By Michael O'Neill (NVIDIA), Brandon Harrell (IBM), Deana Coble (IBM)
The word strength isn’t typically mentioned as a component when designing and building information systems. Words and phrases like fault tolerance, stability, consistency, robustness, or speed are often used to describe the goals of a design. However, this concept of strength is a characteristic that should be included in everything we do.
When architects build a skyscraper or a bridge, they consider and test the strength of the materials to insure against failures under heavy load and provide longevity for the infrastructure. IBM NeXtScale System and NVIDIA GRID bring this kind of strength to datacenter design. Together, let’s build something strong for your future enterprise!
The PCIe Native Expansion tray with two NVIDIA Tesla GPU cards in a single node
Future generations will look back and deem the era we currently live in as “The Information Revolution”. We collect, analyze, and use more data for general, daily decision making than any generation previous. The steady flow of information is what makes us more productive and allows us to make better decisions. The proliferation of compute heavy and graphical content continuously drive us forward to break new ground for more processing power in a smaller footprint.
The previous 10 years have shown us the rise of server, application, and desktop virtualization. This rise is widely contributed to the desire to more fully utilize servers and datacenters running completely occupied but at less than capacity. Now with the popularity of user-facing workloads being virtualized via companies like Citrix, Microsoft, and VMware, success is defined by an entirely new dynamic – user acceptance.
The user experience of today hinges on crisp graphics, snappy response, and multiple sources of content running simultaneously. Today’s general office productivity computing environment leverages technologies like DirectX, OpenGL and CUDA to deliver the crisp, responsive experience that users demand. These technologies depend on NVIDIA GPUs to drive the experience. Tomorrow’s requirements will most certainly demand more and NVIDIA has built strength into its design with technologies like GRID and vGPU.
How does IBM NeXtScale System build strength for the future? The same as when you are looking at a building for your business - you evaluate the rooms, the infrastructure of the building, its ability to flex with your needs, and its ability to grow with your business.
IBM NeXtScale System Chassis doubles the density for compute nodes
The architectural design and capabilities of the IBM NeXtScale System and chassis is flexible, scalable, and focused on performance, availability, and serviceability. The strength of this system is proven in these features:
- The PCIe Native Expansion Tray can be added to the nx360 M4 to form a powerful compute engine, supporting two NVIDIA GPUs within a single node
- Adding compute capability, or later storage or acceleration capability, is as simple as adding specific nodes to the chassis. A typical rack holds only 42 1U systems, but this chassis doubles the density (up to 84 compute nodes within the same footprint)
- Greater system uptime is created as the NeXtScale n1200 chassis supports N+N and N+1 power policies for its six power supplies, and reliability is enhanced by using SSD’s instead of traditional mechanical HDDs.
- All components can be removed from the front of the rack by sliding out the trays or the chassis for easy, quick servicing
- The nx360 M4 offers memory mirroring for redundancy if there is a non-correctable memory failure, and optional RAID arrays enable the server to keep operating if there is a failure of any one drive
When thinking of strength, the tendency is to consider it something concrete, physical, and tangible. However, IBM and NVIDIA recognize and create it in the virtual world as well. The IBM NeXtScale System uses Intel Virtualization Technology to integrate hardware-level virtualization hooks that allow operating system vendors to better use the hardware for virtualization workloads.
NVIDIA GRID and vGPU are key components in building strength into your design and providing future proofing for the unknown workloads of tomorrow. NVIDIA has engineered virtualization specific features into the GRID K1 and K2 cards. These features allow you to inject and share GPU resources into virtual environments and also allow for providing more (or less) resources relatively easily.
NVIDIA GRID K1 and K2 cards
Whatever your enterprise need is, IBM and NVIDIA can work with you to build a strong hardware and virtual solution for today and for the future.
Next week will be the last week of this blog series on GPU accelerated data centers, our recent enhancements, and what does it mean to you – but even when the series is done – we are still here to help and are available for your questions at the IBM support portal and NVIDIA support.
Michael O’Neill is an established leader for NVIDIA and provides specialized technical guidance to customers on NVIDIA GRID and Tesla GPUs in virtualized environments. He works closely with IBM to develop innovative solutions for graphical and compute heavy workloads.
Brandon Harrell is an IBM Technical Solutions Leader. He specializes in x86 Desktop and Application Virtualization as part of the Advanced Technical Sales team. Follow Brandon on Twitter at @BrandonEHarrell.
Deana Coble is an IBM Redbooks Technical Writer. She has worked over 20 years in the field of technology and collaborates with IBM subject matter experts to create IBM Redbooks publications on a variety of systems. Follow Deana on Twitter at @DeanaCoble.