DYNAMIC NETWORK LOAD FORECASTING

• US 20130080617A1
• Filed: 09/22/2011
• Published: 03/28/2013
• Est. Priority Date: 09/22/2011
• Status: Active Grant

• Current Assignee: SAP SE
• Original Assignee: SAP AG (SAP SE)
• Inventors: Driesen, Volker, Eberlein, Peter
• Granted Patent: US 9,137,130 B2
• Time in Patent Office: Days
• US Class Current: 709/224
• CPC Class Codes: G06F 11/3442 for planning or managing the needed capacity
• CPC Class Codes: H04L 41/147 for predicting network behaviour
• CPC Class Codes: H04L 43/0876 Network utilisation, e.g. volume of load or congestion level
• CPC Class Codes: H04L 47/822 Collecting or measuring resource availability data
• CPC Class Codes: H04L 47/823 Prediction of resource usage
• CPC Class Codes: H04L 47/83 based on usage prediction
• CPC Class Codes: H04L 67/306 User profiles

Abstract

The present disclosure involves computer-implemented methods, software, and systems for optimizing the utilization of network infrastructure resources through predictive analysis. A computer-implemented method includes calculating, using at least one hardware processor, at least one user network load profile for each user profile of a set of user profiles, calculating a predicted dynamic network load using the at least one user network load profile for each user profile of the set of user profiles, transferring the predicted dynamic network load to an at least one network infrastructure resource, and adjusting the capacity of the at least one available network infrastructure resource.

Claims

1. A computer-implemented dynamic network load forecast method, comprising:

calculating, using at least one hardware processor, at least one user network load profile for each user profile of a set of user profiles;

calculating a predicted dynamic network load using the at least one user network load profile for each user profile of the set of user profiles;

transferring the predicted dynamic network load to an at least one network infrastructure resource; and

adjusting a capacity of the least one available network infrastructure resource.

2. The computer-implemented method of claim 1, wherein the calculation of the at least one user network load profile includes:

storing at least one resource usage time measurement associated with the user profile for a defined first time unit as defined by at least one resource measurement value;

calculating at least one first average resource usage time measurement and at least one first resource usage time standard deviation using the stored at least one resource usage time measurement and the at least one resource measurement value;

calculating at least one second average resource usage time measurement utilizing the at least one first average resource usage time measurement in relation to a second defined time unit;

calculating at least one integrated capacity resource usage measurement for the at least one second defined time unit utilizing the at least one second average resource usage time measurement; and

calculating an overall capacity demand change utilizing the at least one integrated capacity resource usage measurement for at least one identically defined second time unit.

3. The computer-implemented method of claim 2, wherein the stored at least one resource usage time measurement is normalized to a geographic reference associated with the user profile.

4. The computer-implemented method of claim 3, wherein the geographical reference is a time zone.

5. The computer-implemented method of claim 1, wherein the predicted dynamic network load is based upon parameters including an over-provisioning factor and the sum as a factor of time of at the least one user network load profile for each user profile of the set of user profiles, a standard deviation parameter and a user-profile-specific standard deviation parameter associated with each user profile of the set of user profiles.

6. The computer-implemented method of claim 5, wherein the over-provisioning factor may be reduced over time.

7. The computer-implemented method of claim 5, wherein the user-profile-specific standard deviation parameter is associated with user profiles of the set of user profiles having a non-regular network demand schedule.

8. The computer-implemented method of claim 1, wherein the adjustment of the capacity of the available network infrastructure resource is performed by one of a virtual machine manager or an application server.

9. The system of claim 5, further comprising:

recording the capacity of the at least one available network infrastructure resource;

recording the calculated predicted dynamic network load; and

calculating an over-provisioning buffer value as the difference of the capacity of the at least one available network infrastructure resource and the predicted dynamic network load.

10. The system of claim 9, further comprising adjusting the over-provisioning factor based upon the calculated value of the over-provisioning buffer value.

11. A computer-program product for dynamic forecast of a network load, the computer program product comprising computer-readable instructions embodied on tangible, non-transitory media, the instructions operable when executed to:

calculate, using at least one hardware processor, at least one user network load profile for each user profile of a set of user profiles;

calculate a predicted dynamic network load using the at least one user network load profile for each user profile of the set of user profiles;

transfer the predicted dynamic network load to an at least one network infrastructure resource; and

adjust a capacity of at least one available network infrastructure resource.

12. The computer-program product of claim 11, wherein the calculation of the at least one user network load profile includes:

storing at least one resource usage time measurement associated with the user profile for a defined first time unit as defined by at least one resource measurement value;

calculating at least one first average resource usage time measurement and at least one first resource usage time standard deviation using the stored at least one resource usage time measurement and the at least one resource measurement value;

calculating at least one second average resource usage time measurement utilizing the at least one first average resource usage time measurement in relation to a second defined time unit;

calculating at least one integrated capacity resource usage measurement for the at least one second defined time unit utilizing the at least one second average resource usage time measurement; and

calculating an overall capacity demand change utilizing the at least one integrated capacity resource usage measurement for at least one identically defined second time unit.

13. The computer-program product of claim 12, wherein the stored at least one resource usage time measurement is normalized to a geographic reference associated with the user profile.

14. The computer-program product of claim 13, wherein the geographical reference is a time zone.

15. The computer-implemented method of claim 11, wherein the predicted dynamic network load is based upon parameters including an over-provisioning factor and the sum as a factor of time of at the least one user network load profile for each user profile of the set of user profiles, a standard deviation parameter and a user-profile-specific standard deviation parameter associated with each user profile of the set of user profiles.

16. The computer-program product of claim 15, wherein the over-provisioning factor may be reduced over time.

17. The computer-program product of claim 15, wherein the user-profile-specific standard deviation parameter is associated with user profiles of the set of user profiles having a non-regular network demand schedule.

18. The computer-program product of claim 11, wherein the adjustment of the capacity of the available network infrastructure resource is performed by one of a virtual machine manager or an application server.

19. The system of claim 15, further comprising:

recording the capacity of the at least one available network infrastructure resource;

recording the calculated predicted dynamic network load; and

calculating an over-provisioning buffer value as the difference of the capacity of the at least one available network infrastructure resource and the predicted dynamic network load.

20. The system of claim 19, further comprising adjusting the over-provisioning factor based upon the calculated value of the over-provisioning buffer value.

21. A system for dynamic forecast of a network load, comprising:

memory operable to store a set of user profiles; and

at least one hardware processor operable to;

calculate at least one user network load profile for each user profile of a set of user profiles;

calculate a predicted dynamic network load using the at least one user network load profile for each user profile of the set of user profiles;

transfer the predicted dynamic network load to an at least one network infrastructure resource; and

adjust a capacity of at least one available network infrastructure resource.

22. The system of claim 21, wherein the calculation of the at least one user network load profile includes:

storing at least one resource usage time measurement associated with the user profile for a defined first time unit as defined by at least one resource measurement value;

calculating at least one first average resource usage time measurement and at least one first resource usage time standard deviation using the stored at least one resource usage time measurement and the at least one resource measurement value;

calculating at least one second average resource usage time measurement utilizing the at least one first average resource usage time measurement in relation to a second defined time unit;

calculating at least one integrated capacity resource usage measurement for the at least one second defined time unit utilizing the at least one second average resource usage time measurement; and

calculating an overall capacity demand change utilizing the at least one integrated capacity resource usage measurement for at least one identically defined second time unit.

23. The system of claim 22, wherein the stored at least one resource usage time measurement is normalized to a geographic reference associated with the user profile.

24. The system of claim 23, wherein the geographical reference is a time zone.

25. The computer-implemented method of claim 21, wherein the predicted dynamic network load is based upon parameters including an over-provisioning factor and the sum as a factor of time of at the least one user network load profile for each user profile of the set of user profiles, a standard deviation parameter and a user-profile-specific standard deviation parameter associated with each user profile of the set of user profiles.

26. The system of claim 25, wherein the over-provisioning factor may be reduced over time.

27. The system of claim 25, wherein the user-profile-specific standard deviation parameter is associated with user profiles of the set of user profiles having a non-regular network demand schedule.

28. The system of claim 21, wherein the adjustment of the capacity of the available network infrastructure resource is performed by one of a virtual machine manager or an application server.

29. The system of claim 25, further operable to:

record the capacity of the at least one available network infrastructure resource;

record the calculated predicted dynamic network load; and

calculate an over-provisioning buffer value as the difference of the capacity of the at least one available network infrastructure resource and the predicted dynamic network load.

30. The system of claim 29, further operable to adjust the over-provisioning factor based upon the calculated value of the over-provisioning buffer value.