Systems and methods for dynamically managing virtual machines

Abstract

Techniques for dynamic management of virtual machine environments are disclosed. For example, a technique for automatically managing a first set of virtual machines being hosted by a second set of physical machines comprises the following steps/operations. An alert is obtained that a service level agreement (SLA) pertaining to at least one application being hosted by at least one of the virtual machines in the first set of virtual machines is being violated. Upon obtaining the SLA violation alert, the technique obtains at least one performance measurement for at least a portion of the machines in at least one of the first set of virtual machines and the second set of physical machines, and a cost of migration for at least a portion of the virtual machines in the first set of virtual machines. Based on the obtained performance measurements and the obtained migration costs, an optimal migration policy is determined for moving the virtual machine hosting the at least one application to another physical machine.

1 . A method of automatically managing a first set of virtual machines being hosted by a second set of physical machines, comprising the steps of: 
 obtaining an alert that a service level agreement (SLA) pertaining to at least one application being hosted by at least one of the virtual machines in the first set of virtual machines is being violated;    upon obtaining the SLA violation alert: 
 obtaining at least one performance measurement for at least a portion of the machines in at least one of the first set of virtual machines and the second set of physical machines;  
 obtaining a cost of migration for at least a portion of the virtual machines in the first set of virtual machines; and  
 determining, based on the obtained performance measurements and the obtained migration costs, an optimal migration policy for moving the virtual machine hosting the at least one application to another physical machine.  
   
   
   
       2 . The method of  claim 1 , wherein the optimal policy determining step further comprises the step of selecting a virtual machine from the first set of virtual machines with the lowest migration cost.  
   
   
       3 . The method of  claim 2 , wherein the optimal policy determining step further comprises the step of selecting a physical machine from the second set of physical machines that has a resource residue that is the lowest among the physical machines and that can accommodate the selected virtual machine.  
   
   
       4 . The method of  claim 3 , wherein the optimal policy determining step further comprises the step of generating an instruction to move the selected virtual machine to the selected physical machine.  
   
   
       5 . The method of  claim 4 , wherein the optimal policy determining step further comprises the step of recalculating resource residues for the second set of physical machines.  
   
   
       6 . The method of  claim 5 , wherein the optimal policy determining step further comprises the step of sorting the first set of virtual machines according to migration costs.  
   
   
       7 . The method of  claim 3 , wherein when the second set of physical machines does not include a physical machine that can accommodate the selected virtual machine, mapping the selected virtual machine to a physical machine that is not in the second set of physical machines.  
   
   
       8 . The method of  claim 1 , wherein at least a portion of the steps of the management method are iteratively performed until the SLA violation is remedied.  
   
   
       9 . Apparatus for automatically managing a first set of virtual machines being hosted by a second set of physical machines, comprising: 
 a memory; and    at least one processor coupled to the memory and operative to: (i) obtain an alert that a service level agreement (SLA) pertaining to at least one application being hosted by at least one of the virtual machines in the first set of virtual machines is being violated; and (ii) upon obtaining the SLA violation alert: obtain at least one performance measurement for at least a portion of the machines in at least one of the first set of virtual machines and the second set of physical machines; obtain a cost of migration for at least a portion of the virtual machines in the first set of virtual machines, and determine, based on the obtained performance measurements and the obtained migration costs, an optimal migration policy for moving the virtual machine hosting the at least one application to another physical machine.    
   
   
       10 . The apparatus of  claim 9 , wherein the optimal policy determining operation further comprises selecting a virtual machine from the first set of virtual machines with the lowest migration cost.  
   
   
       11 . The apparatus of  claim 10 , wherein the optimal policy determining operation further comprises selecting a physical machine from the second set of physical machines that has a resource residue that is the lowest among the physical machines and that can accommodate the selected virtual machine.  
   
   
       12 . The apparatus of  claim 11 , wherein the optimal policy determining operation further comprises generating an instruction to move the selected virtual machine to the selected physical machine.  
   
   
       13 . The apparatus of  claim 12 , wherein the optimal policy determining operation further comprises recalculating resource residues for the second set of physical machines.  
   
   
       14 . The apparatus of  claim 13 , wherein the optimal policy determining operation further comprises sorting the first set of virtual machines according to migration costs.  
   
   
       15 . The apparatus of  claim 11 , wherein when the second set of physical machines does not include a physical machine that can accommodate the selected virtual machine, mapping the selected virtual machine to a physical machine that is not in the second set of physical machines.  
   
   
       16 . The apparatus of  claim 9 , wherein at least a portion of the operations of the management apparatus are iteratively performed until the SLA violation is remedied.  
   
   
       17 . An article of manufacture for automatically managing a first set of virtual machines being hosted by a second set of physical machines, comprising a machine readable medium containing one or more programs which when executed implement the steps of: 
 obtaining an alert that a service level agreement (SLA) pertaining to at least one application being hosted by at least one of the virtual machines in the first set of virtual machines is being violated;    upon obtaining the SLA violation alert: 
 obtaining at least one performance measurement for at least a portion of the machines in at least one of the first set of virtual machines and the second set of physical machines;  
 obtaining a cost of migration for at least a portion of the virtual machines in the first set of virtual machines; and  
   determining, based on the obtained performance measurements and the obtained migration costs, an optimal migration policy for moving the virtual machine hosting the at least one application to another physical machine.