Ivan Velikic

A generic model-based test case generator

Statistical usage testing, also referred to as the statistical testing or behavioral testing, is today recognized as a de facto industry standard for quality assessment of large-scale embedded systems. Statistical testing is based on operational profile, essentially Markov process of a system under test. The measure of the quality is reliability of a system, i.e. mean-time-to-failure. The key part of statistical testing is automatic production of test cases in accordance to a given operational profile of the tested system. This problem has been widely addressed in literature, however there is still the need for additional research efforts in this area. In this paper we present the original model-based working environment for production of statistical test cases. The front-end of the environment is the generic modeling environment with our operational profile modeling paradigm registered to it. The back-end of the environment consists of operational profile model interpreter and a generic test case generator. The generic test case generator is the key component of the environment. It consists of previously developed test case generator engine and a newly developed generic emulator. The generic emulator is capable of emulating any system under test. The complete back-end of the environment has been written in Java programming environment.

A formal software verification concept based on automated theorem proving and reverse engineering

Formal software verification typically involves some levels of static theorem proving which is a mathematical process of proving that the function computed by a program match the function specified. A theorem prover, such as THEO, automates this process. On the other hand, reverse engineering is a process inverse to traditional engineering. An example extracts the software specification from its source code. In this paper we present a formal software verification concept, which is based on automated theorem proving and reverse engineering. We are mainly concerned with communications software and with software for families of communication protocols in particular. In the paper we describe how to: (1) model a group of finite state machines using predicate calculus; (2) extract axiomatic software specification from the source code and log files; and (3) formally verify software for a given operational profile (set of test cases). The concept has been successfully applied to a call processing software for systems, which are installed and fully operational in Moscow and Saint Petersburg, Russia.

Use of Finite State Machine Based Framework in Implementation of Communication Protocols A Case Study

In this paper we present a finite state machine (FSM) based implementation of SIP protocol stack for SIP user agent. Design decisions made during the development process, and the resulting characteristics of software product are discussed. Some characteristics of FSM-based framework that has been used in development are presented. We compare quantitatively our implementation with a well known open source SIP implementation, resiprocate. Comparison is focused on the following parameters: lines of code (LOC) and cyclomatic complexity. Results show viability of FSM-based approach, most importantly its support for modularization.

A concept of an integrated development environment for reactive systems

One of the main application areas of modern engineering in computer based systems is development, production, deployment and maintenance of reactive systems. Today, the majority of functions in such systems are implemented in software, and modeled as a group of finite state machines (automata). The past decade has seen many R&D efforts in development of tool sets that support software production processes, in area of reactive systems, in both industry and academia. This paper is our contribution to this movement. We present a concept of an integrated development environment for a class of problems known as finite state machines. The concept is suitable for small and medium size high-tech software R&D teams, where most of the team members participate in majority of phases of a software development cycle.

A model-based statistical usage testing of communication protocols

In this paper, we present our original approach to the model-based statistical usage testing of a class of communication protocol implementations that are based on the state design pattern and Java programming environment augmented with the class FSMSystem. The approach is based on the working environment that has been proven on a number of real-world projects. The working environment is created with the following set of tools: generic modeling environment with the operational profile model paradigm registered to it, operational profile model interpreter, generic test case generator, test case compiler, and the unit testing framework JUnit extended with the class TestBed that acts as both test driver and stub thus providing the complete test harness. In the paper, we present the methodology of the model-based statistical usage testing of a class of communication protocol implementations, the tools that support this methodology, and the case study - the model based statistical usage testing of SIP INVITE client transaction, a part of the SIP protocol stack

Retrieving the useful information from the binary files compiled by C compiler

In order to reuse a software component or verify it, the information about compiler, that is used for compilation of the component, is very significant. Due to unknown build flags and compiler information, third party components usually cannot be reused. When it comes to the matter of software components developed in C language, the only place to look for this information is in compiled binary file. Since there are no standards for C language that bind compiler to leave any information about itself in binary file, the information is not expected to be found. However, the information can be found in most of the binary files. This paper investigates what is the possibility of retrieving the information about compiler name and version based on the content of the binary code, without decompiling the binary code. The information retrieving opens up new possibilities for reusing the components and verifying the software requirements as well.

A Design and Implementation of ASN.1 Parser for H.323 Protocol Stack

This paper presents an object oriented implementation of library for generation and parsing of messages in H.323 protocol stack. Design decisions and rationale behind them are discussed. UML diagrams of important aspects of software are provided. The library is implemented in C++ programming language. Several metrics of the software are presented, and discussed. Implementation included a significant effort in the testing phase. Automated testing using CPPUnit was conducted, and an overview of activities in this phase is given too. The library has performed well in all of the tests, and is ready for use in development of H.323 based applications.

Calculation of number of integration cycles for systems synchronized using the AS6802 standard

Correctly maintained synchronization between devices is very important in deterministic networks like TTEthernet where is used AS6802 standard for this purposes. In this paper is depicted how to calculate number of integration cycles per cluster period depending on accuracy of clocks implemented in network devices. Influence of the number of integration cycles to bandwidth utilization is analyzed too. Correctly calculated number of integration cycles provides stable synchronization in the network together with optimal bandwidth utilization.

Dataflow integration in mixedcritical networks

In mixed-critical networks is very important to provide appropriate mechanisms which prevent lost of data during communication, especially if that are high-priority frames. As a basis for analysis, in this paper we observed principles of multy-priority based traffic integration in TTEthernet networks that can be configured as strictly deterministic, or mixed-critical. It is analyzed influence of the traffic integration to delay, bandwidth utilization and scheduling in TTEthernet network.

Bandwidth utilization in deterministic networks

In this paper we analyzed how a frame size type, period and number of virtual links affect bandwidth utilization in deterministic networks. For services which combine avionics applications and consumer electronics, such as In-flight entertainment, this analysis is very important. We analyzed one topology of TTEthernet deterministic network, and influence of the TT, RC, and synchronization traffic classes to total bandwidth utilization. Bandwidth that is not used for this traffic classes, can be exploited for consumer electronics applications traffic.