Adaptive centralized collection of performance management data using a metamodel

1. A method of obtaining performance measurements via metamodels and metadata streamed from multiple types, models, and firmware versions of network devices connected to a network, comprising

installing a generic collection agent on a server connected to said network, wherein byte code enhancement in real-time is used to add support to said generic collection agent to implement said metamodels for at least one of a new performance measurement and a redefined performance measurement;

retrieving said metamodels from said network devices by said collection agent, wherein said metamodels define information used by said collection agent to decode and interpret said metadata streamed from said network devices;

storing said metamodels in a collection of metamodels on said server to be accessed by said collection agent;

streaming said performance measurements as metadata defined by said metamodels from said network devices to said collection agent; and

analyzing and storing said performance measurements by said collection agent for later use,

wherein said collection agent uses real-time parsing routines to create any missing or incomplete methods, objects, or tools necessary to handle said performance measurements according to instructions defined in said metamodels.

2. The method of claim 1 wherein said collection agent detects said network devices and retrieves said metamodels upon detection of said network devices.

3. The method of claim 1 wherein said collection agent sends addressing information to said network devices, thereby allowing said network devices to stream said performance measurements to said collecting agent.

4. The method of claim 1 wherein each network device's metamodels include one or more management groups.

5. The method of claim 2 wherein said collection agent recognizes a network device upon detection and analyzes only the most recent metamodel from that network device.

6. The method of claim 2 wherein said collection agent recognizes a network device upon detection and analyzes only the most recent metamodel from that network device, and

if said collection agent has already retrieved the most recent metamodel, said collection agent uses relevant metamodel information from said stored collection of metamodels, or

if said collection agent has not retrieved the most recent metamodel, said collection agent retrieves and stores said most recent metamodel in said collection of metamodels.

7. The method of claim 5 wherein if said collection agent has previously retrieved and stored said most recent metamodel, said collection agent uses stored metamodel information from said stored collection of metamodels.

8. The method of claim 5 wherein if said collection agent has not previously retrieved said most recent metamodel, said collection agent retrieves and stores said most recent metamodel in said collection of metamodels.

9. The method of claim 4 wherein each management group contains a set of measurements.

10. The method of claim 4 wherein each management group is defined by a unique identifier.

11. The method of claim 4 wherein each management group is associated with a set of streaming intervals.

12. The method of claim 9 wherein measurements within each management group are defined by a set of attributes including

a display name,

a display unit, and

a count type that defines the encoding of the measurement as one of: an integer, a long integer, a double integer, or a null-terminated string.

13. The method of claim 11 wherein said collection agent activates streaming of a management group associated with a network device after analyzing that network device's metamodels.

14. The method of claim 11 wherein said collection agent selects a streaming interval for each management group to stream said performance measurements to said collection agent.

15. The method of claim 13 wherein said collection agent can dynamically modify said streaming interval for each management group.

16. The method of claim 13 wherein said performance measurements are streamed on a stream between each network device and said collection agent.

17. The method of claim 13 wherein said performance measurements are streamed on a stream between each management group within a network device and said collection agent.

18. A method of obtaining performance measurements via metamodels and metadata streamed from multiple types, models, and firmware versions of network devices connected to a network, comprising:

installing a generic collection agent on a server connected to said network, wherein byte code enhancement in real-time is used to add support to said generic collection agent to implement said metamodels for at least one of a new performance measurement and a redefined performance measurement;

retrieving said metamodels from said network devices by said collection agent;

storing said metamodels in a collection of metamodels on said server to be accessed by said collection agent;

streaming said performance measurements as metadata defined by said metamodels from said network devices to said collection agent; and

analyzing and storing, said performance measurements by said collection agent for later use,

wherein said collection agent uses real-time parsing routines to create any missing or incomplete methods, objects, or tools necessary to handle said performance measurements according to instructions defined in said metamodels.