Hybrid Program Balancing

1 . A method of balancing loads in a system having a plurality of processing elements, the method comprising:

executing a plurality of load balancing algorithms in a dry run on load data from the system;

recording results of each of the load balancing algorithms;

evaluating the results of each of the load balancing algorithms;

selecting the load balancing algorithm providing the best results; and

implementing the results of the selected algorithm on the system,

wherein the plurality of load balancing algorithms are each a set of different rules for solving a problem in a finite number of steps.

2 . The method of claim 1 , wherein the evaluating comprises computing a difference between the monitored load of the processing element having the highest load in the system and the calculated load of the processing element having the highest load in the dry run and dividing the difference by the number of moves of load data among the plurality of processing elements needed to obtain the difference.

3 . The method of claim 1 , wherein the algorithms are executed at a periodic interval.

4 . (canceled)

5 . The method of claim 1 , wherein a gain from implementing the results in the system is measured by a reduction of load on the most loaded processing element and the cost is measured by a number of moves of load data among the plurality of processing elements needed to obtain the reduction.

6 . The method of claim 1 , wherein in selecting the load balancing algorithm providing the best results, a gain from implementing the results in the system represented by a reduction of load on the highest loaded processing element is one of a plurality of determining factors in selecting the load balancing algorithm.

7 . The method of claim 6 , wherein the implemented results provide the gain for a specified number of moves of load data among the plurality of processing elements.

8 . The method of claim 1 , wherein in selecting the load balancing algorithm providing the best results, the implemented results that minimize the load on the processing element having the highest load is one of a plurality of determining factors in selecting the load balancing algorithm.

9 . The method of claim 1 , wherein in selecting the load balancing algorithm providing the best results, the implemented results that maximize the load on the processing element having the lowest load is one of a plurality of determining factors in selecting the load balancing algorithm.

10 . The method of claim 1 , wherein the evaluating comprises computing the difference between the load on the processing element having the highest load and the processing element having the lowest load.

11 . (canceled)

12 . The method of claim 11 , wherein in selecting the load balancing algorithm providing the best results, algorithms that have a low rate of success in balancing loads are excluded from execution, and wherein algorithms that have a low rate of success are algorithms that are not able to balance the load among the plurality of processing elements as well as others among the plurality of load balancing algorithms.

13 . (canceled)

14 . (Originally Filed) The method of claim 11 , wherein the choice is rule based, the rules being set by commands.

15 . The method of claim 11 , wherein the choice is made on an automatic basis by a self-learning system.

16 . The method of claim 1 , wherein the processing elements are in a virtual machine.

17 . The method of claim 1 , wherein the processing elements form a cluster interconnect.

18 . The method of claim 1 , wherein the loads that are balanced are caused by programs.

19 . (canceled)

20 . The method of claim 1 , wherein the loads that are balanced are caused by virtual machines.

21 . The method of claim 1 , wherein the results of a selected algorithm is a placement list.

22 . The method of claim 1 , wherein the processing elements are servers containing content and the loads correspond to traffic interest in content on the servers.

23 . (canceled)

24 . The method of claim 1 , wherein the plurality of load balancing algorithms comprise a distribution algorithm, a move big job algorithm, a greedy moves algorithm, a partitioning algorithm, and a hybrid algorithm comprising one or more of the distribution, move big job, greedy moves and partitioning algorithms.

25 . The method of claim 2 , wherein the load data is software, programs, processes, virtual machines or data that places a load on one or more of the plurality of processing elements.