Alberto Salmerón

Using Model Checking to Generate Test Cases for Android Applications.

The behavior of mobile devices is highly non deterministic and barely predictable due to the interaction of the user with its applications. In consequence, analyzing the correctness of applications running on a smartphone involves dealing with the complexity of its environment. In this paper, we propose the use of model-based testing to describe the potential behaviors of users interacting with mobile applications. These behaviors are modeled by composing specially-designed state machines. These composed state machines can be exhaustively explored using a model checking tool to automatically generate all possible user interactions. Each generated trace model checker can be interpreted as a test case to drive a runtime analysis of actual applications. We have implemented a tool that follows the proposed methodology to analyze Android devices using the model checker Spin as the exhaustive generator of test cases.

Applying MDE Methodologies to Design Communication Protocols for Distributed Systems

Traditionally, protocol engineers have to deal with the design and implementation of complex network services, spending considerable time and effort on creating robust and reliable final source code. Although approaches exist to assist engineers in the development of communication protocols which use several object-oriented frameworks, these do not benefit from new modelling guidelines developed in MDE and UML to exploit automatic code generation from graphical models. This paper introduces a new UML2 profile for communications which guides the construction of software for communications following the client-server architecture. The MDE process is then used to design suitable platform-specific models for the well-known Adaptive Communications Environment (ACE), a high performance C++ toolkit for implementing concurrent and network applications which relies heavily on architectural patterns

Using SPIN for automated debugging of infinite executions of Java programs

This paper presents an approach for the automated debugging of reactive and concurrent Java programs, combining model checking and runtime monitoring. Runtime monitoring is used to transform the Java execution traces into the input for the model checker, the purpose of which is twofold. First, it checks these execution traces against properties written in linear temporal logic (LTL), which represent desirable or undesirable behaviors. Second, it produces several execution traces for a single Java program by generating test inputs and exploring different schedulings in multithreaded programs. As state explosion is the main drawback to model checking, we propose two abstraction approaches to reduce the memory requirements when storing Java states. We also present the formal framework to clarify which kinds of LTL safety and liveness formulas can be correctly analysed with each abstraction for both finite and infinite program executions. A major advantage of our approach comes from the model checker, which stores the trace of each failed execution, allowing the programmer to replay these executions to locate the bugs. Our current implementation, the tool TJT, uses Spin as the model checker and the Java Debug Interface (JDI) for runtime monitoring. TJT is presented as an Eclipse plug-in and it has been successfully applied to debug complex public Java programs.

Obtaining Models for Realistic Mobile Network Simulations using Real Traces

Field tests are vital for characterizing mobile networks and applications under real conditions, but may be too costly. Simulations are an alternative for performing this kind of analysis, but may yield results that are too synthetic. In this paper we present a methodology for obtaining more sophisticated and realistic simulations, based on the use of traffic traces from field tests on live networks and an objective-driven simulation approach for parameter tuning. We use this approach to automatically refine parameterized mathematical jitter models, which can be used in network simulations to replicate communication issues presented in wireless networks.

Combining SPIN with ns-2 for protocol Optimization

This paper presents an approach to integrate the analysis capabilities of the Spin model checker and the ns-2 network simulator into a single framework. The traffic-oriented model of the protocols is managed by ns-2, while Spin automatically generates the most suitable configurations of each ns-2 run in order to meet some designer requirements. These requirements are specified with assertions and with an annotated temporal logic that can be translated into Spins Buchi automata. Spin verification algorithms help us to automatically discard those ns-2 configurations that do not satisfy the expected requirements. With this approach we can automatically obtain the suitable values of parameters like buffer size, timeout to retransmit and window size, to optimize the performance of a protocol implementation in a given scenario. The paper presents the architecture for this integration, the modified temporal logic and its successful application to obtain optimized versions of protocols for videostreaming in wireless networks.

River Basin Management with Spin

This paper presents the use of the Spin model checker as the core engine to build Decision Support Systems (DSSs) to control complex river basins during flood situations. Current DSSs in this domain are mostly based on simulators to predict the rainfall and the water flow along the river basin.

Runtime Verification of Expected Energy Consumption in Smartphones

Smartphones connected to Internet should work properly for days without a reset. One of the most critical non-functional properties to ensure the correct behavior is energy consumption. However, currently there is a lack of automated techniques to check whether the actual mobile consume is within the expected limits. To apply runtime verification techniques in this context, we need a detailed profiles of consumptions for specific actions in apps of interest such as activate GPS, send a data packet to the network, etc.; b a method to automatically generate sufficiently representative use cases of the mobile behavior; c a language to describe the expected behavior in terms of energy consumption energy properties; and d a method to monitor the mobile execution traces and analyze them against the energy properties. We aim to construct a tool chain addressing all these steps. We have already designed and implemented a model-based approach to automatically generate execution traces in mobile devices using Spin. This paper focuses on the formalization and analysis of energy properties with a specification language inspired by the interval logic. The paper presents this logic, the implementation of runtime verification using Buchi automata, and the practical use of the whole tool chain for model-based runtime verification of energy-related properties. Spin is a main ingredient for generating the test cases and checking the properties.

UML-Based Model-Driven Development for HSDPA Design

UML-based model- driven development can reduce the costs of developing, maintaining, and evolving complex software systems. Experiments with a High-Speed Downlink Packet Access design illustrate this approachs effectiveness.

Performance Analysis of Spotify® for Android with Model-Based Testing

This paper presents the foundations and the real use of a tool to automatically detect anomalies in Internet traffic produced by mobile applications. In particular, our MVE tool is focused on analyzing the impact that user interactions have on the traffic produced and received by the smartphones. To make the analysis exhaustive with regard to the potential user behaviors, we follow a model-based approach to automatically generate test cases to be executed on the smartphones. In addition, we make use of a specification language to define traffic patterns to be compared with the actual traffic in the device. MVE also includes monitoring and verification support to detect executions that do not fit the patterns. In these cases, the developer will obtain detailed information on the user actions that produce the anomaly in order to improve the application. To validate the approach, the paper presents an experimental study with the well-known Spotify app for Android, in which we detected some interesting behaviors. For instance, some HTTP connections do not end successfully due to timeout errors from the remote Spotify service.

An end-to-end testing ecosystem for 5G

Connected mobile applications will be a dominant software component in the 5G domain. Ensuring a correct and efficient behavior of the applications and devices becomes a critical factor for the mobile communications market to meet the expectations of final users. The EU project TRIANGLE is building a framework to help app developers and device manufacturers in the evolving 5G sector to test and benchmark new mobile applications, devices, and services utilizing existing and extended FIRE testbeds. This framework will evaluate Quality of Experience and enable certification for new mobile applications and devices.