# US 12,009,989 B2 · On demand synthetic data matrix generation > Agents: this data is available as JSON from the patentgrep API and MCP server — see https://patentgrep.com/llms.txt - Number: US 12,009,989 B2 - Document ID: US12009989B2 - Status: Granted - Kind: B2 - Application number: 17/037,501 - Filed: 2020-09-29 - Published: 2021-01-14 - Granted: 2024-06-11 - Earliest claimed: 2017-11-03 - Assignees: Salesforce, Inc. - Applicants: salesforce.com, inc. - Inventors: Tejaswini Ganapathi; Satish Raghunath; Xu Che; Shauli Gal; Andrey Karapetov - Web page: https://patentgrep.com/patent/US12009989B2 ## CPC - H04L 41/145: {involving simulating, designing, planning or modelling of a network} - G05B 17/02: electric - G06F 16/2477: {Temporal data queries} - G06F 17/16: Matrix or vector computation {, e.g. matrix-matrix or matrix-vector multiplication, matrix factorization (matrix transposition G06F7/78)} - G06N 7/01: Probabilistic graphical models, e.g. probabilistic networks - G06N 20/00: Machine learning - H04L 41/142: using statistical or mathematical methods - H04L 43/0829: Packet loss - H04L 43/0858: {One way delays} - H04L 43/087: Jitter - H04L 43/0888: Throughput - H04L: TRANSMISSION OF DIGITAL INFORMATION, e.g. TELEGRAPHIC COMMUNICATION (arrangements common to telegraphic and telephonic communication H04M) - G05B: CONTROL OR REGULATING SYSTEMS IN GENERAL; FUNCTIONAL ELEMENTS OF SUCH SYSTEMS; MONITORING OR TESTING ARRANGEMENTS FOR SUCH SYSTEMS OR ELEMENTS (systems for controlling or regulating non-electric variables G05D; systems for regulating electric or magnetic variables G05F; control devices or systems insofar as characterised by mechanical features only G05G) - G06F: ELECTRIC DIGITAL DATA PROCESSING (computer systems based on specific computational models G06N) - G06N: COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS ## Abstract An data driven approach to generating synthetic data matrices is presented. By retrieving historical network traffic data, probabilistic models are generated. Optimal distribution families for a set of independent data segments are determined. Applications are tested and performance metrics are determined based on the generated synthetic data matrices. ## Claims ### Claim 1 A method, comprising: accessing historical network traffic data corresponding to data requests sent between a set of client devices and a server system; based on the historical network traffic data, generating a set of synthetic data requests for a particular application; generating one or more synthetic download outcomes for the set of synthetic data requests; determining a set of performance metrics of the particular application by comparing, based on the one or more synthetic download outcomes, performance of the particular application under different combinations of network values; generating TCP parameters for the particular application based on the set of performance metrics, wherein the TCP parameters improve network performance of the particular application; and causing a client device that is executing the particular application to be configured with the TCP parameters. ### Claim 2 (depends on 1) The method of claim 1, wherein the set of performance metrics include inferred performance metrics derived from application performance simulations. ### Claim 3 (depends on 1) The method of claim 1, wherein the one or more synthetic download outcomes are generated from a set of probabilistic models derived based on the historical network traffic data. ### Claim 4 (depends on 1) The method of claim 1, wherein the comparing further includes comparing end user experience and application performance across different combinations of geography and network types. ### Claim 5 (depends on 1) The method of claim 1, further comprising: generating aggregate datasets of performance metrics comparing different combinations of network values that affect application performance. ### Claim 6 (depends on 1) The method of claim 1, wherein the historical network traffic data includes network conditions reported by an agent installed on the client device. ### Claim 7 One or more non-transitory computer-readable storage media, storing one or more sequences of instructions, which when executed by one or more processors cause performance of: accessing historical network traffic data corresponding to data requests sent between a set of client devices and a server system; based on the historical network traffic data, generating a set of synthetic data requests for a particular application; generating one or more synthetic download outcomes for the set of synthetic data requests; determining a set of performance metrics of the particular application by comparing, based on the one or more synthetic download outcomes, performance of the particular application under different combinations of network values; generating TCP parameters for the particular application based on the set of performance metrics, wherein the TCP parameters improve network performance of the particular application; and causing a client device that is executing the particular application to be configured with the TCP parameters. ### Claim 8 (depends on 7) The one or more non-transitory computer-readable storage media of claim 7, wherein the set of performance metrics include inferred performance metrics derived from application performance simulations. ### Claim 9 (depends on 7) The one or more non-transitory computer-readable storage media of claim 7, wherein the set of performance metrics include performance metrics derived from probabilistic modeling. ### Claim 10 (depends on 7) The one or more non-transitory computer-readable storage media of claim 7, wherein the comparing further includes comparing end user experience and application performance across different combinations of geography and network types. ### Claim 11 (depends on 7) The one or more non-transitory computer-readable storage media of claim 7, wherein the one or more sequences of instructions, which when executed by the one or more processors cause further performance of: generating aggregate datasets of performance metrics comparing different combinations of network values that affect application performance. ### Claim 12 (depends on 7) The one or more non-transitory computer-readable storage media of claim 7, wherein the historical network traffic data includes network conditions reported by an agent installed on the client device. ### Claim 13 An apparatus, comprising: one or more processors; and a memory storing instructions, which when executed by the one or more processors, cause the one or more processors to: access historical network traffic data corresponding to data requests sent between a set of client devices and a server system; based on the historical network traffic data, generate a set of synthetic data requests for a particular application; generate one or more synthetic download outcomes for the set of synthetic data requests; determine a set of performance metrics of the particular application by comparing, based on the one or more synthetic download outcomes, performance of the particular application under different combinations of network values; generate TCP parameters for the particular application based on the set of performance metrics, wherein the TCP parameters improve network performance of the particular application; and cause a client device that is executing the particular application to be configured with the TCP parameters. ### Claim 14 (depends on 13) The apparatus of claim 13, wherein the set of performance metrics include inferred performance metrics derived from application performance simulations. ### Claim 15 (depends on 13) The apparatus of claim 13, wherein the set of performance metrics include performance metrics derived from a set of probabilistic models. ### Claim 16 (depends on 13) The apparatus of claim 13, wherein the comparing further includes comparing end user experience and application performance across different combinations of geography and network types. ### Claim 17 (depends on 13) The apparatus of claim 13, wherein the instructions, which when executed by the one or more processors, cause the one or more processors to further: generate aggregate datasets of performance metrics comparing different combinations of network values that affect application performance.