Cloud computing is emerging as an alternative computing platform to bridge the gap between scientists' growing computational demands and their computing capabilities. A scientist who wants to run HPC applications can obtain massive computing resources 'in the cloud' quickly (in minutes), as opposed to days or weeks it normally takes under traditional business processes. Due to the popularity of Amazon EC2, most HPC-in-the-cloud research has been conducted using EC2 as a target platform. Previous work has not investigated how results might depend upon the cloud platform used. In this paper, we extend previous research to three public cloud computing platforms. In addition to running classical benchmarks, we also port a 'full-size' NASA climate prediction application into the cloud, and compare our results with that from dedicated HPC systems. Our results show that 1) virtualization technology, which is widely used by cloud computing, adds little performance overhead; 2) most current public clouds are not designed for running scientific applications primarily due to their poor networking capabilities. However, a cloud with moderately better network (vs. EC2) will deliver a significant performance improvement. Our observations will help to quantify the improvement of using fast networks for running HPC-in-the-cloud, and indicate a promising trend of HPC capability in future private science clouds. We also discuss techniques that will help scientists to best utilize public cloud platforms despite current deficiencies.

One of the world's largest scientific data systems, NASA's Earth observing system data and information system (EOSDIS) has stored over three petabytes of earth science data in a geographically distributed mass storage system. Design for this system began in the early 1990s and included a presentation of the design of the mass storage system at this conference in 1995. Many changes have occurred in the ten years since that presentation, much of it performed while the system was operational. In its first operational year (2000), the EOSDIS system had increased NASA's collection of earth science data holdings eight-fold. Today, EOSDIS collects over 7,000 gigabytes of data per week, almost 60 times more than the hubble space telescope. This load represents major challenges for ingest into the mass storage system, as well as for timely and balanced data distribution out of the mass storage system. This paper discusses the evolution of the EOSDIS archives focusing primarily on the mass storage system component of the archive. We present the lessons that were learned over the years and some directions that we are taking for the future.

NASA Earth and space science applications are currently utilizing geographically distributed computational platforms. These codes typically require the most compute cycles and generate the largest amount of data over any other applications currently supported by NASA. Furthermore, with the development of a leadership class SGI system at NASA Ames (Project Columbia), NASA has created an agency wide computational resource. This resource is heavily employed by Earth and space science users resulting in large amounts of data. The management of this data in a distributed environment requires a significant amount of effort from the users. This paper defines the approach taken to create an enabling infrastructure to help users easily access and move data across distributed computational resources. Specifically, this paper discusses the approach taken to create a wide area storage area network (SAN) using the SGI CXFS file system over standard TCP/IP. In addition, an emerging technology test bed initiative is described to show how NASA is creating an environment to continually evaluate new technology for data intensive computing.