SpyderByte.com ;Technical Portals 
      
 News & Information Related to Linux High Performance Computing, Linux Clustering and Cloud Computing
Home About News Archives Contribute News, Articles, Press Releases Mobile Edition Contact Advertising/Sponsorship Search Privacy
HPC Vendors
Cluster Quoter (HPC Cluster RFQ)
Hardware Vendors
Software Vendors
HPC Consultants
Training Vendors
HPC Resources
Featured Articles
Cluster Builder
Beginners
Whitepapers
Documentation
Software
Lists/Newsgroups
Books
User Groups & Organizations
HP Server Diagrams
HPC News
Latest News
Newsletter
News Archives
Search Archives
HPC Links
ClusterMonkey.net
Scalability.org
HPCCommunity.org

Beowulf.org
HPC Tech Forum (was BW-BUG)
Gelato.org
The Aggregate
Top500.org
Cluster Computing Info Centre
Coyote Gultch
Dr. Robert Brown's Beowulf Page
FreshMeat.net: HPC Software
SuperComputingOnline
HPC User Forum
GridsWatch
HPC Newsletters
Stay current on Linux HPC news, events and information.
LinuxHPC.org Newsletter

Other Mailing Lists:
Linux High Availability
Beowulf Mailing List
Gelato.org (Linux Itanium)

LinuxHPC.org
Home
About
Contact
Mobile Edition
Sponsorship

Latest News

New Technology Enables More Efficient High-Performance Computing Applications
Posted by Staff, Thursday January 08 2004 @ 11:54AM EST

An optical network provisioning protocol to enable more efficient computing applications has been successfully demonstrated by scientists at MCNC Research & Development Institute and N.C. State University.

The demonstration of the Just-in-Time (JIT) protocol for provisioning and managing light path connections in the all-optical Advanced Technology Demonstration Network (ATDnet) in Washington, D.C., confirmed the viability of user-initiated, ultra-fast provisioning of all-optical network connections and marked the transition of the JIT protocol from the laboratory to an operational network. The light paths linked host systems at the U.S. Department of Defense’s Laboratory for Telecommunications Sciences, the Naval Research Laboratory’s Center for Computational Science and the Defense Intelligence Agency.

An overview of the JIT protocol was presented in December at the Globecom 2003 conference’s Optical Networking and Systems Symposium in San Francisco by Dan Stevenson, vice president of MCNC-RDI’s Advanced Network Research Division.

JIT will provide much needed support to U.S. military and civilian researchers to solve real-world problems. In particular, the Naval Research Laboratory is interested in the protocol’s ability to quickly set up and release tens, possibly hundreds, of gigabits of bandwidth for demanding, high-performance computing applications such as immersive real-time visualization of satellite imagery, computational fluid dynamics, ocean and weather modeling, and space physics.

“JIT addresses some very challenging problems in high-performance computing,” said Dr. Hank Dardy, chief scientist for advanced computing at the Naval Research Laboratory’s Center for Computational Science. “It can take weeks to establish an optical connection through a carrier network, and minutes to do so with generalized multi-protocol label switching, the current industry standard. With JIT, we can provision optical connections between sites in a few milliseconds through our microelectromechanical switches, and in a few microseconds when we deploy faster photonic switches.”

The JIT architecture and protocols used in the ATDnet were jointly developed by researchers at MCNC and Professors Paul Franzon, Harry Perros and George Rouskas of North Carolina State University. The research was partially funded by NASA and supported by the Advanced Research and Development Activity, a Department of Defense research and development community for the development of information technologies that current networks, including today’s Internet, do not or cannot support.

“This is the first deployment of its kind in an operational network at greater than gigabit speeds,” Stevenson said. “JIT is especially attractive to government customers because it doesn’t necessarily care about data rate or data format, not even whether the signal is digital or analog. Also, it works with commercial, off-the-shelf equipment from multiple vendors and multiple optical switching technologies.”

Fast, real-time resource provisioning will enable the military, particle physics, and research communities to focus on problems in new ways. Stevenson said that JIT overcomes many limitations and problems inherent with the current Internet. Applications can request, use and release bandwidth when needed, without tying up an optical circuit for days.

“High-speed, on-demand, application-initiated provisioning of bandwidth is also what the grid computing community is demanding,” Stevenson said. Grids connect heterogeneous computing platforms so that they operate, and appear to the user, as a single computing system. This means that computational problems can be directed to a system within the grid that will process it in the quickest and most cost effective manner. Grid computing provides users with unprecedented computing power, services and information, combing the resources of heterogeneous computing resources no matter where they are located.

“Grid resource requirements of big science applications, such as particle physics, are very dynamic,” said Stevenson. “The goal for sparse networks like ATDnet, and the recently announced National Lambda Rail, is to share grid bandwidth the same way you share computing cycles and storage in the grid. You also want to use those resources efficiently. These applications often involve computational steering and cannot afford the latency associated with electronic routers. The applications may require 300 megabits per second, but that’s only a small percentage of a 10-gigabit optical channel. JIT lets you share the remaining 97 percent of that bandwidth with others on the grid without the reduced performance inherent with electronic routing.”

MCNC and the University of North Carolina 16-campus system are jointly developing a statewide grid computing network for North Carolina’s higher education community using the existing statewide North Carolina Research and Education Network, operated by MCNC. The statewide research and education grid will link computing and data resources from multiple institutions in multiple locations with the potential to vastly increase the resources available to individual institutions. When complete, North Carolina will be one of the first states in the nation to deploy a statewide grid infrastructure.

“We intend to move JIT into grid networks,” said Stevenson. “We see the grid as a widely distributed computing system, and optical networks as the backplane for that system. We’re working on several supporting technologies to make that happen, such as protocols for QoS-aware routing, network management, transport, security and authentication, and making JIT OGSI/OGSA (Open Grid Services Infrastructure/Architecture) compliant. We’re also developing JIT-aware network adaptors so that high performance grid servers and hosts can take full advantage of JIT.”

Stevenson also sees other applications. “We believe that JIT will scale to finer timescales, and will support application-initiated provisioning of bandwidth for optical burst switching where a connection is provisioned in nanoseconds and may be released after only a few milliseconds,” he said. Optical burst switching is a high performance networking technology that transports digital and analog data an order of magnitude faster than today’s digital electronic packet switched technologies.

< MSC.Software Announces Release of MSC.Simulation Data Management 2004 | LCI's HPC Revolution 2004 CPF - Submission Deadline is January 30, 2004 >

 

Affiliates

Cluster Monkey

HPC Community


Supercomputing 2010

- Supercomputing 2010 website...

- 2010 Beowulf Bash

- SC10 hits YouTube!

- Louisiana Governor Jindal Proclaims the week of November 14th "Supercomputing Week" in honor of SC10!








Appro: High Performance Computing Resources
IDC: Appro Xtreme-X Supercomputer Blade Solution
Analysis of the Xtreme-X architecture and management system while assessing challenges and opportunities in the technical computing market for blade servers.

Video - The Road to PetaFlop Computing
Explore the Scalable Unit concept where multiple clusters of various sizes can be rapidly built and deployed into production. This new architectural approach yields many subtle benefits to dramatically lower total cost of ownership.
White Paper - Optimized HPC Performance
Multi-core processors provide a unique set of challenges and opportunities for the HPC market. Discover MPI strategies for the Next-Generation Quad-Core Processors.

Appro and the Three National Laboratories
[Appro delivers a new breed of highly scalable, dynamic, reliable and effective Linux clusters to create the next generation of supercomputers for the National Laboratories.

AMD Opteron-based products | Intel Xeon-based products



Home About News Archives Contribute News, Articles, Press Releases Mobile Edition Contact Advertising/Sponsorship Search Privacy
     Copyright © 2001-2013 LinuxHPC.org
Linux is a trademark of Linus Torvalds
All other trademarks are those of their owners.
    
  SpyderByte.com ;Technical Portals