# US 12,192,082 B1 · Telemetry data abstraction > Agents: this data is available as JSON from the patentgrep API and MCP server — see https://patentgrep.com/llms.txt - Number: US 12,192,082 B1 - Document ID: US12192082B1 - Status: Granted - Kind: B1 - Application number: 18/232,299 - Filed: 2023-08-09 - Granted: 2025-01-07 - Earliest claimed: 2023-08-09 - Assignees: Mellanox Technologies, Ltd. - Applicants: Mellanox Technologies, Ltd. - Inventors: Ran Sandhaus; Vladimir Shalikashvili; Ortal Bashan - Web page: https://patentgrep.com/patent/US12192082B1 ## CPC - H04L 43/08: Monitoring or testing based on specific metrics, e.g. QoS, energy consumption or environmental parameters - G06F 11/3409: {for performance assessment} - H04L: TRANSMISSION OF DIGITAL INFORMATION, e.g. TELEGRAPHIC COMMUNICATION (arrangements common to telegraphic and telephonic communication H04M) - G06F: ELECTRIC DIGITAL DATA PROCESSING (computer systems based on specific computational models G06N) ## Abstract Methods, systems, and computer program products to generate a telemetry pipeline. In embodiments, the system includes a communication interface that receives one or more user-defined functions for the telemetry pipeline. The system also includes control logic that implements programmatically the one or more user-defined functions to collect telemetry data at a plurality of layers in the telemetry pipeline based on the one or more user-defined functions and calculate smart metrics at different layers of the plurality of layers in the telemetry pipeline. The smart metrics may be calculated at a layer closest to where associated telemetry data is collected. ## Claims ### Claim 1 A system to generate a telemetry pipeline, the system comprising: an interface that receives one or more user-defined functions for the telemetry pipeline; and a processor including logic that: implements programmatically the one or more user-defined functions to: collect telemetry data at a plurality of Operating System (OS) layers in the telemetry pipeline based on the one or more user-defined functions; and calculate smart metrics at different layers of the plurality of OS layers in the telemetry pipeline, wherein the smart metrics are built in a hierarchal manner with increasing complexity from a lower layer to an upper layer, wherein the smart metrics are calculated at a layer closest to where associated telemetry data is collected, and wherein the smart metrics aggregate the collected telemetry data. ### Claim 2 (depends on 1) The system of claim 1, wherein calculating the smart metrics comprises taking a mathematical function of a collected metric. ### Claim 3 (depends on 2) The system of claim 2, wherein the collected metric comprises at least one of a counter, a state, and an event. ### Claim 4 (depends on 2) The system of claim 2, wherein the collected metric comprises a scalar metric (collected at a point in time). ### Claim 5 (depends on 2) The system of claim 2, wherein the collected metric comprises a sequence/time-based metric. ### Claim 6 (depends on 1) The system of claim 1, wherein the smart metrics are transferred between the plurality of OS layers in the telemetry pipeline. ### Claim 7 (depends on 1) The system of claim 1, wherein the telemetry pipeline spans across each OS layer of the system. ### Claim 8 (depends on 1) The system of claim 1, wherein the plurality of OS layers includes a hardware layer. ### Claim 9 (depends on 1) The system of claim 1, wherein the plurality of OS layers includes a firmware layer. ### Claim 10 (depends on 1) The system of claim 1, wherein the plurality of OS layers includes a software layer. ### Claim 11 (depends on 1) The system of claim 1, wherein the plurality of OS layers includes a user space layer. ### Claim 12 (depends on 1) The system of claim 1, wherein the plurality of OS layers includes a telemetry collector layer. ### Claim 13 (depends on 1) The system of claim 1, wherein the telemetry pipeline uses domain specific language (DSL). ### Claim 14 (depends on 1) The system of claim 1, wherein the logic further: aggregates the telemetry data and the smart metrics across the plurality of OS layers in the telemetry pipeline. ### Claim 15 A system comprising: an interface that receives one or more user-defined functions for the system; and one or more processors to: implement programmatically the one or more user-defined functions; collect telemetry data at a plurality of Operating System (OS) layers in a telemetry pipeline based on the one or more user-defined functions; calculate smart metrics at different layers in the plurality of OS layers in the telemetry pipeline, wherein the smart metrics are built in a hierarchal manner with increasing complexity from a lower layer to an upper layer, and wherein the smart metrics are calculated at a layer closest to where associated telemetry data is collected; and aggregate the telemetry data and the smart metrics across the plurality of OS layers in the telemetry pipeline. ### Claim 16 (depends on 15) The system of claim 15, wherein calculating the smart metrics comprise taking a mathematical function of a collected metric, and wherein the collected metric comprises at least one of a counter, a state, and an event. ### Claim 17 (depends on 16) The system of claim 16, wherein the collected metric comprises a scalar metric (collected at a point in time) or a sequence/time-based metric. ### Claim 18 (depends on 15) The system of claim 15, wherein the plurality of OS layers includes: a hardware layer, a firmware layer, a software layer, a user space layer, and a telemetry collector layer. ### Claim 19 (depends on 15) The system of claim 15, wherein the telemetry pipeline uses domain specific language (DSL). ### Claim 20 A method of using a telemetry pipeline to abstract telemetry data, the method comprising: receiving one or more user-defined functions for a system; implementing programmatically the one or more user-defined functions; collecting telemetry data at a plurality of Operating System (OS) layers in the telemetry pipeline based on the one or more user-defined functions; calculating smart metrics at different layers in the plurality of OS layers in the telemetry pipeline, wherein calculating the smart metrics comprises taking a mathematical function of the telemetry data, wherein the smart metrics are built in a hierarchal manner with increasing complexity from a lower layer to an upper layer, and wherein the smart metrics are calculated at a layer closest to where respective telemetry data is collected; and aggregating the telemetry data and the smart metrics across the plurality of OS layers in the telemetry pipeline.