The main goal of the project is to organize a DevOps software ecosystem to support the collaboration between software developers and IT operations staff, in the context of a startup-centric social environment. In particular, the project will address security, tool interoperability and data portability issues to enable data-driven knowledge transfer between development and operations staff.
The specific objectives of the proposal are described in the following:
– Obj1: Federated cloud platform Economic and social issues in software engineering have been pointed out as industrial challenges for the next years, especially due to the software intensive systems and systems of systems scenarios. In parallel, many technologies emerge from networks of artifacts and stakeholders forming software ecosystems, which can be a fruitful environment to foster co-innovation and startup economy and help alleviating the economic and social issues in software engineering. Particularly, in this project we are interested in developing a federated architecture for a cloud based platform to support startups in facilitating the development-operations (DevOps) link. A key challenge will be stakeholders’ identification and requirements coordination to define characteristics, components and functions of the federated cloud platform for both generic and domain-specific scenarios.
– Obj2: Supporting virtual DevOps collaboration in the cloud Growing startups typically have developers in-house but tend to lack DevOps skills, thus being forced to acquire outside such missing competences. We aim to enrich the DevOps software ecosystem with communication, coordination and information sharing tools to let the development staff on the startup side and operations personnel on the cloud provider side to form a virtual team that breaks down the development/operations separation. Open-source DevOps-oriented tools will be integrated in the federated cloud platform in order to create a consistent environment that will enable startups to easily manage and integrate ‘dev’ and ‘ops’ through a cohesive user interface for making the interaction between them as smooth as possible.
– Obj3: Big data analytics for scientific continuous improvement Startups have a little room for errors in decision-making. Thus, they should apply the scientific decision making process that relies on evidence rather than hunches. Such decision-making relies on data and data-analytics about the DevOps process itself, system performance in real world conditions, and the usage profiles of the features of the system, e.g. which features are popular, and which features are easy/difficult to use. The task is not easy as the amounts of data are collected in various software logs are huge. The focus will be on collecting and analyzing the data with a federated cloud platform that will allow startups to fine-tune both their services and processes. The platform is going to provide even small startups with enough computing power, storage, and tools for simplifying the analysis of the Big Data that will be collected through continuous monitoring. The main goal of performing analyses on the huge amount of monitored data is to provide the DevOps team on taking only data- driven decisions when dealing with, feature development, maintenance, optimization or fixing tasks. This data will be first collected from DevOps tools and then processed and presented by the federated cloud platform to enable easy scientific decision-making.
– Obj4: Conducting software engineering research in the Cloud with DevOps and startups This project also advances the conduct of scientific research in software engineering. First, current data driven software engineering research relies mainly on mining rather small, and slowly changing software repositories such as version control and defect databases. This project will analyze software logs and produce plugins for DevOps tools that provide Big Data of software usage that is rapidly changing due to new product releases, add campaigns, and even variation based on the time of day. DOGS 4 Second, the object of study in the proposal are startups when traditional software engineering has focused mainly on large company context. Research on startups is becoming more meaningful for all software companies as they are accelerating their own software development by following the Lean Startup philosophy. Third, the DevOps software development process, which is highly suitable for startups, has received a limited number of studies in software engineering research that has recently mainly focused on agile software development.
– Obj5: Security The project will develop technologies and best practices for creating a secure, trustworthy, and high assurance federated cloud platform. A special challenge will be the management of security components established in federated clouds. Scalable enforcement of security policies across various geographical, organizational, platform, and administrative boundaries will be in the focus of the project. For achieving such a scalable enforcement, we need a corresponding enforcement infrastructure, whose security concepts will be elaborated in the project. Another challenge, for both DevOps and Big Data, will be the trade-off between security of intellectual property on the one side and the freedom in exchange and processing on the other side. Security policies need to fulfil security requirements, but still need to provide a certain level of freedom in order to be usable and acceptable.