Systems, apparatus and methods for cost and performance-based management of resources in a cloud environment

A computer system for managing allocation of computational workloads in a computer system, comprising:

• a processor;
• a computer memory including stored instructions, executable by the processor, for causing the processor to:
• determine a set of resources for running a workload required by a component of the computer system, the resources including a needed quantity of one or more of CPUs, memory, databases, network bandwidth, and/or input-output capacity;
• identify a first resource offering by a first provider including the needed quantity of resources individually on an on-demand basis and determining a cost for provisioning the needed quantity of resources to the computer system on the on-demand basis;
• identify a second resource offering by a second provider different from the first resource offering and consisting of a reserved instance, the reserved instance specifying a quantity of one or more resources selected from CPUs, memory, databases, network bandwidth, and/or input-output capacity for a fixed time period and exceeding the needed quantity of resources, the reserved instance having a utilization level including a need by one or more other components of the computer system for an excess portion of the reserved instance beyond the needed quantity of resources;
• determine a cost charged for the reserved instance;
• select to utilize the reserved instance to run the workload based at least in part on a comparison of the determined cost for running the workload on the first provider given the utilization level with the determined cost of the reserved instance;
• determine, after passage of a predetermined period of time no greater than the fixed time period associated with the reserved instance, a cost for running the workload on a third provider;
• determine a cost of moving the workload to the third provider;
• compute a utilization value for running the workload on the third provider based at least in part on the determined cost for running the workload on the third provider, the determined cost of moving the workload to the third provider, and a need by one or more other components of the computer system for the remainder of the reserved instance;
• and
• cause the workload to the third provider based at least in part on the utilization value after the utilization value has surpassed a predetermined value.