Context: Several examples of successful migrations of systems to Service Oriented Architecture (SOA) are presented in the literature. Some of the approaches also try to structure the process of migration to SOA. The reported migration attempts underline the role of architecture of migrated system, but they do not explore the architectural patterns applied in architecture of migrated systems while proper usage of patterns may simplify and improve quality of migration. Objectives: This work is aimed at elaborating guidelines that support migration from a system that is based on a chosen architectural pattern towards a system based on Service Oriented Architecture. Methods: Literature review is used as a basic method in the initial steps of the research, that is during investigation of existing techniques of migration to SOA, establishing procedure for selection of the migrated pattern and identifying building blocks of the target architecture. Results of the literature reviews are further analyzed in order to select the migrated architectural pattern and to elaborate the target architecture. The guidelines for migration are the result of the synthesis of the analyzed information. Results: The migration is realized as a translation between two pattern languages: the first pattern language describes the chosen architectural pattern–Model–View–Controller and the second pattern language describes SOA target architecture, expressed using SOA architectural patterns. The translation is defined by a set of migration guidelines. The approach is also illustrated with migrating an example student project. Conclusion: The study shows that the usage of an architectural pattern during migration allows to define the migration in a simple, structured and precise way using guidelines that represent a set of subsequent well defined steps that should be applied in order to migrate a specific type of legacy system.

Made available in DSpace on 2014-12-17T15:48:08Z (GMT). No. of bitstreams: 1 FabioAGS_DISSERT.pdf: 1631308 bytes, checksum: cb5c8aa720129c7e7b32a24acd2a4b2c (MD5) Previous issue date: 2013-07-30 / The importance of non-functional requirements for computer systems is increasing. Satisfying these requirements requires special attention to the software architecture, since an unsuitable architecture introduces greater complexity in addition to the intrinsic complexity of the system. Some studies have shown that, despite requirements engineering and software architecture activities act on different aspects of development, they must be performed iteratively and intertwined to produce satisfactory software systems. The STREAM process presents a systematic approach to reduce the gap between requirements and architecture development, emphasizing the functional requirements, but using the non-functional requirements in an ad hoc way. However, non-functional requirements typically influence the system as a whole. Thus, the STREAM uses Architectural Patterns to refine the software architecture. These patterns are chosen by using non-functional requirements in an ad hoc way. This master thesis presents a process to improve STREAM in making the choice of architectural patterns systematic by using non-functional requirements, in order to guide the refinement of a software architecture / Cada vez mais os sistemas computacionais precisam satisfazer requisitos n?ofuncionais. No entanto a satisfa??o destes requisitos exige uma aten??o especial j? que s?o requisitos que devem ser levados em considera??o pelas arquiteturas de software que representam esses sistemas. Uma arquitetura de software inadequada introduz uma complexidade maior, al?m da complexidade intr?nseca do sistema. Al?m disso, estudos mostram que apesar das atividades de engenharia de requisitos e de arquitetura de software atuarem em diferentes aspectos do desenvolvimento, ? preciso execut?- las de forma iterativa e entrela?ada para produzir sistemas computacionais satisfat?rios. Com o objetivo de diminuir a lacuna existente entre requisitos de software e o desenvolvimento de uma arquitetura de software que reflita esses requisitos, foi proposto o processo STREAM (Strategy for Transition between REquirements models and Architectural Models). O STREAM consiste de uma abordagem sistem?tica que enfatiza os requisitos funcionais, sendo os n?o-funcionais usados de forma ad hoc. Entretanto, os requisitos n?o-funcionais normalmente impactam o sistema como um todo. Assim, a maneira proposta pelo STREAM para fazer com que a arquitetura de software desenvolvida reflita os Requisitos N?o-Funcionais elicitados no modelo de requisitos ? atrav?s do refinamento da arquitetura usando padr?es arquiteturais escolhidos com base nos requisitos n?o-funcionais. No STREAM a escolha desses padr?es ? realizada de forma ad hoc. Com o objetivo de sistematizar a forma de escolha dos padr?es arquiteturais, esta disserta??o prop?e um processo para sistematizar a escolha desses padr?es, a partir de requisitos n?o-funcionais racionalizados, que ser?o usados para refinar arquitetura de software

Röststyrning och rösttolkning är ett gränssnitt mellan användare och dator som blivit vanligare i kommersiella produkter. Gränssnittet används i digitala assistenter, det vill säga en mjukvarubaserad tjänst utformad för att hjälpa användare att utföra digitala uppgifter. Dessa uppgifter inkluderar att svara på frågor, hantera scheman, hemkontroll, spela musik och mycket mer. Eftersom digitala assistenter är relativt nya så finns det ett behov av mer kunskap om hur man kan skapa applikationer för plattformen.Rapporten ger kunskap om hur utvecklingsprocessen ser ut för en röststyrd ap- plikation till Google Assistant. Detta görs med en fallstudie som ger en inblick i de olika design- och arkitekturval som ingår i mjukvaruutvecklingen. Resultatet beskriver lämpliga konversationsmönster för röstgränssnitt i röststyrda applikationer och ett lämpligt arkitekturmönster för kodbasen. Med hjälp av studien drogs slutsatser om vilka begränsningar som finns hos röststyrda applikationer. / Voice control and voice interpretation is an interface between users and computers that have become more common in commercial products. The interface is used in digital assistants, which is a software-based service designed to help users perform digital tasks. These tasks include answering questions, managing their schedules, home control, playing music and more. Because digital assistants are relatively new, there is a need for more knowledge about how to create applications for the platform.The report provides information on how the development process looks like for a voice-controlled application for Google Assistant. This is done by a case study that provides insight into the various design and architecture choices that are included in the software development. The result describes a suitable conversation pattern for voice interfaces in voice-controlled applications and an appropriate architecture for the codebase. The study draws conclusions about the limitations of voice-controlled applications.

Ce mémoire présente un patron d’architecture permettant, dans un contexte orientéobjet, l’exploitation d’objets appartenant simultanément à plusieurs hiérarchies fonctionnelles. Ce patron utilise un reasoner basé sur les logiques de description (web sémantique) pour procéder à la classification des objets dans les hiérarchies. La création des objets est simplifiée par l’utilisation d’un ORM (Object Relational Mapper). Ce patron permet l’utilisation effective du raisonnement automatique dans un contexte d’applications d’entreprise. Les concepts requis pour la compréhension du patron et des outils sont présentés. Les conditions d’utilisation du patron sont discutées ainsi que certaines pistes de recherche pour les élargir. Un prototype appliquant le patron dans un cas simple est présenté. Une méthodologie accompagne le patron. Finalement, d’autres utilisations potentielles des logiques de description dans le même contexte sont discutées. / This master thesis presents a software architectural pattern for use in an object oriented environment to simultaneously access objects in multiple functional hierarchies. A Description Logics (Semantic Web) reasoner is used to classify the objects in the hierarchies. Object creation is simplifed by the use of an ORM - Object Relational Mapper. The pattern effectively allows automatic reasoning procedures to be used in an enterprise application context. All concepts required to understand the architectural pattern and the tools are presented. Usage conditions for the pattern are discussed and research projects are presented to widen the pattern’s applicability. A prototype applying the pattern on a simple problem is presented. A methodology is also presented. Finally, other potential uses of Description Logics based automatic reasoning procedures are discussed.

Pattern matching approaches in reverse engineering aim to incorporate domain knowledge and system documentation in the software architecture extraction process, hence provide a user/tool collaborative environment for architectural design recovery. This thesis presents a model and an environment for recovering the high level design of legacy software systems based on user defined architectural patterns and graph matching techniques. In the proposed model, a high-level view of a software system in terms of the system components and their interactions is represented as a query, using a description language. A query is mapped onto a pattern-graph, where a module and its interactions with other modules are represented as a group of graph-nodes and a group of graph-edges, respectively. Interaction constraints can be modeled by the description language as a part of the query. Such a pattern-graph is applied against an entity-relation graph that represents the information extracted from the source code of the software system. An approximate graph matching process performs a series of graph edit operations (i. e. , node/edge insertion/deletion) on the pattern-graph and uses a ranking mechanism based on data mining association to obtain a sub-optimal solution. The obtained solution corresponds to an extracted architecture that complies with the given query. An interactive prototype toolkit implemented as part of this thesis provides an environment for architecture recovery in two levels. First the system is decomposed into a number of subsystems of files. Second each subsystem can be decomposed into a number of modules of functions, datatypes, and variables.

Pattern matching approaches in reverse engineering aim to incorporate domain knowledge and system documentation in the software architecture extraction process, hence provide a user/tool collaborative environment for architectural design recovery. This thesis presents a model and an environment for recovering the high level design of legacy software systems based on user defined architectural patterns and graph matching techniques. In the proposed model, a high-level view of a software system in terms of the system components and their interactions is represented as a query, using a description language. A query is mapped onto a pattern-graph, where a module and its interactions with other modules are represented as a group of graph-nodes and a group of graph-edges, respectively. Interaction constraints can be modeled by the description language as a part of the query. Such a pattern-graph is applied against an entity-relation graph that represents the information extracted from the source code of the software system. An approximate graph matching process performs a series of graph edit operations (i. e. , node/edge insertion/deletion) on the pattern-graph and uses a ranking mechanism based on data mining association to obtain a sub-optimal solution. The obtained solution corresponds to an extracted architecture that complies with the given query. An interactive prototype toolkit implemented as part of this thesis provides an environment for architecture recovery in two levels. First the system is decomposed into a number of subsystems of files. Second each subsystem can be decomposed into a number of modules of functions, datatypes, and variables.

Care process monitoring is important in healthcare domains to provide precise and detailed analytics on patients, providers, and resources participating in a care process and their status. These analytics are used to keep track of whether the quality of care goals set by healthcare organizations are satisfied and ensure that legislative and organizational guidelines are followed. The complexity of care process monitoring can vary depending on whether the care process takes place in a hospital or out in the community, and it can vary depending on the complexity of the information technology infrastructure that is in place to support the care process. A Care Process Monitoring Application (CPMA) is a software application which collects and integrates data from various sources while a care process is being provided, in order to provide performance reporting of metrics that are used to measure how well the performance goals and guidelines for the care process are being met. In our research, we have studied how CPMAs are built in order to improve the quality of their engineering. The significant challenge in this context is how to engineer a CPMA so that the engineering process is repeatable, produces a CPMA of consistent high quality, and requires less time, less effort and less complexity. This thesis proposes an application framework for care process monitoring that collects and integrates events from event sources, maintains the individual and aggregate states of the care process and populates a metrics data mart to support performance reporting. Our contributions are the following: a state-based application meta-model of care process monitoring, a care process monitoring architectural pattern, and finally, a behavior driven development methodology for CPMAs based on our meta-model and architectural pattern. Our results are validated through three different case studies in which we collaborated with two different health care organizations to build and deploy CPMAs for two different care processes (one hospital-based, the other community-based) in collaboration with healthcare clinicians and researchers.

Ce mémoire présente un patron d’architecture permettant, dans un contexte orientéobjet, l’exploitation d’objets appartenant simultanément à plusieurs hiérarchies fonctionnelles. Ce patron utilise un reasoner basé sur les logiques de description (web sémantique) pour procéder à la classification des objets dans les hiérarchies. La création des objets est simplifiée par l’utilisation d’un ORM (Object Relational Mapper). Ce patron permet l’utilisation effective du raisonnement automatique dans un contexte d’applications d’entreprise. Les concepts requis pour la compréhension du patron et des outils sont présentés. Les conditions d’utilisation du patron sont discutées ainsi que certaines pistes de recherche pour les élargir. Un prototype appliquant le patron dans un cas simple est présenté. Une méthodologie accompagne le patron. Finalement, d’autres utilisations potentielles des logiques de description dans le même contexte sont discutées. / This master thesis presents a software architectural pattern for use in an object oriented environment to simultaneously access objects in multiple functional hierarchies. A Description Logics (Semantic Web) reasoner is used to classify the objects in the hierarchies. Object creation is simplifed by the use of an ORM - Object Relational Mapper. The pattern effectively allows automatic reasoning procedures to be used in an enterprise application context. All concepts required to understand the architectural pattern and the tools are presented. Usage conditions for the pattern are discussed and research projects are presented to widen the pattern’s applicability. A prototype applying the pattern on a simple problem is presented. A methodology is also presented. Finally, other potential uses of Description Logics based automatic reasoning procedures are discussed.

Care process monitoring is important in healthcare domains to provide precise and detailed analytics on patients, providers, and resources participating in a care process and their status. These analytics are used to keep track of whether the quality of care goals set by healthcare organizations are satisfied and ensure that legislative and organizational guidelines are followed. The complexity of care process monitoring can vary depending on whether the care process takes place in a hospital or out in the community, and it can vary depending on the complexity of the information technology infrastructure that is in place to support the care process. A Care Process Monitoring Application (CPMA) is a software application which collects and integrates data from various sources while a care process is being provided, in order to provide performance reporting of metrics that are used to measure how well the performance goals and guidelines for the care process are being met. In our research, we have studied how CPMAs are built in order to improve the quality of their engineering. The significant challenge in this context is how to engineer a CPMA so that the engineering process is repeatable, produces a CPMA of consistent high quality, and requires less time, less effort and less complexity. This thesis proposes an application framework for care process monitoring that collects and integrates events from event sources, maintains the individual and aggregate states of the care process and populates a metrics data mart to support performance reporting. Our contributions are the following: a state-based application meta-model of care process monitoring, a care process monitoring architectural pattern, and finally, a behavior driven development methodology for CPMAs based on our meta-model and architectural pattern. Our results are validated through three different case studies in which we collaborated with two different health care organizations to build and deploy CPMAs for two different care processes (one hospital-based, the other community-based) in collaboration with healthcare clinicians and researchers.