# US 12,602,260 B2 · Computer-based provisioning of cloud resources > Agents: this data is available as JSON from the patentgrep API and MCP server — see https://patentgrep.com/llms.txt - Number: US 12,602,260 B2 - Document ID: US12602260B2 - Status: Granted - Kind: B2 - Application number: 18/185,134 - Filed: 2023-03-16 - Published: 2024-09-19 - Granted: 2026-04-14 - Earliest claimed: 2023-03-16 - Assignees: International Business Machines Corporation - Applicants: International Business Machines Corporation - Inventors: Saurabh Jha; Larisa Shwartz; Frank Bagehorn - Web page: https://patentgrep.com/patent/US12602260B2 ## CPC - G06F 9/505: {considering the load} - G06F 9/5072: {Grid computing} - G06F: ELECTRIC DIGITAL DATA PROCESSING (computer systems based on specific computational models G06N) ## Abstract In an approach to improve enhancing the provisioning cloud resources, embodiments receive a set of potential cloud resource providers and predict a performance for a distributed workload on each potential provider of the set according to a machine learning model. Additionally, embodiments inject fault into the set of potential cloud resource providers and measure an impact of the injected fault upon system performance. Embodiments utilize the injected fault to create a system intervention to ensure that the system intervention is carried out on a network in predetermined system. Further, embodiments provision the distributed workload among the potential cloud resource providers according to dynamic conditions output by the set of potential cloud resources based on the measured impact of the injected fault. ## Claims ### Claim 1 A computer-implemented method for provisioning cloud resources, the computer-implemented method comprising: receiving a set of potential cloud resource providers; predicting a performance for a distributed workload on each potential provider of the set according to a machine learning model; injecting fault into the set of potential cloud resource providers, wherein the fault is utilized to create a system intervention, and wherein the fault ensures that the system intervention is carried out on a network in predetermined system, and wherein the fault is injected to mimic the different cloud provider characteristics, wherein a system-wide intervention alleviates the need to execute the workload on each cloud provider with different fault conditions; measuring an impact of the injected fault upon system performance; and provisioning the distributed workload among the potential cloud resource providers according to dynamic conditions output by the set of potential cloud resources based on the measured impact of the injected fault. ### Claim 2 (depends on 1) The method according to claim 1, further comprising: determining a profile for a new potential cloud resource provider; and predicting the performance for the distributed workload using the profile. ### Claim 3 (depends on 1) The method according to claim 1, further comprising: outputting a ranking of results of the set of potential cloud providers; and selecting a preferred cloud provider or a cloud provider configuration with ranked results that are above a predetermined threshold, wherein the ranked results are based on analyzed benchmarks associated with an executed distributed workload, and wherein the preferred cloud provider may be either a single cloud or multiple cloud computer configuration. ### Claim 4 (depends on 1) The method according to claim 1, further comprising: generating a profile for each known cloud resource provider; training the machine learning model using the impact; and predicting the performance according to the machine learning model. ### Claim 5 (depends on 4) The method according to claim 4, wherein the predicting comprises: collecting data on system utilization during a training job; and predicting the performance using the data and the machine learning model. ### Claim 6 (depends on 1) The method according to claim 1, further comprising: determining that the performance fails to a threshold requirement; and altering the provisioning according to the performance. ### Claim 7 (depends on 1) The method according to claim 1, wherein the machine learning model comprises a system model database and an application sensitivity model. ### Claim 8 A computer system for provisioning cloud resources, the computer system comprising: one or more computer processors; one or more computer readable storage devices; program instructions stored on the one or more computer readable storage devices for execution by at least one of the one or more computer processors, the stored program instructions comprising: program instructions to receive a set of potential cloud resource providers; program instructions to predict a performance for a distributed workload on each potential provider of the set according to a machine learning model; program instructions to inject fault into the set of potential cloud resource providers, wherein the fault is utilized to create a system intervention, and wherein the fault ensures that the system intervention is carried out on a network in predetermined system, and wherein the fault is injected to mimic the different cloud provider characteristics, wherein a system-wide intervention alleviates the need to execute the workload on each cloud provider with different fault conditions; program instructions to measure an impact of the injected fault upon system performance; and program instructions to provision the distributed workload among the potential cloud resource providers according to dynamic conditions output by the set of potential cloud resources based on the measured impact of the injected fault. ### Claim 9 (depends on 8) The computer system according to claim 8, further comprising: program instructions to determine a profile for a new potential cloud resource provider; and program instructions to predict the performance for the distributed workload using the profile. ### Claim 10 (depends on 8) The computer system according to claim 8, further comprising: program instructions to output a ranking of results of the set of potential cloud providers; and program instructions to select a preferred cloud provider or a cloud provider configuration with ranked results that are above a predetermined threshold, wherein the ranked results are based on analyzed benchmarks associated with an executed distributed workload, and wherein the preferred cloud provider may be either a single cloud or multiple cloud computer configuration. ### Claim 11 (depends on 8) The computer system according to claim 8, further comprising: program instructions to generate a profile for each known cloud resource provider; program instructions to train the machine learning model using the impact; and program instructions to predict the performance according to the machine learning model. ### Claim 12 (depends on 11) The computer system according to claim 11, wherein the predicting comprises: program instructions to collect data on system utilization during a training job; and program instructions to predict the performance using the data and the machine learning model. ### Claim 13 (depends on 8) The computer system according to claim 8, further comprising: program instructions to determine that the performance fails to a threshold requirement; and program instructions to alter the provisioning according to the performance. ### Claim 14 (depends on 8) The computer system according to claim 8, wherein the machine learning model comprises a system model database and an application sensitivity model. ### Claim 15 A computer program product for implementing additive manufacturing, the computer program product comprising: one or more computer readable storage devices and program instructions stored on the one or more computer readable storage devices, the stored program instructions comprising: program instructions to receive a set of potential cloud resource providers; program instructions to predict a performance for a distributed workload on each potential provider of the set according to a machine learning model; program instructions to inject fault into the set of potential cloud resource providers, wherein the fault is utilized to create a system intervention, and wherein the fault ensures that the system intervention is carried out on a network in predetermined system, and wherein the fault is injected to mimic the different cloud provider characteristics, wherein a system-wide intervention alleviates the need to execute the workload on each cloud provider with different fault conditions; program instructions to measure an impact of the injected fault upon system performance; and program instructions to provision the distributed workload among the potential cloud resource providers according to dynamic conditions output by the set of potential cloud resources based on the measured impact of the injected fault. ### Claim 16 (depends on 15) The computer program product according to claim 15, further comprising: program instructions to determine a profile for a new potential cloud resource provider; and program instructions to predict the performance for the distributed workload using the profile. ### Claim 17 (depends on 15) The computer program product according to claim 15, further comprising: program instructions to output a ranking of results of the set of potential cloud providers; and program instructions to select a preferred cloud provider or a cloud provider configuration with ranked results that are above a predetermined threshold, wherein the ranked results are based on analyzed benchmarks associated with an executed distributed workload, and wherein the preferred cloud provider may be either a single cloud or multiple cloud computer configuration. ### Claim 18 (depends on 15) The computer program product according to claim 15, further comprising: program instructions to generate a profile for each known cloud resource provider; program instructions to train the machine learning model using the impact; program instructions to predict the performance according to the machine learning model; program instructions to collect data on system utilization during a training job; and program instructions to predict the performance using the data and the machine learning model. ### Claim 19 (depends on 15) The computer program product according to claim 15, further comprising: program instructions to determine that the performance fails to a threshold requirement; and program instructions to alter the provisioning according to the performance. ### Claim 20 (depends on 15) The computer program product according to claim 15, wherein the machine learning model comprises a system model database and an application sensitivity model.