Document details

A Pattern-Based Approach to Scaffold the IT Infrastructure Design Process

Author(s): Silva, Luís Alexandre Ferreira da

Date: 2017

Persistent ID: http://hdl.handle.net/10362/27863

Origin: Repositório Institucional da UNL

Subject(s): Information Technology; IT Infrastructure; UML Profile; Design Patterns; Domínio/Área Científica::Engenharia e Tecnologia::Engenharia Eletrotécnica, Eletrónica e Informática; Domínio/Área Científica::Engenharia e Tecnologia::Engenharia Eletrotécnica, Eletrónica e Informática; Domínio/Área Científica::Engenharia e Tecnologia::Engenharia Eletrotécnica, Eletrónica e Informática


Description

Context. The design of Information Technology (IT) infrastructures is a challenging task since it implies proficiency in several areas that are rarely mastered by a single person, thus raising communication problems among those in charge of conceiving, deploying, operating and maintaining/managing them. Most IT infrastructure designs are based on proprietary models, known as blueprints or product-oriented architectures, defined by vendors to facilitate the configuration of a particular solution, based upon their services and products portfolio. Existing blueprints can be facilitators in the design of solutions for a particular vendor or technology. However, since organizations may have infrastructure components from multiple vendors, the use of blueprints aligned with commercial product(s) may cause integration problems among these components and can lead to vendor lock-in. Additionally, these blueprints have a short lifecycle, due to their association with product version(s) or a specific technology, which hampers their usage as a tool for the reuse of IT infrastructure knowledge. Objectives. The objectives of this dissertation are (i) to mitigate the inability to reuse knowledge in terms of best practices in the design of IT infrastructures and, (ii) to simplify the usage of this knowledge, making the IT infrastructure designs simpler, quicker and better documented, while facilitating the integration of components from different vendors and minimizing the communication problems between teams. Method. We conducted an online survey and performed a systematic literature review to support the state of the art and to provide evidence that this research was relevant and had not been conducted before. A model-driven approach was also used for the formalization and empirical validation of well-formedness rules to enhance the overall process of designing IT infrastructures. To simplify and support the design process, a modeling tool, including its abstract and concrete syntaxes was also extended to include the main contributions of this dissertation. Results. We obtained 123 responses to the online survey. Their majority were from people with more than 15 years experience with IT infrastructures. The respondents confirmed our claims regarding the lack of formality and documentation problems on knowledge transfer and only 19% considered that their current practices to represent IT Infrastructures are efficient. A language for modeling IT Infrastructures including an abstract and concrete syntax is proposed to address the problem of informality in their design. A catalog of IT Infrastructure patterns is also proposed to allow expressing best practices in their design. The modeling tool was also evaluated and according to 84% of the respondents, this approach decreases the effort associated with IT infrastructure design and 89% considered that the use of a repository with infrastructure patterns, will help to improve the overall quality of IT infrastructures representations. A controlled experiment was also performed to assess the effectiveness of both the proposed language and the pattern-based IT infrastructure design process supported by the tool. Conclusion. With this work, we contribute to improve the current state of the art in the design of IT infrastructures replacing the ad-hoc methods with more formal ones to address the problems of ambiguity, traceability and documentation, among others, that characterize most of IT infrastructure representations. Categories and Subject Descriptors:C.0 [Computer Systems Organization]: System architecture; D.2.10 [Software Engineering]: Design-Methodologies; D.2.11 [Software Engineering]: Software Architectures-Patterns.

Document Type Doctoral thesis
Language English
Advisor(s) Abreu, Fernando; Goulão, Miguel
Contributor(s) Silva, Luís Alexandre Ferreira da
facebook logo  linkedin logo  twitter logo 
mendeley logo

Related documents