Windows CCS offers a cost-effective HPC approach that is easy to implement, use and manage so that more users in more organizations can reap the computing benefits without the cost or complexity.
As the benefits of HPC have grown, computing clusters have sprouted in diverse industries. What was once a specialized technology quickly showed benefits for manufacturing and engineering organizations, financial and banking firms, petroleum enterprises and bioscience institutions.
Even unconventional industries are taking advantage of HPC, using it for everything from solubility testing of disposable diapers to tensile strength testing of plastic packaging to improved routing for trucking companies seeking to reduce fuel costs. All of these applications likely would not have been considered in the past as a result of the traditional cost and complexity of HPC.
“Companies in every industry can reap the benefits of high-performance computing capabilities,” says Dan Cox, Manager of High-Performance Computing Programs for the Industry Standard Server group at HP. “Some organizations regularly run into capacity and bandwidth problems and need high-performance computing systems, whereas others have a huge opportunity in front of them but just don’t know it yet.”
Challenges with traditional HPC
“More and more businesses are benefiting from high-performance compute clusters, but traditional high-performance computing is still a dedicated discipline,” warns Ed Turkel, Manager of Product Marketing for the High-Performance Computing Division at HP. “Systems built on Linux® and UNIX® require particular skills to design, deploy, manage and use.”
Not every organization has the necessary skill sets to implement and maintain HPC systems and tools. And of the companies that have successfully deployed Linux- or UNIX-based clusters, not everyone in the organization has easy access to, or knowledge to use those HPC resources.
Regardless of platform, Turkel says the top frustrations of cluster users are:
- cluster management capability and usability
- complexity of system design
- acquisition and deployment
- interconnect performance
- parallel algorithm implementation
- application availability.
HPC moving mainstream
As HPC systems become more commonplace, organizations are realizing that their users want the power these systems provide, but in a familiar operating environment. “Engineers and analysts just want to get their job done,” Cox claims. “Financial analysts are concerned with calculating risk, not learning the intricacies of Linux. Geologists are focused on oil exploration, not how to access or administer a UNIX cluster. In order for high-performance compute systems to be effective, they need to be accessible and easy to use—for everyone.”
Many users are comfortable with the Windows® Server technology platform. This is why the Windows® Compute Cluster Server (CCS) 2003 is becoming a popular solution that alleviates many of the challenges associated with traditional HPC deployments.
“There are a large number of Windows users who are currently running their compute-intensive workloads on their workstations and desktops,” says Norma McKinney, Senior Product Manager, Windows Server division at Microsoft Corp. “Pushing these workloads to high-performance clusters running Windows CCS will significantly improve their productivity; sometimes by an order of magnitude. Also, most users are already familiar with the Windows environment and will need very little, if any, training to get up to speed with Windows CCS tools.”
She adds that there are several new technologies and tools for distributed computing, such as the Microsoft® .NET Framework, Excel Services and third-party development tools, which map well into Windows-based clusters. Altogether, Windows CCS offers a new, cost-effective paradigm for HPC that is easy to implement, use and manage.
Getting the right solution
According to Guodong Zhang, Product Manager for High-Performance Computing at HP, HP can help companies maximize HPC opportunities. For the past three years, HP has ranked tops in HPC, based on revenue.* “Our decades of experience in high-performance computing has enabled us to develop and offer exceptional hardware products, unified cluster solutions, architectural design guidance and support options,” says Zhang.
He adds that HP delivers an extensive choice of industry standard servers, interconnects, integrated cluster hardware platforms as well as a choice of cluster management software suites or all-encompassing cluster solutions through the HP Unified Cluster Portfolio. An extensive, modular package of tested and pre-configured hardware, software and services, the HP Unified Cluster Portfolio provides a wealth of options for scalable computation, data management and visualization.
HP also has strong relationships with industry software platform and application providers. “By working directly with engineering teams from companies such as Microsoft and a host of leading software developers, HP’s cluster offerings deliver a completely tested solution from cluster management to running vertical applications,” Zhang says. “There is no such thing as ‘one-size-fits-all’ when it comes to high-performance computing, so we make sure our technologies operate seamlessly with popular operating environments and software applications.”
As HPC moves further into the mainstream, companies need guidance in selecting solutions, configuration, implementation and management. With decades of experience in helping companies achieve their high-performance computing goals, HP is well prepared to play the role of trusted advisor.
* IDC Tracker Study.
|
Printable version
