Predictive and dynamic resource provisioning with tenancy matching of health metrics in cloud systems

1. A method of dynamically provisioning computing resources to meet changing service level objectives in a cloud computing environment, comprising:

with a processor, determining and monitoring quality of resources available for provisioning to the cloud computing environment in order to expand or contract the cloud computing environment, as required, said determining and monitoring further comprising:

determining a measured health summary of resources available in an initial resource pool using a resource type based parameterized health summary for each individual resource, said parameterized health summary for each individual resource comprising at least number of available central processing unit cores, available random access memory, infrastructure runtime components, network bandwidth, and available disk storage, wherein said measured health for said parameterized health summary for each individual resource varies dynamically to account for resource interaction limitations and change in workload;

determining a provisioning health model based on resource types previously used for similar cloud job tenancy requirements as captured in a cataloged tenancy model, said provisioning health model comprising different resource types, classified based on their possible usage in satisfying a requirement tenancy;

determining a health grading model for each individual resource of a same type to have relative and absolute grade parameters available to determine resource equivalence, wherein said relative grade parameter comprises a score determined to be between a minimum and a maximum range for each individual resource of a same type and said absolute grade parameters comprise scores corresponding to a current value, a critical value, a lower threshold value, and an upper threshold value; and

updating the health grading model and provisioning health model using a health grade monitor to reclassify any said each individual resource in said initial resource pool by filtering out any mismatched resources as determined by said relative and said absolute grade parameters of said health grading model and said resource types used for similar cloud job tenancy requirements as captured in a cataloged tenancy model;

determining, with said processor, current resource needs for a cloud job tenancy; and

dynamically provisioning selected resources from said resources available for provisioning the cloud based on said current resource needs and said quality of said resources available in order to meet said cloud job tenancy and the service level objectives.

2. The method of claim 1 , wherein said determining current resource needs for a cloud job tenancy comprises:

identifying a closest existing tenant provisioned from a pool of tenants in said cataloged tenancy model, said tenancy model comprising resource types previously used for similar cloud job tenancy requirements, based on a modeling of tenancy requirements, as a first match, said tenancy requirements comprising:

functional requirements comprising data parameters of HostingConstraints, TenancyDomain, TenancyLayers, TenancyPriority, Workflow, and the TenancyServiceMode;

non-functional requirements comprising data parameters of ProcessorUtilization, size of PersistentStorageRequired, ResponseTime, AvailabilityFactor, number of ScalabilityNodes, and number of RedundancyNodes; and

constraints comprising DomainConstraints and OperationalConstraints;

identifying a corresponding provision model based on said first match;

identifying a closest candidate provision model as a second match; and

entering said second match into a dynamic provisioning plan customizer and validator.