Kontiki has selected Equinix's network interconnection and exchange services to enhance the performance of its Delivery Management System (DMS).
Kontiki said DMS is one of the industry's first commercial solutions based on Grid architecture and currently serves more than 1.5 million nodes. To provide the scalability, reliability and cost efficiencies unique to its Grid architecture, the company said it must be able to directly connect to a broad range of networks in order bring content as close to the user as possible, avoiding multiple routers and network congestion.
By operating in Equinix's network-rich Internet hubs, Kontiki said it can enhance the speed and reliability of digital media delivery through direct interconnection to the widest choice of networks. This direct interconnection eliminates potential points of service failure by dramatically reducing the number of 'hops' between network destinations, and significantly reducing the costs of reaching these networks through the elimination of local loops, the company said.
"The Kontiki DMS enables enterprises to publish, secure, and deliver digital media such as training videos, product presentations, software and documents to employees, partners, and customers," said Wade Hennessey, CTO and VP of Engineering at Kontiki. "Equinix Internet hubs and Internet exchange services have allowed us to improve the performance of digital media delivery, while lowering our overall operations costs. Equinix services provide a carrier independent solution which allows the Kontiki DMS to intelligently optimize our use of server bandwidth most efficiently."
Gridbus Project Announces Grid Market Directory
The Gridbus Project at the University of Melbourne, Australia has announced the Grid Market Directory (GMD), which enables Grid service providers to publish their services and related costs to the public, so that consumers can browse via Web browser or query using SOAP to find a suitable service meeting their requirements.
"It essentially provides directory for GSPs (Grid service providers) to publish their services and consumers/clients to make a Web service (SOAP) call to identify service details," said project director Raj Buyya. "GSPs can post application details along with prices, etc. In our Gridbus scheduler, we used this for identifying resources and application services and their costs."
The complete code, along with a technical write up, is available for download from: www.gridbus.org/gmd/. The project is also hosting GMD, which can be accessed at http://jarrett.cs.mu.oz.au:8080/gesi/index.htm.
Source Forge Grid Project Unveiled
William Rose, a Dell contractor working at Wal-Mart's Information Systems Division, has created a Grid project on Source Forge. It is a Linux kernel module that moves Grid computing functions into the kernel such that Linux will directly support Grid computing as a native operation, Rose said. The project can be found at http://sourceforge.net/projects/kgrid/.
"This project is a subproject of a much larger entity I am calling 'InterGrid,'" Rose said. "The InterGrid will provide any user that has access to the Internet the means to run Grid software (designed for the InterGrid) simply by sharing the computing resources they have on their computer(s) in return. In effect, a single user at home becomes a node on the InterGrid. Larger entities such as companies and foundations can also access the Grid provided they share resources too, and become a node on the Grid as well."
From the perspective of the InterGrid user, a program is simply executed to connect them to the Grid's resources, and operates almost identically to a Web browser, Rose said. The InterGrid browser will allow the user to execute Grid-enabled applications on the Grid. "In this respect, the browser acts like a desktop environment," he said.
PDCS-2003 Seeks Papers
PDCS-2003: 16th International Conference on Parallel and Distributed Computing Systems will be held August 13-15 in Reno, Nevada.
The forum will address, explore and exchange information on the state-of-the-art in all areas of Parallel and Distributed Computing Systems, their modeling and simulation, design, use and performance, and their impact. The program committee is soliciting papers describing original unpublished research in all areas of parallel processing and distributed computing systems, including theoretical, experimental, and commercial works.
See call for papers at http://www.isca-hq.org/PDCS-2003-call.htm for suggested topics. The deadline for abstracts/proposals is February 24.
Events
IE 7
Firefox 2.0
