Home » High Performance Computing » Hadoop Clusters » Supermicro Hadoop Cluster Shakes up the HPC Industry

Share Button



Hadoop Cluster SystemsThe latest High Performance Computing offerings from Supermicro seem to always fill well-needs with their own respective niches. Supermicro’s latest Hadoop Cluster offering is no exception. An enterprise Hadoop Cluster is the best way ever to harness tremendous high performance computing power. Hadoop Clusters work in tandem by combining the horsepower of several computers to complete the most, compute-intensive challenges around the world. The problem is that all of these computers require energy, space and come at a high cost. Supermicro FatTwin, servers and storage alleviate the sprawl, heat, energy costs and waste of high performance cluster computing by combining several high-performance compute nodes in one, cost-effective, energy-efficient unit. How does Supermicro do it? They do it through the manufacturing of the most innovative, multi-node chassis designs built with the latest, state-of-the-art processors and components.

Hadoop Cluster Servers overview from Supermicro

Supermicro’s latest Hadoop Cluster line-up features 1U Twin, 2U Twin2 and 4U FatTwin chassis which fit up to 2, 4 and 8 Hot-Plug system nodes in each. These new high performance computing nodes not only represent a great savings in real-estate, but also offer revolutionary power-savings and cooling. Supermicro’s entire line-up of Twin and FatTwin Hadoop Clusters are built to reduce TCO in Data Center environments while helping to preserve our environment, 24 x 7. Here is a brief run-down of todays Hadoop Custer-ready offerings.

Hadoop Cluster Models and Specifications

SYS-F617H6-FTPTL+ and SYS-F617H6-FTPT+ feature a total of 4 each, Hot-Swap system nodes inside of a 4U Chassis. These nodes are powered by 1620W Redundant Power supplies. Each node can be configured with Intel’s very latest E5-2600 or E5-2600 V2 Family of processors. Each node supports 256GB of memory for a total of 1TB. For connectivity, the F617H6 FatTwin series is uses Intel’s 10GBase T LAN ports via the X540 controller. ¬†Perhaps, the most attractive capability is that the chassis supports up to 8+4 3.5″ Fixed SAS/SATA / ¬†SATA HDD’s per node and are controlled by LSI’s popular, high-performance 2208/2308 Hardware RAID controller which supports RAID levels 0, 1, 5 , 6, 10, 50 and 60.

The SYS-F617H6-FTL+ and SYS-F617H6-FT+ basically utilize the same chassis with a few minor differences: These units include Intel’s Dual i350 Gigabit Ethernet Controller.

That brings us to one more unit: The SYS-F517H6-FT is specifically designed to be Hadoop Cluster-ready. The F517H6-FT also supports 4 Hot-Plug nodes and 8x 3.5″ Fixed SAS/SATA + 4x 3.5″Fixed SATA + 2 x 2.5″ HD’s per unit. The F517H6 series takes efficiency and savings a step further by utilizing the E3-1200 V2 family of Intel processors. The F517 FatTwin series supports the latest PCI-E 3.0 x 16 and x 8 architecture via expansion slot riser cards. There is onboard, Intel i350 LAN Ports using the as well as Dedicated LAN over IPMI 2.0. These Hadoop Cluster units are driven by LSI’s 2308 AOC 6GBps controller.

All of Supermicro’s Hadoop Cluster-ready servers are able to fit inside of a standard, 19″ Rack Cabinet.

Who uses Hadoop Cluster Systems from Supermicro?

The short answer is that any computing intensive environment is the perfect place for a hadoop cluster solution. The more complete answer is that a number of occupational industries seek and realize the benefits of Supermicro FatTwin architecture and Hadoop-ready systems. The list includes, scientists, engineers, rendering farms, cloud computing and data centers, enterprise IT departments, file systems, storage, GPGPU applications, research, and more. The FatTwin is the answer to today’s latest Hadoop Cluster environment.

About

Leave a Reply

Your email address will not be published. Required fields are marked *

You may use these HTML tags and attributes: <a href="" title=""> <abbr title=""> <acronym title=""> <b> <blockquote cite=""> <cite> <code> <del datetime=""> <em> <i> <q cite=""> <strike> <strong>

Google+ Plus Like On Facebook Linked Follow Email to us

Share

Share Button