Manual deployments of software is a tedious, repetitive and non-scaling method of deploying software.Continuous Delivery is a practice that enables automated deployment of software in a rapid fashion at the click of a button.When deciding whether to start using a new practice, software companies need to make an assessment from a cost-benefit perspective.This thesis compares automated deployments through Continuous Delivery with manual deployments from a cost perspective.The comparison is done at a small software company where two tools for Continuous Delivery are chosen based on requirements imposed by the company. The tools, Octopus Deploy and Azure DevOps, are cost efficient to different degrees.Octopus is cost efficient if several deployments per week are necessary, particularly if many deployment targets are involved.Azure DevOps is quickly cost efficient in most cases due to its pricing scheme, only needing roughly one deployment per week for few deployment targets, and a couple of deployments per year for many deployment targets.The initial cost of having a paid employee set up the tool needs to be paid off, but is easily done within a year using weekly deployments with a small number of deployment targets.

Background : Embedded system consists of both hardware and software. Boththese hardware and software development teams must be worked together to complete the project successfully, which requires a lot of communication among theteams. DevOps is an evolution of agile development that bridges communicationgap between teams. Since Embedded systems are not service-based, it isn’t easy toimplement DevOps. So there is a need to study what challenges are currently facedin implementing embedded application development. Objectives: The focus of this thesis was to gather the challenges and mitigationstrategies to overcome these challenges faced during the implementation of DevOpsin Embedded application development, along with mitigation strategies followed toovercome these challenges. Methods: In this thesis, surveys, and interviews were selected as research methods to identify challenges faced while implementing DevOps in embedded systemapplication development and mitigation strategies, and the survey was considered tocollect demographic questions. Results: A survey was conducted to identify embedded developers with experienceimplementing DevOps for Embedded Application Development. Out of 19 surveyresponses, 12 respondents were further interviewed for detailed information. Datacollected from the interviews are analyzed using thematic analysis and narrativeanalysis. Conclusions: From the analysis, the research concludes that continuous testingis the most challenging phase to implement. Continuous testing and continuous integration are the phases that need to be focused on more for the future implementationof DevOps