CTFR 18/324,340 CTFR 84960 Notice of Pre-AIA or AIA Status 07-03-aia AIA 15-10-aia 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 07-20-aia AIA 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. 07-21-aia AIA Claim (s) 1, 5, 8-9, 13, and 16-17 is/are rejected under 35 U.S.C. 103 as being unpatentable over Vasic, DejaVu: Accelerating Resource Allocation in Virtualized Environments, 2012 in view of Allen-Ware US 20190004579 A1 and further in view of Helmbrecht, Using the IBM Block Storage CSI Driver in a Red Hat OpenShift Environment May 2021 07-22-aia AIA Claim Vasic, DejaVu: Accelerating Resource Allocation in Virtualized Environments, 2012 Analysis 1. A method for adaptively training a machine learning model for estimating energy consumption in a cloud computing system, comprising: p. 1 Cloud computing is rapidly growing in popularity and importance, as an increasing number of enterprises and individuals have been offloading their workloads to cloud service providers, such as Amazon, Microsoft, IBM, and Google However, effective management of virtualized resources is a challenging task for providers, as it often involves selecting the best resource allocation out of a large number of alternatives. Moreover, evaluating each such allocation requires assessing its potential performance, availability, and energy consumption implications detecting an update event in the cloud computing system; p. 6 section 3.5 DejaVu uses the previously identified clusters to label each workload with the cluster number to which it belongs, such that it can train a classifier to quickly recognize newly encountered workloads at runtime. Upon a workload change, DejaVu promptly collects the relevant low-level metrics to form the workload signature of the new workload and queries the DejaVu repository to find the best match p. 7 Along with the preferred resource allocations, the repository also outputs the certainty level with which the repository assigned the new signature to the chosen cluster. If the repository repeatedly outputs low certainty levels, it most likely means that the workload has changed over time and that the current clustering is no longer relevan t. DejaVu can then initiate the clustering and tuning process once again, allowing it to determine new workload classes, conduct the necessary experiments (or modeling activities), and update the Repository . Meanwhile, DejaVu configures the service with the maximum allowed capacity to ensure that the performance is not affected when experiencing non-classified workloads Examiner finds workload change over time that produces new clustering is an update event. collecting, p. 7 If the repository repeatedly outputs low certainty levels, it most likely means that the workload has changed over time and that the current clustering is no longer relevant. DejaVu can then initiate the clustering and tuning process once again, allowing it to determine new workload classes, conduct the necessary experiments (or modeling activities), and update the Repository. Meanwhile, DejaVu configures the service with the maximum allowed capacity to ensure that the performance is not affected when experiencing non-classified workloads. Period of time is the time after the update event and the current clustering is no longer relevant. by one or more sensors, energy consumption data for the cloud computing system for a time period p. 7 If the repository repeatedly outputs low certainty levels, it most likely means that the workload has changed over time and that the current clustering is no longer relevant. DejaVu can then initiate the clustering and tuning process once again, allowing it to determine new workload classes, conduct the necessary experiments (or modeling activities), and update the Repository. Meanwhile, DejaVu configures the service with the maximum allowed capacity to ensure that the performance is not affected when experiencing non-classified workloads. Period of time is the time after the update event and the current clustering is no longer relevant. after an occurrence of the update event;, wherein each sample collected by the one or more sensors includes a workload identifier, a pod identifier, metadata regarding software and hardware, and a current energy usage level for a portion of an identified workload being executed on an identified pod; retraining a power consumption model with the data from the time period p. 2 To enable the cache lookups, DejaVu automatically constructs a classifier that uses off-the-shelf machine learning techniques. The classifier operates on workload clusters that are determined after an initial learning phase. DejaVu clustering has a positive effect on reducing the overall resource management effort and overhead, because it reduces the number of invocations of the tuning process (one per cluster). p. 2 The resource manager can use the output of DejaVu to quickly reallocate resources. T he manager only needs to resort to time consuming modeling, sandboxed experimentation, or on-line experimentation when no previous workload exercises the affected VMs in the same way. When the manager does have to produce a new optimized resource allocation using one of these methods, it stores the allocation into the DejaVu cache for later use. p. 7 If the repository repeatedly outputs low certainty levels, it most likely means that the workload has changed over time and that the current clustering is no longer relevant. D ejaVu can then initiate the clustering and tuning process once again , allowing it to determine new workload classes, conduct the necessary experiments (or modeling activities ), and update the Repository. Meanwhile, DejaVu configures the service with the maximum allowed capacity to ensure that the performance is not affected when experiencing non-classified workloads. Examiner finds initiating the clustering and tuning process again necessarily teaches “retraining,” because, clustering and classification are machine learning models and the tuning clustering processes necessarily require retraining the model. and storing the power consumption model in the cloud computing system. p. 7 If the repository repeatedly outputs low certainty levels, it most likely means that the workload has changed over time and that the current clustering is no longer relevant. DejaVu can then initiate the clustering and tuning process once again, allowing it to determine new workload classes, conduct the necessary experiments (or modeling activities), and update the Repository. Meanwhile, DejaVu configures the service with the maximum allowed capacity to ensure that the performance is not affected when experiencing non-classified workloads. p. 2 The resource manager can use the output of DejaVu to quickly reallocate resources. T he manager only needs to resort to time consuming modeling, sandboxed experimentation, or on-line experimentation when no previous workload exercises the affected VMs in the same way. When the manager does have to produce a new optimized resource allocation using one of these methods, it stores the allocation into the DejaVu cache for later use. Model is stored in cache. Examiner finds this cache is part of the cloud computing system. It appears Vasic fails to explicitly teach “by one or more sensors. . . after an occurrence of the update event. . .” However, Allen-Ware US 20190004579 A1 teaches “by one or more sensors” in para. 75 and para. 79 (power sensors); “after an occurrence of the update event” in para. 79 (update event is change in power output). Allen-Ware and Vasic are analogous art because they are from the same field of endeavor as the claimed invention. It would have been obvious to one skilled in the art before the effective filing date of the invention to modify the collecting in Vasic to include “by one or more sensors” as taught by Allen-Ware. It would have been obvious to one skilled in the art before the effective filing date of the invention to modify the update event in Vasic to include the update event in Allen-Ware and the collection of energy consumption data to include after an occurrence of the update event as taught by Allen-Ware. The motivation would have been to reduce the wasting of power. See Allen-Ware para. 31. It appears Vasic et al. fails to explicitly teach “wherein each sample collected by the one or more sensors includes a workload identifier, a pod identifier, metadata regarding software and hardware.” However, Helmbrecht, teaches “wherein each sample collected by the one or more sensors includes a workload identifier” on p. 80 Example 4-40 (Fs9110 is an example of a workload identifier); “a pod identifier” p. 28 The pod instances must be identifiable _ The distinguished pods maintain individual state across instantiations Identifiable pods means that the pods, although running the same application, can have different roles. For example, a leader might be among them. The etcd cluster that is describe in “Control plane” is an example for that use case. If a pod from an etcd cluster is rescheduled to another node, its database also must go there and it must preserve the state it had on the former node. Distinguishing the pods might also be an application requirement. While in a DaemonSet or a ReplicaSet, the different pod instances do not know each other, and the StatefulSet allows the different instances to identify and communicate with each other. Our example that is shown in Figure 3-8 on page 29 shows a deployed StatefulSet with three pods. p. 29 Fig. 3-8 (Examiner finds my-app-0, 1, and 2 teach pod identifiers; and “metadata regarding software and hardware” pp. 114-115 Example 5-1(Metadata includes software, e.g. “app” and “apiVersion” and Hardware, for example: spec: storageClassName: "ds8k-csi" accessModes: - ReadWriteOnce resources: requests: storage: 1Gi ). Helmbrecht and Vasic et al. are analogous art because they are from the same field of endeavor. It would have been obvious to one skilled in the art before the effective filing date of the invention to modify the energy collected consumption data in Vasic et al to include “wherein each sample collected by the one or more sensors includes a workload identifier, a pod identifier, metadata regarding software and hardware” as taught by Helmbrecht. The motivation would have been to “define the wanted state of the cluster and OCS operators can ensure that the cluster is in that state or approaching it while minimizing manual intervention. For general-purpose persistent storage or dynamic provision requirements, OCS is suitable for consideration of workloads, such as data science and data analytics, artificial intelligence, machine learning, and Internet of Things workloads.” See Helmbrecht p. 7 last two paragraphs. Allen-Ware teaches “and a current energy usage level for a portion of an identified workload being executed on an identified pod” in [0058] Certain embodiments of the present invention manage point of delivery (PoD)/total information technology (IT) current/power load within specified current/power limits. Certain embodiments of the present invention manage point of delivery (PoD)/total information technology (IT) current/power load in excess of specified current/power limits such that the excess is maintained for less than a duration that, if met or exceeded, would result in unacceptable wear or damage to components that receive that excess. Certain embodiments of the present invention recognize that the amount of wear or damage The motivation to add this data to the collected data Vasic is would have been to prevent damage to computer components. See Allen-Ware para. 58. Claim Reference: Vasic Analysis 5. The method of claim 1, wherein the update event includes a deviation of observed workload activity from an expected workload activity. 13. The computing system of claim 9, wherein the update event includes a deviation of observed workload activity from an expected workload activity. p. 7 I f the repository repeatedly outputs low certainty levels, it most likely means that the workload has changed over time and that the current clustering is no longer relevant. DejaVu can then initiate the clustering and tuning process once again, allowing it to determine new workload classes, conduct the necessary experiments (or modeling activities ) , and update the Repository. Meanwhile, DejaVu configures the service with the maximum allowed capacity to ensure that the performance is not affected when experiencing non-classified workloads. p. 2 The resource manager can use the output of DejaVu to quickly reallocate resources. T he manager only needs to resort to time consuming modeling, sandboxed experimentation, or on-line experimentation when no previous workload exercises the affected VMs in the same way. When the manager does have to produce a new optimized resource allocation using one of these methods, it stores the allocation into the DejaVu cache for later use. Workload activity deviates when it has changed over time such that current clustering is no longer relevant. 8. The method of claim 1, wherein a duration of the time period is based at least in part on a type of the update event. 16. The computing system of claim 9, wherein a duration of the time period is based at least in part on a type of the update event. p. 7 . . . If the repository repeatedly outputs low certainty levels, it most likely means that the workload has changed over time and that the current clustering is no longer relevant. DejaVu can then initiate the clustering and tuning process once again, allowing it to determine new workload classes, conduct the necessary experiments (or modeling activities), and update the Repository. Meanwhile, DejaVu configures the service with the maximum allowed capacity to ensure that the performance is not affected when experiencing non-classified workloads Duration of the time period is whatever duration corresponds with workload change over time that results in the current clustering no longer being relevant. Claim(s) 3-4, 11-12, and 19-20 is/are rejected under 35 U.S.C. 103 as being unpatentable over Vasic in view of Allen-Ware and further in view of Helmbrecht as applied to claims 1, 9, and 17 above and further in view of Wang US 20230418688 A1. With respect to claim 3 , Vasic teaches an update event. See above. It appears Vasic fails to explicitly teach “wherein the update event includes a change of one or more of a hardware device and a software within the cloud computing system.” However, Wang US 20230418688 A1 teaches “wherein the update event includes a change of one or more of a hardware device and a software within the cloud computing system” in para. 2 and [0012] Aspects of the disclosure address the above-noted and other deficiencies by providing a workload scheduler using an energy consumption metrics to assist in workload placement on computing nodes of a computing platform. The workload scheduler may obtain an energy consumption profile for hardware types included in the computing nodes of the computing platform. The energy consumption profiles may include power consumption of servers (e.g., hardware) with different workloads at different utilization levels. For example, the energy consumption profiles may indicate power consumption on different hardware with different workloads. The workload scheduler may then generate a correlation model between the energy consumption profiles a nd utilization characteristics of the types of hardware. The correlation model may be an extrapolation of the profile to provide an estimate of energy consumption of the types of hardware at any resource utilization level. [0013] Upon receiving a new workload to be executed by the computing platform, the workload scheduler determines utilization characteristics for computing nodes of the computing platform. The workload scheduler then estimates a current energy consumption based on the utilization characteristics as well as a potential energy consumption of the computing nodes if the new workload were to be allocated to the computing node. For example, knowing the current utilization characteristics and the estimated current energy consumption of each of the computing nodes, the workload scheduler may determine what an expected energy consumption would be were the new workload be allocated to the computing node in addition to the current workloads being performed. In some examples, the workload scheduler allocates the new workload to the compute node that will result in the least amount of energy consumption to perform the workload (e.g., the most efficient energy computing node). Power consumption on different hardware with different workloads teaches “change of one or more of a hardware device and a software within the cloud computing system”.” Wang and Vasic are analogous art because they are from the same field of endeavor as the claimed invention. It would have been obvious to one skilled in the art before the effective filing date of the invention to modify the update event in Vasic to include “wherein the update event includes a change of one or more of a hardware device and a software within the cloud computing system” as taught by Wang. The motivation would have been the following: overall energy consumption of workloads can be reduced and optimized for the computing platform or for workloads of a particular client of the computing platform. Additionally, carbon emissions associated with workloads can be reduced due to the reduced power consumption of the workload. Wang para. 14. Claim 11 and claim 19 are rejected for the same reasons above. With respect to claim 4 , Vasic teaches an update event. See above. It appears fails to teach “wherein the update event includes a deviation of observed energy consumption data from an expected energy consumption data.” However, Okamura teaches “wherein the update event includes a deviation of observed energy consumption data from an expected energy consumption data” in para. 157: [0157] FIGS. 18 and 19 are flowcharts showing the IT workload control processing S22 (divided into two diagrams for convenience of the drawing). In the IT workload control processing, a batch workload to be actually deployed is determined to approach a power consumption target value in each time slot determined by the calculated value of the parameter α. At this time, in addition to the approach to the power consumption target value, in consideration of a deviation from the prediction of the power consumption of each batch workload, the batch workload to be deployed is determined such that a sum of the deviations from the prediction of the power consumption of each batch workload is reduced as much as possible in the time slot with a small risk allowance. [0174] In S 5110 , the parameter determination program 8700 calculates a value of an evaluation function g indicating a degree of seriousness of the deviation from the prediction value of the power consumption of the batch workload based on the power consumption of the batch workloads of the first group and the second group. [0177] The evaluation function described here is one example, and may be a function in consideration of a magnitude of an allowance of the deviation between the prediction value and the actual value. Okamura and Vasic are analogous art because they are from the same field of endeavor as the claimed invention It would have been obvious to one skilled in the art before the effective filing date of the invention to modify the update event in Vasic to include “wherein the update event includes a deviation of observed energy consumption data from an expected energy consumption data” as taught by Okamura. The motivation would have been to accurately predict power consumption thereby decrease carbon emissions. See Okamura paras. 3-4. Claim 12 and claim 20 are rejected for the same reasons above. Claim(s) 7 and 15 is/are rejected under 35 U.S.C. 103 as being unpatentable over Vasic in view of Allen-Ware and further in view of Helmbrecht as applied to claim 1 and claim 9 above and further in view of Fujiwaka US 20130282337 A1 With respect to claim 7 , Vasic teaches collecting metrics for an activity of a workload See above. It appears Vasic fails to explicitly teach “wherein metrics for an activity of a workload executing in the cloud computing system are collected at first sampling rate during the time period and at a second sampling rate, which is less than the first sampling rate, outside of the time period. However, Fujiwaka teaches “wherein metrics for an activity of a workload executing in the cloud computing system are collected at first sampling rate during the time period and at a second sampling rate, which is less than the first sampling rate, outside of the time period” in [0102] As described above, in the second exemplary embodiment, each of the primary performance models is generated on the basis of the plural items of the performance data corresponding to the plural workloads belonging to the same type and having different sizes. Here, with the increase in the number of workloads (the number of samples) generated in relation to the single workload type, more detailed performance data can be acquired, whereby the accuracy of the primary performance models generated increases . [0103] Incidentally, since the combined workload for the combined performance model is formed by combining at least two different types of workloads, the number of samples of the combined workload significantly increases with the increase in the number of samples of the single workload type combined . In other words, in order to generate a highly accurate combined performance model, the period of time required for measurement generally becomes longer. [0104] On the other hand, in the case where the number of samples is reduced to shorten the period of time required for measurement, the amount of performance data obtained also reduces, which results in deterioration in accuracy of the primary performance model and the combined performance model Examiner finds first time period is the longer period of time. Vasic and Fujiwaka are analogous art because they are from the same field of endeavor as the claimed invention. It would have been obvious to one skilled in the art before the effective filing date of the invention to modify “wherein metrics for an activity of a workload executing in the cloud computing system are collected” Vasic to include “wherein metrics for an activity of a workload executing in the cloud computing system are collected at first sampling rate during the time period and at a second sampling rate, which is less than the first sampling rate, outside of the time period” as taught by Fujiwaka The motivation would have been to “generate a highly accurate combined performance model.” Fujiwaka para. 103. Claim 15 is rejected for the same reasons give above . Allowable Subject Matter 12-151-08 AIA 07-43 12-51-08 Claim s 2, 6, 10, 14, 18, are objected to as being dependent upon a rejected base claim, but would be allowable if rewritten in independent form including all of the limitations of the base claim and any intervening claims. Response to Arguments Applicant’s arguments have been considered but are rendered moot by the new grounds of rejection above. Conclusion 07-40 AIA Applicant's amendment necessitated the new ground(s) of rejection presented in this Office action. Accordingly, THIS ACTION IS MADE FINAL . See MPEP § 706.07(a). Applicant is reminded of the extension of time policy as set forth in 37 CFR 1.136(a). A shortened statutory period for reply to this final action is set to expire THREE MONTHS from the mailing date of this action. In the event a first reply is filed within TWO MONTHS of the mailing date of this final action and the advisory action is not mailed until after the end of the THREE-MONTH shortened statutory period, then the shortened statutory period will expire on the date the advisory action is mailed, and any nonprovisional extension fee (37 CFR 1.17(a)) pursuant to 37 CFR 1.136(a) will be calculated from the mailing date of the advisory action. In no event, however, will the statutory period for reply expire later than SIX MONTHS from the mailing date of this final action. Any inquiry concerning this communication or earlier communications from the examiner should be directed to ALBERT M PHILLIPS, III whose telephone number is (571)270-3256. The examiner can normally be reached 10a-6:30pm EST M-F. 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, Ann J Lo can be reached at (571) 272-9767. 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. /ALBERT M PHILLIPS, III/Primary Examiner, Art Unit 2159 Application/Control Number: 18/324,340 Page 2 Art Unit: 2159 Application/Control Number: 18/324,340 Page 3 Art Unit: 2159 Application/Control Number: 18/324,340 Page 4 Art Unit: 2159 Application/Control Number: 18/324,340 Page 5 Art Unit: 2159 Application/Control Number: 18/324,340 Page 6 Art Unit: 2159 Application/Control Number: 18/324,340 Page 7 Art Unit: 2159 Application/Control Number: 18/324,340 Page 8 Art Unit: 2159 Application/Control Number: 18/324,340 Page 9 Art Unit: 2159 Application/Control Number: 18/324,340 Page 10 Art Unit: 2159 Application/Control Number: 18/324,340 Page 11 Art Unit: 2159 Application/Control Number: 18/324,340 Page 12 Art Unit: 2159 Application/Control Number: 18/324,340 Page 13 Art Unit: 2159 Application/Control Number: 18/324,340 Page 14 Art Unit: 2159 Application/Control Number: 18/324,340 Page 15 Art Unit: 2159 Application/Control Number: 18/324,340 Page 16 Art Unit: 2159 Application/Control Number: 18/324,340 Page 17 Art Unit: 2159 Application/Control Number: 18/324,340 Page 18 Art Unit: 2159 Application/Control Number: 18/324,340 Page 19 Art Unit: 2159 Application/Control Number: 18/324,340 Page 20 Art Unit: 2159 Application/Control Number: 18/324,340 Page 21 Art Unit: 2159 Application/Control Number: 18/324,340 Page 22 Art Unit: 2159 Application/Control Number: 18/324,340 Page 23 Art Unit: 2159 Application/Control Number: 18/324,340 Page 24 Art Unit: 2159 Application/Control Number: 18/324,340 Page 25 Art Unit: 2159 Application/Control Number: 18/324,340 Page 26 Art Unit: 2159 Application/Control Number: 18/324,340 Page 27 Art Unit: 2159 Application/Control Number: 18/324,340 Page 28 Art Unit: 2159 Application/Control Number: 18/324,340 Page 29 Art Unit: 2159 Application/Control Number: 18/324,340 Page 30 Art Unit: 2159 Application/Control Number: 18/324,340 Page 31 Art Unit: 2159
Prosecution Insights