FAQ

1. What is Platform Open Cluster Stack (OCS) 5?

2. What is Intel Cluster Ready?

A result of collaboration between Intel, independent software vendors (ISVs), and original equipment manufacturers (OEMs) such as Dell™ Inc., Intel® Cluster Ready includes a set of minimum standards for software and hardware components of a clustered system to allow for an unprecedented level of interoperability. Intel Cluster Ready is designed to make it easier to run off-the-shelf High Performance Computing (HPC) software applications. It also includes a software registration process for compatible HPC applications and a complimentary hardware cluster certification tool, Intel® Cluster Checker. This certification tool checks the cluster hardware and software components to ensure they interoperate properly, facilitating fault isolation, helping to improve early detection of cluster problems and reduce support costs. 
 

3. What is the Platform OCS 5 positioning with Intel® Cluster Ready?

Platform OCS 5 is the first commercial software system used to create certified Intel Cluster Ready systems that enable the consistent delivery of scale-out application clusters. Intel selected Platform OCS 5 to create the first certified reference platforms. Intel Cluster Ready represents the next significant milestone and proof-point for the Platform OCS 5 family to enable the delivery of open clusters that are tested and delivered to Intel cluster level specification.

4. What is the relationship between Intel and Platform?

Platform and Intel have long been strategic alliance partners. The two companies have a strong track record of cooperation and have collaborated in the past on a number of key market initiatives. Intel® Cluster Ready is a natural evolution of that relationship to bring market leadership and expertise for successfully supporting the program.

5. What is the business problem that ‘Intel® Cluster Ready’ is trying to solve?

The unsolved problem for clusters is lack of standardization. Readily available commodity components are decoupled and complex, making Linux clusters difficult to build and use. Getting a cluster up and running requires additional skilled resources, services and time. To bring clusters to the mainstream and stimulate growth through the adoption of more applications for workgroups and departmental clusters, the process of building and managing Linux clusters needs to be simplified, minimizing the cost and time to get a cluster fully operational. In order for this to happen there needs to be a comprehensive set of reference architectures, implementations and technology standards that will enable interoperable components and a ubiquitous computing experience.

Intel® Cluster Ready enables a faster deployment out-of-the-box, higher reliability and lower overhead cost.

6. How do I build an Intel Cluster Ready certified system?

A participating vendor is required to physically build a minimum 4 node reference cluster complete with a software stack that meets the Intel® Cluster Ready specification. The vendor then completes a certification form and sends it to Intel along with the output files of the Intel® Cluster Checker test suite software.

Platform OCS 5 can be used to meet the software stack requirements and enable rapid time to market. Platform OCS 5 complies with the Intel Cluster Ready Specification, and together with other hardware and software components, forms a reference implementation for Linux cluster application, deployment and management delivering a complete Intel Cluster Certified solution.

7. What are the key benefits of using Platform OCS for Intel Cluster Ready? 

Please click here to be directed to a list of features and benefits.
 

8. What is Annual Cluster Care?

Annual Cluster Care is a Platform OCS 5 annual support "subscription". Annual Cluster Care includes support, maintenance, upgrades, fixes, access to resources and other value-add services.

9. How is Platform Open Cluster Stack Annual Cluster Care priced?

Platform OCS 5 Annual Cluster Care SRP is US$150 per node, per year.

10. How does a customer get Platform LSF® HPC?

A customer may purchase Platform LSF licenses and deploy them with Platform OCS 5. Standard Platform LSF pricing applies.

11. What network interconnects are supported?

Gigabit, Myrinet®, Ethernet and Infiniband™ are integrated into Platform OCS 5.

12. What is Platform LAVA?

Platform LAVA is an open source workload management solution based on the industry leading workload management product LSF. Platform LAVA is included as part of Platform OCS 5 to enable users to run application in a cluster environment. For users with complex or mission critical workload, we recommend upgrading to Platform LSF.

13. Is the Platform LSF plug-in scheduler facility available in Platform Lava?

Yes it is available, however, Platform LSF plug-ins will not work in Platform Lava.

14. Does Platform OCS 5 or Platform Lava support Platform LSF MultiCluster?

Platform Lava is not compatible with Platform LSF MultiCluster – customers must upgrade to Platform LSF in order to use Platform LSF MultiCluster. However, Platform OCS 5 can be used to deploy Platform LSF MultiCluster.

15. What components will of Platform OCS 5 will Platform Support?

Platform will support all of the components of Platform OCS 5 that install and configure the cluster, in addition to Platform Lava and Platform LSF.

16. What’s contained in Platform Open Cluster Stack 5?

• Red Hat® Enterprise Linux® or CentOS Linux®
• Ganglia
• mySQL® database
• Multiple MPI libraries (MPICH, MPICH-GM, MVAPICH)
• Platform Lava (for workload management)
• Volcano. A simple batch system web portal
• PVFS2
• HPL (High Performance Linpack)
• CluMon
• ntop network monitoring tool
• OFED (OpenFabrics Enterprise Distribution)
• TotalView Debugger
• Nagios
• Other miscellaneous tools and applications.

16. What are Platform OCS "Kits"?

Platform OCS 5 has a set of 'base' components that are always installed on the Cluster. Kits extend the capability of Platform OCS 5 and provide customers the flexibility of choosing the modules that they require to operate their cluster. Kits can be installed at anytime.