Forensics for network switching diagnosis

1. A system, comprising:

memory;

a sensor coupled with network switch circuitry, the sensor configured to generate data related to a switch attribute of the network switch circuitry; and

a processor coupled to the sensor and the memory, the processor configured to:

analyze the data from the sensor to monitor operation of the network switch circuitry for a determined condition of the switch attribute of the network switch circuitry;

detect the determined condition of the switch attribute during the operation of the network switch circuitry;

generate an informational packet responsive to detecting the determined condition, the informational packet including information on the determined condition;

generate an informational packet aggregate by grouping the informational packet with a selected additional informational packet already present in the memory;

using the informational packet aggregate as an input, track the determined condition over time; and

analyze the determined condition over time to diagnose a cause of the determined condition.

2. The system of claim 1 , where the processor is further configured to adjust the switch attribute to reduce a likelihood of reoccurrence of the determined condition by adjusting an operational range of the switch attribute.

3. The system of claim 1 , where the processor is further configured to adjust the switch attribute to reduce a likelihood of reoccurrence of the determined condition by adjusting a ratio of shared space to reserved space in a buffer.

4. The system of claim 1 , where:

the determined condition is one of multiple determined conditions; and

the processor is configured to:

generate a performance report for the network switch circuitry responsive to the multiple determined conditions; and

send the performance report to a terminal for presentation to an operator.

5. The system of claim 1 , where the processor is further configured to execute a forensic command in response to the determined condition, where executing the command causes generation of the informational packet.

6. The system of claim 1 , where the processor is further configured to send the informational packet to a specified target for receipt.

7. The system of claim 1 , where the switch attribute comprises: buffer fill level, buffer counter, power consumption, queue length, temperature, flow identifier, delay, link state, level of congestion, or any combination thereof.

8. The system of claim 1 , where the determined condition comprises: a packet discard, power consumption greater than a determined consumption level, a buffer fill level above a determined level, or any combination thereof.

9. The system of claim 1 , where the informational packet includes a size snapshot comprising a measured memory level for a buffer, a queue, or both.

10. A method, comprising:

monitoring, at a processor, data generated by a sensor coupled to a network switch circuitry;

generating a size snapshot for a buffer of the network switch circuitry, a queue of the network switch circuitry, or both;

detecting a discard threshold crossing for the buffer, the queue, or both responsive to the data;

generating an event trigger packet comprising the size snapshot in response to detecting the discard threshold crossing; and

sending the event trigger packet to a determined storage location to generate aggregated data on the discard threshold crossing for analysis of the discard threshold crossing over time to diagnose a cause of the discard threshold crossing.

11. The method of claim 10 , further comprising adjusting an allocation of space in the buffer of the network switch circuitry responsive to an observed average use of reserve space in the buffer, shared space in the buffer, or both.

12. The method of claim 10 , further comprising:

adjusting an operational parameter of the network switch circuitry to reduce a likelihood of reoccurrence of the discard threshold crossing.

13. The method of claim 10 , wherein generating the size snapshot comprises generating, at the time of the discard threshold crossing, size snapshots for multiple buffers, multiple queues, or both.

14. The method of claim 10 , further comprising adjusting a queue length of the queue of the network switch circuitry responsive to the size snapshot.

15. The method of claim 10 , where sending the event trigger packet comprises sending the event trigger packet to a determined media access control (MAC) layer address for aggregation.

16. A method comprising:

monitoring data generated by a network switch circuitry of a network, the data characterizing an attribute of the network switch circuitry;

establishing a set of predetermined operating conditions for which to detect in the network switch circuitry;

detecting a specific operating condition among the set of predetermined operating conditions that exists at a given time in the network switch circuitry by analyzing the data;

generating, via a processor, an event trigger;

in response to the event trigger, generating an informational packet including information characterizing the specific operating condition;

generating an aggregate of informational packets by sending the informational packet to a specific address designated for the specific operating condition, the aggregate comprising data tracking the specific operating condition over time;

analyzing the data tracking the specific operating condition over time to produce a diagnosis of a cause of the specific operating condition; and

adjusting the attribute responsive to the diagnosis to reduce a likelihood of reoccurrence of the specific operating condition.

17. The method of claim 16 , where the specific address comprises a pre-determined media access control (MAC) layer address of a device assigned to generate the aggregate.

18. The method of claim 16 , where adjusting comprises:

adjusting reserve buffer space, shared buffer space or both in the network switch circuitry.

19. The method of claim 16 , where monitoring comprises:

monitoring buffer fill level, queue length, power consumption, flow identifier, delay, temperature, port link state, queue link state, congestion level, or any combination thereof.

20. The method of claim 16 , where the specific operating condition comprises:

a packet discard, power consumption greater than a determined consumption level, a buffer fill level above a determined level, or any combination thereof.