Efficiently repairing virtual machines by means of DVFS-aware proactive scheduling

A method for efficiently repairing virtual machines by means of DVFS-aware proactive scheduling, the method comprising:

• receiving notice that a job has failed due to a failure of a failed virtual machine provisioned on a first active-mode physical platform configured to run in active mode, where the failed job resides in a queue of jobs that had been scheduled to be performed by the failed virtual machine;
• initially identifying:
• a Buffer Time B(t) that specifies a duration of time available for completing performance of the job,
• a Repair Time R(t) that specifies an estimated duration of time required to repair the failed virtual machine,
• an Active-Mode Running Time E(x) that specifies an estimated duration of time required to perform the failed job by an active-mode virtual machine provisioned on the first active-mode physical platform, and
• a Transition Time X(t) that specifies an estimated duration of time required to transfer the failed job to a queue of a functioning virtual machine that is distinct from the failed virtual machine;
• determining whether the Buffer Time is less than the Repair Time;
• attempting to reassign the failed job to a queue of the functioning virtual machine if the Buffer Time is less than the Repair time;
• and
• attempting to repair the failed virtual machine if the Buffer Time is not less than the Repair time.

A computer program product, comprising a computer-readable hardware storage device having a computer-readable program code stored therein, the program code configured to be executed by a failure-recovery system comprising a processor, a memory coupled to the processor, and a computer-readable hardware storage device coupled to the processor, the storage device containing program code configured to be run by the processor via the memory to implement a method for efficiently repairing virtual machines by means of DVFS-aware proactive scheduling, the method comprising

• the system receiving notice that a job has failed due to a failure of a failed virtual machine provisioned on a first active-mode physical platform configured to run in active mode, where the failed job resides in a queue of jobs that had been scheduled to be performed by the failed virtual machine;
• the system initially identifying:
• a Buffer Time B(t) that specifies a duration of time available for completing performance of the job,
• a Repair Time R(t) that specifies an estimated duration of time required to repair the failed virtual machine,
• an Active-Mode Running Time E(x) that specifies an estimated duration of time required to perform the failed job by an active-mode virtual machine provisioned on the first active-mode physical platform, and
• a Transition Time X(t) that specifies an estimated duration of time required to transfer the failed job to a queue of a functioning virtual machine that is distinct from the failed virtual machine;
• the system determining whether the Buffer Time is less than the Repair Time;
• the system attempting to reassign the failed job to a queue of the functioning virtual machine if the Buffer Time is less than the Repair time;
• and
• the system attempting to repair the failed virtual machine if the Buffer Time is not less than the Repair time.