Self-learning correlation of network patterns for agile network operations

1. A method, comprising:

receiving a stream of network messages, the stream of network messages comprising a variety of network events for various network devices;

dividing the stream of network messages into contiguous and non-overlapping time slices, each time slice containing a subset of the stream of network messages;

graphically displaying, via a user interface, the stream of messages;

receiving, via the user interface, an identification of patterns within the stream of network messages, the patterns comprising groupings of the variety of network events;

receiving, via the user interface, a determination of a set of operationalization scenarios for each pattern, wherein determining an operationalization scenario comprises manipulating a subset of the variety of network events; and

operationalizing the patterns as correlation rules in a correlation engine to automatically detect or predict network alarms from input provided by a fault management system.

2. The method of claim 1 , further comprising:

normalizing the stream of network messages.

3. The method of claim 2 , wherein the normalizing comprises:

removing identifying information of the various network devices;

removing timestamps associated with generation of the stream of network messages; and

removing location information of various network devices.

4. The method of claim 1 , further comprising:

analyzing the patterns graphically by an expert user to identify the patterns and the set of operationalization scenarios.

5. The method of claim 1 , wherein manipulating the subset of the variety of network events comprises:

determining that the subset of the variety of network events are repetitive; and

suppressing the subset of the variety of network events by removing the subset of the variety of network events from the stream of network messages.

6. The method of claim 1 , wherein manipulating the subset of the variety of network events comprises:

determining that the subset of the variety of network events are repetitive; and

compressing the subset of the variety of network events by maintaining only one occurrence of a network event of the subset.

7. The method of claim 1 , further comprising:

generating an alarm in response to an identified pattern that correlates to a hidden network issue or anticipating a future problem by triggering an alert or a ticket.

8. The method of claim 1 , further comprising:

increasing a time shift for each time slice to create multiple contiguous network message sequences for identifying the patterns.

9. A system, comprising:

a memory; and

a processor executing instructions from the memory to:

receive a stream of network messages, the stream of network messages comprising a variety of network events for various network devices;

dividing the stream of network messages into contiguous and non-overlapping time slices, each time slice containing a subset of the stream of network messages;

graphically display, via a user interface, the stream of messages;

receive, via the user interface, identification of patterns within the stream of network messages, the patterns comprising groupings of the variety of network events;

receive, via the user interface, a determination of a set of operationalization scenarios for each pattern, wherein determining an operationalization scenario comprises manipulating a subset of the variety of network events; and

operationalize the patterns as correlation rules in a correlation engine to automatically detect or predict network alarms from input provided by a fault management system.

10. The system of claim 9 , wherein the processor further executes the instructions from the memory to:

normalize the stream of network messages.

11. The system of claim 10 , wherein the processor further executes the instructions from the memory to:

remove identifying information of the various network devices;

remove timestamps associated with generation of the stream of network messages; and

remove location information of various network devices.

12. The system of claim 9 , wherein the processor further executes the instructions from the memory to:

analyze the patterns graphically by an expert user to identify the set of operationalization scenarios.

13. The system of claim 9 , wherein manipulating the subset of the variety of network events comprises:

determining that the subset of the variety of network events are repetitive; and

suppressing the subset of the variety of network events by removing the subset of the variety of network events from the stream of network messages.

14. The system of claim 9 , wherein manipulating the subset of the variety of network events comprises:

determining that the subset of the variety of network events are repetitive; and

compressing the subset of the variety of network events by maintaining only one occurrence of a network event of the subset.

15. The system of claim 9 , wherein the processor further executes the instructions from the memory to:

generate an alarm in response to an identified pattern that correlates to a hidden network issue or anticipating a future problem by triggering an alert or a ticket.

16. The system of claim 9 , wherein the processor further executes the instructions from the memory to:

increase a time shift for each time slice to create multiple contiguous network message sequences for identifying the patterns.

17. A non-transitory machine-readable storage medium encoded with instructions executable by at least one hardware processor of a network device, the non-transitory machine-readable storage medium comprising instructions to:

receive a stream of network messages, the stream of network messages comprising a variety of network events for various network devices;

dividing the stream of network messages into contiguous and non-overlapping time slices, each time slice containing a subset of the stream of network messages;

graphically displaying, via a user interface, the stream of messages;

receiving, via the graphical user interface, an indication of patterns within the stream of network messages, the patterns comprising groupings of the variety of network events;

receiving, via the user interface, a determination of a set of operationalization scenarios for each pattern, wherein determining an operationalization scenario comprises manipulating a subset of the variety of network events; and

operationalize the patterns as correlation rules in a correlation engine to automatically detect or predict network alarms from input provided by a fault management system.

18. The method of claim 1 , further comprising:

assigning a first subset of the stream of network messages into a first family of equipment into equivalence classes; and

assigning a second subset of the stream of network messages into a second family of equipment into equivalence classes,

wherein the first family of equipment and the second family of equipment are assigned according to a Locality Sensitive Hashing (LSH) algorithm.

19. The method of claim 1 , wherein the patterns within the stream of network messages are identified using a Frequent Pattern Growth feature of an Frequent Pattern (FP) Growth algorithm.

20. The method of claim 1 , wherein the patterns within the stream of network messages are identified in parallel using a Frequent Pattern Growth feature of a Parallel Frequent Pattern (PFP) Growth algorithm.

21. The method of claim 1 , wherein the set of operationalization comprises at least one of: one of suppression, compression, problem detection, or problem prediction.