DETAILED ACTION
Claims 1-20 are pending.
Notice of Pre-AIA or AIA Status
The present application, filed on or after March 16, 2013, is being examined under the first inventor to file provisions of the AIA .
Claim Rejections - 35 USC § 103
The following is a quotation of 35 U.S.C. 103 which forms the basis for all obviousness rejections set forth in this Office action:
A patent for a claimed invention may not be obtained, notwithstanding that the claimed invention is not identically disclosed as set forth in section 102, if the differences between the claimed invention and the prior art are such that the claimed invention as a whole would have been obvious before the effective filing date of the claimed invention to a person having ordinary skill in the art to which the claimed invention pertains. Patentability shall not be negated by the manner in which the invention was made.
Claims 1-4, 7-8, 11-14, and 16-19 are rejected under 35 U.S.C. 103 as being unpatentable over Gonzalez (US 12124835 B2) in view of Kersch (US 20190044830 A1), Caspi (US 20080065755 A1), and Rosenberg (US 20210026707 A1).
Regarding Claim 1, Gonzalez teaches a method for managing a process performed by a host device, the method comprising:
obtaining at least two status information taken at a first point in time (
Gonzalez discloses, “FIG. 7 is a flowchart that describes an example process 700 for performing real-time updates to a likelihood determination, in accordance with one or more embodiments. In some embodiments, process 700 may begin at operation 702. At operation 702, data of an update occurring at a given time may be obtained. For example, an update to a system (or a process performed by the system) may be obtained. In operation 704, descriptors may be retrieved/received. For example, descriptors describing fixed information about the system (or process performed by the system) may be obtained from descriptor database 136. In operation 706, status updates from prior times may be obtained. For example, with reference to FIG. 6A, at the fourth status update, each update to the system (or process performed by the system) previously obtained may be obtained, e.g., first thru third status updates. In operation 708, a failure/success score (or other performance metric indicating the likelihood that a particular outcome will occur at an expiration time of a process performed by a system) may be computed based on the obtained data, the retrieved/received descriptors, and the obtained prior status updates. In operation 710, the failure/success score (or other performance metric) may be stored, e.g., in results database 516. In operation 712, a determination may be made as to whether the expiration time has been reached. The expiration time refers to a predefined time when a particular process (the process with which the trained machine learning model is configured to predict the likelihood of) ends,” Col 16, Lines 8-20, “Computing system 800 may include one or more processors (e.g., processors 810-1-810-N) coupled to system memory 820, an input/output I/O device interface 830, and a network interface 840 via an input/output (I/O) interface 850,” Col 16, Lines 54-58, “I/O devices may include, for example, graphical user interface presented on display,” Col 17, Lines 26-27, and “Processes, such as logic flows, described herein may be performed by one or more programmable processors executing one or more computer programs to perform functions by operating on input data and generating corresponding output. Processes described herein may be performed by, and apparatus can also be implemented as, special purpose logic circuitry, e.g., an FPGA (field programmable gate array) or an ASIC (application specific integrated circuit),” Col 17, Lines 11-19.
The claimed “process state” is mapped to the disclosed “status update”. Here, it can be seen that multiple (at least two) status updates can be used for the machine learning model. A first status update is taken at a certain point in time before a second status update.
The claimed “process” is mapped to the disclosed “process performed by the system”.);
obtaining a first inference (
Gonzalez discloses, “In operation 708, a failure/success score (or other performance metric indicating the likelihood that a particular outcome will occur at an expiration time of a process performed by a system) may be computed based on the obtained data, the retrieved/received descriptors, and the obtained prior status updates. In operation 710, the failure/success score (or other performance metric) may be stored, e.g., in results database 516. In operation 712, a determination may be made as to whether the expiration time has been reached. The expiration time refers to a predefined time when a particular process (the process with which the trained machine learning model is configured to predict the likelihood of) ends,” Col 15, Lines 60-67 and Col 16, Lines 1-20.
The claimed “inference” is mapped to the disclosed “failure/success score” that indicates the likelihood that the process will either fail/succeed upon completion. This is calculated using the obtained prior status updates.);
Gonzalez does not teach that the process states are inferred, the process only providing information regarding progress of the process via an interface, the process only being locally manageable using hardware resources of the host device, a first inferred process state being based on a first screenshot of the interface taken at a first point in time; a first inference that indicates progression of the process in the future; managing continued performance of the process based on the progression of the process to facilitate completion of the process to obtain an updated host device; and providing computer-implemented services using the updated host device.
However, Kersch teaches that the process states are inferred (
Kersch discloses, “The training data center may provide the correlated infrastructure performance indicators and the local performance indicators to the learning machine. The learning machine then calculates the performance indicator for the transport layer and lower layer metrics received from the dispatcher and the correlation received from the training data center. The performance indicator, i.e. QoE estimate, may be forwarded to the verification unit in order to verify that the learning machine calculated the QoE estimate correctly,” ¶ 0053, and “For example, the processing unit 111-2 may determine that a spinner 20b is present in the captured display content for some time and may thus determine a quite specific local performance indicator (i.e. long spinning, low performance). The analysis of a progress bar 20a may even provide more detailed information on the actual process. Further, the appearance of and/or change in text, graphic and video elements 20f, 20e, 20d, 20c may be used for determining a sufficiently good local performance indicator,” ¶ 0078.
Here, process states (indicators) are not received directly, but instead inferred from analysis of processes using a machine learning model (learning machine).),
the process (
Kersch discloses, “With respect to privacy, raw screen content may potentially include sensitive information. However, the information may be processed only locally and may never leave the device. That is, only condensed QoE reports may be sent to the network management system containing user ID, timestamp and QoE metrics for foreground applications. In addition, the screen to QoE app may be installed and used in an opt-in way where the user actively decides to allow such processing of data,” ¶ 0072.),
the process only being locally manageable using hardware resources of the host device (
Kersch discloses, “With respect to privacy, raw screen content may potentially include sensitive information. However, the information may be processed only locally and may never leave the device. That is, only condensed QoE reports may be sent to the network management system containing user ID, timestamp and QoE metrics for foreground applications. In addition, the screen to QoE app may be installed and used in an opt-in way where the user actively decides to allow such processing of data,” ¶ 0072.),
an inferred state being based on a screenshot, and a first inference that indicates progression of the process in the future (
Kersch discloses, “That is, raw screen content such as the screenshot data may be analyzed in parallel by different image processing modules focusing on generic user interface building blocks, e.g., progress or stall indicators,” ¶ 0065, and “For example, the processing unit 111-2 may determine that a spinner 20b is present in the captured display content for some time and may thus determine a quite specific local performance indicator (i.e. long spinning, low performance). The analysis of a progress bar 20a may even provide more detailed information on the actual process,” ¶ 0078.
Here, an inferred state (mapped to disclosed “indicator”, which measures progress or performance) is based on a screenshot. An inference can be determined based on the indicators, which indicates whether or not the process associated with the progress bar will complete within a reasonable amount of time.
After the combination of Gonzalez with Kersch, the process state from Gonzalez, is inferred based on a screenshot/image of a graphical user interface, as specified by Kersch. Furthermore, the first inference from Gonzalez now indicates progression of the process in the future based based on the indicators from Kersch.).
Gonzalez and Kersch are both considered to be analogous to the claimed invention because they are in the same field of machine learning. Therefore, it would have been obvious to someone of ordinary skill in the art before the effective filing date of the claimed invention to have modified Gonzalez to incorporate the teachings of Kersch and provide that the process states are inferred, the process providing information regarding progress of the process via an interface, the process only being locally manageable using hardware resources of the host device, a first inferred process state of the at least two inferred process states being based on a first screenshot of the interface taken at a first point in time, and a first inference that indicates progression of the process in the future. Doing so would help allow for utilizing image data in order to determine whether a process is progressing normally or has stalled. (Kersch discloses, “That is, raw screen content such as the screenshot data may be analyzed in parallel by different image processing modules focusing on generic user interface building blocks, e.g., progress or stall indicators,” ¶ 0065.).
Gonzalez in view of Kersch does not teach the process only providing information regarding progress of the process via an interface, managing continued performance of the process based on the progression of the process to facilitate completion of the process to obtain an updated host device; and providing computer-implemented services using the updated host device.
However, Caspi teaches the process only providing information regarding progress of the process via an interface (
Caspi discloses, “However, presently, access to the presence and identity information of a colleague is only accessible from a graphical user interface (GUI), and only for people served by the same service provider,” ¶ 0005.).
Gonzalez in view of Kersch, and Caspi are both considered to be analogous to the claimed invention because they are in the same field of data collection. Therefore, it would have been obvious to someone of ordinary skill in the art before the effective filing date of the claimed invention to have modified Gonzalez in view of Kersch to incorporate the teachings of Caspi and provide the process only providing information regarding progress of the process via an interface. Doing so would help ensure greater security of the data and prevent unauthorized access. (Caspi discloses, “However, presently, access to the presence and identity information of a colleague is only accessible from a graphical user interface (GUI), and only for people served by the same service provider,” ¶ 0005.).
Gonzalez in view of Kersch and Caspi does not teach managing continued performance of the process based on the progression of the process to facilitate completion of the process to obtain an updated host device; and providing computer-implemented services using the updated host device.
However, Rosenberg teaches managing continued performance of the process based on the progression of the process to facilitate completion of the process to obtain an updated host device (
Rosenberg discloses, “For example, migration may be used for load balancing. In load balancing, guest processes (e.g., virtual machines) are moved to host physical machines with lower usage when their host physical machine becomes overloaded, or another host physical machine is under-utilized,” ¶ 0012.);
and providing computer-implemented services using the updated host device (
Rosenberg discloses, “The process manager (e.g., hypervisor) also estimates transfer speed during migration, so when the remaining amount of data to transfer will take a certain configurable period of time (for example, 10 milliseconds), the process manager (e.g., hypervisor) suspends the original guest process (e.g., virtual machine), transfers the remaining data, and resumes the same guest process (e.g., virtual machine) on the destination host physical machine,” ¶ 0014.
The claimed “computer-implemented services” is mapped to the disclosed “same guest process” that is resumed on the destination host physical machine, after being suspended on the original host machine.).
Gonzalez in view of Kersch and Caspi, and Rosenberg are both considered to be analogous to the claimed invention because they are in the same field of computer processes. Therefore, it would have been obvious to someone of ordinary skill in the art before the effective filing date of the claimed invention to have modified Gonzalez in view of Kersch and Caspi to incorporate the teachings of Rosenberg and provide managing continued performance of the process based on the progression of the process to facilitate completion of the process to obtain an updated host device; and providing computer-implemented services using the updated host device. Doing so would help provide increased resilience of the processes (Rosenberg discloses, “In load balancing, guest processes (e.g., virtual machines) are moved to host physical machines with lower usage when their host physical machine becomes overloaded, or another host physical machine is under-utilized. Migration may also be used to achieve hardware independence. For example, when upgrading, adding, or removing hardware devices on the host physical machine, guest processes (e.g., virtual machines) may be safely relocated to other host physical machines. This means that guest processes (e.g., virtual machines) do not experience any downtime for hardware improvements,” ¶ 0012.).
Claims 11 and 16 are a non-transitory machine-readable medium claim and a data processing system claim, respectively, corresponding to the method Claim 1 (Col 19, Lines 32-56 of Gonzalez.). Therefore, Claims 11 and 16 are rejected for the same reasons set forth in the rejection of Claim 1.
Regarding Claim 2, Gonzalez in view of Kersch, Caspi, and Rosenberg teaches the method of claim 1, wherein the first screenshot is obtained by a management controller of the host device, and the first screenshot being obtained while the host device is performing the process (
Kersch discloses, “For example, the processing unit 111-2 may determine that a spinner 20b is present in the captured display content for some time and may thus determine a quite specific local performance indicator (i.e. long spinning, low performance). The analysis of a progress bar 20a may even provide more detailed information on the actual process. Further, the appearance of and/or change in text, graphic and video elements 20f, 20e, 20d, 20c may be used for determining a sufficiently good local performance indicator,” ¶ 0031.
The claimed “management controller” is mapped to the disclosed “processing unit”, which belongs to the host of the operating system. The processing unit processes the raw screen content (screenshot) while the process is ongoing, in order to infer indicators.).
Gonzalez and Kersch are both considered to be analogous to the claimed invention because they are in the same field of machine learning. Therefore, it would have been obvious to someone of ordinary skill in the art before the effective filing date of the claimed invention to have modified Gonzalez to incorporate the teachings of Kersch and provide wherein the first screenshot is obtained by a management controller of the host device, and the first screenshot being obtained while the host device is performing the process. Doing so would help allow for utilizing image data in order to determine whether a process is progressing normally or has stalled. (Kersch discloses, “That is, raw screen content such as the screenshot data may be analyzed in parallel by different image processing modules focusing on generic user interface building blocks, e.g., progress or stall indicators,” ¶ 0065.).
Claims 12 and 17 are a non-transitory machine-readable medium claim and a data processing system claim, respectively, corresponding to the method Claim 2 (Col 19, Lines 32-56 of Gonzalez.). Therefore, Claims 12 and 17 are rejected for the same reasons set forth in the rejection of Claim 2.
Regarding Claim 3, Gonzalez in view of Kersch, Caspi, and Rosenberg teaches the method of claim 2, wherein a second of the at least two inferred process states is based on a second screenshot of the interface taken at a second point in time, and the second screenshot being obtained by the management controller of the host device (
Gonzalez discloses, “FIG. 7 is a flowchart that describes an example process 700 for performing real-time updates to a likelihood determination, in accordance with one or more embodiments. In some embodiments, process 700 may begin at operation 702. At operation 702, data of an update occurring at a given time may be obtained. For example, an update to a system (or a process performed by the system) may be obtained. … In operation 706, status updates from prior times may be obtained. For example, with reference to FIG. 6A, at the fourth status update, each update to the system (or process performed by the system) previously obtained may be obtained, e.g., first thru third status updates,” Col 15, Lines 60-67 and Col 16, Lines 1-8.
In Gonzalez, a second status update can be obtained at a later time than that of a previous, first status update, as evident by the disclosed “status updates from prior times”.
Kersch also discloses, “For example, the processing unit 111-2 may determine that a spinner 20b is present in the captured display content for some time and may thus determine a quite specific local performance indicator (i.e. long spinning, low performance). The analysis of a progress bar 20a may even provide more detailed information on the actual process. Further, the appearance of and/or change in text, graphic and video elements 20f, 20e, 20d, 20c may be used for determining a sufficiently good local performance indicator,” ¶ 0031.
In Kersch, the disclosed “processing unit” (mapped to claimed “management controller”) takes a screenshot of the interface as the process is still running.
After the combination of Gonzalez with Kersch, the second status update from Gonzalez is inferred from a second screenshot of the interface as the process is still running, as specified by Kersch.).
Gonzalez and Kersch are both considered to be analogous to the claimed invention because they are in the same field of machine learning. Therefore, it would have been obvious to someone of ordinary skill in the art before the effective filing date of the claimed invention to have modified Gonzalez to incorporate the teachings of Kersch and provide wherein a second of the at least two inferred process states is based on a second screenshot of the interface taken at a second point in time, and the second screenshot being obtained by the management controller of the host device. Doing so would help allow for utilizing image data in order to determine whether a process is progressing normally or has stalled. (Kersch discloses, “That is, raw screen content such as the screenshot data may be analyzed in parallel by different image processing modules focusing on generic user interface building blocks, e.g., progress or stall indicators,” ¶ 0065.).
Claims 13 and 18 are a non-transitory machine-readable medium claim and a data processing system claim, respectively, corresponding to the method Claim 3 (Col 19, Lines 32-56 of Gonzalez.). Therefore, Claims 13 and 18 are rejected for the same reasons set forth in the rejection of Claim 3.
Regarding Claim 4, Gonzalez in view of Kersch, Caspi, and Rosenberg teaches the method of claim 1, wherein obtaining the at least two inferred process states comprises:
obtaining the first inferred process state, the first inferred process state indicating a current state of the process, and the first inferred process state being obtained using:
a first inference model, and at least the first screenshot (
Kersch discloses, “The training data center may provide the correlated infrastructure performance indicators and the local performance indicators to the learning machine. The learning machine then calculates the performance indicator for the transport layer and lower layer metrics received from the dispatcher and the correlation received from the training data center. The performance indicator, i.e. QoE estimate, may be forwarded to the verification unit in order to verify that the learning machine calculated the QoE estimate correctly,” ¶ 0053, and “For example, the processing unit 111-2 may determine that a spinner 20b is present in the captured display content for some time and may thus determine a quite specific local performance indicator (i.e. long spinning, low performance). The analysis of a progress bar 20a may even provide more detailed information on the actual process. Further, the appearance of and/or change in text, graphic and video elements 20f, 20e, 20d, 20c may be used for determining a sufficiently good local performance indicator,” ¶ 0078.
The claimed “first inference model” is mapped to Kersch’s disclosed learning machine.
Here, a first process state (indicator) is inferred from a screenshot using the disclosed “learning machine”.),
wherein the first inference is obtained using a second inference model (
Gonzalez discloses, “FIG. 7 is a flowchart that describes an example process 700 for performing real-time updates to a likelihood determination, in accordance with one or more embodiments. In some embodiments, process 700 may begin at operation 702. At operation 702, data of an update occurring at a given time may be obtained. For example, an update to a system (or a process performed by the system) may be obtained. … In operation 706, status updates from prior times may be obtained. For example, with reference to FIG. 6A, at the fourth status update, each update to the system (or process performed by the system) previously obtained may be obtained, e.g., first thru third status updates… In operation 710, the failure/success score (or other performance metric) may be stored, e.g., in results database 516. In operation 712, a determination may be made as to whether the expiration time has been reached. The expiration time refers to a predefined time when a particular process (the process with which the trained machine learning model is configured to predict the likelihood of) ends,” Col 15, Lines 60-67 and Col 16, Lines 1-20.
The claimed “second inference model” is mapped to Gonzalez’s “trained machine learning model”, which produces the failure/success score based on the status updates.
After the combination of Gonzalez with Kersch, the status updates are inferred by using Kersch’s machine learning model (learning machine) on screenshots to produce the status updates, and then Gonzalez’s own machine learning model takes in these status updates to produce the inference, which is replaced from a failure/success score with an indicator from Kersch.).
Gonzalez and Kersch are both considered to be analogous to the claimed invention because they are in the same field of machine learning. Therefore, it would have been obvious to someone of ordinary skill in the art before the effective filing date of the claimed invention to have modified Gonzalez to incorporate the teachings of Kersch and provide wherein obtaining the at least two inferred process states comprises: obtaining the first inferred process state, the first inferred process state indicating a current state of the process, and the first inferred process state being obtained using: a first inference model, and at least the first screenshot, wherein the first inference is obtained using a second inference model. Doing so would help allow for utilizing image data with the machine learning models in order to determine whether a process is progressing normally or has stalled. (Kersch discloses, “That is, raw screen content such as the screenshot data may be analyzed in parallel by different image processing modules focusing on generic user interface building blocks, e.g., progress or stall indicators,” ¶ 0065.).
Claims 14 and 19 are a non-transitory machine-readable medium claim and a data processing system claim, respectively, corresponding to the method Claim 4 (Col 19, Lines 32-56 of Gonzalez.). Therefore, Claims 14 and 19 are rejected for the same reasons set forth in the rejection of Claim 4.
Regarding Claim 7, Gonzalez in view of Kersch, Caspi, and Rosenberg teaches the method of claim 4, wherein the first inference model is trained to generate inferences indicating a current progression status of the process (
Kersch discloses, “The training data center may provide the correlated infrastructure performance indicators and the local performance indicators to the learning machine. The learning machine then calculates the performance indicator for the transport layer and lower layer metrics received from the dispatcher and the correlation received from the training data center. The performance indicator, i.e. QoE estimate, may be forwarded to the verification unit in order to verify that the learning machine calculated the QoE estimate correctly,” ¶ 0053, “That is, raw screen content such as the screenshot data may be analyzed in parallel by different image processing modules focusing on generic user interface building blocks, e.g., progress or stall indicators,” ¶ 0063, and “For example, the processing unit 111-2 may determine that a spinner 20b is present in the captured display content for some time and may thus determine a quite specific local performance indicator (i.e. long spinning, low performance). The analysis of a progress bar 20a may even provide more detailed information on the actual process. Further, the appearance of and/or change in text, graphic and video elements 20f, 20e, 20d, 20c may be used for determining a sufficiently good local performance indicator,” ¶ 0078.
Here, the disclosed “progress/stall indicators” represent a current progression status of a process. Kersch’s learning machine is trained to generate such indicators.
After the combination of Gonzalez with Kersch, the process state from Gonzalez, is inferred based on a screenshot of the user interface taken at a first point in time, as specified by Kersch. Furthermore, the first inference from Gonzalez now indicates progression of the process in the future based on the “progress/stall indicators” from Kersch.),
the current progression status indicating whether the current progress of the process is nominal, and a progress status of the process indicates how much of the process has been completed (
Kersch discloses, “That is, raw screen content such as the screenshot data may be analyzed in parallel by different image processing modules focusing on generic user interface building blocks, e.g., progress or stall indicators,” ¶ 0063, and “For example, the processing unit 111-2 may determine that a spinner 20b is present in the captured display content for some time and may thus determine a quite specific local performance indicator (i.e. long spinning, low performance). The analysis of a progress bar 20a may even provide more detailed information on the actual process. Further, the appearance of and/or change in text, graphic and video elements 20f, 20e, 20d, 20c may be used for determining a sufficiently good local performance indicator,” ¶ 0078.
Here, it can be determined if the progress of the process is nominal from the stall indicator. The progress indicator can also indicate how much of the process has been completed.
This aligns with paragraph 67 of the present application’s specification, which states “For example, the inferences (e.g., generated by trained inference model 212) may indicate whether the progression of the process is performing nominally (e.g., as expected for the type of process) at corresponding points in time. A process that has stalled when compared to typical performances of the process may not be progressing nominally, while a process that has reached various checkpoints throughout the process in durations of time that are typical for the type of the process and the hardware resources for use in the process may be progressing nominally.”).
Gonzalez and Kersch are both considered to be analogous to the claimed invention because they are in the same field of computer processes. Therefore, it would have been obvious to someone of ordinary skill in the art before the effective filing date of the claimed invention to have modified Gonzalez to incorporate the teachings of Kersch and provide wherein the first inference model is trained to generate inferences indicating a current progression status of the process, the current progression status indicating whether the current progress of the process is nominal, and a progress status of the process indicates how much of the process has been completed. Doing so would help allow for determining if a process has stalled or not in order to take remedial actions. (Kersch discloses, “For example, the processing unit 111-2 may determine that a spinner 20b is present in the captured display content for some time and may thus determine a quite specific local performance indicator (i.e. long spinning, low performance),” ¶ 0078.).
Regarding Claim 8, Gonzalez in view of Kersch, Caspi, and Rosenberg teaches the method of claim 1, wherein obtaining the first inference for the progression comprises: ingesting the at least two inferred process states in an inference model to obtain the progression of the process, wherein the progression being a predicted progression status for the process at a point in time in the future, the predicted progression status indicating whether performance of the process is complete at the point in time in the future (
Gonzalez discloses, “In operation 708, a failure/success score (or other performance metric indicating the likelihood that a particular outcome will occur at an expiration time of a process performed by a system) may be computed based on the obtained data, the retrieved/received descriptors, and the obtained prior status updates. In operation 710, the failure/success score (or other performance metric) may be stored, e.g., in results database 516. In operation 712, a determination may be made as to whether the expiration time has been reached. The expiration time refers to a predefined time when a particular process (the process with which the trained machine learning model is configured to predict the likelihood of) ends,” Col 15, Lines 60-67 and Col 16, Lines 1-20.
Here, the disclosed “failure/success score” from Gonzalez is determined from at least two process states in the form of at least two status updates taken from at least two different points in time, which are inputted into a machine learning model to obtain the score.
After the combination of Gonzalez with Kersch, said “failure/success score” from Gonzalez is replaced with inferred “indicators” of the process, from Kersch. Kerch’s progress/stall indicators would produce equivalent results to the failure/success score from Gonzalez.).
Claims 5-6, 15, and 20 are rejected under 35 U.S.C. 103 as being unpatentable over Gonzalez (US 12124835 B2) in view of Kersch (US 20190044830 A1), Caspi (US 20080065755 A1), Rosenberg (US 20210026707 A1), and Ngo (US 20220327405 A1).
Regarding Claim 5, Gonzalez in view of Kersch, Caspi, and Rosenberg teaches the method of claim 4. Gonzalez in view of Kersch, Caspi, and Rosenberg does not teach wherein obtaining the at least two inferred process states further comprises: obtaining telemetry data for the hardware resources, the telemetry data comprising measurements of characteristics of the hardware resources while the host device is performing the process; and obtaining hardware data for the hardware resources, the hardware data specifying the hardware resources that are contributing to performance of the process, wherein the first inferred process state is also obtained using the telemetry data and the hardware data.
However, Ngo teaches wherein obtaining the at least two inferred process states further comprises:
obtaining telemetry data for the hardware resources, the telemetry data comprising measurements of characteristics of the hardware resources while the host device is performing the process (
Ngo discloses, “For example, estimates such as the respective processing speeds of the inference processing apparatus 1 in the first and second operation modes T.sub.M1 and T.sub.M2 can be obtained in advance based on the hardware used, the clock frequency, the bit accuracy of processing calculation, or the like,” ¶ 0138.
The claimed “telemetry data” is mapped to the disclosed “clock frequency, the bit accuracy of processing calculation”.
This is consistent with paragraph 70 of the present application’s specification, which states “For example, during telemetry data process 216, information regarding operation of a processor (e.g., temperature data, processing speed, etc.) while the process is being performed may be collected and transmitted to management controller 152 (shown in FIG. 1B). Telemetry data 218 may include any type of information usable to identify performance related issues, operating conditions, and/or other types of information for in-band components of a data processing system (e.g., data processing system 100).”);
and obtaining hardware data for the hardware resources, the hardware data specifying the hardware resources that are contributing to performance of the process, wherein the first inferred process state is also obtained using the telemetry data and the hardware data (
“For example, estimates such as the respective processing speeds of the inference processing apparatus 1 in the first and second operation modes T.sub.M1 and T.sub.M2 can be obtained in advance based on the hardware used, the clock frequency, the bit accuracy of processing calculation, or the like,” ¶ 0138.
The claimed “hardware data” is mapped to the disclosed data describing the type of category of “hardware” that is used for estimation.
This is consistent with paragraph 71 of the present application’s specification, which states “For example, during hardware process 222, informational data (e.g., type, size, quantity, etc.) for hardware components (e.g., processors, memory modules, storage devices, etc.) of data management system 100 may be collected and transmitted to management controller 152 (shown in FIG. 1B). Hardware data 224 may include any type of information usable to identify the hardware resources of the data processing system (e.g., 100 shown in FIG. 1B) that may affect the performance of the processes being performed.”
After the combination of Gonzalez in view of Kersch, Caspi, and Rosenberg, with Ngo, the telemetry data and hardware data from Ngo is used to obtain the first inferred process state from Gonzalez in view of Kersch, Caspi, and Rosenberg.).
Gonzalez in view of Kersch, Caspi, and Rosenberg, and Ngo are both considered to be analogous to the claimed invention because they are in the same field of computer processes. Therefore, it would have been obvious to someone of ordinary skill in the art before the effective filing date of the claimed invention to have modified Gonzalez in view of Kersch, Caspi, and Rosenberg to incorporate the teachings of Ngo and provide wherein obtaining the at least two inferred process states further comprises: obtaining telemetry data for the hardware resources, the telemetry data comprising measurements of characteristics of the hardware resources while the host device is performing the process; and obtaining hardware data for the hardware resources, the hardware data specifying the hardware resources that are contributing to performance of the process, wherein the first inferred process state is also obtained using the telemetry data and the hardware data. Doing so would help allow for more efficiently predicting progress completion using the telemetry data and the hardware data without manual calculations (Ngo discloses, “More specifically, in DNN-based inference processing, input data such as time series data or image data is given to a trained neural network model to infer features of the input data,” ¶ 0005.).
Claims 15 and 20 are a non-transitory machine-readable medium claim and a data processing system claim, respectively, corresponding to the method Claim 5 (Col 19, Lines 32-56 of Gonzalez.). Therefore, Claims 15 and 20 are rejected for the same reasons set forth in the rejection of Claim 5.
Regarding Claim 6, Gonzalez in view of Kersch, Caspi, Rosenberg, and Ngo teaches the method of claim 5, wherein obtaining the first inferred process state comprises:
ingesting at least the first screenshot, the telemetry data, and the hardware data into the first inference model, the first inference model generating the first inferred process state based on the first screenshot, the telemetry data, and the hardware data (
Ngo discloses, “For example, estimates such as the respective processing speeds of the inference processing apparatus 1 in the first and second operation modes T.sub.M1 and T.sub.M2 can be obtained in advance based on the hardware used, the clock frequency, the bit accuracy of processing calculation, or the like,” ¶ 0138.
Kersch already teaches ingesting screenshot data into a model to generate an inferred process state, and Gonzalez already teaches ingesting data into a model to generate an inference. After the combination of Gonzalez in view of Kersch, Caspi, and Rosenberg, with Ngo, the screenshot from Kersch, as well as the telemetry data and hardware data from Ngo, are ingested into the model from Kersch to produce the first inferred process state, which is then ingested into the model from Gonzalez to produce the first inference.).
Gonzalez in view of Kersch, Caspi, and Rosenberg, and Ngo are both considered to be analogous to the claimed invention because they are in the same field of computer processes. Therefore, it would have been obvious to someone of ordinary skill in the art before the effective filing date of the claimed invention to have modified Gonzalez in view of Kersch, Caspi, and Rosenberg to incorporate the teachings of Ngo and provide wherein obtaining the first inferred process state comprises: ingesting at least the first screenshot, the telemetry data, and the hardware data into the first inference model, the first inference model generating the first inferred process state based on the first screenshot, the telemetry data, and the hardware data. Doing so would help allow for more efficiently predicting progress completion using the telemetry data and the hardware data without manual calculations (Ngo discloses, “More specifically, in DNN-based inference processing, input data such as time series data or image data is given to a trained neural network model to infer features of the input data,” ¶ 0005.).
Claims 9-10 are rejected under 35 U.S.C. 103 as being unpatentable over Gonzalez (US 12124835 B2) in view of Kersch (US 20190044830 A1), Caspi (US 20080065755 A1), Rosenberg (US 20210026707 A1), and Boyd (US 20240201767 A1).
Regarding Claim 9, Gonzalez in view of Kersch, Caspi, and Rosenberg teaches the method of claim 8. Gonzalez in view of Kersch, Caspi, and Rosenberg does not teach wherein managing continued performance of the process comprises: in an instance of the managing of the continued performance where the progression indicates that the process is not complete at the point in time in the future: obtaining instructions for a management controller of the host device to perform a restart of the host device, the instructions indicating actions performable by the management controller to complete the restart; and performing, by the management controller, the actions to restart the performance of the process.
However, Boyd teaches wherein managing continued performance of the process comprises: in an instance of the managing of the continued performance where the progression indicates that the process is not complete at the point in time in the future:
obtaining instructions for a management controller of the host device to perform a restart of the host device, the instructions indicating actions performable by the management controller to complete the restart (
Boyd discloses, “For instance, a management entity provides instructions in-band or out-of-band (e.g., via an auxiliary service controller) to the host device to perform one or more processes to improve the health of the host device. In some implementations, the recovery process includes rebooting or restarting the host device,” ¶ 0031.);
and performing, by the management controller, the actions to restart the performance of the process (
Boyd discloses, “In some implementations, the recovery process includes restarting one or more virtual machines or applications running on the host device,” ¶ 0031.
Here, restarting applications or processes running on the host device would occur after the host device has been restarted.).
Gonzalez in view of Kersch, Caspi, and Rosenberg, and Boyd are both considered to be analogous to the claimed invention because they are in the same field of computer processes. Therefore, it would have been obvious to someone of ordinary skill in the art before the effective filing date of the claimed invention to have modified Gonzalez in view of Kersch, Caspi, and Rosenberg to incorporate the teachings of Ngo and provide wherein managing continued performance of the process comprises: in an instance of the managing of the continued performance where the progression indicates that the process is not complete at the point in time in the future: obtaining instructions for a management controller of the host device to perform a restart of the host device, the instructions indicating actions performable by the management controller to complete the restart; and performing, by the management controller, the actions to restart the performance of the process. Doing so would help provide increased resilience of the host device and the processes running on it (Boyd discloses, “For example, as used in this document, the term ‘recovery process’ refers to improving the health of a host server by repairing one or more elements of the host device,” ¶ 0031.).
Regarding Claim 10, Gonzalez in view of Kersch, Caspi, Rosenberg, and Boyd teaches the method of claim 8, wherein managing continued performance of the process further comprises:
in an instance of the managing of the continued performance where the progression indicates that the process is not complete at the point in time in the future:
obtaining management actions for a management controller of the host device, the management actions being actions performable by the management controller to modify operation of the hardware resources (
Rosenberg discloses, “For example, migration may be used for load balancing. In load balancing, guest processes (e.g., virtual machines) are moved to host physical machines with lower usage when their host physical machine becomes overloaded, or another host physical machine is under-utilized,” ¶ 0012.);
and performing, by the management controller, the management actions to obtain updated hardware resources to facilitate continued performance of the process (
Rosenberg discloses, “The process manager (e.g., hypervisor) also estimates transfer speed during migration, so when the remaining amount of data to transfer will take a certain configurable period of time (for example, 10 milliseconds), the process manager (e.g., hypervisor) suspends the original guest process (e.g., virtual machine), transfers the remaining data, and resumes the same guest process (e.g., virtual machine) on the destination host physical machine,” ¶ 0014.).
Gonzalez in view of Kersch and Caspi, and Rosenberg are both considered to be analogous to the claimed invention because they are in the same field of computer processes. Therefore, it would have been obvious to someone of ordinary skill in the art before the effective filing date of the claimed invention to have modified Gonzalez in view of Kersch and Caspi to incorporate the teachings of Rosenberg and provide wherein managing continued performance of the process further comprises: in an instance of the managing of the continued performance where the progression indicates that the process is not complete at the point in time in the future: obtaining management actions for a management controller of the host device, the management actions being actions performable by the management controller to modify operation of the hardware resources; and performing, by the management controller, the management actions to obtain updated hardware resources to facilitate continued performance of the process. Doing so would help provide increased resilience of the processes (Rosenberg discloses, “In load balancing, guest processes (e.g., virtual machines) are moved to host physical machines with lower usage when their host physical machine becomes overloaded, or another host physical machine is under-utilized. Migration may also be used to achieve hardware independence. For example, when upgrading, adding, or removing hardware devices on the host physical machine, guest processes (e.g., virtual machines) may be safely relocated to other host physical machines. This means that guest processes (e.g., virtual machines) do not experience any downtime for hardware improvements,” ¶ 0012.).
Conclusion
The prior art made of record and not relied upon is considered pertinent to applicant's disclosure.
Namiki et al. (US 20240095494 A1): Information Processing Apparatus, Determination Method, and Storage Medium
Any inquiry concerning this communication or earlier communications from the examiner should be directed to ANDREW SUN whose telephone number is (571)272-6735. The examiner can normally be reached Monday-Friday 8:00-5:00.
Examiner interviews are available via telephone, in-person, and video conferencing using a USPTO supplied web-based collaboration tool. To schedule an interview, applicant is encouraged to use the USPTO Automated Interview Request (AIR) at http://www.uspto.gov/interviewpractice.
If attempts to reach the examiner by telephone are unsuccessful, the examiner’s supervisor, Aimee Li can be reached at (571) 272-4169. The fax phone number for the organization where this application or proceeding is assigned is 571-273-8300.
Information regarding the status of published or unpublished applications may be obtained from Patent Center. Unpublished application information in Patent Center is available to registered users. To file and manage patent submissions in Patent Center, visit: https://patentcenter.uspto.gov. Visit https://www.uspto.gov/patents/apply/patent-center for more information about Patent Center and https://www.uspto.gov/patents/docx for information about filing in DOCX format. For additional questions, contact the Electronic Business Center (EBC) at 866-217-9197 (toll-free). If you would like assistance from a USPTO Customer Service Representative, call 800-786-9199 (IN USA OR CANADA) or 571-272-1000.
/ANDREW NMN SUN/Examiner, Art Unit 2195
/Aimee Li/Supervisory Patent Examiner, Art Unit 2195
Prosecution Insights