Self-modulation in a model-based automated management framework

A method comprising:

• calculating confidence values of models;
• selecting a first model based on said confidence values;
• processing said first model through a constraint solver to produce first workload throttling values;
• and
• repeatedly processing workloads through a feedback-based execution engine,
• wherein said feedback-based execution engine is controlled by said first workload throttling values and said first workload throttling values are applied incrementally to said feedback-based execution engine during repetitions of said processing of said workloads,
• wherein said processing of said workloads is repeated until an objective function is maximized, wherein said objective function specifies performance goals of said workloads;
• and
• outputting said workloads to a computer display.

A system comprising:

• at least one processor connected to a storage device;
• one or more models, wherein a first model is selected from said models based on confidence values of said models;
• a constraint solver for processing said first model;
• first workload throttling values, wherein said first workload throttling values are produced by said processing of said first model through said constraint solver;
• and
• a feedback-based execution engine for repeatedly processing workloads, wherein said feedback-based execution engine is controlled by said first workload throttling values and said first workload throttling values are applied incrementally to said feedback-based execution engine during repetitions of said processing of said workloads,
• wherein said feedback-based execution engine detects divergent behavior and trigger re-calculation of workload throttling values,
• wherein said processing of said workloads is repeated until an objective function is maximized, wherein said objective function specifies performance goals of said workloads;
• and
• wherein said feedback-based execution engine outputs said workloads to a computer display.

A computer storage device comprising a computer usable medium storing a computer readable program, wherein the computer readable program, when executed on a computer, causes the computer to perform a method comprising:

• calculating confidence values of models;
• selecting a first model based on said confidence values;
• processing said first model through a constraint solver to produce first workload throttling values;
• repeatedly processing workloads through a feedback-based execution engine,
• wherein said feedback-based execution engine is controlled by said first workload throttling values and said first workload throttling values are applied incrementally to said feedback-based execution engine during repetitions of said processing of said workloads,
• and wherein said processing of said workloads is repeated until an objective function is maximized, wherein said objective function specifies performance goals of said workloads;
• and
• outputting said workloads to a computer display.