Umberto Souza da Costa

Specification and Runtime Verification of Java Card Programs

Java Card is a version of Java developed to run on devices with severe storage and processing restrictions. The applets that run on these devices are frequently intended for use in critical, highly distributed, mobile conditions. They are required to be portable and safe. Often, the requirements of the application impose the use of dynamic, on-card verifications, but most of the research developed to improve safety of Java Card applets concentrates on static verification methods. This work presents a runtime verification approach based on Design by Contract to improve the safety of Java Card applications. To this end, we propose JCML (Java Card Modeling Language) a specification language derived from JML (Java Modeling Language) and its implementation: a compiler that generates runtime verification code. We also present some experiments and quality indicators.

JCML: A specification language for the runtime verification of Java Card programs

Java Card is a version of Java developed to run on devices with severe storage and processing restrictions. The applets that run on these devices are frequently intended for use in critical, highly distributed, mobile conditions. They are required to be portable and safe. Often, the requirements of the application impose the use of dynamic, on-card verifications, but most of the research developed to improve the safety of Java Card applets concentrates on static verification methods. This work presents a runtime verification approach based on Design by Contract to improve the safety of Java Card applications. To this end, we propose JCML (Java Card Modelling Language) a specification language derived from JML (Java Modelling Language) and its implementation: a compiler that generates runtime verification code. We also present some experiments and quality indicators. This paper extends previous published work from the authors with a more complete and precise definition of the JCML language and new experiments and results.

Designing service-based applications in the presence of non-functional properties

ContextThe development of distributed software systems has become an important problem for the software engineering community. Service-based applications are a common solution for this kind of systems. Services provide a uniform mechanism for discovering, integrating and using these resources. In the development of service based applications not only the functionality of services and compositions should be considered, but also conditions in which the system operates. These conditions are called non-functional requirements (NFR). The conformance of applications to NFR is crucial to deliver software that meets the expectations of its users. ObjectiveThis paper presents the results of a systematic mapping carried out to analyze how NFR have been addressed in the development of service-based applications in the last years, according to different points of view. MethodOur analysis applies the systematic mapping approach. It focuses on the analysis of publications organized by categories called facets, which are combined to answer specific research questions. The facets compose a classification schema which is part of the contribution and results. ResultsThis paper presents our findings on how NFR have been supported in the development of service-based applications by proposing a classification scheme consisting in five facets: (i) programming paradigm (object/service oriented); (ii) contribution (methodology, system, middleware); (iii) software process phase; (iv) technique or mathematical model used for expressing NFR; and (v) the types of NFR addressed by the papers, based on the classification proposed by the ISO/IEC 9126 specification. The results of our systematic mapping are presented as bubble charts that provide a quantitative analysis to show the frequencies of publications for each facet. The paper also proposes a qualitative analysis based on these plots. This analysis discusses how NFR (quality properties) have been addressed in the design and development of service-based applications, including methodologies, languages and tools devised to support different phases of the software process. ConclusionThis systematic mapping showed that NFR are not fully considered in all software engineering phases for building service based applications. The study also let us conclude that work has been done for providing models and languages for expressing NFR and associated middleware for enforcing them at run time. An important finding is that NFR are not fully considered along all software engineering phases and this opens room for proposing methodologies that fully model NFR. The data collected by our work and used for this systematic mapping are available in https://github.com/placidoneto/systematic-mapping_service-based-app_nfr.

Automatic refinement of service compositions

We propose a method for the automatic refinement of web service compositions: given a composite web service specification over abstract modules, our method generates lower-level versions of this composition. The refinement process is based on query rewriting techniques extended to take into account not only functional and non-functional requirements but also semantic information. Experimental results illustrate the performance and scalability of the method.

Efficient evaluation of context-free path queries for graph databases

We present a context-free path query evaluation algorithm inspired by top-down parsing techniques. Given a graph and a query defined over a context-free grammar, our algorithm identifies paths on the graph which form words of the language generated by the grammar. We show that our algorithm is correct. We conduct performance evaluation experiments with some popular ontologies and synthetic databases to endorse the efficiency of our approach. The algorithm presents a cubic worst-case runtime complexity in terms of the number of nodes in the graph, which is an improvement over previous work.

Experiments on service composition refinement on the basis of preference-driven recommendation

The service oriented computing paradigm proposes the construction of applications by integrating pre-existent services. Since a large number of services may be available in the cloud, service recommendation and selection are key problems. The paper proposes the automation of the selection task that takes the abstract specification of a composition, the definition of concrete services and the preferences obtained from a recommendation routine to produce a list of refinements in preference order. This paper presents three versions of our approach. They differ in implementation options or in the methods used to combine preferences. Experiment results are provided as a benchmark for ensuring the efficiency of our proposal in practice.

Supporting Non-functional Requirements in Services Software Development Process: An MDD Approach

This paper presents the πSOD-M method, an extension to the Service-Oriented Development Method (SOD-M) to support the development of services software by considering their functional and non-functional requirements. Specifically, πSOD-M proposes: (i) meta-models for representing non-functional requirements at different abstraction levels; (ii) model-to-model transformation rules, useful to semi-automatically refine Platform Independent Models into Platform Specific Models; and (iii) rules to transform Platform Specific Models into concrete implementations. In order to illustrate our proposal, the paper also describes how to apply the methodology to develop a proof of concept.

A Bottom-Up Algorithm for Answering Context-Free Path Queries in Graph Databases

Many computing applications require processing of data that are directly collected from the Internet. In this context, the use of the Resource Description Framework (RDF) has became a common feature. The query and analysis of RDF data is paramount to explore the full potential of the data available on the Web. Query languages for RDF graph databases rely on the use of regular expressions to identify paths over the data. Some interesting queries, such as same-generation queries, cannot be expressed by regular expressions. We are interested in extending the expressiveness of queries over graph databases by using paths defined by context-free grammars. We introduce a new query algorithm to process context-free path queries over graph databases. Our approach is inspired by the LR(1) parsing techniques. A prototype was implemented and experiments were conducted to validate and compare the results of our algorithm with those obtained by similar approaches.

Automating systematic mappings, adding quality to quantity: poster

N/A.

An abstract machine for integrating heterogeneous web applications

The adoption of Cloud Computing technologies by the organizations has profound consequences on the way software applications are developed and used. The migration to the Cloud may be accompanied by the revision of the business process, to integrate tasks over big data. In this way, existing workflow implementations may be extended with calls to operations in Hadoop or other tools. In this work, we define a framework to implement business process over heterogeneous technologies. Our framework is based on a novel workflow engine, called μBP-AM. Workflow execution in μBP-AM is performed by successively transforming the graph obtained from a workflow definition. μBP-AM has a formal semantics, which gives a precise definition of how the workflow is implemented. μBP-AM is at the core of an extensible framework capable of supporting not only Web service operations but also Hadoop operation calls (among others). The tool described here was conceived to increase reliability and to promote interoperability. We describe a prototype implementation of our framework, as well as some experimental results. Experiments using this prototype show that compositions run in μBP-AM using about the same resources as those run by using other tools.