Identifying optimal platforms for workload placement in a networked computing environment

摘要

Embodiments of the present invention provide a workload optimization approach that measures workload performance across combinations of hardware (platform, network configuration, storage configuration, etc.) and operating systems, and which provides a workload placement on the platforms where jobs perform most efficiently. This type of placement may be based on performance measurements (e.g., throughput, response, and other such service levels), but it can also be based on other factors such as power consumption or reliability. In a typical embodiment, ideal platforms are identified for handling workloads based on performance measurements and any applicable service level agreement (SLA) terms.

主权项

1. A method for optimizing workload placement in a networked cloud computing environment, comprising:
defining the networked cloud computing environment as comprising a plurality of independent cloud networks; monitoring a workload; determining, based on the monitoring, whether the workload is new or previously-exists; if the workload is new, determining where to initially place the workload, and placing the workload onto an initial platform based on the determination, the initial platform comprising a first hardware and software combination; if the workload previously-exists, locating the workload on a current platform for further evaluation, the current platform and the initial platform comprising the same platform; taking a measurement of a performance of the workload on a combination of software and hardware of a current platform of an independent cloud network using a set of performance metrics, the set of performance metrics comprising at least one of throughput and energy consumption; identifying an optimal platform for handling the workload based on a set of service level agreement (SLA) terms and a comparison of the measurement of the performance of the workload on the combination of software and hardware of the current platform to measurements of a performance of the workload on a combination of software and hardware of each of all of a plurality of platforms of independent cloud networks where the workload has been measured; determining whether the optimal platform has availability to handle the workload; evaluating if a scarce resource is needed; if a scarce resource is needed, evaluating an effect of swapping the workload with active resources on the optimal platform; if the swapping is not effective, fetching a next workload; if it is determined that the optimal platform has the availability, recommending to migrate the workload to the optimal platform; and migrating the workload from the current platform to the optimal platform if it is recommended.