Abstract: Sessions states within virtual machine (VM) environments are maintained. Each VM environment hosts a guest operating system (OS) and one or more application programs running. The guest OS and the application computer programs of each VM environment constitute a session. A connection broker maintains a state machine as to states of the sessions and permits transitions among the states in response to messages, commands, and internal decisions. Different types of agents may send the messages, and the commands. The states of the sessions may include a pending state and a number of other states. The pending state is an interim state that indicates a session is being transitioned from one of the other states to another of the other states. These other states may include an offline state, an online-down state, an online-up state, a suspended state, an active state, an idle state, a disconnected state, and a failed state.

Abstract: Direct memory access (DMA) is provided between input/output (I/O) devices and memory within virtual machine (VM) environments. A computing device includes an I/O device, an operating system (OS) running on a VM of the computing device, a device driver for the I/O device, a VM manager (VMM), I/O translation hardware, and a hardware abstraction component for the OS. The I/O translation hardware is for translating physical addresses of the computing device assigned to the OS to machine addresses of the I/O device. The hardware abstraction component and the VMM cooperatively interact to enable the device driver to initiate DMA between the I/O device and memory via the translation hardware. The OS may be unmodified to run on the VM of the computing device, except that the hardware abstraction component is particularly capable of cooperatively interacting with the VMM to enable the device driver to receive DMA from the I/O device.

Abstract: Virtual machines are managed on the basis of data obtained from a management information database of a network switch having a plurality of Ethernet links coupled to compute nodes running a plurality of virtual machines. A management entity, such as a provisioning manager, determines the amount of network bandwidth being utilized through each of the first and second Ethernet links and the amount of network bandwidth being utilized by the Internet Protocol addresses attributable to each of the virtual machines. Accordingly, one of the virtual machines may be migrated from one compute node to another compute node coupled to an Ethernet link having a greater amount of unutilized network bandwidth. Virtual machines may be dynamically migrated in order to provide each virtual machine with a required amount of network bandwidth.

Abstract: Method and computer program for managing the network response times experienced by virtual machines. Traffic is routed within an Ethernet network through a network switch according to IP addresses identified within Ethernet frames. A plurality of compute nodes are each coupled to the network switch via a separate Ethernet link. A response time is determined for each virtual machine running on the compute nodes, wherein the response time for a particular virtual machine is the difference between a time stamp in a TCP/IP Request message and a time stamp in a TCP/IP Response message for an IP address assigned to the particular virtual machine. The particular virtual machine may then be migrated to a target compute node in response to a particular one of the virtual machines on a particular one of the compute nodes having a response time that exceeds a response time setpoint.

Abstract: Virtual machines are provisioned computers in a computer environment based on input/output (I/O) requirements of software tasks. A workload request, requesting the execution of a software task on a virtual machine, is received. The I/O requirements of the software task are matched to an optimal computer, in the computer environment, that has an I/O bandwidth capability that best matches the I/O requirements of the software task. The software task is then routed to a virtual machine, on the optimal computer, for execution of the software task.

Abstract: Sessions states within virtual machine (VM) environments are maintained. Each VM environment hosts a guest operating system (OS) and one or more application programs running. The guest OS and the application computer programs of each VM environment constitute a session. A connection broker maintains a state machine as to states of the sessions and permits transitions among the states in response to messages, commands, and internal decisions. Different types of agents may send the messages, and the commands. The states of the sessions may include a pending state and a number of other states. The pending state is an interim state that indicates a session is being transitioned from one of the other states to another of the other states. These other states may include an offline state, an online-down state, an online-up state, a suspended state, an active state, an idle state, a disconnected state, and a failed state.

Abstract: Sessions states within virtual machine (VM) environments are maintained. Each VM environment hosts a guest operating system (OS) and one or more application programs running. The guest OS and the application computer programs of each VM environment constitute a session. A connection broker maintains a state machine as to states of the sessions and permits transitions among the states in response to messages, commands, and internal decisions. Different types of agents may send the messages, and the commands. The states of the sessions may include a pending state and a number of other states. The pending state is an interim state that indicates a session is being transitioned from one of the other states to another of the other states. These other states may include an offline state, an online-down state, an online-up state, a suspended state, an active state, an idle state, a disconnected state, and a failed state.

Abstract: Power management for a virtual machine farm in which each hypervisor respectively serving each virtual machine platform in the farm is provided with an extended hypervisor function coacts with functions provided by the connection broker and the manual configuration interface of the virtual machine farm management server for managing each respective virtual machine platform to maximize the time that each platform is in the reduced power state.

Abstract: Assigning interrupts for I/O devices among the nodes of NUMA systems is disclosed. At least one of the following is performed. First, interrupts for the devices are assigned among the nodes based on at least one of: the nodes to which the devices are connected, the nodes at which interrupt service routines for the devices reside, and the processors of the nodes. Second, for each node, the interrupts for the devices that are performance critical and that have been assigned to the node are assigned to the processors of the node in a round-robin manner. Third, assignments of the interrupts among the nodes of the system are dynamically modified based on actual performance characteristics of the assignments. Fourth, for each node, assignments of the interrupts that are performance critical and that have been assigned to the node are dynamically modified based on actual performance characteristics of the assignments.

Abstract: A target server is queried to build an inventory of components installed within the target server. The target server has a boot process by which it is employed within a production environment. Updated software is received for each component from an update server, based on the inventory built. The boot process is configured so that the target server installs the updated software for each component the next time it boots, instead of being employed within the production environment. The target server is caused to boot such that it installs the updated software for each component. Upon completion of the target server installing the update software for each component, the boot process is reconfigured so that the next time the target server boots it is again employed within the production environment. The target server is again caused to boot such that it is again employed within the production environment.

Abstract: A method of efficiently increasing intra-nodal page allocations and increasing processor cache utilization in a multiprocessing environment is provided. Each physical page of memory is identified and organized according to the page color as well as the nodal identifier. The organization process places each of the pages into a logical data structure. Both the identification and organization steps allow the system to properly select a page based upon both page color and nodal locale. In addition, the selection process updates the preferred page color for a subsequent page allocation. Accordingly, the method of identifying and organizing each page of memory enables an application to efficiently select pages with good processor cache distribution.

Abstract: A method of efficiently increasing intra-nodal page allocations and increasing processor cache utilization in a multiprocessing environment is provided. Each physical page of memory is identified and organized according to the page color as well as the nodal identifier. The organization process places each of the pages into a logical data structure. Both the identification and organization steps allow the system to properly select a page based upon both page color and nodal locale. In addition, the selection process updates the preferred page color for a subsequent page allocation. Accordingly, the method of identifying and organizing each page of memory enables an application to efficiently select pages with good processor cache distribution.