US 11,915,061 B2
Datacenter efficiency management system for migrating workload across nodes based on workload performance efficiency ranking
Rishi Mukherjee, Bangalore (IN); Ravishankar Kanakapura Nanjundaswamy, Bangalore (IN); Prasoon Sinha, Bangalore (IN); and Raveendra Babu Madala, Bangalore (IN)
Assigned to Dell Products L.P., Round Rock, TX (US)
Filed by Dell Products L.P., Round Rock, TX (US)
Filed on Oct. 26, 2021, as Appl. No. 17/510,601.
Prior Publication US 2023/0129548 A1, Apr. 27, 2023
Int. Cl. G06F 9/50 (2006.01); G06F 9/48 (2006.01)
CPC G06F 9/5094 (2013.01) [G06F 9/4856 (2013.01); G06F 9/5055 (2013.01); G06F 9/5077 (2013.01); G06F 2209/501 (2013.01); G06F 2209/508 (2013.01); G06F 2209/5022 (2013.01); Y02D 10/00 (2018.01)] 20 Claims
OG exemplary drawing
 
1. A datacenter, comprising:
a plurality of node devices; and
a datacenter efficiency management system that is coupled to each of the node devices and that is configured to:
generate, for each of the plurality of node devices based on a power consumption estimation for that node device and a performance estimation for that node device, a node group ranking;
group, based on the node group ranking generated for each of the plurality of node devices, subsets of the plurality of node devices into respective homogenous node groups that each include node devices that are homogenous with respect to their power consumption estimation and performance estimation such that the subset of the plurality of node devices in any one of the homogenous node groups include node devices that are estimated to consume power in a power consumption range for that homogeneous node group that is different from the other homogeneous node groups, and that are estimated to perform in a performance range for that homogeneous node group that is different from the other homogeneous node groups;
power cap at least one of the homogenous node groups based on a node component inventory that identifies the components included in the subset of the plurality of node devices in that homogeneous node group, and a node component health of the components included in the subset of the plurality of node devices in that homogeneous node group;
deploy a respective workload on each node device in each of the homogenous node groups;
generate, based on each of a node workload bandwidth, a node power consumption, and a node health of each node device that is indicative of a level of performance of that node device in performing the workload deployed on that node device, a workload performance efficiency ranking for that node device that identifies a relative efficiency of that node device in performing its respective workload relative to the other node devices in performing their respective workloads; and
migrate, based on a first workload performance efficiency ranking of a first node device that is included in the plurality of node devices and that is part of a first homogenous node group and a second workload performance efficiency ranking of a second node device that is included in the plurality of node devices and that is part of a second homogenous node group, a first workload between the first node device and the second node device.