February 13, 2012
The National Science Foundation (NSF) describes its support for cloud computing research in a 21-page report, submitted to Congress last week.
The "NSF Report on Support for Cloud Computing" was undertaken in response to section 524.c of the America Competes Reauthorization Act of 2010, which permits the NSF to define programs of research and gives NSF's Program Officers the power to seek and manage research projects in cloud computing. The Act also specifies that a report be submitted to Congress annually for a minimum of five years "on the outcomes of National Science Foundation investments in cloud computing research, recommendations for research focus and program improvements, or other related recommendations."
According to the report, the NSF is "committed to providing the science and engineering communities with the opportunity to conduct research and education activities in cloud and data-intensive computing and their applications."
Section 524 of the America Competes Reauthorization Act of 2010 states:
The Director may support a national research agenda in key areas affected by the increased use of public and private cloud computing, including new approaches, techniques, technologies, and tools for (A) optimizing the effectiveness and efficiency of cloud computing environments; and (B) mitigating security, identity, privacy, reliability, and manageability risks in cloud-based environments, including as they differ from traditional data centers.
According to the same section, the Director of the National Institute of Standards and Technology shall
(1) collaborate with industry in the development of standards supporting trusted cloud computing infrastructures, metrics, interoperability, and assurance; and
(2) support standards development with the intent of supporting common goals.
The NSF report recognizes the vital importance of cloud computing, citing the NIST model:
"The Cloud Computing model offers the promise of massive cost savings combined with increased IT agility. It is considered critical that government and industry begin adoption of this technology in response to difficult economic constraints. However, cloud computing technology challenges many traditional approaches to datacenter and enterprise application design and management. Cloud computing is currently being used; however, security, interoperability, and portability are cited as major barriers to broader adoption."
The Computer and Information Science and Engineering (CISE) Directorate of NSF has increasingly channeled its funding to six research areas that are expected to have a significant impact on cloud computing. They are:
1. Computer Systems
2. Computer Networks
3. Security and Privacy
4. Algorithms and data management
5. Applications and software engineering
6. Computer science education
From 2009-2011, CISE gave its approval to 125 research projects spanning topics as diverse as architecture, green computing, big data, security, migration and more. Of these active awards, 76 are managed by the Computer and Network Systems (CNS) Division, 40 are managed by the Computing and Communications Foundations (CCF) Division, and 9 are managed by the Information and Intelligent Systems (liS) Division.
The PROBE project, described as a national facility for hosting cloud test-beds, will interest many in the HPC research community. The project is funded by the CNS division of CISE, and hosted by the New Mexico Consortium. As explained in the report, "the test-beds can be utilized by members of the computer systems research community in the United States to carry out their system experiments on large-scale, data-intensive, and supercomputing machines. ...PROBE will enable critical experimental research in many computing areas, including reliability, correctness, productivity, and energy efficiency."
PROBE relies mainly on decommissioned DOE machines, which makes it less expensive than equivalent supercomputing systems. With its 1,000-node cluster with 2,000 processors, PROBE will support both computing-intensive and disk-intensive experiments. There are already plans for another 1,000-node cluster, with 2,000 to 3,000 processors, to be added. More computing muscle will be added each year as machines are decommissioned from the DOE laboratories. The annual cost of the facility, including staff, operating costs, and infrastructure, is right around three million dollars, which the report notes is "a small fraction of the typical operational costs of a large supercomputing center."
Looking forward, the NSF report underscores the importance of designing correct and efficient future cloud systems as well as improving the utilization of existing cloud systems. To that end it seeks input from the following five technical areas:
1. Computer Systems, which is supported by the CNS program Computer Systems Research (CSR)
2. Networks, which are supported by the CNS program Networking Technology and Systems (NeTS)
3. Security (including cryptography), which is supported by the CISE program Trustworthy Computing (TC) and by the CCF program Algorithmic Foundations (AF)
4. Computer Architecture and Software, which are supported by the CCF program Software and Hardware Foundations (SHF)
5. Databases and Data-Intensive Computing, which are supported by the liS program Information Integration and Informatics (Ill)
"It is believed that a focused effort to bring these research communities together will best serve the need expressed by NIST and others to develop and mature the science of cloud computing," the authors note.
The US government, under the auspices of NIST and the America COMPETES Reauthorization Act, recognizes that "cloud computing is an area vital to the economic growth and competitiveness of the nation." Furthermore, as the report affirms, "the CISE community has responded vigorously to the challenge, as evidenced by the wide range of research efforts underway that are supported by numerous programs throughout the CISE Directorate."
The ever-growing complexity of scientific and engineering problems continues to pose new computational challenges. Thus, we present a novel federation model that enables end-users with the ability to aggregate heterogeneous resource scale problems. The feasibility of this federation model has been proven, in the context of the UberCloud HPC Experiment, by gathering the most comprehensive information to date on the effects of pillars on microfluid channel flow.
Large-scale, worldwide scientific initiatives rely on some cloud-based system to both coordinate efforts and manage computational efforts at peak times that cannot be contained within the combined in-house HPC resources. Last week at Google I/O, Brookhaven National Lab’s Sergey Panitkin discussed the role of the Google Compute Engine in providing computational support to ATLAS, a detector of high-energy particles at the Large Hadron Collider (LHC).
Frank Ding, engineering analysis & technical computing manager at Simpson Strong-Tie, discussed the advantages of utilizing the cloud for occasional scientific computing, identified the obstacles to doing so, and proposed workarounds to some of those obstacles.
05/10/2013 | Cleversafe, Cray, DDN, NetApp, & Panasas | From Wall Street to Hollywood, drug discovery to homeland security, companies and organizations of all sizes and stripes are coming face to face with the challenges – and opportunities – afforded by Big Data. Before anyone can utilize these extraordinary data repositories, however, they must first harness and manage their data stores, and do so utilizing technologies that underscore affordability, security, and scalability.
04/02/2012 | AMD | Developers today are just beginning to explore the potential of heterogeneous computing, but the potential for this new paradigm is huge. This brief article reviews how the technology might impact a range of application development areas, including client experiences and cloud-based data management. As platforms like OpenCL continue to evolve, the benefits of heterogeneous computing will become even more accessible. Use this quick article to jump-start your own thinking on heterogeneous computing.