Introduction
ALI R. BUTT
Virginia Tech
DILMA DA SILVA
IBM Research
Cloud computing is emerging as a disruptive technology that will change the way users, especially scientists and engineers, design, develop, deploy, use, and disseminate their applications and data. By decoupling lower-level computer system details from application development, and freeing users to focus on their technical and scientific missions, cloud computing is likely to have a profound impact on our lives.
Computer-based simulations and applications are considered a “third-pillar” of scientific discovery, which complements the traditional pillars of theory and experimentation. Currently, these simulations and applications, which require significant investment in the acquisition and maintenance of system infrastructure, are used only by seasoned computer scientists. Cloud computing promises to lower the entry barrier and allow for the easy integration of knowledge gained from scientific observation and for predictions of future responses or outcomes.
The speakers in this session highlight some recent advances in technologies that are shaping the modern cloud-computing paradigm. Their talks cover a wide range of “cloud aspects,” from designing innovative computer systems to how such systems can be used and configured in an energy-efficient way.
Armando Fox (UC-Berkeley) begins with an overview of how next-generation clouds should look. Based on user feedback and a survey of requirements, he discusses the major trends as computer scientists work toward realizing future clouds and making them amenable to wide-scale use and adaptation, enabling the democratization of supercomputing. Next, Luiz Andre Barroso (Google) describes the basics of cloud computing—how such systems are realized, the challenges to providing transparent interfaces to users while maintaining unfathomable scale,
and support for user applications in a seamless, world-wide “supercomputer” (i.e., the cloud).
In the third talk, YY Zhou (UC-San Diego) describes the challenges of building robust applications in the cloud. Finally, Parthasarathy Ranganathan (HP Labs) describes the environmental and energy implications of using hundreds of thousands of computing nodes at a central location. He also discusses how building architecture and software design can be done in ways that reduce the carbon footprint of the supporting cloud infrastructure.