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 .
Specification
The disclosure is objected to because of the following informalities: on page 16, in the paragraph describing steps 208-212, the specification refers to "manager 110," whereas the corresponding element is consistently designated "manager 120" elsewhere in the specification (reference character 110 is separately used to designate "network 110"). Appropriate correction is required.
Claim Objections
Claim 8 is objected to because of the following informalities: claim 8 depends from claim 7, a method claim, but recites "the predictive orchestrator further configured to cause the image layer to be removed from the host system...," a structural limitation lacking antecedent basis in claim 7 and inconsistent with the method format of claim 7. Appropriate correction is required.
Claim Rejections - 35 USC § 112
The following is a quotation of the first paragraph of 35 U.S.C. 112(a):
(a) IN GENERAL.—The specification shall contain a written description of the invention, and of the manner and process of making and using it, in such full, clear, concise, and exact terms as to enable any person skilled in the art to which it pertains, or with which it is most nearly connected, to make and use the same, and shall set forth the best mode contemplated by the inventor or joint inventor of carrying out the invention.
The following is a quotation of the first paragraph of pre-AIA 35 U.S.C. 112:
The specification shall contain a written description of the invention, and of the manner and process of making and using it, in such full, clear, concise, and exact terms as to enable any person skilled in the art to which it pertains, or with which it is most nearly connected, to make and use the same, and shall set forth the best mode contemplated by the inventor of carrying out his invention.
Claims 1-18 are rejected under 35 U.S.C. 112(a) as failing to comply with the written description requirement. The claim(s) contains subject matter which was not described in the specification in such a way as to reasonably convey to one skilled in the relevant art that the inventor or a joint inventor, at the time the application was filed, had possession of the claimed invention.
Per claims 1, 7 and 13, each of these independent claim recites, in a distributed ecosystem comprising a plurality of host systems, determining a probability of an image layer executing on a host system of the plurality of host systems and causing the image layer to be pre-loaded on the host system if the probability of the image layer executing on the host system satisfies a threshold probability.
Claim 1 recites these acts as performed by a predictive orchestrator comprising a program of instructions executed by a processor; claim 7 recites them as method steps; and claim 13 recites them as instructions carried on a non-transitory computer-readable medium.
The second act appears adequately described. The specification discloses layer repository 122 and pre-loaded layers 124 (page 12, lines 1–8) and step 210 (page 16, lines 3–12), under which a layer that “meets or exceeds a particular threshold for execution on a particular endpoint” is loaded onto that endpoint so that the layer “may not require download when a workload requiring the layer is later placed on such endpoint.”
The first act is not described. The specification recites the determination of the probability at precisely the same level of generality at which the claims recite it, and nowhere descends below that level:
At step 202 (page 14, lines 1–11), “manager 120 may determine the probability Pa of an application launching based on one or more predictive parameters,” which “may be based on the state of an endpoint …, historic workload placement …, and/or user actions.” This passage names the inputs to the determination and names its output. It describes no operation performed on the inputs to produce Pa.
At step 204 (page 14, lines 12–20), the probability of the application launching on each endpoint (Pae1, Pae2, Pae3) is said to be determined “based on the probability Pa and the one or more predictive parameters.” No functional relationship between Pa and Pae is disclosed.
At step 206 (page 14, lines 21–32), the layer probabilities (PL1e1, PL2e1, PL3e1) are said to be determined “based on the image layers that make up a container for executing the workload.” No relationship between the endpoint probability and the layer probability is disclosed.
FIG. 2 does not supply the missing disclosure. Blocks 202, 204, and 206 are labeled boxes that restate the claimed functions — “DETERMINE PROBABILITY OF APPLICATION LAUNCH,” “DETERMINE PROBABILITY OF APPLICATION LAUNCHING ON ENDPOINT,” and “LAYER PROBABILITIES” — fed by an input labeled “USER PATTERNS / ENDPOINT STATE / HISTORIC WORKLOADS.” No block discloses a computation and the figure contains no equation.
The single quantitative disclosure in the application does not cure the deficiency. At step 207 (page 15, line 22 through page 16, line 2), the specification states that where a layer has a probability P(A) of executing on an endpoint in connection with a first application and a probability P(B) in connection with a second application, “manager 120 may use basic principles of probability and statistics to determine that the probability of the layer executing on the endpoint is P(A) + P(B) − P(A) · P(B).” That formula is an aggregation rule — the inclusion-exclusion expression for the union of two independent events — which operates on probabilities P(A) and P(B) that it presupposes have already been determined by undisclosed means. It describes how two probabilities are combined, not how any probability is determined.
The artificial intelligence and machine learning passage (page 16, lines 18–31) likewise does not cure the deficiency. It states that reinforcement “may be applied (e.g., by modifying associated probability metrics for the layer and/or endpoint) based on a frequency of workloads executing on the endpoint that utilize the layer.” By its terms this passage refines probability metrics that already exist, and so again presupposes the determination that the claims require. It identifies no model class, no feature representation, no reward or objective function, no magnitude or direction of adjustment, and no training procedure.
Under the broadest reasonable interpretation, the limitation determine a probability of an image layer executing on a host system is defined solely by the result it achieves and places no constraint on the manner of determination. It therefore recites a genus encompassing every technique capable of producing such a value, including frequency estimation, Bayesian inference, regression, supervised and unsupervised machine learning, reinforcement learning, heuristic scoring, static rule tables, and techniques not yet devised as of the effective filing date. For a genus defined in functional terms, the specification must describe either a representative number of species or structural features common to the members of the genus. See MPEP § 2163; Abbvie Deutschland GmbH & Co. v. Janssen Biotech, Inc., 759 F.3d 1285, 1300–01 (Fed. Cir. 2014); Carnegie Mellon Univ. v. Hoffmann-La Roche Inc., 541 F.3d 1115, 1126 (Fed. Cir. 2008). The specification does neither. It discloses no species — no equation, decision rule, counting or frequency method, statistical model, or machine-learning architecture — for determining the application-launch probability, the per-endpoint probability, or the image-layer probability, and it identifies no common feature of the genus beyond the recited result.
Identifying the inputs to a computation is not a description of the computation. A recitation that an output is “based on” named parameters delimits the boundary of the operation without describing its contents. Where a claimed function is computer-implemented, it is not sufficient that a person of ordinary skill could devise a suitable algorithm; the specification itself must disclose how the claimed function is achieved. See MPEP § 2161.01(I); Vasudevan Software, Inc. v. MicroStrategy, Inc., 782 F.3d 671, 681–82 (Fed. Cir. 2015). A disclosure that announces a desired result does not convey possession of all means of achieving it. Ariad Pharmaceuticals, Inc. v. Eli Lilly & Co., 598 F.3d 1336, 1349–51 (Fed. Cir. 2010) (en banc).
It is acknowledged that claims 1–18 were present in the application as filed and that original claims may serve as their own written description. See MPEP § 2163(I)(A). That presumption is rebutted here. The original claims recite the determination in the same purely functional terms in which the specification describes it, and therefore add nothing to the description of how the function is performed. A claim cannot establish possession of a functionally defined genus by restating the result that defines the genus. See In re Koller, 613 F.2d 819, 823 (CCPA 1980).
Accordingly, the specification does not reasonably convey to one skilled in the relevant art that the inventors, at the time the application was filed, had possession of the claimed invention.
Per claims 3, 9 and 15, these claims recite that determining the probability of the image layer executing on the host system comprises determining a probability of a workload requiring the image layer executing on the host system. The claims substitute one undescribed determination for another. The specification discloses the workload-to-layer dependency in outline at steps 204–206 (page 14), but it no more describes how the probability of a workload executing on a host system is determined than it describes how the layer probability is determined. The deficiency identified above is relocated rather than cured.
Per claims 4, 5, 10, 11, 16, and 17, these claims recite that the probability is based on one or more predictive parameters comprising one or more of a state of the host system, historic workload placement …, and user actions …. The three named parameters do appear in the specification at step 202 (page 14, lines 4–11), so the inputs themselves are described. The claims nevertheless remain deficient because naming the inputs to a determination is not a description of the determination, as explained above. Moreover, because “one or more of” permits reliance on any single listed parameter, these claims encompass determining the probability from host-system state alone, or from historic workload placement alone, or from user actions alone, and the specification describes no mechanism by which any one of these parameters, standing alone, yields the recited probability.
Per claims 6, 12, and 18, these claims recite that determining the probability comprises determining at least two probabilities, namely a probability of a first workload and a probability of a second workload requiring the image layer executing on the host system. The premise is described at page 15, lines 14–21 (“a first layer may have multiple probabilities PL1e1 of executing on endpoint 1 — each probability associated with a different application”), and the specification supplies the union rule for combining them at step 207. However, the claims require determining each of the two constituent probabilities, and it is precisely those constituent determinations that the specification presupposes rather than describes. The disclosed aggregation formula operates on inputs whose derivation is undescribed and therefore cannot supply possession of the recited determinations.
Per claims 2, 8, and 14, these claims recite causing the image layer to be removed from the host system if the layer is already pre-loaded on the host system and the probability of the image layer executing on the host system fails to satisfy the threshold probability. The removal operation itself is described at step 212 (page 16, lines 12–17). The claims are nevertheless rejected because they incorporate every limitation of their respective parent claims, including the undescribed probability determination, and the added operation is conditioned on that same unsupported probability.
Claims 2–6, 8–12, and 14–18 are further rejected as depending from a rejected base claim. The additional limitations of these claims do not supply the missing description of how the recited probability is determined.
Claim Rejections - 35 USC § 112
The following is a quotation of 35 U.S.C. 112(b):
(b) CONCLUSION.—The specification shall conclude with one or more claims particularly pointing out and distinctly claiming the subject matter which the inventor or a joint inventor regards as the invention.
The following is a quotation of 35 U.S.C. 112 (pre-AIA ), second paragraph:
The specification shall conclude with one or more claims particularly pointing out and distinctly claiming the subject matter which the applicant regards as his invention.
Claims 3, 4, 5, 6, 8, 9, 10, 11, 12, 15, 16, 17, and 18 are rejected under 35 U.S.C. 112(b) as being indefinite for failing to particularly point out and distinctly claim the subject matter which the inventor or a joint inventor regards as the invention.
Claims 3, 9, and 15 each recite "a probability of a workload requiring the image layer executing on the host system." This phrase is susceptible to at least two materially different readings: (i) the probability that a workload -- which requires the image layer -- itself executes on the host system, or (ii) the probability that a workload requires the image layer to execute on the host system. Comparison to parallel claims 5, 11, and 17, which use the unambiguous phrase "workloads requiring execution of the image layer," confirms that the claim language of claims 3, 9, and 15 does not clearly convey which event the recited probability measures. A person of ordinary skill in the art therefore cannot determine, with reasonable certainty, the scope of this limitation. For purposes of examination, this limitation is interpreted under BRI as the probability that a workload which requires the image layer will itself execute on the host system.
Claims 4, 10, and 16 each recite "the probability of the workload executing on the host system," which inherits the antecedent ambiguity of parent claims 3, 9, and 15, respectively, discussed above. For purposes of examination, this limitation is interpreted under BRI consistently with the interpretation applied to claims 3, 9, and 15, above.
Claims 4, 5, 10, 11, 16, and 17 each recite predictive parameters "comprising one or more of" a list of three items joined by "and" before the final item. This "one or more of A, B, and C" format creates ambiguity as to whether the claim requires any single listed parameter, any subset of the parameters, or the conjunctive combination of all three parameters, in the same manner as the "at least one of A, B, and C" construction identified as ambiguous in MPEP § 2173.05(h). For purposes of examination, this limitation is interpreted under BRI to mean any one or more of the listed parameters, singly or in any combination.
Claims 6, 12, and 18 each recite "a probability of a first workload requiring the image layer executing on the host system" and "a probability of a second workload requiring the image layer executing on the host system." Each instance is subject to the same two-reading ambiguity discussed above with respect to claims 3, 9, and 15. For purposes of examination, each limitation is interpreted under BRI as the probability that the respective (first or second) workload -- which requires the image layer -- itself executes on the host system, consistent with the interpretation applied to claims 3, 9, and 15, above.
Claim 8 recites the limitation "the predictive orchestrator" in line 1. There is insufficient antecedent basis for this limitation in the claim. Claim 8 depends from method claim 7, which does not recite a predictive orchestrator or any structural element of that name; the term appears to have been carried over from system claim 2 without adjustment for claim 8's method context. For purposes of examination, "the predictive orchestrator" is interpreted under BRI as referring to whatever entity performs the steps of parent claim 7.
Claim 8 is further rejected as being indefinite for improperly mixing method and apparatus limitations. Claim 8 depends from method claim 7 but recites its added limitation as an apparatus configuration ("further configured to") tied to a structural element ("the predictive orchestrator"), rather than as an active method step. It is therefore unclear whether infringement of claim 8 occurs upon performance of the recited removal alone, or requires in addition the presence of a specifically configured "predictive orchestrator." See IPXL Holdings, L.L.C. v. Amazon.com, Inc., 430 F.3d 1377, 1384 (Fed. Cir. 2005). For purposes of examination, this limitation is interpreted under BRI as if it recited the active method step "further comprising causing the image layer to be removed from the host system if the layer is already pre-loaded on the host system and the probability of the image layer executing on the host system fails to satisfy the threshold probability".
Claim 8 is further rejected as failing to set forth the subject matter which the inventor or a joint inventor regards as the invention. Unlike parent claim 7 and sibling method claims 9-12, which add limitations phrased as active method steps, claim 8's added limitation is phrased entirely as an apparatus configuration rather than as a step of the method, and does not correspond to what the specification's separately claimed method (claim 7) or system (claims 1-2) individually regard as the invention. For purposes of examination, this limitation is interpreted under BRI as if it recited an active method step, as set forth above.
Appropriate correction is required.
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.
The factual inquiries for establishing a background for determining obviousness under 35 U.S.C. 103 are summarized as follows:
1. Determining the scope and contents of the prior art.
2. Ascertaining the differences between the prior art and the claims at issue.
3. Resolving the level of ordinary skill in the pertinent art.
4. Considering objective evidence present in the application indicating obviousness or nonobviousness.
This application currently names joint inventors. In considering patentability of the claims the examiner presumes that the subject matter of the various claims was commonly owned as of the effective filing date of the claimed invention(s) absent any evidence to the contrary. Applicant is advised of the obligation under 37 CFR 1.56 to point out the inventor and effective filing dates of each claim that was not commonly owned as of the effective filing date of the later invention in order for the examiner to consider the applicability of 35 U.S.C. 102(b)(2)(C) for any potential 35 U.S.C. 102(a)(2) prior art against the later invention.
Claims 1, 3-7, 9-13 and 15-18 are rejected under 35 USC 103 as being unpatentable over US Pat. Pub. No. 2018/0039524 to Dettori et al. (hereinafter Dettori) in view of US Pat. No. 11,451,615 to Scrivano (hereinafter Scrivano).
Per claim 1, Dettori discloses An information handling system (Dettori: ¶[0028]…Dettori's embodiment is a computer system built from one or more processors, one or more memories and one or more storage devices carrying the program instructions the processors execute, which constitutes an information handling system under BRI, "An embodiment includes a computer system. The computer system includes one or more processors, one or more computer-readable memories, and one or more computer-readable storage devices, and program instructions stored on at least one of the one or more storage devices for execution by at least one of the one or more processors via at least one of the one or more memories"), comprising:
a processor; and (Dettori: ¶[0028]…the same computer system recites one or more processors that execute the stored program instructions, which constitutes a processor under BRI, "An embodiment includes a computer system. The computer system includes one or more processors, one or more computer-readable memories, and one or more computer-readable storage devices, and program instructions stored on at least one of the one or more storage devices for execution by at least one of the one or more processors via at least one of the one or more memories"; ¶[0075]…Dettori's data processing system 200 embodies that recitation in processing unit 206, "Processing unit 206 may contain one or more processors and may be implemented using one or more heterogeneous processor systems").
a predictive orchestrator comprising a program of instructions configured to, when read and executed by the processor, in a distributed ecosystem comprising a plurality of host systems (Dettori: ¶[0068]…Dettori's application 105, acting through provisioning system 107, predicts which container image layers are needed and provisions them across the several hosts 144A-144D of managed data processing environment 142, which constitutes a predictive orchestrator operating in a distributed ecosystem comprising a plurality of host systems under BRI, "Application 105 implements an embodiment described herein. Provisioning system 107 provisions layers for deploying containers on one or more hosts in managed data processing environment 142, such as hosts 144A, 144B, 144C, and 144D"; ¶[0066]…Dettori states that the embodiment is distributed across several data processing systems joined by a data network, "an embodiment can be distributed across several data processing systems and a data network as shown"; ¶[0027]…the embodiment is carried as program instructions on storage devices for execution, which constitutes a program of instructions configured to be read and executed by the processor, "An embodiment includes a computer usable program product. The computer usable program product includes one or more computer-readable storage devices, and program instructions stored on at least one of the one or more storage devices"):
determine a probability of an image layer executing on a host system of the plurality of host systems; and (Dettori: ¶[0038]…Dettori predicts, for each host of the environment, which container image layers are likely to be needed at that host, which constitutes determining a probability of an image layer executing on a host system of the plurality of host systems under BRI, "An embodiment predicts which layers are likely to be needed at which host in the data processing environment"; ¶[0046]…Dettori computes that per-layer, per-host prediction from a quantified level-of-participation factor, so the prediction is a measure of how likely the layer is to be needed at that host, "An embodiment uses the level of participation of a layer as a factor in predicting whether the layer is likely to be needed by a host to configure a container during a period").
cause the image layer to be pre-loaded on the host system… (Dettori: ¶[0019]…Dettori causes the predicted subset of container image layers to be placed on the node in advance of the future period in which a container needing them will be configured there, which constitutes causing the image layer to be pre-loaded on the host system under BRI, "The embodiment causes the adjusted subset of layers to be provisioned on the node prior to the future time"; ¶[0023]…Dettori conditions that advance placement on the layer's computed level exceeding a threshold, "An embodiment further selects from the set of levels a subset of levels, each level in the subset of levels exceeding a threshold level of participation. The embodiment selects the subset of layers corresponding to the selected subset of levels").
To the extent it is argued that Dettori's prediction of whether a layer is likely to be needed at a host does not expressly constitute a probability, Scrivano expressly teaches determine a probability of an image layer executing on a host system of the plurality of host systems (Scrivano: 5:11-13…Scrivano's value generation component computes, for each unit of container image content, a probability value expressing how likely that content is to be used by a container running on the host system, "The value generation component 164 may then generate probability values for each of the files of the container image"; 5:50-51…Scrivano makes that determination for content destined for any of several host systems, "In embodiments, containers 136a-d may be supported by multiple host systems").
Dettori does not expressly disclose, but Scrivano does teach:
…if the probability of the image layer executing on the host system satisfies a threshold probability. (Scrivano: 5:14-23…Scrivano identifies a threshold value against which those probability values are compared and requests only the content whose probability value satisfies that threshold, which constitutes the satisfies a threshold probability condition under BRI, "The file requesting component 162 may identify a threshold value associated with the probability values determined by the value generation component 164 that is used to determine which files are to be pulled from a container image of the file system 130. The file requesting component 162 may generate a request for files of the container image that have a probability value that satisfies the threshold and provide the request to the file system 130"; 8:30-35…only the content whose probability value meets or exceeds the threshold is in fact delivered to and stored on the host system, "Because FILE A, FILE B, and FILE C have corresponding access scores that are greater than or equal to the threshold 529, file system 130 may provide FILE A, FILE B, and FILE C to host system 110. Upon receipt of the files, host system 110 may store the files in locally in memory 170 that is accessibly by container 527").
Dettori and Scrivano are analogous art because they are from the same field of endeavor, specifically the predictive placement of container image content onto host systems in a container-orchestration environment. They address the same problem of reducing container start-up latency by placing on a host system, in advance, only that container image content which the host system is expected to need.
Before the effective filing date of the claimed invention, it would have been obvious to a PHOSITA to modify Dettori's predictive layer pre-provisioning so that the predicted need for an image layer at a host system is expressed as a probability value and the layer is pre-loaded onto that host system only when the probability value satisfies a threshold probability.
The suggestion/motivation for doing so would have been provided by Dettori itself, which already gates pre-provisioning on a computed prediction factor measured against a numerical threshold and therefore invites the substitution of an express probability value for that factor, "An embodiment further selects from the set of levels a subset of levels, each level in the subset of levels exceeding a threshold level of participation. The embodiment selects the subset of layers corresponding to the selected subset of levels" (Dettori: ¶[0023]). Scrivano supplies the express probabilistic form of the very same threshold comparison and teaches that scoring container image content by probability lets the content placed on the host system be selected intelligently rather than statically, "By utilizing probabilistic per-file image preloading, the files of a container image may be intelligently selected for preloading for a container based on the probability that the files will be used by the container" (Scrivano: 3:30-33). Furthermore, this is the application of a known technique, probability-valued prediction compared against a threshold, to a known method ready for improvement, Dettori's prediction-factor-and-threshold layer pre-provisioning, to yield the predictable result of pre-loading exactly the layers the host is most likely to need, the rationale of MPEP § 2143(D).
Per claim 3, Dettori combined with Scrivano discloses claim 1. Dettori further teaches wherein determining the probability of the image layer executing on the host system comprises determining a probability of a workload requiring the image layer executing on the host system (Dettori: ¶[0038]…Dettori derives its layer-level prediction for a host from the expectation that a workload whose container requires those layers will be sent to that host during the period, which constitutes determining a probability of a workload requiring the image layer executing on the host system under BRI, "One non-limiting example of the conditions is a period during which a certain type of workload is expected to be sent to the host for which a container will likely have to be configured at the host during the period"; ¶[0052]…Dettori expressly determines the host at which the workload is likely to be processed, "Based on the analysis, the embodiment determines a host where the workload is likely to be processed").
Per claim 4, Dettori combined with Scrivano discloses claim 3. Dettori further teaches wherein the probability of the workload executing on the host system is based on one or more predictive parameters comprising one or more of a state of the host system, historic workload placement of the workload, and user actions associated with the workload (Dettori: ¶[0041]…Dettori conditions the prediction on the availability of the host during the period, which constitutes a state of the host system and satisfies the one or more of alternative the claim recites, "Another non-limiting example of the conditions is the availability of the host or other hosts during a period or for an expected workload"; ¶[0049]…Dettori additionally draws on a repository of historical usage of and requests for layers in evaluating the prediction factors, which constitutes historic workload placement of the workload, "The historical usage of specific layers in container images, historical requests for specific layers, and other similarly purposed data can be stored and made available from a historical data repository").
Per claim 5, Dettori combined with Scrivano discloses claim 1. Dettori further teaches wherein the probability of the image layer executing on the host system is based on one or more predictive parameters comprising one or more of a state of the host system, historic workload placement of one or more workloads requiring execution of the image layer, and user actions associated with the one or more workloads requiring execution of the image layer (Dettori: ¶[0049]…Dettori evaluates the factors feeding its layer prediction from stored records of how specific layers were historically used and requested in container images, which constitutes historic workload placement of one or more workloads requiring execution of the image layer and satisfies the one or more of alternative the claim recites, "The historical usage of specific layers in container images, historical requests for specific layers, and other similarly purposed data can be stored and made available from a historical data repository"; ¶[0041]…Dettori also conditions the same layer prediction on host availability, which constitutes a state of the host system, "Another non-limiting example of the conditions is the availability of the host or other hosts during a period or for an expected workload").
Per claim 6, Dettori combined with Scrivano discloses claim 1. Dettori further teaches wherein determining the probability of the image layer executing on the host system comprises determining at least two probabilities comprising: a probability of a first workload requiring the image layer executing on the host system; and a probability of a second workload requiring the image layer executing on the host system (Dettori: ¶[0040]…Dettori evaluates, for one and the same host, other workloads likely to be processed there whose containers share one or more of the very layers already predicted for that host, so the same layer's need at that host is assessed against a first workload and against a second workload, which constitutes determining at least two probabilities under BRI, "Depending on the period for which the prediction is being made, the host may already be configured with other containers for such other workloads, some of which might share one or more layers with the predicted layers"; ¶[0051]…Dettori performs that assessment workload by workload, looking ahead into the workload plan and analyzing each planned workload in turn, "If workload planning is available for a prediction period, the embodiment looks ahead into the planning and analyzes a workload that is planned during the period"; ¶[0052]…and for each such workload Dettori determines the host at which it is likely to be processed, "Based on the analysis, the embodiment determines a host where the workload is likely to be processed").
Per claim 7, Dettori discloses A method comprising, in a distributed ecosystem comprising a plurality of host systems (Dettori: ¶[0019]…Dettori's embodiment is expressly a method of predicting and pre-provisioning container image layers, which constitutes the recited method under BRI, "An embodiment includes a method that computes, using a metadata of a layer, a prediction factor comprising a level of participation of the layer in a set of container images"; ¶[0068]…that method operates across the several hosts 144A-144D of managed data processing environment 142, which constitutes a distributed ecosystem comprising a plurality of host systems, "Application 105 implements an embodiment described herein. Provisioning system 107 provisions layers for deploying containers on one or more hosts in managed data processing environment 142, such as hosts 144A, 144B, 144C, and 144D"):
determining a probability of an image layer executing on a host system of the plurality of host systems; and (Dettori: ¶[0038]…Dettori's method predicts, for each host of the environment, which container image layers are likely to be needed at that host, which constitutes determining a probability of an image layer executing on a host system under BRI, "An embodiment predicts which layers are likely to be needed at which host in the data processing environment"; ¶[0046]…that prediction is computed from a quantified level-of-participation factor and is therefore a measure of how likely the layer is to be needed at that host, "An embodiment uses the level of participation of a layer as a factor in predicting whether the layer is likely to be needed by a host to configure a container during a period").
causing the image layer to be pre-loaded on the host system… (Dettori: ¶[0019]…Dettori's method causes the predicted subset of layers to be placed on the node before the future period in which they will be needed there, which constitutes causing the image layer to be pre-loaded on the host system under BRI, "The embodiment causes the adjusted subset of layers to be provisioned on the node prior to the future time").
To the extent it is argued that Dettori's prediction does not expressly constitute a probability, Scrivano expressly teaches determining a probability of an image layer executing on a host system of the plurality of host systems (Scrivano: 5:11-13…Scrivano computes for each unit of container image content a probability value expressing how likely that content is to be used by a container running on the host system, "The value generation component 164 may then generate probability values for each of the files of the container image").
Dettori does not expressly disclose, but Scrivano does teach:
…if the probability of the image layer executing on the host system satisfies a threshold probability. (Scrivano: 5:14-23…Scrivano compares those probability values against an identified threshold and requests only the content whose probability value satisfies it, which constitutes the satisfies a threshold probability condition under BRI, "The file requesting component 162 may identify a threshold value associated with the probability values determined by the value generation component 164 that is used to determine which files are to be pulled from a container image of the file system 130. The file requesting component 162 may generate a request for files of the container image that have a probability value that satisfies the threshold and provide the request to the file system 130"; 8:22-25…Scrivano defines satisfaction of the threshold as the score being greater than or equal to the threshold value, "Processing logic of the file system 130 may read the corresponding metadata 402a-d for each of the files to determine which of the files have access scores that satisfy the threshold 529. In embodiments, the access score may satisfy the threshold 529 if the access score is greater than or equal to the value of threshold 529").
Dettori and Scrivano are analogous art because they are from the same field of endeavor, specifically the predictive placement of container image content onto host systems in a container-orchestration environment. They address the same problem of reducing container start-up latency by placing on a host system, in advance, only that container image content which the host system is expected to need.
Before the effective filing date of the claimed invention, it would have been obvious to a PHOSITA to carry out Dettori's predictive layer pre-provisioning method so that the predicted need for an image layer at a host system is expressed as a probability value and the layer is pre-loaded onto that host system only when the probability value satisfies a threshold probability.
The suggestion/motivation for doing so would have been provided by Dettori itself, whose method already gates pre-provisioning on a computed prediction factor measured against a numerical threshold and therefore invites the substitution of an express probability value for that factor, "An embodiment further selects from the set of levels a subset of levels, each level in the subset of levels exceeding a threshold level of participation. The embodiment selects the subset of layers corresponding to the selected subset of levels" (Dettori: ¶[0023]). Scrivano supplies the express probabilistic form of the same threshold comparison and teaches that scoring container image content by probability lets the content placed on the host system be selected intelligently rather than statically, "By utilizing probabilistic per-file image preloading, the files of a container image may be intelligently selected for preloading for a container based on the probability that the files will be used by the container" (Scrivano: 3:30-33). Furthermore, this is the application of a known technique, probability-valued prediction compared against a threshold, to a known method ready for improvement, Dettori's prediction-factor-and-threshold layer pre-provisioning method, to yield the predictable result of pre-loading exactly the layers the host is most likely to need, the rationale of MPEP § 2143(D).
Claims 9-12 are substantially similar in scope and spirit as claims 3-6. Therefore the rejections of claims 3-6 are applied accordingly.
Per claim 13, Dettori discloses An article of manufacture comprising: a non-transitory computer-readable medium; and computer-executable instructions carried on the computer-readable medium, the instructions readable by a processor, the instructions, when read and executed, for causing the processor to, in a distributed ecosystem comprising a plurality of host systems (Dettori: ¶[0116]…Dettori's computer readable storage medium bearing the stored program instructions is expressly characterized as an article of manufacture whose instructions, when executed by a processor, implement the disclosed functions, which constitutes the recited article of manufacture and its carried computer-executable instructions under BRI, "the computer readable storage medium having instructions stored therein comprises an article of manufacture including instructions which implement aspects of the function/act specified in the flowchart and/or block diagram block or blocks"; ¶[0112]…Dettori excludes transitory signals from that medium, which constitutes a non-transitory computer-readable medium, "A computer readable storage medium, as used herein, is not to be construed as being transitory signals per se, such as radio waves or other freely propagating electromagnetic waves"; ¶[0068]…the instructions so carried operate across the several hosts 144A-144D of managed data processing environment 142, which constitutes a distributed ecosystem comprising a plurality of host systems, "Application 105 implements an embodiment described herein. Provisioning system 107 provisions layers for deploying containers on one or more hosts in managed data processing environment 142, such as hosts 144A, 144B, 144C, and 144D"):
determine a probability of an image layer executing on a host system of the plurality of host systems; and (Dettori: ¶[0038]…the instructions predict, for each host of the environment, which container image layers are likely to be needed at that host, which constitutes determining a probability of an image layer executing on a host system under BRI, "An embodiment predicts which layers are likely to be needed at which host in the data processing environment"; ¶[0046]…that prediction is computed from a quantified level-of-participation factor and is therefore a measure of how likely the layer is to be needed at that host, "An embodiment uses the level of participation of a layer as a factor in predicting whether the layer is likely to be needed by a host to configure a container during a period").
cause the image layer to be pre-loaded on the host system… (Dettori: ¶[0019]…the instructions cause the predicted subset of container image layers to be placed on the node in advance of the future period in which they will be needed there, which constitutes causing the image layer to be pre-loaded on the host system under BRI, "The embodiment causes the adjusted subset of layers to be provisioned on the node prior to the future time"; ¶[0023]…that advance placement is gated on the layer's computed level exceeding a threshold, "An embodiment further selects from the set of levels a subset of levels, each level in the subset of levels exceeding a threshold level of participation. The embodiment selects the subset of layers corresponding to the selected subset of levels").
To the extent it is argued that Dettori's prediction does not expressly constitute a probability, Scrivano expressly teaches determine a probability of an image layer executing on a host system of the plurality of host systems (Scrivano: 5:11-13…Scrivano computes for each unit of container image content a probability value expressing how likely that content is to be used by a container running on the host system, "The value generation component 164 may then generate probability values for each of the files of the container image").
Dettori does not expressly disclose, but Scrivano does teach:
…if the probability of the image layer executing on the host system satisfies a threshold probability. (Scrivano: 5:14-23…Scrivano compares those probability values against an identified threshold and requests only the content whose probability value satisfies it, which constitutes the satisfies a threshold probability condition under BRI, "The file requesting component 162 may identify a threshold value associated with the probability values determined by the value generation component 164 that is used to determine which files are to be pulled from a container image of the file system 130. The file requesting component 162 may generate a request for files of the container image that have a probability value that satisfies the threshold and provide the request to the file system 130"; 8:30-35…only the content whose probability value meets or exceeds the threshold is in fact delivered to and stored on the host system, "Because FILE A, FILE B, and FILE C have corresponding access scores that are greater than or equal to the threshold 529, file system 130 may provide FILE A, FILE B, and FILE C to host system 110. Upon receipt of the files, host system 110 may store the files in locally in memory 170 that is accessibly by container 527").
Dettori and Scrivano are analogous art because they are from the same field of endeavor, specifically the predictive placement of container image content onto host systems in a container-orchestration environment. They address the same problem of reducing container start-up latency by placing on a host system, in advance, only that container image content which the host system is expected to need.
Before the effective filing date of the claimed invention, it would have been obvious to a PHOSITA to embody Dettori's predictive layer pre-provisioning on the non-transitory computer-readable medium Dettori already describes so that the predicted need for an image layer at a host system is expressed as a probability value and the layer is pre-loaded onto that host system only when the probability value satisfies a threshold probability.
The suggestion/motivation for doing so would have been provided by Dettori itself, which already gates pre-provisioning on a computed prediction factor measured against a numerical threshold and therefore invites the substitution of an express probability value for that factor, "An embodiment further selects from the set of levels a subset of levels, each level in the subset of levels exceeding a threshold level of participation. The embodiment selects the subset of layers corresponding to the selected subset of levels" (Dettori: ¶[0023]). Scrivano supplies the express probabilistic form of the same threshold comparison and teaches that scoring container image content by probability lets the content placed on the host system be selected intelligently rather than statically, "By utilizing probabilistic per-file image preloading, the files of a container image may be intelligently selected for preloading for a container based on the probability that the files will be used by the container" (Scrivano: 3:30-33). Furthermore, this is the application of a known technique, probability-valued prediction compared against a threshold, to a known product ready for improvement, Dettori's program product for prediction-factor-and-threshold layer pre-provisioning, to yield the predictable result of pre-loading exactly the layers the host is most likely to need, the rationale of MPEP § 2143(D).
Claims 15-18 are substantially similar in scope and spirit as claims 3-6. Therefore the rejections of claims 3-6 are applied accordingly.
Claims 2, 8 and 14 are rejected under 35 USC 103 as being unpatentable over Dettori in view of Scrivano, as applied in the rejection of claims 1, 7 and 13 above, and further in view of US Pat. Pub. No. 2014/0136792 to Frachtenberg (hereinafter Frachtenberg).
Per claim 2, Dettori combined with Scrivano discloses claim 1, and Scrivano further discloses …the probability of the image layer executing on the host system fails to satisfy the threshold probability (Scrivano: 8:22-25…Scrivano defines satisfaction of the threshold as the probability value being greater than or equal to the threshold, so a layer whose probability falls below that value necessarily fails to satisfy the threshold probability, which constitutes the recited fails to satisfy condition under BRI, "Processing logic of the file system 130 may read the corresponding metadata 402a-d for each of the files to determine which of the files have access scores that satisfy the threshold 529. In embodiments, the access score may satisfy the threshold 529 if the access score is greater than or equal to the value of threshold 529"; 8:54-57…Scrivano further raises the threshold as the host system's resources are consumed, so content already resident on the host can cease to satisfy the threshold it previously met, "In some embodiments, processing logic of host system 110 may dynamically adjust the threshold 602 based on the resources of host system 110").
Dettori combined with Scrivano does not expressly disclose, but with Frachtenberg does teach:
the predictive orchestrator further configured to cause the image layer to be removed from the host system if the layer is already pre-loaded on the host system… (Frachtenberg: ¶[0010]…Frachtenberg's eviction module removes from a host's local cache content that is already resident there on the basis of its predicted probability of future access, which constitutes causing the image layer to be removed from the host system where the layer is already pre-loaded on it under BRI, "The eviction module can be configured to evict data from the cache which has the lowest probability of future access as predicted by the statistically aggregated prediction model"; ¶[0007]…Frachtenberg makes that removal decision by identifying the resident content whose probability of future access has fallen relative to the content retained, "By identifying the probability of access on the future data as being lower than any other data based on the set of access patterns, a determination can be made as to which data should be evicted from the cache").
Dettori, Scrivano and Frachtenberg are analogous art. Dettori and Scrivano are from the same field of endeavor, specifically the predictive placement of container image content onto host systems in a container-orchestration environment. Frachtenberg is reasonably pertinent to the particular problem with which the claimed invention is concerned, namely deciding on the basis of a predicted likelihood of future need which content to retain in, and which content to remove from, the limited local storage of a host system.
Before the effective filing date of the claimed invention, it would have been obvious to a PHOSITA to apply Frachtenberg's probability-driven removal of already-resident content to the image layers that Dettori combined with Scrivano has pre-loaded onto a host system, so that a layer already pre-loaded on the host system is removed from it when the probability of that layer executing on the host system fails to satisfy the threshold probability.
The suggestion/motivation for doing so would have been provided by Dettori itself, which teaches that the layer set predicted for a host is to be trimmed of layers the host will not in fact need, "The embodiment changes the prediction for the host for the period by removing from provisioning a layer that may not be needed for the planned workload" (Dettori: ¶[0052]), and by Scrivano, which teaches that the storage and other resources of a host system are finite and that the governing threshold is to be raised as those resources are consumed, "In some embodiments, processing logic of host system 110 may dynamically adjust the threshold 602 based on the resources of host system 110" (Scrivano: 8:54-57) - a teaching that necessarily leaves previously-placed content below the governing threshold and therefore supplies the occasion for its removal. Frachtenberg supplies the reason to act on that occasion by removal rather than by mere non-placement, explaining that retention policies which ignore predicted future need evict content that is in fact more likely to be accessed, "While there have been many efforts to come up with better ways to determine which items to store in the cache and which to evict to make room for more likely items, the traditional caching policies have been typically based on coarse locality metrics (e.g., oldest items are always evicted). Consequently, data which is more likely to be accessed may be evicted" (Frachtenberg: ¶[0004]). Furthermore, this is the application of a known technique, probability-driven eviction of already-cached content, to a known device ready for improvement, the host system onto which Dettori and Scrivano pre-load container image layers, to yield the predictable result of reclaiming host storage occupied by layers no longer likely to be needed, the rationale of MPEP § 2143(D).
Claims 8 and 14 are substantially similar in scope and spirit as claim 2. Therefore the rejection of claim 2 is applied accordingly.
Conclusion
Any inquiry concerning this communication or earlier communications from the examiner should be directed to ALAN CHEN whose telephone number is (571) 272-4143. The examiner can normally be reached M-F 10-7.
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, Kamran Afshar can be reached at (571) 272-7796. 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.
/ALAN CHEN/ Primary Examiner, Art Unit 2125
Prosecution Insights