Performance assurance using workload phase detection

A controller comprising:

• communication circuitry;
• and
• processing circuitry configured to:
• detect that a compute node operating in a network controlled by the controller is unable to satisfy a service level objective (SLO) of the network;
• identify a resource of the compute node that will mitigate the SLO deficiency;
• determine whether the resource is available in a resource pool accessible by the controller;
• and
• allocate the resource to the compute node when the resource is available in the resource pool.

A machine-readable medium storing instructions that, when executed by processing circuitry, cause the processing circuitry to perform operations comprising:

• generating a phase sequence matrix that indicates an identity and number of phases of a workload by measuring resources used in the compute node during execution of the workload throughout a lifetime of the workload;
• generating a workload fingerprint, the workload fingerprint including the phase sequence matrix and a phase residency matrix, the phase residency matrix indicating a fraction of execution time of the workload spends in each phase identified in the phase sequence matrix;
• receiving a value corresponding to a service level objective (SLO) via the communication circuitry;
• and
• allocating a resource of the compute node based on the workload fingerprint and based on comparisons between the value corresponding to the SLO parameter and measurements made during operation of the compute node in a network.

A compute node comprising:

• means to generate a phase sequence matrix that indicates an identity and number of phases of a workload by measuring resources used in the compute node during execution of the workload throughout a lifetime of the workload;
• means to generate a workload fingerprint, the workload fingerprint including the phase sequence matrix and a phase residency matrix, the phase residency matrix indicating a fraction of execution time of the workload spends in each phase identified in the phase sequence matrix;
• and
• means to allocate a resource of the compute node based on the workload fingerprint and based on comparisons between the value corresponding to the SLO parameter and measurements made during operation of the compute node in a network.