System and Method for a Self-Optimizing Reservation in Time of Compute Resources

Abstract

A system and method of dynamically controlling a reservation of resources within a cluster environment to maximize a response time are disclosed. The method embodiment of the invention comprises receiving from a requestor a request for a reservation of resources in the cluster environment, reserving a first group of resources, evaluating resources within the cluster environment to determine if the response time can be improved and if the response time can be improved, then canceling the reservation for the first group of resources and reserving a second group of resources to process the request at the improved response time.

1 - 20 . (canceled) 
     
     
         21 . A method of operating a compute environment comprising a plurality of compute nodes under common administrative control, the method comprising:
 receiving a request for processing one or more workload items;   causing processing of at least one of the one or more workload items submitted within the request using at least a portion of the compute environment;   monitoring at least one performance attribute after causing the processing; and   based at least on the monitored at least one performance attribute, causing dynamic modification of at least a portion of the compute environment.   
     
     
         22 . The method of  claim 21 , wherein:
 the receiving a request comprises receiving a request having at least one quality of service (QoS) requirement associated therewith;   the monitoring the at least one performance attribute after causing the processing comprises monitoring the processing of the one or more workload items relative to the at least one QoS requirement; and   the causing dynamic modification of the at least portion of the compute comprises causing dynamic modification of the compute environment to increase or enhance performance of at least one aspect of the compute environment.   
     
     
         23 . The method of  claim 21 , further comprising causing utilization of a reservation associated with the processing of the one or more workload items; and
 wherein the causing dynamic modification of the at least portion of the compute environment comprises causing dynamic modification of a priority associated with at least a portion of the reservation.   
     
     
         24 . The method of  claim 21 , further comprising:
 causing utilization of a reservation associated with the processing of the one or more workload items; and   associating a trigger with the reservation, the trigger comprising at least one condition for firing;   wherein the causing dynamic modification of the at least portion of the compute environment comprises causing dynamic modification of a priority of at least one of the one or more workload items based at least on the monitoring indicating that the at least one condition for firing has been met.   
     
     
         25 . The method of  claim 21 , further comprising:
 causing utilization of a software object associated with scheduling of the processing of the one or more workload items; and   associating a trigger with the software object, the trigger comprising a second software object specifying at least one condition for firing;   wherein the causing dynamic modification of the at least portion of the compute environment comprises causing dynamic modification of a priority of at least one of the one or more workload items based at least on the monitoring indicating that the at least one condition for firing has been met.   
     
     
         26 . The method of  claim 21 , further comprising causing placing, prior to the causing processing, at least some resources of the compute environment in a reserved state for the processing of the one or more workload items. 
     
     
         27 . The method of  claim 21 , further comprising causing placing at least one of the one or more workload items in a queue for processing; and
 wherein:   the performance attribute relates to at least one QoS requirement; and   the causing dynamic modification of the at least portion of the compute environment comprises causing dynamic modification of a priority of the at least one of the one or more workload items within the queue based at least on the monitoring indicating that the at least one QoS requirement is not being met.   
     
     
         28 . The method of  claim 21 , wherein:
 the receiving a request comprises receiving a request having at least one quality of service (QoS) requirement associated therewith;   the causing processing of the at least one of the one or more workload items comprises assigning a first resource configuration to the at least portion of the compute environment for processing the at least one of the one or more workload items;   the at least one performance attribute relates to processing of the one or more workload items relative to the at least one QoS requirement; and   the causing dynamic modification of the at least portion of the compute environment comprises causing dynamic modification of the compute environment in order to assign a second resource configuration to at least part of the compute environment for processing of the one or more of the workload items.   
     
     
         29 . The method of  claim 21 , wherein:
 the at least one performance attribute comprises a level of processing activity associated with at least one of the plurality of compute nodes; and   the causing dynamic modification of the at least portion of the compute environment comprises causing dynamic modification of the compute environment in order to reduce a likelihood that the at least one of the plurality of compute nodes will be subsequently selected for processing workload.   
     
     
         30 . The method of  claim 29 , wherein:
 the level of processing activity associated with at least one of the plurality of compute nodes comprises a level of processing activity associated with the at least one of the plurality of compute nodes during processing of the one or more workload items; and   the causing dynamic modification of the compute environment in order to reduce a likelihood that the at least one of the plurality of compute nodes will be subsequently selected for processing workload comprises reducing a priority associated with the at least one compute node relative to others of the plurality of compute nodes.   
     
     
         31 . The method of  claim 29 , wherein the level of processing activity associated with at least one of the plurality of compute nodes comprises at least one of: (i) a level of data transmission associated with the at least one compute node, or (ii) a level of data swapping associated with the at least one compute node. 
     
     
         32 . The method of  claim 21 , wherein:
 the method further comprises associating a trigger with at least one of the plurality of compute nodes, the trigger comprising at least one condition for firing, the at least one condition for firing comprising a level of processing activity associated with at least one compute node;   the at least one performance attribute comprises the level of processing activity associated with at least one compute node; and   the causing dynamic modification of the at least portion of the compute environment comprises causing reduction of a likelihood that the at least one compute node will be subsequently selected for processing workload, the causing reduction responsive to the monitoring determining that the level of processing activity associated with at least one compute node has been met or exceeded.   
     
     
         33 . A method of operating a compute environment comprising a plurality of compute nodes under common administrative control, the method comprising:
 receiving a request for processing one or more workload items, the request comprising data indicative of a required level of processing performance;   causing processing of at least one of the one or more workload items submitted within the request using at least a portion of the compute environment;   monitoring one or more events after causing the processing; and   based at least on the monitored one or more events, causing dynamic modification of at least a priority associated with the request.   
     
     
         34 . The method of  claim 33 , wherein:
 the compute environment further comprises at least one of a computerized workload manager process or computerized scheduler process; and   the priority associated with the request is based at least on a user-specified parameter relating to configuration of the at least one of a computerized workload manager process or computerized scheduler process.   
     
     
         35 . The method of  claim 33 , wherein the priority associated with the request is based at least on parameter relating to a required level of performance for processing the request. 
     
     
         36 . The method of  claim 35 , wherein the parameter relating to a required level of performance for processing the request comprises a user-specified parameter which is set using at least a computerized user process in data communication with the compute environment. 
     
     
         37 . The method of  claim 33 , wherein:
 the receiving a request comprises receiving a request having at least one quality of service (QoS) requirement associated therewith;   the causing processing of the at least one of the one or more workload items comprises assigning a first priority to the at least one of the one or more workload items;   the monitoring one or more events after causing the processing comprises monitoring the processing of the one or more workload items relative to the at least one QoS requirement; and   the causing dynamic modification of at least a priority associated with the request comprises causing dynamic modification of the first priority.   
     
     
         38 . The method of  claim 37 , wherein the causing dynamic modification of the first priority comprises increasing the first priority of the at least one of the one or more workload items to a second priority, the second priority greater than the first priority. 
     
     
         39 . The method of  claim 37 , wherein the causing dynamic modification of the first priority comprises dynamic modification of at least part of a computerized scheduler process associated with the compute environment. 
     
     
         40 . The method of  claim 33 , further comprising causing utilization of a reservation associated with the processing of the one or more workload items; and
 wherein the causing dynamic modification of the at least priority associated with the request comprises causing dynamic modification of a priority associated with the reservation.   
     
     
         41 . The method of  claim 33 , further comprising:
 causing utilization of a reservation associated with the processing of the one or more workload items; and   associating a trigger with the reservation, the trigger comprising at least one condition for firing;   wherein the causing dynamic modification of the at least priority associated with the request based at least on the monitored one or more events, comprises causing dynamic modification of a priority of at least one of the one or more workload items based at least on the monitoring indicating that the at least one condition for firing has been met.   
     
     
         42 . The method of  claim 21 , further comprising:
 causing utilization of a software object associated with scheduling of the processing of the one or more workload items; and   associating a trigger with the software object, the trigger comprising a second software object specifying at least one condition for firing;   wherein the causing dynamic modification of at least one of the compute environment or a priority associated with the request based at least on the monitored one or more events, comprises causing dynamic modification of a priority of at least one of the one or more workload items based at least on the monitoring indicating that the at least one condition for firing has been met.   
     
     
         43 . A method of operating a compute environment comprising a plurality of resources including a plurality of compute nodes, the method comprising:
 creating a trigger, the trigger comprising at least one attribute relating to a level of compute node processing activity;   associating the trigger with at least one of the plurality of compute nodes;   monitoring the level of compute node processing activity for the at least one compute node; and   based at least on the monitored level of compute node processing activity meeting at least one prescribed criterion, causing the execution of the trigger, the execution causing one or more modifications to the compute environment to be initiated;   wherein the one or more modifications to the compute environment comprise modification in order to reduce a likelihood that the at least one compute node will be subsequently selected for processing workload.   
     
     
         44 . The method of  claim 43 , wherein the method further comprises:
 receiving a request for processing, by the compute environment, of one or more compute jobs; and   based at least on the received request, causing processing of at least one of the one or more compute jobs on one or more selected compute nodes of the compute environment, a selection of the one or more selected compute nodes being based at least on the modification.   
     
     
         45 . The method of  claim 44 , wherein the modification in order to reduce the likelihood that the at least one compute node will be subsequently selected for processing workload comprises a modification of a priority associated with the at least one compute node. 
     
     
         46 . The method of  claim 44 , wherein the monitoring of a level of processing activity associated with at least one of the plurality of compute nodes comprises monitoring at least one of: (i) a level of data transmission associated with the at least one compute node, or (ii) a level of data swapping associated with the at least one compute node. 
     
     
         47 . The method of  claim 43 , wherein the creating a trigger comprises creating a software function via a user input to a computerized user process in data communication with the compute environment. 
     
     
         48 . The method of  claim 43 , wherein the modification in order to reduce a likelihood that the at least one compute node will be subsequently selected for processing workload comprises a dynamic modification to one of a (i) computerized workload manager process of the compute environment, or (ii) computerized scheduler process of the compute environment. 
     
     
         49 . The method of  claim 43 , wherein the modification in order to reduce a likelihood that the at least one compute node will be subsequently selected for processing workload comprises a modification in order to reduce a likelihood that the at least one compute node will be subsequently selected for processing of a particular type of workload. 
     
     
         50 . The method of  claim 49 , wherein the particular type of workload comprises a batch workload.