No abstract

Software development processes are fundamentally based on efficient and effective communication. Communication between engineers, between engineers and managers, and between teams and clients are all essential components of a successful project. Requirements must be effectively transferred from client to engineer, specifications must be transitioned from architect to engineer, and constant communication between project team members, managers, and clients throughout the project life cycle is critical to the success of projects of any complexity. To succeed in a world where technologies, requirements, ideas, tools, and timelines are constantly changing, information must be accurate, readily available, easily found, and ideally delivered constantly, in real-time, to all team members. To meet these challenges, modern software development has evolved to encompass key concepts of adaptability to change and data-driven project management. A recent movement dubbed DevOps has attempted to use automated systems to bridge the information gap between project team entities and to enforce rigorous processes to ensure real-time communications.In this paper, the authors frame this challenge as a communications problem that can be addressed by the introduction of specifically designed autonomous system actors and processes. Successful implementation of such a methodology will enable efficient, effective, and immediate data collection, synthesis, and transfer of information between all requisite entities within the software project. A generalized model of DevOps will be presented and analyzed, offering a formalization of the communications and actors requisite to any effective software development process. These concepts will be further developed to illustrate the information flow between human and system actors, and explore how this model can be used to optimize the processes of a software development team to maximize productivity and quality of work products.

No abstract

Today's incident response training, architectures, and methodologies are all built upon disconnected siloes of domain expertise, but attacks upon an organization's critical information systems are not done in a disjointed way. Attacks on critical information systems and infrastructure are not solely network, or malware, or single disks; they are coordinated, large-scale multisite attacks done in an organized manner. With the increase in frequency and sophistication of these attacks, it is not enough to rely on intrusion detection systems, trusted IT staff, or organizational information security divisions. The velocity of a cyber attack should be met with an equally coordinated response. There is a need to develop a platform that enables responders to establish trust and develop an effective collaborative response plan and investigation process across multiple organizations and legal bodies to track adversaries, mitigate the threat, get critical systems back online, and pursue legal action against the offenders. In this work we propose such a platform for efficient collaboration. Our work is informed by our practices in supporting law enforcement organizations dealing with largescale distributed attacks on critical information systems and infrastructure and by an examination of Stuxnet, a computer worm discovered in June 2010 that is believed to have been created by the United States and Israel to attack Iran's nuclear facilities. Based on these experiences of operational support, the authors propose Cerebro, an Extensible Large-Scale Analysis Platform designed to fuse structured domain specific information, decision support, and collaboration in an automated fashion, to effectively detect and respond to such attacks.