Predictive analytics for virtual network functions

1. A method comprising,

with an analytics component, receiving performance data from a Virtual Network Function (VNF), the VNF including a plurality of VNF components running on a plurality of virtual machines;

with the analytics component, performing an analytics function on the performance data, wherein the analytics function detects how changes in a number of one type of VNF component affects demand for a different type of VNF component based on performance data for different types of VNF components;

with the analytics component, based on the analytics function determining a prediction event in response to determining that a set of conditions within the performance data is present; and

with the analytics component, notifying a VNF manager of the prediction event.

2. The method of claim 1 , wherein the performance data includes network conditions of networks associated with the plurality of virtual machines and utilization of the VNF components.

3. The method of claim 1 , wherein the analytics function includes at least one of: a linear regression analysis, a neural network analysis, and a Pearson correlation coefficient.

4. The method of claim 1 , wherein the prediction event indicates that the VNF is projected to reach a threshold level of capacity within a predetermined period of time.

5. The method of claim 1 , wherein the set of conditions includes a current utilization of the VNF, an increase of utilization, and a confidence value being above respective thresholds.

6. The method of claim 1 , further comprising, after notifying the VNF manager of the prediction event, receiving additional performance data from a newly provisioned VNF component.

7. The method of claim 1 , further comprising:

receiving external data that includes factors that affect utilization of the VNF; and

using the external data with the analytics function to determine the prediction event.

8. The method of claim 1 , wherein the different type of VNF component includes a load-balancing VNF component.

9. The method of claim 1 , wherein the different type of VNF component includes a database component.

10. A method comprising:

with a Virtual Network Function (VNF) manager that manages a VNF that includes a plurality of VNF components running on a plurality of virtual machines, sending performance data to an analytics component;

with the VNF manager, receiving a notification from the analytics component, the notification indicating that a prediction event has occurred; and

with the VNF manager, in response to the notification, instructing a virtual infrastructure manager to adjust a number of VNF components within the VNF;

wherein the prediction event is based on analysis of the performance data, the analytics component detects how changes in a number of one type of VNF component affects demand for a different type of VNF component based on the performance data from the one type of VNF component and the different type of VNF component.

11. The method of claim 10 , wherein the prediction event indicates that the VNF is projected to reach a threshold level of capacity within a predetermined period of time.

12. The method of claim 11 , wherein instructing the virtual infrastructure manager to adjust the number of VNF components comprises instructing the virtual infrastructure manager to increase the number of VNF components.

13. The method of claim 10 , further comprising, in response to the notification, instructing the virtual infrastructure manager to provision an additional virtual machine.

14. The method of claim 10 , wherein the prediction event indicates that the VNF is projected to fall below a threshold level of capacity within a predetermined period of time.

15. The method of claim 14 , further comprising, wherein instructing the virtual infrastructure manager to adjust the number of VNF components comprises instructing the virtual infrastructure manager to decrease the number of VNF components.

16. The method of claim 10 , wherein the analytics component uses machine-learning techniques.

17. The method of claim 10 , wherein the analytics component uses linear regression and Pearson correlation coefficient techniques.

18. A system comprising:

a Virtual Network Function (VNF) that includes a VNF manager and a plurality of VNF components running on a plurality of virtual machines; and

an analytics component in communication with the VNF, wherein the analytics component includes a processor and a memory having machine readable instructions that when executed by the processor, cause the system to:

receive performance data from the VNF;

perform an analytics function on the performance data;

determine a prediction event in response to determining that a set of conditions within the performance data is present; and

notify the VNF manager of the prediction event;

wherein the analytics function detects how changes in a number of one type of VNF component affects demand for a different type of VNF component based on the performance data.

19. The system of claim 18 , wherein the VNF manager is configured to increase a number of VNF components in response to being notified of the prediction event.

20. The system of claim 18 , wherein the VNF manager is configured to define prediction events for the analytics component.