Predictive scaling of datacenters

Abstract

Examples described herein include systems and methods for efficiently scaling an SDDC. An example method can include storing resource utilization information for a variety of resources of the SDDC. The example method can also include predicting a future resource utilization rate for the resources and determining that a predicted utilization rate is outside of a desired range. The system can determine how long it would take to perform the scaling, including adding or removing a host and performing related functions such as load balancing or data transfers. The system can also determine how long the scaling is predicted to benefit the SDDC to ensure that the benefit is sufficient to undergo the scaling operation. If the expected benefit is greater than the benefit threshold, the system can perform the scaling operation.

1 - 20 . (canceled) 
     
     
         21 . A system for efficiently scaling a software defined datacenter (SDDC), comprising:
 a memory storage including a non-transitory, computer-readable medium comprising instructions; and   a computing device including a hardware-based processor that executes the instructions to carry out stages comprising:
 predicting a future resource utilization rate for a resource of the SDDC based on resource utilization information for the resource; 
 determining that the future resource utilization rate of the resource is predicted to be outside a desired range for a first time period; 
 determining a scaling time required to scale the SDDC; 
 determining a beneficial time period by subtracting the scaling time from the first time period; 
 comparing the beneficial time period to a time threshold; and 
 based on the beneficial time period being greater than the time threshold, scaling the SDDC. 
   
     
     
         22 . The system of  claim 21 , wherein scaling the SDDC comprises adding or removing a host from the SDDC. 
     
     
         23 . The system of  claim 21 , wherein the resource utilization information comprises at least one of a processing resource, a memory resource, a storage resource, and an input-output (I/O) resource. 
     
     
         24 . The system of  claim 21 , wherein the scaling time includes time required to add a new host to the SDDC and time required to load balance the SDDC based on the new host. 
     
     
         25 . The system of  claim 21 , wherein the scaling time includes time required to remove a virtual machine from a host and time required to move data to a different host. 
     
     
         26 . The system of  claim 21 , wherein the time threshold is set by a customer of the SDDC through use of a graphical user interface (GUI). 
     
     
         27 . The system of  claim 21 , wherein predicting the future resource utilization rate is performed by a machine learning model trained using historical resource utilization information of the SDDC. 
     
     
         28 . A non-transitory, computer-readable medium containing instructions that, when executed by a hardware-based processor, performs stages for efficiently scaling a software defined datacenter (SDDC), the stages comprising:
 determining that a resource utilization rate for a resource of the SDDC is above a desired range;   predicting that the resource utilization rate for the resource will remain above the desired range for a period of time greater than the time required to instantiate a new host in the SDDC;   predicting a beneficial time period after which the new host will no longer be needed to maintain the resource utilization rate within the desired range; and   based on the beneficial time period being greater than a benefit threshold, instantiating the new host in the SDDC.   
     
     
         29 . The non-transitory, computer-readable medium of  claim 28 , wherein instantiating a new host comprises instantiating at least one virtual machine on the new host. 
     
     
         30 . The non-transitory, computer-readable medium of  claim 28 , wherein predicting of the resource utilization rate is performed by a machine learning model trained using historical resource utilization information of the SDDC. 
     
     
         31 . The non-transitory, computer-readable medium of  claim 30 , wherein the resource utilization information comprises at least one of a processing resource, a memory resource, a storage resource, and an input-output (I/O) resource 
     
     
         32 . The non-transitory, computer-readable medium of  claim 28 , wherein the time required to instantiate the new host includes a time required to load balance the SDDC based on the new host. 
     
     
         33 . The non-transitory, computer-readable medium of  claim 28 , wherein the time required to instantiate the new host includes a time required to remove a virtual machine from an existing host and move data to the new host. 
     
     
         34 . The non-transitory, computer-readable medium of  claim 28 , wherein the benefit threshold is set by a customer of the SDDC through use of a graphical user interface (GUI). 
     
     
         35 . A method for efficiently scaling a software defined datacenter (SDDC), comprising:
 determining that a resource utilization rate for a resource of the SDDC is above a desired range;   predicting that the resource utilization rate for the resource will remain above the desired range for a period of time greater than the time required to instantiate a new host in the SDDC and load balance the SDDC;   predicting a beneficial time period after which the new host will no longer be needed to maintain the resource utilization rate within the desired range; and   based on the beneficial time period being greater than a benefit threshold, instantiating the new host in the SDDC.   
     
     
         36 . The method of  claim 35 , wherein instantiating a new host comprises instantiating at least one virtual machine on the new host. 
     
     
         37 . The method of  claim 35 , wherein predicting of the resource utilization rate is performed by a machine learning model trained using historical resource utilization information of the SDDC. 
     
     
         38 . The method of  claim 37 , wherein the resource utilization information comprises at least one of a processing resource, a memory resource, a storage resource, and an input-output (I/O) resource. 
     
     
         39 . The method of  claim 35 , wherein the time required to instantiate the new host includes time required to remove a virtual machine from an existing host and move data to the new host. 
     
     
         40 . The method of  claim 35 , wherein the benefit threshold is set by a customer of the SDDC through use of a graphical user interface (GUI).