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 .
This communication is in response to the application filed on 10/29/2024.
Claims 1-20 are pending and rejected.
Priority
Acknowledgment is made of applicant’s claim for foreign priority under 35 U.S.C. 119 (a)-(d). The certified copy has been filed in parent Application No. IN202141013580 , filed on 03/26/2021.
Claim Rejections - 35 USC § 103
In the event the determination of the status of the application as subject to AIA 35 U.S.C. 102 and 103 (or as subject to pre-AIA 35 U.S.C. 102 and 103) is incorrect, any correction of the statutory basis (i.e., changing from AIA to pre-AIA ) for the rejection will not be considered a new ground of rejection if the prior art relied upon, and the rationale supporting the rejection, would be the same under either status.
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 1-2, 4-6, 8-9, 11-13, 15-16, and 18-20 are rejected under 35 U.S.C. 103 as being unpatentable over Samuel et. al, (US 20220269537 A1), hereafter Samuel in view of Morgan (US 20120311154 A1).
Regarding claim 1, Samuel teaches a computerized method of managing artificial intelligence (AI) workloads on a cloud infrastructure platform, the method comprising:
integrating distributed infrastructure resources into the cloud infrastructure platform via native support interfaces of the distributed infrastructure resources ([0017] AI workloads is migrated to other IHSs within a networked group of Information Handling Systems (IHSs), such as a trusted workgroup of an organization);
receiving the AI workloads for execution on the cloud infrastructure platform, the received AI workloads including training workloads and inferencing workloads ([0018], fig. 1, at a later point in time, the target IHS 102 receives a processed AI workload 114 from the other IHS 104. The processed AI workload 114 includes one or more profile recommendations (e.g., inferences) that may be applied for optimizing a performance of the target IHS 102; [0020] an AI workload 112 may include a set of input data (e.g., telemetry data, past profile recommendations, machine learning hints from other AI services, etc.) that may be processed to generate one or more inferences (e.g., profile recommendations);
assigning the received AI workloads to the integrated distributed infrastructure resources for execution, the assigning including determining resource requirements of the received Al workloads and identifying a portion of the integrated distributed infrastructure resources that satisfy the determined resource requirements ([0006] determining that an AI workload of the AI service exceeds a specified threshold, selecting a second IHS to perform at least a portion of the AI workload, and transmitting the at least one portion of the AI workload to the second IHS; [0053] the AI workload sharing method determines whether or not to allow access to the cloud-based AI service, it may access the policy associated with that feature to obtain the necessary information. If the AI workload sharing method determines that use of the cloud-based service is allowed, send AI workload to the cloud-based AI service; otherwise, the AI workload sharing method continues to use the resources of the target IHS for performing the AI workload);
scheduling the Al workloads for execution on the identified portion of the integrated distributed infrastructure resources in accordance with priorities of the received AI workloads ([0057] each of the target IHS, lower priority IHS 102, 104b and higher priority IHS 102, 104c publish their own resource information to one another. For example, each IHS 102, 104 may transmit information associated with its current workload along with its available capacity to handle additional workload. Using this information received from the other IHSs, the target IHS 102 determines higher priority IHS 102, 104c has a relatively higher priority than lower priority IHS 102, 104b);
executing the received AI workloads in accordance with the scheduling ([0059] receives the processed workload, and applies the processed workload to its resource).
Samuel does not explicitly teach
scheduling the Al workloads for execution; and
Morgan teaches
scheduling the Al workloads for execution ([0042] Upon determining that one or more policies reflected in the migration policy stack permit or require the migration of the one or more workloads to one or more clouds in the one or more target clouds, the policy management tool and/or other logic can interact with a workload migration scheduler to schedule one or more movements or migrations of the one or more workloads to those host target(s)).
It would have been obvious to a person of ordinary skill in the art before the effective filling date of the claimed invention made to include in the Samuel disclosure, the schedule to migrate workload to cloud, as taught by Morgan. One would be motivated to do so that potentially reduced costs and/or other benefits can be achieved based on time of usage.
Regarding claims 2, 9, and 16, Samuel and Morgan teach all limitations of parent claims 1, 8, and 15, wherein Samuel further teaches the native support interfaces include interfaces and libraries of resource providers, the providers including tenants of the cloud infrastructure platform ([0042] , fig. 1, the AI workload sharing manager 304 may be provided as a subscription service, in which users of IHS 102 may register for providing the AI workload sharing system).
Regarding claims 4, 11, and 18, Samuel and Morgan teach all limitations of parent claims 1, 8, and 15, wherein Morgan further teaches a global scheduling subsystem generates a schedule for the AI workloads, scheduling the execution of the AI workloads includes a schedule for training workloads and another schedule for inferencing workloads on the same infrastructure resources, the training workloads and inferencing workloads being multiplexed on the same infrastructure resources ([0055-0056] Scheduling allows to migrate data automatically at set time, so one can easily run migration in intervals or at off-hours ([0056] Two options are available for migration scheduler. One option is, to schedule the migration for one time only, for a certain date and time. Or to schedule a schedule backup, which means every certain day during the week, at a certain time migration will be running).
It would have been obvious to a person of ordinary skill in the art before the effective filling date of the claimed invention made to include in the Samuel disclosure, the schedule to migrate workload to cloud, as taught by Morgan. One would be motivated to do so that potentially reduced costs and/or other benefits can be achieved based on time of usage.
Regarding claims 5, 12, and 19, Samuel and Morgan teach all limitations of parent claims 1, 8, and 15, wherein Samuel further teaches the priorities of the AI workloads comprise a higher tier AI workload being scheduled for a greater share of resource usage time than a lower tier AI workload ([0056], fig. 1, the priority-based method 500 is shown with a target IHS 102, and two other IHSs, namely a lower priority IHS 102, 104b, and a higher priority IHS 102, 104c).
Regarding claims 6, 13, and 20, Samuel and Morgan teach all limitations of parent claims 1, 8, and 15, wherein Samuel further teaches scheduling the AI workloads for execution on the identified portion of the integrated distributed infrastructure resources includes isolating the AI workloads from each other in secure containers, and scheduling AI workloads associated with different tenants to run alongside each other on resources associated with a same server ([0047] AI workload sharing manager is stored and executed on the IHS it is configured to provide AI workload sharing services for; [0053] During configuration, for example, target IHS 102 may either be configured (e.g., registered) or not to use a cloud-based AI service 108 based on certain factors, such as expected security exposure, anticipated workload sharing requirements, logistics associated with maintaining a subscription to the cloud-based service).
Regarding claim 8, Samuel teaches a system for managing artificial intelligence (AI) workloads on a cloud infrastructure platform, the system comprising:
a processor; and a memory storing computer-executable instructions that, in response to execution by the processor ([0022], fig. 2, HS 200 includes one or more processors 201, such as a Central Processing Unit (CPU), that execute code retrieved from system memory), cause the processor to:
integrate distributed infrastructure resources into the cloud infrastructure platform via native support interfaces of the distributed infrastructure resources ([0017] AI workloads is migrated to other IHSs within a networked group of Information Handling Systems (IHSs), such as a trusted workgroup of an organization);
receive the AI workloads for execution on the cloud infrastructure platform, the received AI workloads including training workloads and inferencing workloads ([0018], fig. 1, at a later point in time, the target IHS 102 receives a processed AI workload 114 from the other IHS 104. The processed AI workload 114 includes one or more profile recommendations (e.g., inferences) that may be applied for optimizing a performance of the target IHS 102; [0020] an AI workload 112 may include a set of input data (e.g., telemetry data, past profile recommendations, machine learning hints from other AI services, etc.) that may be processed to generate one or more inferences (e.g., profile recommendations);
assign the received AI workloads to the integrated distributed infrastructure resources for execution, the assigning including determining resource requirements of the received AI workloads and identifying a portion of the integrated distributed infrastructure resources that satisfy the determined resource requirements ([0006] determining that an AI workload of the AI service exceeds a specified threshold, selecting a second IHS to perform at least a portion of the AI workload, and transmitting the at least one portion of the AI workload to the second IHS; [0053] the AI workload sharing method determines whether or not to allow access to the cloud-based AI service, it may access the policy associated with that feature to obtain the necessary information. If the AI workload sharing method determines that use of the cloud-based service is allowed, send AI workload to the cloud-based AI service; otherwise, the AI workload sharing method continues to use the resources of the target IHS for performing the AI workload);
schedule the AI workloads for execution on the identified portion of the integrated distributed infrastructure resources in accordance with priorities of the received AI workloads ([0057] each of the target IHS, lower priority IHS 102, 104b and higher priority IHS 102, 104c publish their own resource information to one another. For example, each IHS 102, 104 may transmit information associated with its current workload along with its available capacity to handle additional workload. Using this information received from the other IHSs, the target IHS 102 determines higher priority IHS 102, 104c has a relatively higher priority than lower priority IHS 102, 104b); and
execute the received Al workloads in accordance with the scheduling ([0059] receives the processed workload, and applies the processed workload to its resource).
Samuel does not explicitly teach
scheduling the Al workloads for execution; and
Morgan teaches
scheduling the Al workloads for execution ([0042] Upon determining that one or more policies reflected in the migration policy stack permit or require the migration of the one or more workloads to one or more clouds in the one or more target clouds, the policy management tool and/or other logic can interact with a workload migration scheduler to schedule one or more movements or migrations of the one or more workloads to those host target(s)).
It would have been obvious to a person of ordinary skill in the art before the effective filling date of the claimed invention made to include in the Samuel disclosure, the schedule to migrate workload to cloud, as taught by Morgan. One would be motivated to do so that potentially reduced costs and/or other benefits can be achieved based on time of usage.
Regarding claim 15, Samuel teaches a memory device storing computer-executable instructions, that when executed by a processor, cause the processor to perform operations comprising:
integrating distributed infrastructure resources into a cloud infrastructure platform via native support interfaces of the distributed infrastructure resources ([0017] AI workloads is migrated to other IHSs within a networked group of Information Handling Systems (IHSs), such as a trusted workgroup of an organization);
receiving artificial intelligence (AI) workloads for execution on the cloud infrastructure platform, the received AI workloads including training workloads and inferencing workloads ([0018], fig. 1, at a later point in time, the target IHS 102 receives a processed AI workload 114 from the other IHS 104. The processed AI workload 114 includes one or more profile recommendations (e.g., inferences) that may be applied for optimizing a performance of the target IHS 102; [0020] an AI workload 112 may include a set of input data (e.g., telemetry data, past profile recommendations, machine learning hints from other AI services, etc.) that may be processed to generate one or more inferences (e.g., profile recommendations);
assigning the received AI workloads to the integrated distributed infrastructure resources for execution, the assigning including determining resource requirements of the received AI workloads and identifying a portion of the integrated distributed infrastructure resources that satisfy the determined resource requirements ([0006] determining that an AI workload of the AI service exceeds a specified threshold, selecting a second IHS to perform at least a portion of the AI workload, and transmitting the at least one portion of the AI workload to the second IHS; [0053] the AI workload sharing method determines whether or not to allow access to the cloud-based AI service, it may access the policy associated with that feature to obtain the necessary information. If the AI workload sharing method determines that use of the cloud-based service is allowed, send AI workload to the cloud-based AI service; otherwise, the AI workload sharing method continues to use the resources of the target IHS for performing the AI workload);
scheduling the Al workloads for execution on the identified portion of the integrated distributed infrastructure resources in accordance with priorities of the received AI workloads ([0057] each of the target IHS, lower priority IHS 102, 104b and higher priority IHS 102, 104c publish their own resource information to one another. For example, each IHS 102, 104 may transmit information associated with its current workload along with its available capacity to handle additional workload. Using this information received from the other IHSs, the target IHS 102 determines higher priority IHS 102, 104c has a relatively higher priority than lower priority IHS 102, 104b); and
executing the received AI workloads in accordance with the scheduling ([0059] receives the processed workload, and applies the processed workload to its resource).
Samuel does not explicitly teach
scheduling the Al workloads for execution; and
Morgan teaches
scheduling the Al workloads for execution ([0042] Upon determining that one or more policies reflected in the migration policy stack permit or require the migration of the one or more workloads to one or more clouds in the one or more target clouds, the policy management tool and/or other logic can interact with a workload migration scheduler to schedule one or more movements or migrations of the one or more workloads to those host target(s)); and
It would have been obvious to a person of ordinary skill in the art before the effective filling date of the claimed invention made to include in the Samuel disclosure, the schedule to migrate workload to cloud, as taught by Morgan. One would be motivated to do so that potentially reduced costs and/or other benefits can be achieved based on time of usage.
Claims 3, 10, and 17 are rejected under 35 U.S.C. 103 as being unpatentable over Samuel (US 20220269537 A1) in view of Morgan (US 20120311154 A1) and further in view of McClure (US 20220012104 A1).
Regarding claims 3, 10, and 17, Samuel and Morgan teach all limitations of parent claims 1, 8, and 15, wherein the portion of the integrated distributed infrastructure resources is assigned to an AI workload, the assigning including saving a state checkpoint of an AI workload that is currently being executed on a first resource, migrating the AI workload to a second resource, restoring the saved state checkpoint of the migrated AI workload on the second resource, and subsequently assigning at least a portion of the first resource to another AI workload.
McClure teaches
wherein the portion of the integrated distributed infrastructure resources is assigned to an AI workload, the assigning including saving a state checkpoint of an AI workload that is currently being executed on a first resource, migrating the AI workload to a second resource, restoring the saved state checkpoint of the migrated AI workload on the second resource, and subsequently assigning at least a portion of the first resource to another AI workload ([0033] The management service 31 can augment the original compute kernel with instructions that can save its execution state into the offline register file at the halting points; [0034] The augmented compute kernel can be suspended and subsequently resumed on the same host, or the augmented compute kernel can be suspended on one host and resumed on a different host).
It would have been obvious to a person of ordinary skill in the art before the effective filling date of the claimed invention made to include in the Samuel disclosure, resume workload to another node with the same state of the original node, as taught by McClure. One would be motivated to do so that the workload can continue on the destination host while the source host being halted.
Claims 7 and 14 are rejected under 35 U.S.C. 103 as being unpatentable over Samuel (US 20220269537 A1) in view of Morgan (US 20120311154 A1) and further in view of Wang (US 20210103468 A1).
Regarding claims 7 and 14, Samuel and Morgan teach all limitations of parent claims 1 and 8, Samuel does not explicitly teach:
monitoring a performance of the cloud infrastructure platform and, based on the monitoring, adjusting the scheduling of the AI workloads.
monitoring a performance of the cloud infrastructure platform and, based on the monitoring, adjusting the scheduling of the AI workloads ([0064] performance component 108 can assign the performance points based on one or more defined workload execution objectives defined by AI; [0110] modify a scheduling decision (e.g., a scheduler ranking) to run a workload type on a node based on the performance points).
It would have been obvious to a person of ordinary skill in the art before the effective filling date of the claimed invention made to include in the Samuel disclosure, modify the schedule base on workload performance, as taught by Wang. One would be motivated to do so to performance biased resource scheduling based on runtime performance of a certain workload type on one or more nodes.
Conclusion
The prior art made of record and not relied upon is considered pertinent to applicant's disclosure.
Aronovich (US 20220179693 A1) and JHA (US 20200151018 A1).
Any inquiry concerning this communication or earlier communications from the examiner should be directed to ANH NGUYEN whose telephone number is (571)270-0657. The examiner can normally be reached 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, Umar Cheema can be reached at 5712703037. 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.
/ANH NGUYEN/Primary Examiner, Art Unit 2458
Prosecution Insights