Core Grid Services Specification Moves Forward
Page 1 of 1
The core specification for the convergence of Grid computing with Web services is on track to become a standard.
The Open Grid Services Infrastructure (OGSI) version 1 specification, the basis for the Open Grid Services Architecture, went into "last call" within the OGSI working group at last week's Global Grid Forum 7 in Tokyo.
"In other words, the spec is considered 'done' by the editors, and this was a final opportunity for comment by working group participants before it is submitted into the GGF recommendation process," said Steve Tuecke of Argonne National Laboratory, a co-leader of the open source Globus Project. "The last call went smoothly, resulting in just a few small changes and clarifications. We are integrating these changes and will be submitting the final document to GGF shortly."
The GGF document process is loosely modeled after the Internet Standards Process, and the recommendation process is analogous to an Internet Standards track document.
World's Fastest Network Launched to Connect TeraGrid Sites
Fiber optic links between Los Angeles and Chicago have been "lit up" to form the cross-country network backbone for the National Science Foundation's $88 million TeraGrid project.
Technicians are sending the first test data packets racing across the network, which boasts unprecedented bandwidth of roughly one million times faster than a typical dial-up Internet connection and four times faster than existing research networks.
At 40 gigabits per second, the new "backplane," developed in partnership with Qwest Communications, will connect the resources of the TeraGrid, a multiyear effort to build and deploy the world's largest, fastest, distributed computing infrastructure for open scientific research. Scientists will use the TeraGrid to make fundamental discoveries in fields as varied as biomedicine, global climate, and astrophysics. The first applications will begin to use the TeraGrid capabilities from all sites this spring.
The TeraGrid partners are the National Center for Supercomputing Applications; the San Diego Supercomputer Center; Argonne National Laboratory; the Caltech Center for Advanced Computing Research; and the Pittsburgh Supercomputing Center. When completed, the TeraGrid will include 20 teraflops of computing power, facilities capable of managing and storing nearly one petabyte of data, high-resolution visualization environments, and toolkits for Grid computing.
The backplane consists of four 10 Gb/s optical fiber "lambdas" (light pipelines) running from a major Internet hub in Los Angeles to the StarLight hub in Chicago and three 10 Gb/s lambdas to each site. Juniper Networks provided the routers, the first in the nation available to handle the combined 40 Gb/s fiber traffic.
Argonne's Linda Winkler, TeraGrid network architect, led the effort to design, select, and install the network. "A key design criterion was that all the TeraGrid resources appear to be part of a single facility," she said. "The vendors were selected to help us stay at the cutting edge. The design allows new sites to be connected to the TeraGrid network in the future."
Altair Acquires PBS From Veridian
Altair Engineering Inc., an engineering, consulting and high-performance computing firm, has acquired portable batch system (PBS) technology, including PBS Pro and Open PBS workload management software and related assets from Veridian Corp.
PBS Pro is the commercial version of the Portable Batch System (PBS), originally developed to manage aerospace computing resources at the NASA Ames Research Center. PBS Pro is a leader in high performance workload management and batch queuing on LINUX clusters.
The acquisition allows Altair to broaden its product suite to address new markets requiring greater computational resources in the life sciences, energy, research, logistics, manufacturing and financial industries. PBS Pro technology will become the flagship product of Altair's Enterprise Computing business line, which provides software and consulting services to help companies with high-performance computing requirements achieve maximum return on their technology investments.
"This acquisition is an extension of Altair's business strategy to provide a variety of industries with solutions for high performance computing and innovative product engineering and design," said Jim Scapa, president and chief executive officer of Altair. "Automation, numerical methods for optimization and robust design are dramatically putting a constraint on computing resources in almost every industry, from product development to bioinformatics. PBS Pro technology addresses these market constraints and will help to grow the Altair customer base accordingly."
Altair has been a value-added reseller and implementation partner of Veridian's PBS Pro. Clients using PBS Pro include the U.S. Department of Defense, NASA, the U.S. Department of Energy, Lockheed-Martin, Boeing and the Ford Motor Company.
As part of the agreement, the PBS development team, software and technical development will remain in Mountain View, Calif., headed by Bill Nitzberg, managing director and chief technology officer for Altair Grid Technologies. Nitzberg joins the company with over 20 years of experience in industry and research computing, and is a member of the Steering Group and Director of the Scheduling and Resource Management Area of the Global Grid Forum.
Sales, marketing, customer support and distribution of PBS products will be managed from Altair's world headquarters in Troy, Mich., by Michael Humphrey, vice president of Altair's Enterprise Computing Business Line.
Platform Symphony 1.5 Released
Platform Computing has announced the general availability of Platform Symphony 1.5, an application infrastructure solution designed to accelerate online and real-time distributed processing.
Initially launched in the financial services market, Platform Symphony uses an open, extendable architecture that enables commercial institutions to integrate all distributed IT assets, creating a single, virtual environment, the company said. Through this virtual environment, institutions can transparently share applications across the enterprise, transforming a traditional 'siloed' compute environment into a shared, service-oriented model.
"With Platform Symphony 1.5, commercial institutions across all industries can improve application performance 100-fold, dramatically reducing processing times from hours to minutes," Platform said. "This highly efficient and cost-effective solution is designed to help enterprises develop, run and manage Grid-enabled applications, and to accelerate results by dynamically and intelligently distributing and executing work in parallel, leveraging existing hardware resources and legacy applications. With Platform Symphony, larger volumes and increasingly complex problems can be handled using the increased capacity and hardware utilization."
JPMorgan Chase is using Platform Symphony to manage selected, compute-intensive portions of its application workload across its globally distributed technology infrastructure, and several other global commercial institutions are deploying Platform Symphony on departmental and enterprise levels, allowing them to complete business critical analysis in hours instead of days, Platform said.
"With Platform Symphony, commercial enterprises can manage more transactions and get faster results, but with fewer operational resources and at a lower cost," said Ian Baird, vice president of marketing and sales operations for Platform. "With Symphony as the next generation of computing infrastructure, enterprises and ISVs can easily build, manage, and run Grid applications as they move to Grid computing and Web services technologies."
Axceleon Launches Academic High-Performance Grid Computing Package
Axceleon is offering special pricing to the academic community.
Axceleon's EnFuzion software utilizes the combined power of networked computers to significantly reduce application execution time. EnFuzion allows research departments to run multiple sets of input data across their Grid or network to achieve results much faster and with greater accuracy. EnFuzion works with Windows, Linux and Unix based operating systems, and utilizes all available computing hardware, including idle desktops, dedicated servers, rack mounted servers and blade servers.
Several universities are currently using EnFuzion to address a variety of their clustering needs, including Monash University, Yale University, Griffith University, and University of Chile.
"We have been using EnFuzion for several years now at Monash University," said David Abramson, professor of computer science at Monash University. "EnFuzion serves as a great tool in the academic arena. At Monash we have used EnFuzion to support a number of research projects, both in computer science and other varied disciplines, from mechanical engineering to public health policy. Also, we have used EnFuzion as a learning tool to teach our students about concepts such as parametric studies and high-performance computing."
Avaki Appoints Vice President of Marketing
Avaki has appointed Craig Muzilla as vice president of marketing. Based at Avaki's Burlington, Mass. headquarters, Muzilla will be responsible for leading the company's strategic direction and managing overall marketing functions.
"Craig is a valuable addition to the management team," said Tim Yeaton, Avaki president. "With over 20 years of experience, Craig has a proven track record of establishing new markets with emerging technologies through his marketing, product management, and business development experience in companies such as Bowstreet, Openpages, Samsung and Lotus Development. He is both a skilled marketing executive and a skilled strategist, and will play a key role in driving our business going forward."
Most recently, Muzilla was vice president of marketing at Openpages Inc., a Westford, Mass.-based provider of enterprise content management systems for the financial services and media markets, where he led marketing, product management and business development activities.
"Commercial Grid software technology has attracted significant attention in the market over the past year," said Muzilla. "Avaki is leading the market in delivering commercial Grid solutions that solve complex data access and other integration problems for businesses. I'm thrilled to join this team and committed to bringing my experience to bear as we continue to leverage the company's strengths and reinforce its leadership position."
Gridbus Meets Google
The Gridbus Project at the University of Melbourne has developed a guided meta-search engine called "Guided Google" that serves as an interface to the actual Google.com search engine, using Google Web Services.
The project has released a number of papers recently, including one on brain activity analysis. For more information, visit Gridbus.org.