HPCwire
 The global publication of record for High Performance Computing - LIVEwire Edition / November 18, 2003: Vol. 10, No. 1

  |  Table of Contents  |  

Features:

INTERVIEW WITH DANIEL KIM, CEO, APPRO
By Alan Beck, Editor-in-Chief, HPCwire

HPCwire: Appro has had some pretty exciting announcements lately with the new Texas A&M, Nasa JPL and other major wins. How do these deals position Appro in the market place?

DANIEL KIM: Appro has made significant deals in the supercomputing market that have demonstrated the company's innovation in commodity cluster computing. For example, Appro built this past month two Appro HyperBlade Linux Clusters to support multiple research activities at Texas A&M University for hydrology, groundwater and oil and gas reservoir simulations. The clusters consisted of 130 nodes with a total of 260 AMD Opteron processors, including remote management capabilities. Our BladeDome Remote Management software provides node power control and platform monitoring making it easy for them to manage and control individual nodes. Appro is well positioned to deliver open standards commodity based, flexible and scalable HPC solutions based on the latest processor technologies from AMD and Intel to our targeted customers.

HPC: How about the NASA JPL win, and how does this win relate to Raytheon Company?

DK: Raytheon HPC Solutions Group added Appro HyperBlade cluster solutions to its HPC offerings this April. This arrangement was based on a common goal of selling and deploying Appro's integrated HPC solutions and Raytheon's value added services to address multiple vertical markets. This marks another milestone in delivering enterprise class products and services into this rapidly growing market. For the NASA JPL win, a combined team of Raytheon and JPL engineers determined the optimum architecture based on software and throughput requirements to be Opteron chips running Linux in a cluster. Raytheon openly competed the procurement and chose Appro 160-node 1U Linux clusters based on dual AMD Opteron processors using Gigabit Ethernet interconnect as the best value vendor for this configuration.

HPC: What are Appro's vertical markets?

DK: Appro's vertical market includes government, universities, petroleum, bio/life sciences, digital content creation and enterprise computing for a wide variety of applications that require real-time and computationally intense performance. We distribute our products through dual channels. We sell directly to enterprise and educational end-users and through channel partners such as value-added reseller (VAR), strategic partners and system integrators. Appro's customers include Fortune 1000 companies, major universities, large system integrators in the government market and international channel partners.

HPC: How does Appro differentiate itself from other cluster vendors?

DK: There are three things that differentiate Appro: Product innovation, strategic technology partnerships, and leadership in price/performance value. We offer superior engineering capabilities and clever implementation of commodity parts. Our strategic partnerships with AMD, Intel, Infiniband Consortium, Raytheon, Computer Science Corporation and Uniwide Technologies allows us to expand our product footprint and address the needs of different industry verticals. We're always looking ahead to what's next in the HPC market and what innovations would help make our customers more successful.

HPC: We hear so much about the Appro HyperBlade Cluster solutions. What impact does this product have in the High Performance Computing industry?

DK: Our Appro HyperBlade Cluster Solutions are raising the bar for HPC commodity clusters. With support for up to 80 compute blades and up to 160 Intel Xeon or AMD Opteron Processors, Appro HyperBlade architecture doubles the current rack density using 1U servers. Density-managed architecture is achieved by highly effective thermal design and using commodity x86 components in a single cluster. Because Appro HyperBlade Clusters are flexible, modular scalable architecture, they are easy to deploy in any data center. In addition to the HyperBlade, Appro offers BladeDome Cluster Management Solution, a hardware and software monitoring and management tool that gives customers in-depth visual status on all their hardware subsystems. As options, we offer high-speed interconnect and a choice of Linux or Windows operating systems. Appro is constantly pushing the limits of delivering uncompromising quality while offering an excellent price to performance ratio.

HPC: Appro has been demonstrating the BladeDome Remote Management Software with the Appro HyperBlade clusters in many trade shows. How do you compare Appro's Cluster Management Solution with other similar server monitoring and management tools available?

DK: Appro BladeDome remote management software is integrated with Appro Command Center, cluster management appliance. Appro Command Center combines the functionalities of serial console server and intelligent power management PDU, to provide both serial and power management capabilities in one. For example, for a typical 80-node cluster, a customer would need to purchase remote management software, 2x APC NetShelter VX42U, 2x console server, 5x APC Master Switch and an AMI server management card. With an 80-node Appro HyperBlade Cluster, customers will only need to purchase the one Command Center appliance and the 80 Appro BladeDome software license fee. Most of the console server functionality needed for the typical 80-node cluster is already included and integrated in the Appro Command Center. Appro's BladeDome monitoring and management solution delivers cluster management functions, reduces IT complexity and improves administration efficiency while lowering the total cost of ownership.

HPC: How does Appro ensure its hardware is compatible with Linux and Open Source software?

DK: We test each component before we configure them for use. The components are tested for full interoperability with each other and with Linux. We use open source stress test software that also runs on Linux. We run stress tests on each server for at least 48 hours to make sure the server is stable and reliable before it ships.

HPC: What challenges did Appro face to get where you are today?

DK: The main challenge Appro faced was to change our business model. Previously, we were one of the largest OEM rackmount computer systems manufacturers in the United States. Now, we produce and market our own line of Appro-branded high-density servers for the high performance computing and Internet computing markets. Today, customers rely on Appro as a leading developer of high-performance computing solutions. Our customer base now consisted of government, universities, petroleum, bio/life sciences, digital content creation and enterprise computing markets.

HPC: So what's next for Appro?

DK: Appro will continue to play a key role in the high performance computing market. We also expect that Linux use will increase Appro's market share. Linux will empower users to develop their cluster systems in an affordable ownership model. The benefits of server clustering include higher performance and scalability and to a lesser extent, investment protection and simplified administration. Appro will continue to deliver cutting edge server cluster technology to our customers. Early next year, Appro will also offer storage solutions to further enhance our overall high-performance computing product offerings.


Top of Page

  |  Table of Contents  |