The term “Globalization” brought many challenges with itself in the field of software development. The challenge of accurate effort estimation in GSD is one among them. When talking about effort estimation, the discussion starts for effort estimation methods. There are a number of effort estimation methods available. Existing effort estimation methods used for co-located projects are might not enough capable to estimate effort for distributed projects. This is why; ratio of failure of GSD projects is high. It is important to calibrate existing methods or invent new with respect to GSD environment. This thesis is an attempt to explore the accuracy of effort estimation methods for distributed projects. For this purpose, the authors selected three estimation approaches: COCOMO II, SLIM and ISBSG. COCOMO II and SLIM are two well known effort estimation methods, whereas, ISBSG is used to check the trend of a project depending upon its (ISBSG’s) repository. The selection of the methods and approaches was based on their popularity and advantages over other methods/approaches. Two finished projects from two different organizations were selected and analyzed as case studies. The results indicated that effort estimation with COCOMO II deviated 15.97 % for project A and 9.71% for project B. Whereas, SLIM showed the deviation of 4.17% for project A and 10.86 % for project B. Thus, the authors concluded that both methods underestimated the effort in the studied cases. Furthermore, factors that might cause deviation are discussed and several solutions are recommended. Particularly, the authors state that existing effort estimation methods can be used for GSD projects but they need calibration by considering GSD factors to achieve accurate results. This calibration will help in process improvement of effort estimation.

Uncertainty in software development and business environment and the need to increase thespeed of development have driven organizations to search for methods that are responsive toboth change and speed. Providing iterative development, agile development involvescustomers and users through different phases of development, and delivers frequent releasesof software to customer while receives the corresponding feedback. Using this approach, agiledevelopment thus aims at addressing mentioned issues of speed and uncertainty whiledeveloping only what customer needs from the beginning of the project. On the other hand,distributed software development is used in many organizations to reach global talent andglobal market. The problems associated with distributed software development such as lack ofenough communication and team coherency, have forced project managers to combine it withagile to mitigate these social problems. This study focuses on distributed agile development,its suitability for a typical project and its challenges and deficiencies. Text analysis andinterviews using qualitative methods are used in this scientific research work. From thetheoretical view point, different text covering agile methodology, distributed development andcombination of them were considered. This study covered two parts: first, an evaluation ofagile and distributed development opportunities and problems to help determine whether ornot distributed development is suitable for a project and second, considering the challengesonce starting to use this method and practices required to regard them. For the empirical part,the focus was put on Volvo IT employees by having seven interviews with members who arecurrently active in distributed agile development. These interviews were used to compare andverify the finding of the theoretical part. The results of the study were categorized into twosections. In the first part, important elements required to verify the suitability of using thismethod are provided. The recommended factors for this evaluation are cost, productivity,customer, team structure, etc. In the second part, the challenges of using distributed agiledevelopment were categorized into four parts: (a) challenges of selected agile method, whichthe focus in this study is Scrum, (b) challenges with time-zone differences, (c) communicationchallenges and (d) finally team building challenges. The required practices to address thesechallenges were also provided. / Program: Masterutbildning i Informatik

Context. Global software development (GSD) gained huge recognition in today’s business world. Most of the software companies day by day are striving hard to evolve globally where software is developed in the context of different environmental settings that are distanced on various factors like geography, timezone, culture and language. Coordination is the factor that plays one of the prominent roles in such a setting for effective teamwork and project success. Although numerous efforts has been done in this research area, there has been no proper evidence from industry about the impact of these coordination challenges on the overall quality of the software when being developed in a distributed setting. Objectives. The purpose of this study is to examine and identify the coordination challenges and risks faced in global software development projects that has a negative impact on the quality of software from practitioner’s perspective. It also identify the tools, methods, and techniques that are used in industry to overcome these challenges and maintain quality standards. Methods. The aims and objectives of our study are fulfilled by conducting survey among practitioners working in GSD projects all around the globe. Further, 10 interviews are conducted with practitioners working in different companies and geographical locations in order to gain a detailed understanding of the impact of identified coordination challenges on the quality of software in GSD projects. Results. A total of 50 survey responses are recorded, out of which 48 respondents specify that coordination challenges has a negative impact on software quality in GSD context. By the ratings given by the participants, we identified the challenges and risks that had a major impact. Mixed results are obtained during interviews where most of them prioritized coordination as a major problem in GSD projects. It also included that use of some tools, methods and processes help them in overcoming this issue. The quality attributes that are mostly affected due to the challenges in GSD projects are also identified. Conclusions. After the analysis of survey results, the coordination challenges and associated risks in GSD projects are identified. They were found to havemostly negative impact on software quality. After thematic analysis of interview results, we observed that though the impact of coordination challenges is negative, its extent of implication is moderate in most cases.

In recent years, software tools used for Global Software Development (GSD) processes (e.g., continuous integration, version control and bug tracking) are increasingly being deployed in the cloud to serve multiple users. Multitenancy is an important architectural property in cloud computing in which a single instance of an application is used to serve multiple users. There are two key challenges of implementing multitenancy: (i) ensuring isolation either between multiple tenants accessing the service or components designed (or integrated) with the service; and (ii) resolving trade-offs between varying degrees of isolation between tenants or components. The aim of this thesis is to investigate how to architect the deployment of cloud-hosted service while guaranteeing the required degree of multitenancy isolation. Existing approaches for architecting the deployment of cloud-hosted services to serve multiple users have paid little attention to evaluating the effect of the varying degrees of multitenancy isolation on the required performance, resource consumption and access privilege of tenants (or components). Approaches for isolating tenants (or components) are usually implemented at lower layers of the cloud stack and often apply to the entire system and not to individual tenants (or components). This thesis adopts a multimethod research strategy to providing a set of novel approaches for addressing these problems. Firstly, a taxonomy of deployment patterns and a general process, CLIP (CLoud-based Identification process for deployment Patterns) was developed for guiding architects in selecting applicable cloud deployment patterns (together with the supporting technologies) using the taxonomy for deploying services to the cloud. Secondly, an approach named COMITRE (COmponent-based approach to Multitenancy Isolation Through request RE-routing) was developed together with supporting algorithms and then applied to three case studies to empirically evaluate the varying degrees of isolation between tenants enabled by multitenancy patterns for three different cloud-hosted GSD processes, namely-continuous integration, version control, and bug tracking. After that, a synthesis of findings from the three case studies was carried out to provide an explanatory framework and new insights about varying degrees of multitenancy isolation. Thirdly, a model-based decision support system together with four variants of a metaheuristic solution was developed for solving the model to provide an optimal solution for deploying components of a cloud-hosted application with guarantees for multitenancy isolation. By creating and applying the taxonomy, it was learnt that most deployment patterns are related and can be implemented by combining with others, for example, in hybrid deployment scenarios to integrate data residing in multiple clouds. It has been argued that the shared component is better for reducing resource consumption while the dedicated component is better in avoiding performance interference. However, as the experimental results show, there are certain GSD processes where that might not necessarily be so, for example, in version control, where additional copies of the files are created in the repository, thus consuming more disk space. Over time, performance begins to degrade as more time is spent searching across many files on the disk. Extensive performance evaluation of the model-based decision support system showed that the optimal solutions obtained had low variability and percent deviation, and were produced with low computational effort when compared to a given target solution.