IP Communications

BLADE Network Technologies, Inc. specializing in data center networking, has reportedly announced that its BLADE RackSwitch G8124 and G8000 10 Gigabit Ethernet or GbE switches have been deployed by The Netherlands’ Advanced School for Computing and Imaging or ASCI. 

The BLADE solution was selected for the Distributed ASCI Supercomputer 3 High Performance Cluster Computing system owing to the performance of the 10GbE solutions that will enhance DAS-3 performance. It includes traffic management features and ultra-low latency and matches the requirements of ASCI for high throughput 10 GbE server connectivity for powering high-definition visualization, said company sources.

ASCI research team leverages the ASCI DAS-3 which runs over a country-wide grid computing access network. Team members dispersed across four universities can collaborate on different aspects of parallel, distributed and grid computing as well as large-scale multimedia content analysis.

The BLADE RackSwitch G8124 24-port 10GbE switch enhances the access network which features StarPlane, a high-speed wide-area interconnect, based on dedicated 10GbE lightpaths from SURFnet. The rich layer 2 and layer 3 network features high throughput and low latency besides offering more control over StarPlane links between the five sites. 

Company officials said that the BLADE RackSwitch Active Multi-Path function will support segregation of traffic for optimized usage of lightpaths eliminating the need to block links and force traffic on intermediary hops.

The VU University of Amsterdam is building a high-end visualization infrastructure which will connect to DAS-3. This forms part of the Intertain project at the university to be served by BLADE RackSwitch G8000 switch connecting to an ultra-HD video system that can be assessed and shared between participating sites over the StarPlane network.

Henri Bal of ASCI said, that DAS-3 is an ambitious High Performance Distributed Computing, or “HPDC,” environment prototyping various novel protocols and systems for the processing and sharing of very large amounts of data on a computational grid.

“In the large-scale use cases investigated on our system, often much higher bandwidths, lower latencies and more consistent network performance are required than would be typical for enterprise environments,” Bal said. “For our project, we chose the BLADE RackSwitch equipment specifically based on their rich functionality combined with an excellent price/performance.”

Charles Ferland, vice president of EMEA, BLADE Network Technologies said that BLADE’s success at ASCI is an example of how BLADE’s RackSwitch family caters to the needs of users who have high bandwidth and application-intensive requirements. It enables them to fully realize the benefits of HPC-class 10GbE networks more affordably and efficiently.