Resource management using reliable and efficient delivery of application performance information in a cloud computing system
System and method and computer program product implemented for a cloud computing infrastructure that allows a hypervisor to optimize resource allocation in a cloud computing environment by exploiting the application-level performance, O/S system performance, and hypervisor performance information through a reliable and efficient channel.
1. A method for reliable data delivery in a cloud computing environment running multiple virtual machines, a virtual machine running one or more applications, said method comprising:
generating, by a kernel level process, a system call providing operating system level performance data regarding performance metrics of a running virtual machine;
receiving, at a host machine hypervisor level process, said performance data via said system call;
generating, by a running application process, a further system call providing application performance metrics data regarding performance metrics of one or more said running applications;
receiving, at said kernel level process of said running virtual machine, said application performance metrics data of said further system call,
said kernel level process responsively generating a hypercall relaying said application performance metrics data to a host machine hypervisor; and,
said host machine hypervisor delivering said application performance metrics data and said system level performance data regarding said running virtual machine to a cloud controller for managing resources of said cloud computing environment.
2. The method of claim 1 , wherein said cloud controller for managing resources is a software management stack of a cloud computing environment.
3. The method of claim 1 , wherein said further system call provides application performance status information.
4. The method of claim 1 , wherein said cloud controller for managing resources manages hypervisor resources comprising: CPU processing allocation, memory allocation, I/O and network resources allocation by adding or reducing those resources.
5. The method of claim 1 , wherein said host machine hypervisor guarantees delivery of said application performance metrics data and system level performance data to said cloud controller for managing resources.
6. The method of claim 1 , wherein said cloud controller includes a cloud customer utilizing cloud resources.
7. The method of claim 3 , further comprising:
generating, at said host machine hypervisor, hypervisor performance data, said host machine hypervisor pushing said application performance metrics and performance status information, said system level performance data and hypervisor performance data to said cloud controller.
8. The method of claim 7 , wherein said method further comprising:
merging said application performance and status information, system level performance data and hypervisor performance data into a single message to be delivered to the cloud controller.
9. The method of claim 7 , further comprising:
correlating the application performance and status information, system level performance data and hypervisor performance data to manage allocation of resources or for policy enforcement.
10. The method of claim 1 , wherein said system call and further system call providing each said application performance metrics data, status information and/or system level performance data includes a key-value pair identifying an originating VM device or application and an associated measurement or status value, said key-value pair configured to fit inside arguments of the system call or further system call.
11. The method of claim 10 , wherein said system call and further system call providing said key value pair are configured to be delivered to said host machine hypervisor without the use of memory allocation at said hypervisor.