TMCnet LOGIN
Webinars
TMCnews Featured Article
November 24, 2009
NextIO's IBM Cluster Solutions to Sport Virtualization Technology
By Carolyn J Dawson, TMCnet Contributor
NextIO, a provider of next-generation I/O virtualization solutions for modern day data centers in a multitude of industries including enterprise, telecommunications, high performance computing, digital media and financial services, recently announced that it is collaborating with IBM (News - Alert) to deliver an integrated set of cluster solutions.
The new set of cluster solutions will bring together NextIO technology and virtualization technology. The company has set its sights on making the product available by 2010. The solution will help provide reconfigurable and on demand GPU compute technologies for IBM iDataPlex users. NextIO had made this announcement during the 2009 edition of the Supercomputing Show which was held at Portland, Oregon. NextIO was demonstrating its offerings at IBM booth number 1335 and its own booth number 2585 during the show.
The solution has been built using the reference architecture derived from NextIO’s I/O virtualization technology. This makes the platform more dynamic, proactive, scalable, open and energy optimized as it will be able to prop up several kinds of GPUs. IBM and NextIO will provide a modular “hot plug” sleds for clearing anomalies or upgrading GPUs without allowing the system to have a downtime.
This is only one of the several technologies planned to be incorporated by NextIO. The GPU virtualization offering will provide the capability for one IBM iDataPlex server to gain entry to anything between one and eight double wide GPUs or even up to 16 single wide GPUs inbuilt in the device. Customers will also be able to rapidly and easily facilitate more or less GPU resources on demand, in keeping with the system requirements. Each iDataPlex array will be able to prop up 10 GPU devices resulting in performance levels of up to 160 GPUs and more than 80TFlops of compute processing per rack.
IBM has reaffirmed its commitment to delivering its iDataPlex customers with the most optimal HPC solution for safeguarding their investment keeping the overall cost of ownership as low as possible.
In a release, Alex Yost, vice president of IBM Systems & Technology Group said “Customers in HPC now have the performance, density and power savings of iDataPlex available for applications that require GPU computing.”
K.C. Murphy, founder and CEO of NextIO, remarked about the collaboration with IBM, “This helps validate NextIO’s leadership in I/O virtualization and will help drive rapid adoption of NextIO-based solutions in the HPC market. NextIO looks forward to continuing our work with IBM in rolling out a joint solution in 2010.”
The new set of cluster solutions will bring together NextIO technology and virtualization technology. The company has set its sights on making the product available by 2010. The solution will help provide reconfigurable and on demand GPU compute technologies for IBM iDataPlex users. NextIO had made this announcement during the 2009 edition of the Supercomputing Show which was held at Portland, Oregon. NextIO was demonstrating its offerings at IBM booth number 1335 and its own booth number 2585 during the show.
The solution has been built using the reference architecture derived from NextIO’s I/O virtualization technology. This makes the platform more dynamic, proactive, scalable, open and energy optimized as it will be able to prop up several kinds of GPUs. IBM and NextIO will provide a modular “hot plug” sleds for clearing anomalies or upgrading GPUs without allowing the system to have a downtime.
This is only one of the several technologies planned to be incorporated by NextIO. The GPU virtualization offering will provide the capability for one IBM iDataPlex server to gain entry to anything between one and eight double wide GPUs or even up to 16 single wide GPUs inbuilt in the device. Customers will also be able to rapidly and easily facilitate more or less GPU resources on demand, in keeping with the system requirements. Each iDataPlex array will be able to prop up 10 GPU devices resulting in performance levels of up to 160 GPUs and more than 80TFlops of compute processing per rack.
IBM has reaffirmed its commitment to delivering its iDataPlex customers with the most optimal HPC solution for safeguarding their investment keeping the overall cost of ownership as low as possible.
In a release, Alex Yost, vice president of IBM Systems & Technology Group said “Customers in HPC now have the performance, density and power savings of iDataPlex available for applications that require GPU computing.”
K.C. Murphy, founder and CEO of NextIO, remarked about the collaboration with IBM, “This helps validate NextIO’s leadership in I/O virtualization and will help drive rapid adoption of NextIO-based solutions in the HPC market. NextIO looks forward to continuing our work with IBM in rolling out a joint solution in 2010.”
Carolyn John is a Contributor to TMCnet. To read more of her articles, please columnist page.
Edited by Patrick Barnard
