Foundations for Tools as a Service Workspace: A Reference Architecture: Design Space, Architecture Models, Design Tactics, Prototype and Experiences

Nowadays, on-demand provisioning of computing resources following a payper-
use service model have enabled client organizations to have easy and on the fly access to the resources. The resources that are provisioned using the service model are generally characterized as resource as a Service (*aaS). However, providing the resources corresponding to a domain following *aaS requires specific challenges associated with the domain to be addressed. Software as a Service (SaaS) model enables software vendors to offer their software solutions to end users following pay-per-use model. SaaS also enables end users to have access to the software system without being bound into long-term license commitments and without incurring additional infrastructure and maintenance overheads. Though SaaS has been successful in providing stand-alone applications where users can perform a specific set of activities using an application, applicability of SaaS for scenarios where users need to use a number of software systems to perform activities and associated tasks is limited. A typical such use case is the activities associated with software engineering domains in general and software architecting domain in particular.

This dissertation presents an approach to provide Tools as a service workSPACE (TSPACE), which is characterized as provisioning of a bundled suite of Tools as a Service (TaaS) required to perform activities associated with a specific domain as part of a cloud-enabled workspace. The presented approach is focused on addressing the challenges using an architecture centric solution by providing a Software Reference Architecture for TSPACE. As a specific case, software architecting domain and the tools used for software architecting have been focused in this dissertation.

This dissertation explores (a) the challenges associated with software architecting of cloud-enabled systems, (b) TSPACE reference architecture design guidelines, (c) TSPACE requirements, (d) information structuring needs and methods of TSPACE reference architecture, (e) a detailed description of TSPACE reference architecture sub-systems and components, and (f) TSPACE reference architecture evaluation.

The research goals are achieved by applying systematic literature review method, general literature surveys, reference architecture design methodologies, cloud and general architecture styles and patterns, architecture
prototyping and architecture evaluation methods as tools.

The main results of this dissertation are (a) a systematic review of the literature that identifies the challenges associated with software architecting of cloud-enabled systems, (b) a framework that guides the development of iv
TSPACE reference architecture and its evaluation, (c) the business, functional, quality and software architecture significant requirements of TSPACE, (d) TSPACE reference architecture meta-models and a set of ontologies to structure concepts and elements of TSPACE (e) a detailed description of the reference architecture based on established design practices and architecture patterns and (f) the detailed process of TSPACE reference architecture evaluation based on using architecture evaluation methods and a prototype of TSPACE reference architecture along with integrated tools to analyze applicability, effectiveness and usefulness of TSPACE reference architecture.

The main conclusions of this dissertation are (a) as the nature of the tools that are used in a specific domain vary, the domain models play a vital role in design of the TSPACE reference architecture for the respective domain (b) TSPACE reference architecture should have the capability to identify tools needed for the specific tasks (e.g. software architecting tasks and activities) and be able to bundle the needed tools in a suite of tools to provision TaaS, (c) as the nature of the tools vary in terms of the activities that are supported (e.g. specifying architecture significant requirements, capturing architecture knowledge and modeling architecture components), abstraction level of the artifacts that are produced and the way artifacts are exchanged among the tools and support for semantic as well as process-centric integration among the tools play a vital role in the adaptation of the reference architecture and (d) the reference architecture should be flexible enough to accommodate a wide variety of the tools used in a domain and the operations that are performed on the artifacts using the tools.