HPCwire: Enabling Low-TCO Supercomputers Through 'Scalable Units'

The Leading Source for Global News and Information Covering the Ecosystem of High Productivity Computing

Leading HPC
Solution Providers

HPCwire >> Off the Wire

November 09, 2007

Enabling Low-TCO Supercomputers Through 'Scalable Units'

The quest for greater performance at lower costs has propelled clusters to the forefront of high-performance computing. Still, HPC users are demanding continually better reliability, availability, flexibility, compatibility, power efficiency – and of course price/performance. Responding to these demands, Appro has introduced the Appro Xtreme-X(TM) Supercomputing Series, which takes advantage of the concept of "Scalable Units" to create next-generation clusters with supercomputer performance.

"High-performance clustering technologies are maturing, and scalability and availability are becoming crucial elements for critical business functions in HPC organizations," says Jim Ballew, Appro's CTO. "In designing our new Xtreme-X Supercomputer Series, Appro has taken the cluster concept to the next plane of both performance and cost advantages."

Question: What is a Scalable Unit for HPC – and how can you call this a Supercomputer?

Ballew: Scalable Units (SUs) are basic building blocks needed to meet specific computational requirements. With this basic modular concept, clusters of various sizes can be built and deployed very rapidly. The SU concept is based on a flexible architecture that groups together clusters with the same configuration to make a unified, fully integrated single "entity" that can be provisioned and managed as a stand-alone supercomputer system to deliver performance up to 1000TF.

Question: What are the advantages of this approach?

Ballew: One main advantage is total cost of ownership (TCO). By making a volume purchase of SUs and using them for a specific job, you can achieve economies of scale that exceed the purchase of a single large system.

Question: How are Scalable Units different from "regular" clusters?

Ballew: It has to do with the way SUs can be grouped together – assembled to make a unified, fully integrated single entity. At the same time, you gain all the benefits of clusters. You will see results in real time. The sum of multiple SUs unified in this way is greater than the value of its cluster parts added together separately.

Question: How does Appro apply this Scalable Unit concept in the Xtreme-X Supercomputer series?

Ballew: The Appro Xtreme-X Supercomputer series offer a next-generation platform based on SU clusters, a "building block" IT infrastructure with a cost-effective, energy-efficient and scalable architecture. The Appro Xtreme-X series is designed to scale out data centers for medium to large-scale enterprises and HPC deployments ranging from 6TF to more than 1000TF of computing power. It arrives as a fully integrated solution, completely pre-packaged, which simplifies the configuration for a particular computational requirement.

Question: How does the Xtreme-X series address the growing HPC market?

Ballew: It's important to note that the HPC market is not only growing, but also maturing, so customers are demanding more of everything – performance, manageability, reliability, availability, flexibility and TCO.

The Appro Xtreme-X Supercomputing Series addresses the need for higher performance and higher reliability by offering redundant (Dual Rail) InfiniBand and Ethernet management fabrics and network switches. All critical components are easily accessible, hot-swappable and redundant, for better reliability and easy serviceability.

We have leveraged blades as the node of choice and built the Appro Xtreme-X on standards-based technology. In particular our first model, the Xtreme-X1 supercomputer, is based on x86 Intel(R) Xeon(R) processors to deliver significantly reduced TCO. Also, these supercomputers can be configured with more processors per rack, which reduces a system's footprint.

Our energy-efficient (green) clusters come pre-installed with a complete lights-out management solution – the Appro Cluster Engine(TM) Software Management – to meet the demands of high-capacity computing and enterprise-level system availability. The management modules include Fabric Management, Server Management, Cluster Management and Resource Management. In addition, the management software supports multiple networking topologies to achieve maximum reliability, performance and high availability, including multi-rail networks with load balancing and nonstop network failover.

The Appro Xtreme-X Supercomputer is ideal for demanding HPC applications that require high reliability and excellent system management.

Question: How do you see the adoption of Windows in large high-performance computing environments?

Ballew: Clustering applications have traditionally evolved from the UNIX environment, where in the old days, every clustering application ran on a UNIX machine. Now, users can migrate their cluster applications from UNIX to Linux with minimal effort. This capability has fueled the rapid adoption of Linux clustering in HPC. Windows traditionally comes from the desktop, but Microsoft has continued to invest in HPC. Windows is more popular than Linux in the enterprise market. As enterprise applications begin to take advantage of clustering, we will see more implementations of Windows-based cluster systems in enterprise HPC.

Question: According to market analysts, InfiniBand saw rapid adoption in HPC markets in 2007 and is expected to continue to grow tremendously throughout 2008 and beyond. How does clustering with InfiniBand compare to other types of clustering solutions, and what are its advantages and disadvantages?

Ballew: InfiniBand provides a high-performance interconnect built on open standards that are supported by multiple vendors. It offers high bandwidth and low latency for cluster computing and is running on systems with thousands of nodes. As an enterprise interconnect, it is able to run multiple protocols on the same connections, thus providing connectivity to existing storage and network equipment in the data center.

InfiniBand is currently a less expensive solution than 10GbE, with twice the bandwidth. So the user benefits from a high-speed network without having to entirely upgrade the data center.

On the other hand, bulky cables and high cost have been disadvantages for InfiniBand. However, cost is decreasing as the market starts to embrace this technology, and cabling can be minimized or replaced with low-cost fiber optic connections – which is what Appro has done in the Xtreme-X series. The Xtreme-X combines InfiniBand, PCI-Express, Mellanox ConnectX adapters, Intel fiber optic interconnects between racks and multi-core processor technologies to deliver outstanding performance, scalability and reliability.

Question: In addition to the economies of scale of SUs and blade architecture, how else does the Xtreme-X architecture lower TCO for customers?

Ballew: Appro's Xtreme-X Supercomputer reduces TCO through things like using less floor space and improved power and cooling efficiency, all of which translate into cost savings over time.

The Xtreme-X features a new Directed Airflow(TM) cooling configuration designed to reduce data center floor space while maximizing cooling efficiency and power consumption. This cooling system features front access for maintenance, allowing racks to be mounted back-to-back to take advantage of the under-floor cooling available in most data centers and delivering the cold air directly to the equipment to be cooled. Delivering cold under-floor air directly to the equipment eliminates the aisle between the racks. This improves density by 30 percent – which saves 30 percent on floor space.

Within the blades themselves, board components are positioned so that processor-heated air flows over the fewest components and is ejected from the blade through the front access panel. More than half of this front panel is perforated to minimize turbulence and reduce pressure. Independent of the rack configuration, Xtreme-X also has three independently controlled cooling fans per node for redundancy and to ensure airflow efficiency.

And speaking of efficiency, Appro Xtreme-X also comes with "Power-Factor-Corrected" power supplies that are 90 percent efficient compared to the 70 percent to 80 percent efficiency typical of power supplies used in current equipment.

Question: When and how will the Xtreme-X series be available?

Ballew: The first model in the series – the Appro Xtreme-X1 – is based on dual-socket, Quad-Core Intel Xeon processors. It features a reliable and scalable architecture with 128 nodes/512 Xeon processors and 6TF of computing power in a single 42U equipment rack. It offers up to 4TB of memory, hot-swappable blades and redundant cooling fans and a robust management system. Appro will showcase its new supercomputer at the SC07 conference and expo, in booth #714 at the Reno-Sparks Convention Center.

In the first half of 2008, we plan to introduce the X2 model, based on AMD Opteron(TM) processors.

In addition, as part of the Xtreme-X series and based on the same concept of Scalable Units for building Supercomputers Linux clusters, Appro offers the Xtreme-X supercomputer configuration based on the Tri-Labs (TLCC07) awarded procurement. For more information about this configuration, please contact an Appro sales representative.

To accompany this new series, Appro is also introducing Xtreme-X Services, delivering closed-loop solution management, up-front consulting, pre-integration and testing with pre-sales and post-sales support. Appro's Xtreme-X Services will support Linux operating environments, with configurations that provide exceptional performance across a wide range of HPC applications.

Question: How do you see Appro's future in the HPC market?

Ballew: IDC projects that the overall technical server market will continue to show strong growth over the next five years, but at a more moderate level of 8 percent to 12 percent a year. They say the cluster portion of the HPC market will likely grow at a higher rate during this period, in the 14 percent to 20 percent range, topping out at 70 percent to 80 percent of the HPC market.

Appro has established a good foothold and reputation in the HPC market. With the introduction of the Xtreme-X Supercomputer Series, Appro is building on our strong track record. By marrying a rich feature set and management capabilities with strong price/performance, the Xtreme-X offers an affordable supercomputer with considerable flexibility and innovation.

Appro will continue to focus on HPC. We believe that the Xtreme-X Supercomputer Series will go a long way toward addressing the needs for greater performance, reliability, availability and TCO for compute-intensive applications for both HPC and commercial markets.

Article Tools

Share & Save Options

Discussion

There are 0 discussion items posted.

Join the Discussion

You must be a registered user in order to comment on HPCWire stories.

Members enjoy:

	Access to Special Content, including Podcasts, Webcasts, and Whitepapers and more!
	Ability to take an active role in the discussions that are shaping the HPC technology discipline!
	Special offers available only to HPCWire members!

Become a Registered User Today!

Registered Users Log in Here to Comment

Email Address:	Password (case sensitive)

Remember me	Forgot Password?

RSS Feeds

Feeds by Topic

Feeds by Industry

Feeds by Content Type

Feature Articles

The Linux HPC Empire Strikes Back

While the Microsoft juggernaut has been touting the joys of its new Windows HPC Server 2008, the Linux HPC contingent has been somewhat less vocal of late. But now Red Hat has come up with its version of an integrated cluster solution.
Read More...

Nexsan Looks to Scare Up HPC Customers With Storage Beast

Even though the cost of servers still dominates the datacenter budget, storage is actually on a steeper growth curve. HPC storage, in particular, is being singled out as high-growth opportunity. Vendors are scrambling to keep up.
Read More...

The Week in Review

Google datacenters most energy efficient; Cluster Resources to demo Moab Hybrid Cluster; Red Hat Linux releases HPC distro. John West recaps those stories and more in our weekly wrap-up.
Read More...

Read more HPCwire features...

Top Headlines

Oracle and HP's Database Machine Predicated on Voltaire

Oct 06 | The Register | Does the HP Oracle Database Machine represent InfiniBand's big chance to break out its HPC niche? Read more...

3D Imaging Spreads to Fashion and Beyond

Oct 06 | BusinessWeek | A body scan can save a lot of time in the fitting room, and fields from medicine to architecture are adopting 3D computing applications. Read more...

Structural Engineers and Computer Scientists Hope to Integrate Disciplines to 'Revolutionize Building Construction'

Oct 03 | UCSD News | Despite the evolution of computer science over the past 30 years, structural engineering -- hindered by a reluctance to adapt to digital innovations -- has remained relatively unchanged as a discipline. Read more...

Credit Crisis Spreads a Pall Over Silicon Valley

Oct 02 | New York Times | Silcon Valley is starting to feel the effects of the credit crunch. Read more...

Google: 'The World's Most Efficient Data Centers'

Oct 01 | Data Center Knowledge | Google today disclosed details of its data center energy usage, confirming that it operates some of the most efficient facilities in the world. Read more...

Read more headlines...

Featured Whitepapers

Panasas® Tiered Parity™ Architecture

Sep 04 | | Disk drives are approximately 250 times denser today than a decade ago. This is good news for users who are creating, manipulating and storing more data than ever before. It gives them an opportunity to derive more value from their stored data and lowers the capital acquisition and operating expense associated with that data.

View the White Paper Library

Multimedia

Video White Paper: Architecting a Better Network Storage Solution

BlueArc's Titan architecture represents an evolutionary step in file servers by creating a hardware-based file system that can scale bandwidth, IOPS, and overall data capacity well beyond conventional software-based devices. With its ability to virtualize a massive storage pool of up to four usable petabytes of tiered storage, Titan can scale with growing data requirements, offering a competitive advantage for businesses, researchers, or other enterprises seeking to better manage data growth while still ensuring optimal performance.