Network telemetry based on application-level information

Patent metadata

Field	Value
Publication number	US-11848837-B2
Application number	US-202117504517-A
Country	US
Kind code	B2
Filing date	Oct 19, 2021
Priority date	Oct 19, 2021
Publication date	Dec 19, 2023
Grant date	Dec 19, 2023

Abstract

A network device includes processing circuitry and one or more ports. The one or more ports are configured to connect to a communication network. The processing circuitry is configured to receive a packet originating from a network node running an application program, the packet includes application-level metadata relating to the application program, to generate telemetry data based at least on the application-level metadata, and to transmit the telemetry data via one of the ports, over the communication network.

First claim

The invention claimed is:

1. A network device, comprising:
a plurality of ports, to connect to a communication network;
packet processing circuitry, to forward packets between the plurality of ports, including receiving, over the communication network, a packet originating from a remote host running an application program, the packet comprising application-level metadata relating to the application program, and forwarding the packet over the communication network to another remote host running the application program; and
a telemetry processing circuitry, to:
hold a telemetry policy that specifies conditions for triggering telemetry data generation;
receive the application-level metadata of the received packet from the packet processing circuitry;
generate, in response to a match between the application-level metadata in the received packet and one or more of the conditions specified in the telemetry policy, telemetry data indicative of a network traffic condition of the network device affecting the received packet;
tag the generated telemetry data with a tag that associates the network traffic condition of the network device with the application-level metadata of the application running on the remote host; and
transmit the tagged telemetry data via one of the plurality of ports, over the communication network to a system management node.

2. The network device according to claim 1 , wherein the application-level metadata comprises an application identifier pre-assigned to the application program.

3. The network device according to claim 1 , wherein the application program has multiple execution states, and wherein the application-level metadata is indicative of a selected execution state among the multiple execution states.

4. The network device according to claim 3 , wherein the telemetry processing circuitry is to generate the telemetry data in response to detecting that the application-level metadata comprises the selected execution state and not any other execution state.

5. The network device according to claim 1 , wherein the telemetry processing circuitry is to tag one or more parts of the telemetry data with respective tags related to the application program, based on the application-level metadata.

6. The network device according to claim 1 , wherein the telemetry processing circuitry is to trigger generation of the telemetry data in response to the application-level metadata.

7. The network device according to claim 1 , wherein the telemetry processing circuitry is to generate at least part of the telemetry data independently of the application-level metadata.

8. The network device according to claim 1 , wherein the telemetry processing circuitry is to receive telemetry data collected by one or more other network elements coupled to the communication network, and to analyze the received telemetry data for presentation.

9. The network device according to claim 1 , wherein, based on the application-level metadata, the processing circuitry is to mirror at least part of a packet comprising the telemetry data back to the remote network node that has sent the packet comprising the application-level metadata, for analysis.

10. The network device according to claim 1 , further comprising a processor to run at least part of an analyzer program that analyzes the telemetry data based at least on parts in the telemetry data that are tagged in relation to the application program.

11. The network device according to claim 1 , wherein the packet is comprised in a message that the application program sends over the communication network to a peer network node, wherein the application-level metadata is indicative of a position of the packet within the message, and wherein the telemetry processing circuitry is to generate the telemetry data based on the position of the packet within the message.

12. The network device according to claim 1 , wherein the network traffic condition, indicated by the telemetry data, comprises a local traffic bottleneck in the network device.

13. The network device according to claim 1 , wherein the network traffic condition, indicated by the telemetry data, comprises ingress-to-egress latency in the network device.

14. The network device according to claim 1 , wherein the network traffic condition, indicated by the telemetry data, comprises a condition of one or more queues in the network device.

15. A method, comprising:
in a network device that connects to a communication network by a plurality of ports, forwarding packets between the plurality of ports, including receiving over the communication network a packet originating from a remote host running an application program, the packet comprising application-level metadata relating to the application program, and forwarding the packet over the communication network to another remote host running the application program;
holding in the network device a telemetry policy that specifies conditions for triggering telemetry data generation;
generating, in response to a match between the application-level metadata in the received packet and one or more of the conditions specified in the telemetry policy, telemetry data indicative of a network traffic condition of the network device affecting the received packet;
tagging the generated telemetry data with a tag that associates the network traffic condition of the network device with the application-level metadata of the application running on the remote host; and
transmitting the tagged telemetry data, via one of the plurality of ports, over the communication network to a system management node.

16. The method according to claim 15 , wherein the application-level metadata comprises an application identifier pre-assigned to the application program.

17. The method according to claim 15 , wherein the application program has multiple execution states, and wherein the application-level metadata is indicative of a selected execution state among the multiple execution states.

18. The method according to claim 17 , wherein generating the telemetry data comprises generating the telemetry data in response to detecting that the application-level metadata comprises the selected execution state and not any other execution state.

19. The method according to claim 15 , wherein generating the telemetry data comprises tagging one or more parts of the telemetry data, with respective tags related to the application program, based on the application-level metadata.

20. The method according to claim 15 , wherein generating the telemetry data comprises triggering generation of the telemetry data in response to the application-level metadata.

21. The method according to claim 15 , wherein generating the telemetry data comprises generating at least part of the telemetry data independently of the application-level metadata.

22. The method according to claim 15 , and comprising receiving telemetry data collected by one or more other network elements coupled to the communication network, and analyzing the received telemetry data for presentation.

23. The method according to claim 15 , and comprising, based on the application-level metadata, mirroring at least part of a packet comprising the telemetry data back to the remote network node that has sent the packet comprising the application-level metadata, for analysis.

24. The method according to claim 15 , and comprising running at least part of an analyzer program that analyzes the telemetry data

Assignees

• Mellanox Technologies Ltd

Inventors

• Levy Aviad
• Levi Lion
• Bloch Noam
• Bashan Ortal

Classifications

• H04L43/06 Primary

  Generation of reports · CPC title

• H04L43/04

  Processing captured monitoring data, e.g. for logfile generation · CPC title

• H04L67/025

  for remote control or remote monitoring of applications · CPC title

• H04L67/12 Primary

  specially adapted for proprietary or special-purpose networking environments, e.g. medical networks, sensor networks, networks in vehicles or remote metering networks · CPC title