This session will examine the research environment in which there is a rapidly-growing amount of scientific data, improved access to computing power that can be provisioned in the cloud, and better access and discovery of tools for coordination of research. Voluminous data products, incompatible data analysis tools, and inadequate access to high-productivity computing platforms expose inherent inefficiencies in science research ecosystems that make external collaborations difficult. This session will focus on strategies for optimizing the process of data analysis and will propose a reference architecture for analytic systems to address such inefficiencies, and will include talks ranging from existing systems, technology reviews, and science applications.