitable for the requirement of the conformance test of the SCA core frame and the test automation is favorably realized;

step five, the constructed model can be further reduced according to the following rules:

rule one is as follows: for any required test, constructing a transition from an initial state to a final state, merging a trigger event, an execution condition and an action on the transition from the original initial state to the execution action state and the transition from the execution action state to the final state, and placing the merged trigger event, execution condition and action on a new transition. For the transition of other state pointing to the action execution state, constructing the transition of the state pointing to the corresponding last state, and the operation is similar;

rule two: for any required test, adding self-circulation migration to the initial state, merging the migration of the original initial state to the action execution state and the triggering event, the execution condition and the action of the migration of the action execution state to the initial state, and placing the merged events, the execution conditions and the actions on the new migration. For the transition of other states pointing to the action execution state, constructing the self-circulation transition on the state, and the operation is similar;

rule three: and deleting all transitions pointing to the action execution state and the end state pointed by the action execution state, and then deleting all states without transitions. The result is to remove all states in the model that describe the action and their migration.

According to the steps and the reduction rules, a global EFSM model for testing the conformance of the SCA core framework can be finally constructed, and the scale and the complexity of the model are greatly reduced compared with those of the original model. The specific demand test in this model will be represented by a single migration, and the system failure is also represented by a self-migration on the state. On the basis, a specific demand set can be input, the model is extracted to obtain the submodel, and the steps are as follows:

(1) inputting a demand set;

(2) selecting demand nodes in the set from the test dependency tree model generated in the step five;

(3) for each selected demand node, sequentially selecting the upper-level node until the root node, extracting all the selected nodes, and constructing a new demand set;

(4) extracting the migration represented by each demand test in the demand set, and the initial state and the final state of the migration in sequence from the original model to obtain a sub-model;

(5) adding migration between the end states of the demand tests belonging to the brother nodes in the test dependency tree model in the sub-model, wherein each label on the migration is the same as the migration of the demand test of the brother nodes;

(6) if the initial state of the submodel is not the initial state s0 of the original model, adding s0 the transition and state between the initial states of the submodel to the submodel, thus obtaining a complete submodel.

The invention has the following gains for the conformance test of the SCA core framework:

(1) the invention provides a modeling method of an EFSM model based on SCA core frame conformance test, and the model can visually display the whole test flow;

(2) based on the model, the test conditions can be judged, the feasibility of a test path is judged, a feasible test sequence is constructed, and the construction of a test case is guided;

(3) the construction requirements of complex test cases can be met;

(4) the proposal of the personalized submodel is provided.

Drawings

FIG. 1 is a diagram of a state transition model for single demand testing,

figure 2 is a diagram of a test dependency tree model,

FIG. 3 is a diagram of an EFSM model based on SCA core framework conformance testing,

FIG. 4 is a diagram of a reduction process of a single demand test,

FIG. 5 is a diagram of a reduced EFSM model based on SCA core framework conformance testing,

FIG. 6 is a diagram of a sub-model,

FIG. 7 is a flow chart of test model generation.

Detailed Description

The invention will be further described with reference to the accompanying drawings:

FIG. 1 is a state transition diagram of a single requirement test according to the present invention, which illustrates the test process and the output result of any requirement in the SCA compliance test. Before testing, the system is in an initial state of testing, such as "initialized", "started", "connected", etc.; in the test execution process, the system executes corresponding actions and is in an execution state; if the test passes, the state of the system may reach the next state due to the performed operation, or may return to the initial state. Thus, an initial state s1 of the initial test, a test execution action state s2, a final state s3 of the test, a final state of the model may be the same as the initial state, and transitions t1, t2, and t3, where a tag event [ condition ]/action exists at any transition t, which represents a trigger event, an execution condition, and an execution operation, respectively, and the tag may be empty. At the beginning of the test, the system under test enters the test execution state from the initial state via t1, if the system can operate normally and the condition "is satisfied, the" success "event is triggered, and the corresponding" action "operation is executed, and the system reaches the final state s3 via t 2. If the system has an error in the test process, or the condition "is not satisfied, the system cannot successfully reach the end state, then a" failure "event is triggered, and via t3, the system returns to the initial state s1, and the cause of the abnormality is thrown out.

Fig. 2 is a test dependency tree diagram of an SCA compliance test, and in the SCA compliance test process, some requirement tests have dependency relationships therebetween, as shown in fig. 2, this example is formed by partial requirement tests of basic components in a core framework, any node in the model represents a requirement test, and the requirement test of any child node can be performed only when the requirement test of its parent node passes. The test sequence is constructed according to the model, so that unnecessary tests can be reduced to a certain extent, and the test time is reduced.

FIG. 3 is an EFSM model diagram based on SCA core framework conformance testing, as shown in FIG. 3, this example represents a state transition diagram of partial requirement testing of basic components in the core framework, which is formed by combining a plurality of state transition models of single requirement testing according to a test-dependent tree model. Firstly, merging the repeated states existing in the demand test model, and secondly, constructing all parent nodes according to the test dependency tree model to point to the child nodes to execute migration between action states, such as t12, t13 and the like in the graph. In the test dependence tree model, the requirement tests of the siblings can be executed only under the condition that the requirement tests of the parents of the siblings pass, so that the requirement tests of the siblings are mutually reachable, and for the requirement test of each node, the migration of the execution action state of the requirement test of which the test tail state points to the siblings of the node is constructed.

A test path represents the set of transitions from s0, through multiple transitions, to eventually reach the "Released" state or to re-reach the "Initialized" state through the "Released" state. In the SCA conformance test, a test for one component is generally composed of a plurality of paths. In the graph, the dependency relationship among the tests can be represented by judging the migration condition, so that the feasibility of the test path can be judged, and a feasible test path can be generated, thereby guiding the construction of the test case.

FIG. 4 is a simplified process diagram of a single demand test. As shown in the figure, firstly, a transition t4 between an initial state s1 and a final state s3 is constructed, and a trigger event, an execution condition and an action on t1 and t2 are merged and placed on a new transition t 4; secondly, a self-circulation migration t5 of an initial state s1 is built, and the trigger event, the execution condition and the action on t1 and t3 are combined and placed on a new migration t 5. T1, t2, t3 are then deleted, and the action state s2 is executed, as indicated by the dashed line.

FIG. 5 is a diagram of an EFSM model based on SCA core framework conformance testing after reduction. The number of states and transitions is less in fig. five relative to fig. 4. In the figure, self-circulation error reporting migration generated by test errors is omitted, each migration represents a required test, and the construction of a test case can be better guided.

FIG. 6 is a diagram of a submodel of FIG. 5, which may be extracted to obtain a submodel based on the input requirement set. The extraction of the submodels aims at the test of different conditions, and provides guidance for personalized test.

FIG. 7 is a flow chart of test model generation, the process of test model generation:

(1) analyzing the starting of the test process of each requirement in the SCA specification, determining a state set, a variable set and a migration set of a single requirement test, and constructing a single requirement test state migration model by combining an EFSM principle;

(2) constructing a test dependence tree model according to the SCA specification;

(3) integrating the state transition models of the single requirement tests according to the test dependence model to obtain an initial EFSM model;

(4) reducing the model according to a certain reduction rule to obtain a final test model diagram;

(5) the sub-models can be extracted according to a certain rule by inputting a demand set according to the demand.

In summary, the invention provides the EFSM model based on the SCA core framework conformance test and the modeling method thereof, the model can visually display the operation flow of the software radio system in the test process, improve the readability of the test program, effectively judge the feasibility of the test path, play a supporting role in the construction of test cases with complex requirements, make up for the domestic vacancy of SCA conformance test research, and lay a good foundation for the subsequent research of test automation.

Claims

1. A software communication system architecture conformance test modeling method based on an extended finite-state machine is based on an SCA core framework conformance test to perform a modeling model, and is characterized by comprising the following steps:

step one, determining an initial state S0 of a test, constructing a state set S and a migration set T according to SCA specifications, abstracting each required test into an EFSM model in an SCA conformance test, wherein the state existing in the required test is an initial state si, executing an action state sj, and a test passing state after the required test passes, namely a test final state sz; there is a migration ta where si points to sj, a migration tb where sj points to sz, and a migration tc where sj points to si,

analyzing the test process of each requirement, and respectively adding the state and the transition existing in the test process of each requirement into a state set S and a transition set T;

step two, constructing a variable set V, and adding variables existing in the test process of each requirement into the variable set V;

constructing a test dependence tree diagram, wherein the requirement test of each node corresponds to an EFSM model;

fourthly, constructing an EFSM model for SCA conformance testing according to the testing dependence tree model and the defined state set S, the migration set T and the variable set V;

and step five, reducing the constructed model according to rules on the basis of the EFSM model given in the step four.

2. The extended finite state machine-based software communication architecture conformance test modeling method according to claim 1, wherein the fourth step specifically comprises the following steps:

4.1) each requirement test corresponds to different EFSM models, the models have the same state, the models are connected by taking the same state as a node, and the repeated state existing in S is removed;

4.2) specifying that the initial state s0 points to a migration of the execution state of the root node in the tree graph;

4.3) for the requirement test of each node, constructing the migration of the execution action state of the requirement test of which the test end state points to the child node;

4.4) for each node's requirement test, constructing a migration of the execution action state of the requirement test whose test end state points to its siblings.

3. The extended finite state machine-based software communication system architecture conformance test modeling method according to claim 1, wherein the rule in the fifth step is specifically:

rule one is as follows: for any required test, constructing a transition between an initial state and a final state, merging a transition of the original initial state pointing to an execution action state and a trigger event, an execution condition and an action of the transition of the execution action state pointing to the final state, placing the merged events on a new transition, constructing a transition of the state pointing to the corresponding final state for the transition of other state pointing to the execution action state, and similarly operating;

rule two: for any required test, adding self-circulation migration to the initial state, merging the migration of the original initial state pointing to the action execution state and the triggering event, the execution condition and the action of the migration of the action execution state pointing to the initial state, placing the merged events on new migration, and constructing the self-circulation migration of the state for the migration of other states pointing to the action execution state, wherein the operation is similar;

rule three: deleting all transitions pointing to the action-executing state and the end state pointed by the action-executing state, and then deleting all states without transitions, which results in deleting all states describing actions and transitions thereof in the model.