Test development for communication protocols: towards automation

1993

Dans Ie cadre de mes etudes~l'Ecole Polytechnique d'Eindhoven il est necessaire d'effectuer un stage de fin d'etudes concretisant rna formation d'ingenieur. J'ai effectue Ie premier mois (decembre 1992)~Philips Research Laboratories en Eindhoven, et les huit autres mois Ganvier 1993 jusqu'

Protocol Test Systems, 1995

As the OSI protocols come to be widely adopted in various communication systems, the testing of OSI protocol implementations becomes important. In this paper, we propose the OSI protocol testing system with user friendly test scenario specification functions aiming to support both the state transition oriented testing and the parameter oriented testing for the Transport through Application layer protocols. Our system introduces the separate tools with friendly human interface for specifying the test scenarios, for the test execution and for analyzing the test results. This paper describes the function and implementation of our system and shows its usefulness compared with the conventional testing system based on formal specification language.

1993

An approach for conformance testing of protocols specified as a collection of communicating finite state machines (FSMs) with two parts, pruning and a guided random walk procedure, is presented. First the protocol is pruned to various sets of machines; each set provides only one service. This significantly reduces the test sequence length. Then a guided random walk procedure that attempts to cover all transitions in the component FSMs is used. The results of applying the procedure to the full-duplex alternating bit protocol and the asynchronous transfer mode (ATM) adaptation layer convergence protocol are presented. For the ATM adaptation layer, 99% of component FSMs' edges can be covered in a test with 11692 input steps. Previous approaches cannot generate conformance tests for standard protocols (such as ATM adaptation layer) specified as a collection of communicating FSMs

AT&T Technical Journal, 1990

In this paper, we present a generic approach for modeling a communications protocol with state transitions and Sherlf is a member of ~dvanced Digital Signal window and timer mechanisms, and for generating con-Uyar is a member of formance test sequences automatically. Based on this model, minimum-cost (i.e., minimum run time) confor-Processing Department, and M. h i t technical staff in the Standards Deoaflment. method based On Unique input/ou@ut sequences and mance test sequences can be generated by using a Network lnterfaces and hey are with AT&T ~r. Sherif works on the Rural Chinese Postman tours. As a case study, this method is applied to the Integrated Services Digital Network link-access protocol on the D channel. Laboratories in Holmdel, New Jersey. systems engineering and standardization of wideband packet protocols. He joined AT&T Bell Laboratories in 1983 and has both a B. Sc. in electronics and communications and an M.

Computer Communications, 2003

Reliable protocols require early-stage validation and testing. Due to the state explosion problem in validation methods such as model checking [IEEE Trans. Software Engng 19 (1993) 24], sometimes it is not possible to test all the system states. We apply our state-of-the-art algorithm in computing the most critical states and branches to be tested. We prioritize this information to guide the validation of the protocol. We implemented this technology in a tool that visualizes the specifications of protocols with their testing priorities. Such a tool can also be used to identify faulted place in the protocol when some tests failed. It provides information such as where in the protocol is most likely to have bugs. Our tool provides many benefits, including (1) early detection and recovery of protocol faults, (2) visualization and simulation of the protocol specifications, (3) quantification of the reliability confidence of protocols, (4) making code generation directly from protocol specifications more possible, and (5) reduction of the number of introduced faults. This paper considers the case when the specification of the protocol is given in Specification and Description Language (International Telecommunication Union standard). Our technology is based on both the control flow and the data flow of the specifications. It first generates a control flow diagram from the specification and then automatically analyses the coverage features of the diagram. It collects the corresponding flow data during the simulation time to be mapped to the control flow diagram. The coverage information for the original specification is then obtained from the coverage information of the flow diagram.

Proceedings of the IEEE, 1990

The wide range and high complexity of services expected from a communication protocol have made the process ofprotocol conformance testing a challenging task. Formal methods are required to thoroughly test these complex protocols. This paper reviews the four major methods of conformance test generation reported in the literature: transition tours, distinguishing sequences, characterizing sequences, and unique input/output sequences. These methods are used to test the control portion of a protocol specification. The conformance testing concepts developed in the standards world are also summarized. Their relationship with the four formal methods is discussed.

Proceedings of the 9th International Conference on Evaluation of Novel Approaches to Software Engineering, 2014

IEEE/ACM Transactions on Networking, 1994

A theoretical analysis of fault coverage of conformance test sequences for communication protocols specified as finite state machines (FSM's) is presented. Faults of different types are considered and their effect on testing analyzed. The interaction between faults of different categories and the impact it has on conformance testing is investigated. Fault coverage is defined for testing of incompletely specified machines and also for testing of completely specified machines. An algorithm is presented to generate test sequences with maximal fault coverage for testing of incompletely specified machines. It is then augmented for testing of completely specified machines and finally a technique for generating test sequences which provide guaranteed maximal fault coverage for conformance testing of communication protocols is presented.

ACM SIGCOMM Computer Communication Review, 1988

This paper presents results on the application of four protocol test sequence generation techniques (T-, U-, D-, and W-methods) to the NBS Class 4 transport protocol (TP4). The ability of a test sequence to decide whether a protocol implementation conforms to its specification depend on the range of faults that it can capture. The study shows that a test sequence produced by the T-method has a poor fault detection capability whereas test sequences produced by the U-, D- and W-methods have comparable (superior to that for T-method) fault coverage on several classes of randomly generated machines. The lengths of test sequences produced by the four methods tend to be different. The length of a test sequence produced by the T-method (W-method) is the smallest (largest). The length of a test sequence from the U-method is smaller than that for the D-method and lengths for both are greater than that for the T-method and less than that for the W-method.

IEEE Transactions on Communications, 1994

Protocols are large and complex software systems. Complete conformance testing of an implementation against its standard may not be feasible in terms of the resources available. This paper discusses a new approach, the P-method, to the testing of meaningful subsets of communication protocols for an asynchronous model of communication. The approach is based on the probabilistic verification of protocols, which is carried out on the more probable part of the protocol first. The technique can be used for generating probabilistic test sequences for the conformance testing of communication protocols to standards. The proposed method yields meaningful protocol test sequences which test the most probable behaviors of a protocol when the testing of the complete protocol is not feasible. Probabilistic test sequences can be categorized into different classes. The higher the class a probabilistic test sequence is in, the larger the extent of the protocol it covers, and the better is the fault coverage. If the class of a test sequence is high enough, its fault coverage is comparable to the fault coverage of test sequences generated by other methods. Results from a study of the P-method, using alternating bit protocol (ABP) and a subset of NBS TP4 as examples, support the claims above. It can also be shown that if errors are introduced only to the more probable part of the protocol, the fault coverage of P-method is also comparable to other methods.