Abstract: Aspects of the disclosure relate to learning the most useful backup tasks when processing a job in a distributed computing application. For example, training data may be generated by running a job and running backup tasks for some of the tasks for that job. List of features may be generated for each backup task at different times. A representative list of features may be labeled according to the backup tasks usefulness to the job. The lists and usefulness labels may be used to generate a model of usefulness. The model may then be used to evaluate the usefulness of the tasks for the next job. This information may then be used to schedule backup tasks for execution. In addition, lists of features for these backup tasks may be generated and then used to update the model in a continuous learning loop.

Abstract: Functionality is disclosed herein for managing resource bursting and other operational characteristics of software executing on a virtual machine instance. The mechanism disclosed herein may receive a request to initialize a virtual machine instance. The request might include a virtual machine instance type and a weight value. The virtual machine instance may be instantiated by establishing a virtual machine of the virtual machine instance type and associating the weight value therewith. The utilization of resources by the virtual machine instance may be monitored and the weight value may be adjusted based upon the monitoring. The adjusted weight value might also be utilized when instantiating new virtual machine instances for the same user and/or customer of a service provider network.

Abstract: A computer manages methods for determining workload placement in a computing environment. The computer receives a plurality of workloads with associated information, wherein the associated information for each workload contains at least: contract information, billing information, and resource availability information. The computer determines a profitability factor for each workload of the plurality of workloads, wherein the profitability factor is at least based on the billing information. The computer determines a penalty factor for each workload of the plurality of workloads, wherein the penalty factor is at least based on the contract information. The computer determines a preference factor for each workload of the plurality of workloads, wherein the preference factor is at least based on the resource availability information. The computer assigns a priority ordering for each of the workloads from the plurality of workloads.

Abstract: A cyber-physical system and a method of monitoring a virtual machine thereof are provided. The cyber-physical system includes a plurality of target controllers that includes middleware operating based on different operating systems and that control different targets, and a system monitor that includes integrated middleware of analyzing and synthesizing information collected from the middleware. First virtual machines corresponding to virtualization of the target controllers and a second virtual machine corresponding to virtualization of the system monitor are independently formed through a virtual machine monitor.

Abstract: Interfaces to storage devices that employ storage space optimization technologies, such as thin provisioning, are configured to enable the benefits gained from such technologies to be sustained. Such an interface may be provided in a hypervisor of a virtualized computer system to enable the hypervisor to discover features of a logical unit number (LUN), such as whether or not the LUN is thinly provisioned, and also in a virtual machine (VM) of the virtualized computer system to enable the VM to discover features of a virtual disk, such as whether or not the virtual disk is thinly provisioned. The discovery of these features enables the hypervisor or the VM to instruct the underlying storage device to carry out certain operations such as an operation to deallocate blocks previously allocated to a logical block device, so that the storage device can continue to benefit from storage space optimization technologies implemented therein.

Abstract: A device may receive a chart generated via a technical computing environment, where the chart includes a textual portion and a graphical portion, and the graphical portion includes state information. The device may parse the chart into the textual portion and the graphical portion, and may process the textual portion with a textual engine of the technical computing environment to generate textual results. The device may process the graphical portion with a graphical engine of the technical computing environment to generate graphical results, and may combine the textual results with the graphical results to generate chart results. The device may output or store the chart results.

Abstract: A system and method for managing resources in a distributed computer system that includes at least one resource pool for a set of virtual machines (VMs) utilizes a set of desired individual VM-level resource settings that corresponds to target resource allocations for observed performance of an application running in the distributed computer system. The set of desired individual VM-level resource settings are determined by constructing a model for the observed application performance as a function of current VM-level resource allocations and then inverting the function to compute the target resource allocations in order to meet at least one user-defined service level objective (SLO). The set of desired individual VM-level resource settings are used to determine final RP-level resource settings for a resource pool to which the application belongs and final VM-level resource settings for the VMs running under the resource pool, which are then selectively applied.

Abstract: An apparatus for a hypervisor to obtain a faulting instruction, wherein the hypervisor runs between a physical machine including a central processing unit (CPU) and a virtual machine includes a content addressable memory (CAM); a special-purpose register (SPR) which is accessible by the hypervisor; and a control logic circuit with an input terminal connected to the CPU and an output terminal connected to the CAM, the input terminal receiving data from an instruction fetching (IF) stage and a write-back (WB) stage of a CPU instruction pipeline respectively, the output terminal causing instructions from the IF stage of the CPU instruction pipeline to be stored into the CAM and triggering the CAM to output a faulting instruction among the instructions stored therein to the SPR.

Abstract: For multi-threaded network processing, a real-time task is scheduled in a network server by placing the task on a real-time task list, activating the task when an event occurs during operation of the network server, and executing the task when scanning the task list and finding that the task has been activated. For example, an application or kernel process sends to a real-time network thread manager a request to register a task specified by an address of a task service routine. The thread manager places the address in an entry on the task list and returns a cookie identifying the entry. Upon occurrence of the event, an interrupt routine sends to a task activation routine a request to activate the task specified by the cookie. At periodic intervals, if there are any activated tasks on the task list, then a real-time scheduler scans the task list.

Abstract: A method for scheduling MapReduce tasks includes receiving a set of task statistics corresponding to task execution within a MapReduce job, estimating a completion time for a set of tasks to be executed to provide an estimated completion time, calculating a soft decision point based on a convergence of a workload distribution corresponding to a set of executed tasks, calculating a hard decision point based on the estimated completion time for the set of tasks to be executed, determining a selected decision point based on the soft decision point and the hard decision point, and scheduling upcoming tasks for execution based on the selected decision point. The method may also include estimating a map task completion time and estimating a shuffle operation completion time. A computer program product and computer system corresponding to the method are also disclosed.

Abstract: Method and system for determining the number of computing resources for executing a plurality of jobs. In accordance with the method and the system, the plurality of jobs along with a plurality of variables associated to the plurality of jobs may be received. The plurality of variables may include a mean processing time (?), a standard deviation (?), a service level agreement (?), and a confidence level (1-?). After the receipt of the plurality of variables, an inverse parameter (?) may be determined using the mean processing time (?), the standard deviation (?), and the confidence level (1-?). Further, a queue length (k) may be computed based upon the total processing time (T), the mean processing time (?), the standard deviation (?), and the inverse parameter (?). Finally, the queue length (k) and the number of jobs (n) may be processed in order to determine the number of computing resources.

Abstract: Systems, methods, and machine-readable and executable instructions are provided for dynamically ranking and scheduling monitoring tasks. Dynamically ranking and scheduling monitoring tasks can include determining an updated ranking for each of a number of monitoring tasks, where the updated ranking can include analyzing historical measurements of each of the number of monitoring tasks. An order of execution can be scheduled for each of the number of monitoring tasks based on the updated ranking for each of the number of monitoring tasks.

Abstract: In a distributed server storage environment, a set of like tasks to be performed is organized into a first group, and a last used processing group associated with the like tasks is stored. Upon a subsequent dispatch, the last used processing group is compared to other processing groups and the tasks are assigned to a processing group based upon a predetermined threshold.

Abstract: A method and system for intelligent tiering is provided. The method includes receiving a request for enabling a tiering process with respect to data. The computer processor retrieves a migration list indicating migration engines associated with the data. Additionally, an entity list of migration entities is retrieved and each migration entity is compared to associated policy conditions. In response, it is determined if matches exist between the migration entities and the associated policy conditions and a consolidated entity list is generated.

Abstract: An apparatus is described having multiple cores, each core having: a) an accelerator; and, b) a general purpose CPU coupled to the accelerator. The general purpose CPU has functional unit logic circuitry to execute an instruction that returns an amount of storage space to store context information of the accelerator.

Abstract: In the various aspects, virtualization techniques may be used to improve performance and reduce the amount of power consumed by selectively enabling a hypervisor operating on a computing device during sandbox sessions. In the various aspects, a high-level operating system may allocate memory such that its intermediate physical addresses are equal to the physical addresses. When the hypervisor is disabled, the hypervisor may suspend second stage translations from intermediate physical addresses to physical addresses. During a sandbox session, the hypervisor may be enabled and resume performing second stage translations.

Abstract: A set of like tasks to be performed is organized into a first group, and a last used processing group associated with the like tasks is stored. Upon a subsequent dispatch, the last used processing group is compared to other processing groups and the tasks are assigned to a processing group based upon a predetermined threshold.

Abstract: A processor device executes program code in one or more threads. The processor device detects a call of a function in one of the threads and executes the function in a further thread. Further, the processor device performs a selection between saving a state of the processor device when starting execution of the function in the further thread and not saving the state of the processor device when starting execution of the function in the further thread. In response to a conflict related to the execution of the function in the further thread, the processor device may perform a rollback to a last saved state of the processor device and execute the function in the thread in which it was called.

Abstract: A distributed execution system includes an output-side pipe worker that operates on a node same as an output-side worker realized by a first distributed program, and an input-side pipe worker that operates on a node same as an input-side worker realized by a second distributed program, receives output data on the output-side worker from the output-side pipe worker, and transfers it to the input-side worker, in which the output-side pipe worker acquires, from the output-side worker, output data together with a sequence number indicating an order of the output data to be transmitted to the input-side worker, acquires a restore sequence number corresponding to an execution state of the input-side worker, compares the sequence number and the restore sequence number, and does not forward, to the input-side pipe worker, the output data acquired together with the sequence number indicating the order equal to or earlier than the restore sequence number.

Abstract: A system, method and computer program product for implementing a video camera that is accessible by different Virtual Machine GOSs and the Host system. A standard USB video camera is build into a Host system. A VM running a Guest OS is launched on the Host system. The Guest OS has virtual driver for the Host video camera. The virtual driver is connected to a virtual USB bus, which communicates with a special Host library over a Virtual Machine Monitor (VMM). A special Host side VM application acquires and processes video data from a standard Host system video processing API that communicates directly with a Host-side virtual camera driver. Video frames are provided via the Host-side VM application to the virtual USB bus and to the virtual driver. Host applications or other VMs can acquire video streams from the Host USB camera simultaneously.