Abstract: In one embodiment, a method for on-demand instantiation in a high-performance computing (HPC) system includes receiving a connection request from a client specifying a first port number and a first host name advertised externally with respect to an HPC server including a cluster of nodes, identifying a service at the HPC server corresponding to the first port number and the first host name, determining whether the identified service is available, and, if the identified service is available, instantiating a host providing the identified service at one or more nodes in the cluster.

Abstract: In one embodiment, a method for fault tolerance and recovery in a high-performance computing (HPC) system includes monitoring a currently running node in an HPC system including multiple nodes. A fabric coupling the multiple nodes to each other and coupling the multiple nodes to storage accessible to each of the multiple nodes and capable of storing multiple hosts that are each executable at any of the multiple nodes. The method includes, if a fault occurs at the currently running node, discontinuing operation of the currently running node and booting the host at a free node in the HPC system from the storage.

Abstract: Cluster management software comprises a plurality of cluster agents, with each cluster agent associated with an HPC node including an integrated fabric and the cluster agent operable to determine a status of the associated HPC node. The software further includes a cluster management engine communicably coupled with the plurality of the HPC nodes and operable to execute an HPC job using a dynamically allocated subset of the plurality of HPC nodes based on the determined status of the plurality of HPC nodes.

Abstract: A method for managing HPC node failure includes determining that one of a plurality of HPC nodes has failed, with each HPC node comprising an integrated fabric. The failed node is then removed from a virtual list of HPC nodes, with the virtual list comprising one logical entry for each of the plurality of HPC nodes.

Abstract: A method for computer cluster virtualization includes selecting a distributed application. A policy associated with the distributed application is retrieved. One of a plurality of nodes is dynamically selected. Then, a boot image of the selected node is reset based, at least in part, on the retrieved policy, with the boot image being compatible with the distributed application. Then, a virtual disk image is associated with the node. At least a portion of the distributed application is then executed on the reset node using the associated virtual disk image.

Abstract: A method for providing a graphical user interface in a High Performance Computing (HPC) environment includes collecting information on a plurality of HPC nodes, with each node comprising an integrated fabric. A plurality of graphical elements are generated based, at least in part, on the collected information. At least a portion of the graphical elements are presented to a user.

Abstract: A method for job management in an HPC environment includes determining an unallocated subset from a plurality of HPC nodes, with each of the unallocated HPC nodes comprising an integrated fabric. An HPC job is selected from a job queue and executed using at least a portion of the unallocated subset of nodes.