Abstract: A method of autoscaling in a datacenter includes receiving one or more datacenter metrics at a control service, comparing the one or more datacenter metrics to a threshold value, selecting at least one component of a server computer to scale-up, and sending an instruction to the server computer to scale-up the at least one component.

Abstract: A method for dynamically adjusting a number of virtual machines for a workload, includes: receiving a probability indicator for each of a plurality of N sequential stages, where N is a natural number greater than 1, of a likelihood that a virtual machine assigned to a workload will be evicted during the N sequential stages; predicting a target number of virtual machines to configure in a current stage for a subsequent stage from among the plurality of N sequential stages based on the probability indicator, a target capacity for the workload, and a current price for maintaining a virtual machine; and configuring a number of virtual machines for the workload during the current stage based on the target number to be loaded for the workload for the subsequent stage.

Abstract: Systems and methods are provided for scheduling a virtual machine (VM) to host a workload in a cloud system. In particular, the disclosed technology schedules an evicted VM for redeploying an interruptible workload. The scheduling is based on capacity prediction and inference data associated with a type of the evicted VM. Capacity signal predictor generates training data for training a machine learning model using capacity signal history data of the cloud system. The machine-learning model, once trained, predicts capacity including a rate of evictions for the types of the evicted VM. The predicted data is based on at least the current status of available computing resources. Upon receiving a notice associated with a workload interruption, the intelligent scheduler prioritizes the evicted VM for scheduling and determines whether to defer redeploying the evicted VM based on the rate of eviction for the type of the evicted VM.

Abstract: The present disclosure relates to systems, methods, and computer readable media for optimizing resource allocation on a computing zone based on bursting data and in accordance with an allocation policy. For example, systems disclosed herein may implement a lifetime-aware allocation of computing resources to facilitate computational bursting on respective server nodes to prioritize different aspects of performance on the server nodes of a cloud computing system. The systems described herein provide a number of benefits, including, by way of example, facilitating an allocation policy that enables more densely packing virtual machines on respective server nodes, boosting individual virtual machine performance, and reducing a buffer of empty nodes that need to be maintained on one or more node clusters. This may include more densely packing virtual machines on respective server nodes.

Abstract: Systems and methods related to harvesting of the unused resources in a distributed computing system are described. An example method, in a system including a host server, having a first instruction set architecture (ISA), and an interface card, having a second ISA is provided. The method includes designating at least one type of resource, associated with the host server for harvesting by compute entities configured for execution using the processor having the second ISA, where the host server is configured to execute compute entities requiring execution by the processor having the first ISA. The method further includes in response to a request for accessing the at least one type of resource by a compute entity, executing on the processor having the second ISA, automatically allowing the compute entity to access the at least one type of resource associated with the host server.

Abstract: Systems and methods related to harvesting of the unused resources in a distributed computing system are described. An example method, in a system including a host server, having a first instruction set architecture (ISA), and an interface card, having a second ISA is provided. The method includes designating at least one type of resource, associated with the host server for harvesting by compute entities configured for execution using the processor having the second ISA, where the host server is configured to execute compute entities requiring execution by the processor having the first ISA. The method further includes in response to a request for accessing the at least one type of resource by a compute entity, executing on the processor having the second ISA, automatically allowing the compute entity to access the at least one type of resource associated with the host server.

Abstract: The present disclosure describes a type of virtual machine, which the present disclosure may refer to as a harvest virtual machine, that may allow improved utilization of physical computing resources on a cloud-computing system. First, the harvest virtual machine may be evictable. In other words, higher priority virtual machines may preempt the harvest virtual machine's access to physical computing resources. Second, the harvest virtual machine may receive access to a dynamic amount of physical computing resources during the course of its operating life. Third, the harvest virtual machine may have a minimum size (in terms of an amount of physical computing resources) and may terminate whenever the harvest virtual machine has access to an amount of physical computing resources less than the minimum size.

Abstract: A method and system for selecting servers for storage of replicas of a block of data is provided. The system selects a first server for storage of the data. The first server has a first processor utilization classification and a first reimaging rate classification. The system then selects a second server for storage of the data. The second server has a second processor utilization classification and a second reimaging rate classification. The system selects the second server so that the second processor utilization classification is different from the first processor utilization classification and the second reimaging rate classification is different from the first reimaging rate classification.

Abstract: A method and system for selecting servers for storage of replicas of a block of data is provided. The system selects a first server for storage of the data. The first server has a first processor utilization classification and a first reimaging rate classification. The system then selects a second server for storage of the data. The second server has a second processor utilization classification and a second reimaging rate classification. The system selects the second server so that the second processor utilization classification is different from the first processor utilization classification and the second reimaging rate classification is different from the first reimaging rate classification.