Unidata's vision for Transforming Geoscience by moving data services and software to the Cloud

- Indicates paper has been withdrawn from meeting
- Indicates an Award Winner
Tuesday, 4 February 2014: 3:45 PM
Room C105 (The Georgia World Congress Center )
Mohan Ramamurthy, University Corporation for Atmospheric Research, Boulder, CO

Universities are facing many challenges: shrinking budgets, rapidly evolving information technologies, growing data volumes, multidisciplinary science requirements, and high student expectations. Most faculty and researchers would prefer to focus on teaching and doing science rather than setting up computer systems. These changes are upending traditional approaches to accessing and using data and software;

Unidata has been providing data services to the atmospheric sciences community for more than a quarter century. However, Unidata's products and services must evolve to support modern approaches to research and education. After years of hype and ambiguity, cloud computing technologies have matured. Their promise is now being realized in many areas of commerce, science, and education, bringing the benefits of virtualized and elastic remote services to infrastructure, software, computation, and data. Cloud environments can reduce the amount of time and money spent to procure, install, and maintain new hardware and software, reduce costs through resource pooling and shared infrastructure, and provide greater security. Cloud services aimed at providing any resource, at any time, from any place, using any device are increasingly being embraced by all types of organizations. NOAA, NASA, and other federal science agencies are establishing cloud computing services. Universities are no exception; the University of Washington, University of Illinois, Cornell University, and George Washington University are some of the universities that have already set up cloud services for scientific and academic computing.

Given this trend and the enormous potential of cloud-based services, we propose to gradually augment Unidata's products and services to align with the cloud-computing paradigm. Specifically, we will work to establish a community-based development environment that supports the creation and use of software services to build end-to-end data workflows. The design will encourage the creation of services that can be broken into small, independent chunks that provide simple capabilities. Chunks could be used individually to perform a task, or chained into simple or elaborate workflows. The services we envision will be loosely coupled to meet user needs rather than tightly coupled into a monolithic system. The services will be portable, allowing their use in researchers' own cloud-based computing environments. In this talk, I will present a vision for Unidata's future that will provide a transformative community platform for collaborative development and an array of innovative data services to our users. Through these efforts, Unidata will empower its users to tackle multidisciplinary grand challenge problems, develop the profession's human capacity, and transform the conduct of science.