DETAILED ACTION
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 office correspondence is in response to “Amendment and Response under 37 C.F.R. 1.111 filed on June 22, 2026 in response to a non-final office action issued on April 1, 2026,
Claims 1 – 20 are pending.
Claims 1 – 2, 4 – 6, 11, 14, and 15 are amended,
Claims 16 – 20 are added.
Claims 1 – 20 are rejected.
Applicant’s arguments filed on 6/22/2026 have been fully considered:
In regard to claims 1 – 15, which were rejected under 35 U.S.C. 103, at least one argument is persuasive to the rejection of claims from the last office action and said rejections are withdrawn, but applicant’s amendment necessitated a new search and consideration resulting in a new grounds of rejections for claims 1 – 20 under 35 U.S.C. 103. The examiner here now responds to each argument. Underlined text indicates claim language that was amended since the last office action.
In regard to claims 1 – 6 and 14 – 15, the applicant argues that the prior art combination of Sarabi, Matefi, and Gupta fails to anticipate, disclose or teach:
“ obtaining routing information of a device, wherein the routing information is obtained based on one or more probe packets sent by the device to a server that is connectable to the device via the Internet, whereby a series of packet hops was implemented to route the one or more probe packets to the server, the routing information includes a series of Internet Protocol (IP) addresses of the series of packet hops along a path to the server;” (as recited in claim 1 and substantially replicated on claims 14 and 15)
The applicant states:
“ . . . Applicant respectfully submits that the cited art does not teach or suggest this feature. Specifically, the cited art does not teach or suggest the recited routing information, which includes IP addresses of series of packet hops that were implemented to route one or more probe packets to a server. The Office Action relied on Paragraphs [0021], [0022], and [0024]-[0027] of Sarabi for allegedly teaching the recited routing information (Pages 4-6 of the Office Action). Conversely, in the sections of Sarabi cited by the Office, Sarabi merely describes performing scans of a computer network by sending probe packets to a target IP address of a host in order to determine whether the host is responsive. Specifically, each probe packet is sent to the IP address if the predicted probability of obtaining a response to the probe exceeds a threshold ([0026]-[0027] of Sarabi). Applicant asserts that this description of Sarabi fails to relate to any series of packet hops that route the one or more probe packets to the server. Conversely, Sarabi merely describes sending probe packets to a target IP address of a host (Paragraphs [0022], and [0024]-[0027]), without any disclosure regarding packet hops or routing paths of such probe packets. More specifically, since Sarabi fails to relate to any packet hops, Sarabi inherently fails to teach or suggest obtaining routing information that includes IP addresses of the series of packet hops used to route the probe packets to the server. Accordingly, Sarabi fails to teach the recited routing information. . . .” (Applicant’s remarks page 9 – 10)
In response to the applicant’s argument:
The applicant’s argument is persuasive that the cited prior art fails to teach the amended limitation. The applicant amendment triggered a new search and consideration which found a new grounds of rejection using prior art citation Dave et al. (U.S. 2022/0166722 A1; herein referred to as Dave) to teach the limitation as recited. Dave teaches methods and systems for managing a plurality of network devices, a network device includes processing circuitry configured to: receive packet data corresponding to a network flow originating at a first device, the packet data destined to a second device; generate an entropy label to add to a label stack of the packet data, wherein the entropy label is generated from one or more attributes corresponding to the network flow as measured by packet hops that originated at the first device and is destined to the second device; generate a flow record including the entropy label, wherein the entropy label identifies the network flow amongst a plurality of network flows in the network; and send, to a controller of the network, the flow record, wherein the controller identifies the flow record based on the entropy label corresponding to the network flow originating at the first device and is destined to the second device (see Dave abstract, Fig. 1, ¶ [0022]). Routing information is obtained using end-to-end flow monitoring on a label switched path (LSP) which comprises a series of packet hops. (see Dave ¶ [0021] ).
In regard to claims 1 – 6 and 14 – 15, the applicant argues that the prior art combination of Sarabi, Matefi, and Gupta fails to anticipate, disclose or teach:
“creating, based on the routing information, a fingerprint describing an architecture of the server, the path comprising at least the series of packet hops;” (as recited in claim 1 and substantially replicated on claims 14 and 15).
The applicant states:
“ . . . Applicant respectfully submits that the cited art does not teach or suggest the recited fingerprint. Specifically, the Office Action relied on Paragraphs [0011], and [0028]-[0029] of Sarabi for allegedly teaching the recited fingerprint creation (Pages 6-7 of the Office Action). Conversely, Paragraph [0011] of Sarabi merely describes training a model to predict whether a network address will respond to a probe, based on a data set. Paragraphs [0028]-[0029] of Sarabi describe that the predicted probability of obtaining a response to a probe can take into account features obtained from completed probes, to boost the model's performance.
Applicant asserts that nothing in this description of Sarabi relates to creation of a fingerprint. In this regard, the Office indicated that the data set of Sarabi can allegedly be interpreted as the recited fingerprint (Page 6 of the Office).
Applicant respectfully traverses this contention. Specifically, Paragraph [0011] of Sarabi describes that the data set includes entries, each of which includes a network address to which a probe was sent, and an indicator of the response to the given probe. A feature vector is then constructed for each data set entry.
Applicant asserts that neither the data set entries of Sarabi, nor the respective feature vectors, are created based on the routing information. Specifically, neither the data set entries of Sarabi, nor the respective feature vectors, are created based on a series of IP addresses of a series of packet hops. As discussed above, no packet hops are mentioned by Sarabi. Additionally, neither the data set entries of Sarabi, nor the respective feature vectors, describe an architecture of the path of the device to the server. At most, the data set entries of Sarabi relate to a target network address and whether it was responsive. Applicant asserts that this description of Sarabi with respect to a data set with network addresses and responsiveness indicators is fundamentally different from, and unrelated to, the recited fingerprint, which is created "based on the routing information" and which describes "an architecture of the path of the device to the server, the path comprising at least the series of packet hops". Accordingly, Sarabi fails to teach or suggest the recited fingerprint. Applicant asserts that Matefi and Gupta cannot cure at least this deficiency of Sarabi. . . .” (Applicant’s remarks pages 10 – 11).
In response to the applicant’s argument:
The applicant’s argument is persuasive that the cited prior art fails to teach the amended limitation. The applicant amendment triggered a new search and consideration which found a new grounds of rejection using prior art citation Dave et al. (U.S. 2022/0166722 A1; herein referred to as Dave) to teach the limitation as recited. Dave further teaches the creation of a flow record (which maps to the “fingerprint” of the claimed invention) when packet and next-hop data is received (see Dave ¶ [0007], ¶ [0027]).
In regard to claims 1 – 6 and 14 – 15, the applicant argues that the prior art combination of Sarabi, Matefi, and Gupta fails to anticipate, disclose or teach:
“utilizing a prediction model to determine a label for the fingerprint, wherein the label is indicative of metadata of a user of the device, wherein the prediction model is trained using a training dataset that includes pairs of fingerprints and labels using edge devices having known labels, the fingerprints of the training dataset are indicative of-a routing information of an edge device to the Internet” (as recited in claim 1 and substantially replicated on claims 14 and 15).
The applicant states:
“ . . . Applicant respectfully submits that the cited art does not teach or suggest the recited label. The Office acknowledged that Sarabi fails to teach the feature "utilizing a prediction model to determine a label for the fingerprint", and relies instead on Paragraphs [0011] and [0028] of Matefi
Conversely, Paragraphs [0011] and [0028] of Matefi describe training prediction models to determine a benign or malicious label for a connection. Applicant asserts that the benign/malicious label of Matefi cannot be interpreted as the recited
label, at least since such an interpretation contradicts multiple claim features.
Firstly, in contrast to Claim 1, the label of Matefi is not indicative of metadata of a user of the device. Conversely, the label of Matefi merely indicates whether a connection between devices is benign or malicious. For example, as recited in Claim 6, the metadata of the user can include user information such as a workplace of the user, demographic information, the user's income level, or the like. In contrast to Claim 1, no user metadata is inferred in the cited portions of Matefi.
Secondly, in contrast to Claim 1, the label of Matefi is not assigned to the fingerprint. Instead, the label of Matefi is assigned to a connection,
The connection of Matefi cannot be interpreted as the recited fingerprint, at least since the connection of Matefi does not constitute a fingerprint created based on the routing information, e.g., based on the series of IP addresses of the packet hops. Conversely, in the cited portions of Matefi, the connection is based at most on source information, destination information, protocol, and destination port. Neither of these data items
corresponds to IP addresses of packet hops that route probe packets to a server.
More specifically, even if the connection of Matefi would allegedly include IP addresses of packet hops (a contention that Applicant respectfully traverses), Matefi would still fail to teach that a fingerprint is created based on such connection data, or that a label is determined for such a fingerprint. Applicant notes, in this regard, that Claim 1 recites: "determine a label for the fingerprint", rather than for a fingerprint in general. Thus, in order to establish a prima facie case of obviousness under 35 U.S.C. § 103, the Office must show that the cited prior art teaches or suggests each and every limitation of the claimed invention, including determining a label for the specific fingerprint that is based on the recited routing information.
Applicant asserts that Gupta cannot cure at least this deficiency of Sarabi and Matefi.
Specifically, in the portions of Gupta cited by the Office, Gupta relates to a system for
automatically generating test scripts for testing a process. Specifically, Gupta describes that the system can be updated to adaptively integrate new attributes, validation rules, and operational processes when user interface labels fail to map to an existing database (Paragraph [0042] of Gupta). The updates are performed manually or automatically via a machine learning module.
Firstly, Applicant asserts that the reliance on Gupta for the current rejection is improper, at least since the invention of Gupta is not analogous art to the claimed invention (e.g., see MPEP § 2141.01(a)(I)).
Secondly, the user interface labels of Gupta cannot be interpreted as the recited label, at least since they are not indicative of metadata of a user. Specifically, as is well known in the art, a user interface of a process being tested is not indicative in any way of user metadata such as a user's workplace (e.g., see Claim 6).
Thirdly, the user interface labels of Gupta cannot be interpreted as the recited label, at least since they are not assigned to any fingerprint. As a result, the cited art, alone or in combination, fails to teach or suggest, let alone disclose "utilizing a prediction model to determine a label for the fingerprint, wherein the label is indicative
of metadata of a user of the device", as is explicitly claimed. . . . The Office acknowledged that Sarabi fails to teach this feature, and relies instead on Paragraph
[0043] of Matefi (Page 7 of the Office). Conversely, Paragraph [0043] of Matefi merely describes, with respect to Figure 1, how element 103 records data flow records associated with devices 104a-104b in a data flow dataset, and exports the data flow records. Nothing in this description relates fingerprints generated for the data records, labels of such fingerprints, or to any training dataset used to train a prediction model.
As a result, the cited art, alone or in combination, fails to teach or suggest, let alone disclose "wherein the prediction model is trained using a training dataset that includes pairs of fingerprints and labels using edge devices having known labels, the fingerprints of the training dataset are indicative of routing information of an edge device to the Internet", as is explicitly claimed. . . .”( Applicant’s argument s page 11 -14)
In response to the applicant’s argument:
The applicant’s argument is persuasive that the cited prior art fails to teach the amended limitation. The applicant amendment triggered a new search and consideration which found a new grounds of rejection using prior art citation Dave et al. (U.S. 2022/0166722 A1; herein referred to as Dave) in view of Gentleman et al. (U.S. 2022/0207163 A1; herein referred to as Gentleman) in further view of Farooq et al. (U.S. 2023/0162089 A1; herein referred to as Farooq) to teach the limitation as recited. The cited prior art Gentleman teaches methods and systems for providing for the dynamic data classification of data objects. Examples enable prediction of candidate data classification labels for data objects associated with one or more applications, services, or computing devices. Examples enable the assignment of one or more data classification labels to a data object for transmission to one or more computing devices. Examples enable the interactive and progressive application of machine learning techniques to data classification systems to assign data classification labels with probable certainty. In combination with Dave, Gentleman applies a classification model to generate classification label predictions (see Gentleman Fig. 1 ¶ [0181]) for routing information (see Gentleman ¶ [0084], ¶ [0151], and the label is associated with metadata associated with user data (see Gentleman ¶ [0244]) and creating a training data set (Gentleman ¶¶ [0251-0252] ) The cited prior art Farooq teaches methods and systems for edge device discovery by training a edge machine learning model to create datasets (Farooq ¶ [0037]) with labels for the devices (Farooq ¶ [0047]) and stored routing information for the devices (see Fig. 7, ¶ [0067]).
In regard to claims 1 – 6 and 14 – 15, the applicant argues that the prior art combination of Sarabi, Matefi, and Gupta fails to anticipate, disclose or teach:
“wherein the fingerprint comprises IP addresses associated with N consecutive packet hops from the device in accordance with the routing information.” (as recited in claim 3 and substantially replicated in new claim 17)
The applicant states:
“ . . . The Office relies on Paragraph [0043] of Matefi for allegedly teaching the fingerprint comprising IP addresses associated with N consecutive packet hops from the device (Pages 11-12 of the Office). Conversely, Paragraph [0043] of Matefi merely describes data flow records that log a single data point representing the IP address of the immediate next-hop. This immediate next-hop attribute identifies, at most, a single next hop from a given communication device, and does not disclose consecutive series of packet hops, as recited by Claim 3. Additionally, Matefi does not generate a fingerprint based on this immediate next-hop attribute of the data flow records. Instead, Matefi at most determines a label for a connection strictly defined by source information, destination information, protocol, and destination ports. . . . “ (Applicant’s remarks page 14)
In response to the applicant’s argument:
The applicant’s argument is persuasive that the cited prior art fails to teach the amended limitation. However the search and consideration found a new grounds of rejection using prior art citation Dave et al. (U.S. 2022/0166722 A1; herein referred to as Dave) in view of Gentleman et al. (U.S. 2022/0207163 A1; herein referred to as Gentleman) in further view of Farooq et al. (U.S. 2023/0162089 A1; herein referred to as Farooq) to teach the limitation as recited, as Dave generates flow records with next hop information (see Dave ¶ [0022], ¶ [0027], ¶ [0045]).
In regard to claims 1 – 6 and 14 – 15, the applicant argues that the prior art combination of Sarabi, Matefi, and Gupta fails to anticipate, disclose or teach:
“ wherein the metadata comprises at least one of: a workplace identity, demographic information, a geographic proximity, an income level, an education level, or user interests
The applicant states:
“ . . . Applicant asserts that Sarabi, Matefi and Gupta, alone or in combination, fail to teach or disclose determining a label for a fingerprint that indicates a workplace identity of a user, his demographic information. his geographic proximity, his income level, his education level, or his interests. . . . “ (Applicant’s remarks page 15)
In response to the applicant’s argument:
The applicant’s argument is persuasive that the cited prior art fails to teach the claim limitation. However the search and consideration found a new grounds of rejection using prior art citation Dave et al. (U.S. 2022/0166722 A1; herein referred to as Dave) in view of Gentleman et al. (U.S. 2022/0207163 A1; herein referred to as Gentleman) in further view of Farooq et al. (U.S. 2023/0162089 A1; herein referred to as Farooq) in further view of Alexander et al. (U.S. 2019/0045241 A1; herein referred to as Alexander) to teach the limitation as recited. The prior art Alexander teaches systems and methods for providing content through a terrestrial fiber network wherein a replicated content stream contains multiple packets, each of which may have a header populated by the network device with a routing label corresponding to the path the packet is to follow. Alexander details metadata from the label being used to enhance the transmission of the packets across the system (see Alexander ¶ [0011], ¶ [0031], ¶ [0044]).
Therein while the 35 USC 103 rejections are withdrawn, the search and consideration performed found new grounds of rejection under 35 USC 103:
Claims 1 – 5 and 14 – 19 are rejected under 35 U.S.C. 103 as being un-patentable over Dave et al. (U.S. 2022/0166722 A1; herein referred to as Dave) in view of Gentleman et al. (U.S. 2022/0207163 A1; herein referred to as Gentleman) in further view of Farooq et al. (U.S. 2023/0162089 A1; herein referred to as Farooq).
Claims 6 and 20 are rejected under 35 U.S.C. 103 as being un-patentable over Dave et al. (U.S. 2022/0166722 A1; herein referred to as Dave) in view of Gentleman et al. (U.S. 2022/0207163 A1; herein referred to as Gentleman) in further view of Farooq et al. (U.S. 2023/0162089 A1; herein referred to as Farooq) as applied to claims 1 – 5 and 14 – 19 in further view of Alexander et al. (U.S. 2019/0045241 A1; herein referred to as Alexander).
Claims 7, 8, and 13 are rejected under 35 U.S.C. 103 as being un-patentable over Dave et al. (U.S. 2022/0166722 A1; herein referred to as Dave) in view of Gentleman et al. (U.S. 2022/0207163 A1; herein referred to as Gentleman) in further view of Farooq et al. (U.S. 2023/0162089 A1; herein referred to as Farooq) as applied to claims 1 – 5 and 14 – 19 in further view of Wang et al. (U.S. 2024/0356817 A1; herein referred to as Wang).
Claims 9 – 10 are rejected under 35 U.S.C. 103 as being un-patentable over Dave et al. (U.S. 2022/0166722 A1; herein referred to as Dave) in view of Gentleman et al. (U.S. 2022/0207163 A1; herein referred to as Gentleman) in further view of Farooq et al. (U.S. 2023/0162089 A1; herein referred to as Farooq) as applied to claims 1 – 5 and 14 – 19 in further view of McCourt Jr. (U.S. 2020/0019884 A1; herein referred to as McCourt).
Claims 11 – 12 are rejected under 35 U.S.C. 103 as being un-patentable over Dave et al. (U.S. 2022/0166722 A1; herein referred to as Dave) in view of Gentleman et al. (U.S. 2022/0207163 A1; herein referred to as Gentleman) in further view of Farooq et al. (U.S. 2023/0162089 A1; herein referred to as Farooq) as applied to claims 1 – 5 and 14 – 19 in further view of Alanazi (U.S. 2021/0152523 A1; herein referred to as Alanazi)
The new rejections are described below
The examiner recommends that the applicant review the specification for disclosure that if integrated into the independent claims would distinguish the amended claims from the cited prior art. The applicant is invited to contact the examiner for an interview to discuss how to move the prosecution forward.
Authorization for Internet Communications
The examiner encourages Applicant to submit an authorization to communicate with the examiner via the Internet by making the following statement (from MPEP 502.03):
“Recognizing that Internet communications are not secure, I hereby authorize the USPTO to communicate with the undersigned and practitioners in accordance with 37 CFR 1.33 and 37 CFR 1.34 concerning any subject matter of this application by video conferencing, instant messaging, or electronic mail. I understand that a copy of these communications will be made of record in the application file.”
Please note that the above statement can only be submitted via Central Fax (not Examiner's Fax), Regular postal mail, or EFS Web using PTO/SB/439.
Priority
This Application is the National Stage filing under 35 U.S.C. § 371 of PCT Application Ser. No. PCT/IL2023/050572 filed on June 4, 2023 which claims the benefit of U.S. Provisional Application No. 63/365888 filed on June 6, 2022 entitled “Client-Based Network Fingerprinting.” The applicant is entitled to a priority date of 6/6/2022.
35 USC § 101 Analysis – Judicial Exception
35 U.S.C. 101 reads as follows:
Whoever invents or discovers any new and useful process, machine, manufacture, or composition of matter, or any new and useful improvement thereof, may obtain a patent therefore, subject to the conditions and requirements of this title.
The claimed invention is directed to statutory subject matter and are not rejected under 35 USC 101 because of a judicial exception.. The claimed subject matter is integrated into a practical application under prong 2 of the Step 2A analysis as documented in MPEP 2016.04(d). The claims are directed to non-abstract improvements in computer related technology. A claim is non-statutory when it is directed to a judicial exception (e.g. either one of mathematical concepts, mental processes, or certain methods of organizing human activity) without significantly more. The claimed invention is not directed to a judicial exception. Instead, the claimed invention is directed to a technological improvement to machine learning based on client based fingerprinting wherein the claimed invention teaches an algorithm comprising: obtaining routing information of a device, wherein the routing information is obtained based on one or more probe packets sent by the device to a server that is connectable to the device via the Internet, whereby a series of packet hops was implemented to route the one or more probe packets to the server, the routing information includes a series of Internet Protocol (IP) addresses of the series of packet hops until reaching the Internet, and therein creating, based on the routing information, a fingerprint describing an architecture of connection path of the device to the Internet, then further utilizing a prediction model to determine a label for the fingerprint, wherein the label is indicative of metadata of a user of the device, wherein the prediction model is trained using training dataset that includes pairs of fingerprints and labels using edge devices having known labels, the fingerprints of the training dataset are indicative of a routing information of an edge device to the Internet. The ordered steps of the claim language impose meaningful limits on the scope of the claims, and provides a useful improvement for creating fingerprints of a device’s connection path to the internet, based on the routing information obtained from probe packets sent by the device, without requiring a geographical location for the device, in so preserving the privacy of the data of the device and users thereof, while using device routing information to predict metadata of a user of the device, both in collecting training data, and while applying the prediction model. Therein the claimed invention is statutory under 35 USC 101.
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 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 of this title, 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 – 5 and 14 -19 are rejected under 35 U.S.C. 103 as being un-patentable over Dave et al. (U.S. 2022/0166722 A1; herein referred to as Dave) in view of Gentleman et al. (U.S. 2022/0207163 A1; herein referred to as Gentleman) in further view of Farooq et al. (U.S. 2023/0162089 A1; herein referred to as Farooq).
In regard to claim 1, Dave teaches A method (see abstract “ . . . in a network comprising a plurality of network devices, a network device includes processing circuitry configured to: receive packet data corresponding to a network flow originating at a first device, the packet data destined to a second device; generate an entropy label to add to a label stack of the packet data, wherein the entropy label is generated from one or more attributes corresponding to the network flow that originated at the first device and is destined to the second device; generate a flow record including the entropy label, wherein the entropy label identifies the network flow amongst a plurality of network flows in the network; and send, to a controller of the network, the flow record, wherein the controller identifies the flow record based on the entropy label corresponding to the network flow originating at the first device and is destined to the second device. . .”) comprising:
obtaining routing information of a device (see ¶ [0003]” . . . Certain devices (i.e., nodes), such as routers, maintain routing information that describes routes through the network. A “route” can generally be defined as a path between two locations on the network. Upon receiving an incoming packet, the router examines keying information within the packet to identify the destination for the packet. Based on the destination, the router forwards the packet in accordance with the routing information. . . “) , wherein the routing information is obtained based on one or more probe packets sent hy the device to a server that is connectable to the device via the Internet (see Fig. 1 ¶ [0021]” . . . FIG. 1 is a block diagram illustrating an example system 2 in which client device 12 sends data to server device 14 with end-to-end flow monitoring on a label-switched path (LSP) 16. LSP 16 (which may be herein referred to a tunnel or LSP tunnel) is formed by routers 20-32, in this example, although it should be understood that fewer or additional network devices (not shown in FIG. 1) may form LSP 16. In this example, routers 20-32 are label switching routers (LSRs) along LSP 16. Routers 20-32 execute protocols for establishing paths, such as LSP 16 . . .”; see ¶ [0029]” . . . Some conventional techniques implement network probing where the controller may send probes to determine issues within routers in the MPLS network but are unable to determine which application is affected in performance. Even when the probes clone application packet headers, the probes cannot identify which flows correspond to a particular application. . . “) , whereby a series if packet hops was implemented to route the one or more probe packets to the server (see ¶ [0022]” . . . In general, packets directed along LSP 16 are encapsulated by one or more labels in a label stack, e.g., in accordance with Multi-Protocol Label Switching (MPLS). In this manner, to direct traffic along an LSP, a router along the LSP may use a label of the label stack of a received packet to determine the “next hop” for the packet, that is, the next network device to which to forward the packet. In addition, the router may pop one or more labels from the label stack and push a new label onto the label stack, which causes the next hop to direct the packet along the LSP. . . “) , the routing Information includes a series of Internet Protocol (IP) addresses of the series of packet hops (e.g. LSP) along a path to the server (see ¶ [0004]” . . . Packet-based networks increasingly utilize label switching protocols for traffic engineering and other purposes. Multi-protocol label switching (MPLS) is a mechanism used to engineer traffic patterns within Internet protocol (IP) networks according to the routing information maintained by the routers in the networks. By utilizing MPLS protocols, label switching routers (LSRs) can forward traffic along a particular path, i.e., a label switched path (LSP), through a network to a destination device using labels prepended to the traffic. An LSP defines a distinct path through the network to carry MPLS packets from the source device to a destination device. . . .”);
creating, based on the routing information, a fingerprint (e.g. flow record) describing an architecture of the path of the device to the server (see ¶ [0007]” . . . Some techniques described herein are implemented in one or more of a number of routers that forward packet data along the computer network. Each router may be configured to repurpose entropy labels for end-to-end network flow monitoring. For example, an ingress router may export key information for computing the entropy label, which enables the network controller to distinguish between network flows. When packet data having the entropy label is received, each router in the MPLS network may generate a flow record to include the entropy label (as a primary key) and any useful information associated with the flow. By doing so, each router may be triggered by the entropy label into exporting, to the network controller, their flow record using the entropy label as the record's index key. The network controller may collect (e.g., and coalesce) the flow records from each router and continue to use the same entropy label as the key for those flow records. The network controller may use the flow record associated with a given entropy label to detect failures in the network flow. For example, if the network flow is down, the network controller may identify the routers that previously exported the flow record having the entropy label and identify which of those routers did not currently export the flow record. . . .”) , the path comprising at least the series of packet hops (see ¶ [0027]” . . . Conventional networks may define templates specifying an exporting format for different routers to use when exporting flow records to a network controller. When such a network implements passive flow monitoring and/or active flow monitoring, a router may monitor the traffic flow and export flow information in a format defined by the template to the network controller. For example, the router may typically generate a flow record having information about the following fields: source and destination IP address; total number of bytes and packets sent; start and end times of the data flow; source and destination port numbers; TCP flags; IP protocol and IP type of service; originating Autonomous System of source and destination address; source and destination address prefix mask lengths; next-hop router's IP address; and MPLS label(s) . . “);
Dave fails to explicitly teach
However Gentleman teaches
utilizing a prediction n1odeJ to determine u label for the fingerprint (see ¶ [0181]” . . . FIG. 1 illustrates an example system architecture 100 within which embodiments of the present disclosure may operate. The example system architecture 100 includes a data classification system 105 configured to interact with one or more client devices, such as client device 102A, client device 102B, and client device 102C. The data classification system 105 may be configured to receive data objects from one or more origination sources (e.g., an application, a service, etc.) or data classification label requests from one or more client devices (e.g., client device 102A, etc.) that include a data object identifier associated with a data object stored in a data object repository (e.g., data object repository 107, etc.). The data classification system 105 may process the data classification label requests to generate one or more candidate data classification label predictions, to generated one or more user interfaces (e.g., a truth interface, etc.), or to dynamically associate a data classification label with a received data object. . . .”), wherein the label is indicative of metadata of a user of the device (see ¶ [0084]” . . . Various embodiments discussed herein are directed to improved data classification systems that are configured to retrieve one or more data objects from a data object repository, parse the data objects, or at least a portion thereof (e.g., metadata, programmer comments, file descriptions, file types, API documentation, API calls, API logs, data routing information, associated data objects, or the like), into independent text based data elements. The text based data elements are converted into word based data elements using natural language processing (NLP) tools. In some embodiments, the one or more data objects may be retrieved from at least a temporary data object repository (e.g., temporary memory allocated to hold one or more data objects after generation and before transmission, etc.). Moreover, in some embodiments, a multi-layer service oriented platform is configured to provide for the processing of, for example, API call information using predefined parameter terms (e.g., names or words associated with particular data classification label types, etc.) and/or descriptions (e.g., programmer comments describing the information contained within a data object, etc.) to identify the data within a data object and to apply particular data classification labels to such data objects. In some embodiments, predefined parameter terms may be weighted based on a probability of association with a particular data classification label or a particular data classification label type. . . .” see ¶ [0244]” . . . a data classification model (e.g., as described below with respect to FIG. 10) may perform one or more operations using the metadata contained in data object 905 to apply one or more data classification labels to the comment data object (i.e., the data object that data object 905 is metadata thereto). For example, data classification server 106 may utilize a natural language processing algorithm to detect key words in data object 905. For example, the word ‘status’ in name data element 906A and description data element 908A may be associated with the “Specific/Configuration” data classification label because the status is automatically applied to a user post via the collaboration application's configuration settings. Moreover, for example, the word ‘comment’ in name data element 906B and description data element 908B may be associated with the “UGC/Label” data classification label (e.g., in a data classification repository, etc.) because the natural language processing algorithm is configured to detect the ‘comment’ metadata tag as associated with user generated content. Further, the natural language processing algorithm may detect the phrase ‘added by a user’ within description data element 908B and may further increase the probability that the associated content should be classified under the “UGC/Label” data classification label. . . .”), wherein the prediction model is trained using a training dataset (see ¶¶ [0251-0252]” . . . Example embodiments of the present disclosure provide for a data classification label accuracy score that may be determined based on an F-score measure of a plurality of data objects associated with a data classification model (e.g., generated training data sets, etc.). For example, during a binary classification analysis, the F-score may be used to measure the accuracy of a data classification model based on a plurality of data objects labeled by the data classification model. In some embodiments, the data classification model may be measured via the F-score using a precision value and a recall value associated with a plurality of data objects labeled by the data classification model. The precision value may be a ratio of true positives (t.sub.p) to all predicted positives (t.sub.p+f.sub.p), including false positives (f.sub.p). The precision value may be represented as: precision=(t.sub.p)/(t.sub.p+f.sub.p). In some embodiments, the recall value is the ratio of true positives (t.sub.p) to all actual positives (t.sub.p+f.sub.n), including false negatives (fa). The recall value may be represented as: recall=(t.sub.p)/(t.sub.p+f.sub.n). The plurality of data objects utilized by the F-score may be at least a partial sample of data objects labeled by the data classification model. The sample of data objects labeled by the data classification model may be selected based on one or more predefined criteria (e.g., data objects labeled within a particular time period, associated with a particular data classification label, associated with a particular word and/or word frequency, etc.). In some embodiments, the sample of data objects labeled by the data classification model may be selected based on one or more word distribution charts (e.g., as described in further detail below with respect to FIGS. 12-14 . . .”)).that includes pairs of fingerprints and labels using edge devices having known labels (see ¶ [0025] “ . . . an apparatus is provided for generating notifications based on data object distribution events, the apparatus comprising at least one processor and at least one memory including program code that with the at least one processor, cause the apparatus to receive an interaction input from a computing device, wherein the interaction input defines at least a data object identifier and an application programming interface pathway. In some embodiments, the at least one memory including the program code that with the at least one processor, further cause the apparatus to retrieve a labeled data object from a data object repository based on the data object identifier, wherein the labeled data object comprises the data object identifier and a data classification label set. In some embodiments, the at least one memory including the program code that with the at least one processor, further cause the apparatus to determine a target recipient identifier for the labeled data object based on the application programming interface pathway, wherein the target recipient identifier is one or more of a computing device, a service, an application, or a data object repository. In some embodiments, the at least one memory including the program code that with the at least one processor, further cause the apparatus to generate a data classification label restricted usage notification based on the data classification label set for the labeled data object and the target recipient identifier, wherein the data classification label restricted usage notification comprises the data object identifier, at least one data classification label of the data classification label set, and the target recipient identifier. . . “) ,
It would have been obvious to one skilled in the art, before the effective filing date of the applicant’s claimed invention to incorporate a method and system for providing for the dynamic data classification of data objects. Examples enable prediction of candidate data classification labels for data objects associated with one or more applications, services, or computing devices. Examples enable the assignment of one or more data classification labels to a data object for transmission to one or more computing devices. Examples enable the interactive and progressive application of machine learning techniques to data classification systems to assign data classification labels with probable certainty, as taught by Gentleman, into a method and system for managing a plurality of network devices, a network device includes processing circuitry configured to: receive packet data corresponding to a network flow originating at a first device, the packet data destined to a second device; generate an entropy label to add to a label stack of the packet data, wherein the entropy label is generated from one or more attributes corresponding to the network flow as measured by packet hops that originated at the first device and is destined to the second device; generate a flow record including the entropy label, wherein the entropy label identifies the network flow amongst a plurality of network flows in the network; and send, to a controller of the network, the flow record, wherein the controller identifies the flow record based on the entropy label corresponding to the network flow originating at the first device and is destined to the second device. Such incorporation enables machine learning classification to be used to label predictable message and content flow paths between devices.
The combination of Dave and Gentleman fails to explicitly teach,
However Farooq teaches
the fingerprints of the training dataset (see ¶ [0037]” . . . a method that is based on the notion that if training datasets distributed at different edge devices have the same distribution and the same class of ML models, then the ML models trained on those datasets should have similar optimal hyperparameter values. The embodiments include a method for extracting underlying distribution of the training data (e.g., using statistical measures, parametric, or non-parametric distribution fitting methods) to learn the underlying distribution of the training data available at the edge devices. This knowledge of underlying distribution of the training data along with the hyperparameters used and the resulting performance is sent to the cloud where this tuple (probability distribution, hyperparameter settings, and performance) is used for training a globally shared ML model (hyperparameter ML model). A heuristic search agent consisting of state-of-art heuristic search techniques polls this hyperparameter ML model and returns the optimal hyperparameter values back to the edge devices. . . .”) are indicative of a routing information of an edge device to the Internet (see ¶ [0047] “ . . . The hyperparameter server collects the data from each cluster and/or each edge device (Block 109). For example, the tuple coming from various edge devices in the network can be appended to the training database with training data distribution and hyperparameter values as input features and resulting performance as the output label i.e., [X.sub.train]=[[D.sub.data,c][p]]. The compiled dataset [X.sub.train] is 1 then used by the hyperparameter server to train a globally shared ML hyperparameter model for each cluster . . . “; (see Fig. 7D ¶ [0079] “ . . . where the special-purpose network device 702 is used, the control communication and configuration module(s) 732A-R of the ND control plane 724 typically include a reachability and forwarding information module to implement one or more routing protocols (e.g., an exterior gateway protocol such as Border Gateway Protocol (BGP), Interior Gateway Protocol(s) (IGP) (e.g., Open Shortest Path First (OSPF), Intermediate System to Intermediate System (IS-IS), Routing Information Protocol (RIP), Label Distribution Protocol (LDP), Resource Reservation Protocol (RSVP) (including RSVP-Traffic Engineering (TE): Extensions to RSVP for LSP Tunnels and Generalized Multi-Protocol Label Switching (GMPLS) Signaling RSVP-TE)) that communicate with other NEs to exchange routes, and then selects those routes based on one or more routing metrics. Thus, the NEs 770A-H (e.g., the processor(s) 712 executing the control communication and configuration module(s) 732A-R) perform their responsibility for participating in controlling how data (e.g., packets) is to be routed (e.g., the next hop for the data and the outgoing physical NI for that data) by distributively determining the reachability within the network and calculating their respective forwarding information. Routes and adjacencies are stored in one or more routing structures (e.g., Routing Information Base (RIB), Label Information Base (LIB), one or more adjacency structures) on the ND control plane 724. The ND control plane 724 programs the ND forwarding plane 726 with information (e.g., adjacency and route information) based on the routing structure(s) . . .”).
It would have been obvious to one skilled in the art, before the effective filing date of the applicant’s claimed invention to incorporate a method and system for edge device discovery by training a edge machine learning model to create datasets with labels for the devices and stored routing information for the devices, as taught by Farooq, into a method and system for managing a plurality of network devices, a network device includes processing circuitry configured to: receive packet data corresponding to a network flow originating at a first device, the packet data destined to a second device; generate an entropy label to add to a label stack of the packet data, wherein the entropy label is generated from one or more attributes corresponding to the network flow as measured by packet hops that originated at the first device and is destined to the second device; generate a flow record including the entropy label, wherein the entropy label identifies the network flow amongst a plurality of network flows in the network; and send, to a controller of the network, the flow record, wherein the controller identifies the flow record based on the entropy label corresponding to the network flow originating at the first device and is destined to the second device, for providing for the dynamic data classification of the devices, including prediction of candidate data classification labels for devices associated with one or more applications and services, or edge nodes such that the assignment of one or more data classification labels for transmission to one or more computing devices. Examples enable the interactive and progressive application of machine learning techniques to data classification systems to assign data classification labels with probable certainty, as taught by the combination of Dave and Gentleman. Such incorporation provides dataset creation for each labeled device.
In regard to claim 2, the combination of Dave, Gentleman, and Farooq teaches wherein the fingerprint comprises the series of IP addresses of the series of packet hops or an encoding thereof (see Dave ¶ [0027] “ . . . Conventional networks may define templates specifying an exporting format for different routers to use when exporting flow records to a network controller. When such a network implements passive flow monitoring and/or active flow monitoring, a router may monitor the traffic flow and export flow information in a format defined by the template to the network controller. For example, the router may typically generate a flow record having information about the following fields: source and destination IP address; total number of bytes and packets sent; start and end times of the data flow; source and destination port numbers; TCP flags; IP protocol and IP type of service; originating Autonomous System of source and destination address; source and destination address prefix mask lengths; next-hop router's IP address; and MPLS label(s) . . .”)
In regard to claim 3, the combination of Dave, Gentleman, and Farooq teaches wherein the fingerprint comprises IP addresses associated with N consecutive packet hops from the device in accordance with the routing information (see Dave ¶ [0022] “ . . . packets directed along LSP 16 are encapsulated by one or more labels in a label stack, e.g., in accordance with Multi-Protocol Label Switching (MPLS). In this manner, to direct traffic along an LSP, a router along the LSP may use a label of the label stack of a received packet to determine the “next hop” for the packet, that is, the next network device to which to forward the packet. In addition, the router may pop one or more labels from the label stack and push a new label onto the label stack, which causes the next hop to direct the packet along the LSP . . .”; see Dave ¶ [0045] “ . . . The selected route to reach the destination generally includes an indication of a “next hop” along the route to reach the destination. This next hop typically corresponds to a network device, such as, for example, another router, switch, gateway, or other network device along the route to reach the destination. The next hop device is connected to router 50 via one of IFCs 90. Accordingly, using the selected route to reach a destination, control unit 52 can determine the one of IFCs 90 connected to the next hop along the route to the destination and update forwarding information stored by PFE 60 to indicate the one of IFCs 90 to which to send packets destined for the destination. . . . “).
In regard to claim 4, the combination of Dave, Gentleman, and Farooq teaches wherein the prediction model is configured to predict the labels based on a similarity between the fingerprints wherein a similarity between two fingerprints is determined based on a size of an identical subset of consecutive packet hops (see Gentleman ¶ [0135] “ . . . a plurality of vector data objects may be stored within a repository, or within a partition of a repository, based on a shared data classification label and common data structure. In some embodiments, a data classification vector data set may comprise training data. In some embodiments, a trained data classification vector data set may be accessed to compare with a similarly structured vector data object for determining data classification labels for the similarly structured vector data object . . . “).
The motivation to combine the references is described for the rejection of claim 1 and is incorporated herein. Additionally Gentleman can assign labels of similar objects such as data flows.
In regard to claim 5, the combination of Dave, Gentleman, and Farooq teaches wherein said utilizing the prediction model is performed on the device, to predict the label for the device without exposing the fingerprint to an external device, the method further comprising augmenting prediction of the label for the fingerprint using additional features gathered at the device, wherein the additional features are not available to the external device (see Farooq ¶ [0055] “ . . . FIG. 4 is a diagram of one embodiment of an edge device and hyperparameter server executing aspects of the hyperparameter optimization process. The diagram illustrates an example where the hyperparameter server 201 operates in conjunction with a base station 401 to determine and provide optimal hyperparameters to the base station 401. The hyperparameter server sends instructions to the edge base station 401 to extract distribution features. The base station 401 will find statistical distribution features (meta-features) of their training dataset (real data) and perform hyperparameter tuning with some random configuration. In some embodiments, only one or few trials are performed by the base station 401. The base station 401 then sends this information (dataset meta-features, hyperparameter configuration, performance metrics) to the hyperparameter server (e.g., operating in a cloud) wherein this information coming from all edge base stations is compiled to form a dataset in the training database with dataset meta-features and hyperparameter configuration as input features and corresponding performance metrics (area under curve (AUC)) as the predicted value (regression). This dataset is utilized to train a global hyperparameter (HP) ML model that will be able to predict performance (AUC) for a given dataset meta-features and hyperparameter configuration. Once this dataset is compiled from a sufficient number of edge devices in a cluster then the hyperparameter server can performed training for the HP ML model. Prior to training, the edge base stations will send their dataset meta-features (they will not perform any hyperparameter optimization) and those will be used by the heuristic search algorithm to poll this trained hyperparameter ML model to get optimal hyperparameters that are sent back to the edge base station 401. The frequency with which the edge base station 401 can request optimal hyperparameters can be configured and can have any frequency or interval. . . .”).
The motivation to combine the references is described for the rejection of claim 1 and is incorporated herein. Additionally, Farooq secures a label to the device.
In regard to claim 14, Dave teaches An apparatus (see router 50 Fig. 2 see ¶ [0040] “ . . . FIG. 2 is a block diagram illustrating an example router 50 including a monitoring unit 80 configured according to the techniques of this disclosure. Router 50 may correspond to an ingress router, a transit router, a penultimate router, or an egress router, e.g., one of routers 20, 24, 28, and 32 in FIG. 1. In the example of FIG. 2, router 50 includes interface cards 90A-90N (IFCs 90), and control unit 52. Control unit 52 includes packet forwarding engine (PFE) 60, routing engine (RE) 70, and monitoring unit 80. . . .”) comprising a processor and coupled memory (see ¶ [0042] “ . . Control unit 52 includes processing hardware and, in some examples, software and/or firmware executed by the processing hardware. In various examples, control unit 52 and the various elements thereof, e.g., PFE 60 and RE 70, are implemented in one or more processors, processing units, digital signal processors (DSPs), application specific integrated circuits (ASICs), field programmable gate arrays (FPGAs), or any combination thereof. When implemented in software or firmware, control unit 52 includes one or more processors or processing units for executing instructions for the software or firmware, as well as a computer-readable storage medium for storing the instructions . . .”)., said processor being adapted to:
obtain routing information of a device (see ¶ [0003] as described for the rejection of claim 1 and is incorporated herein) , wherein the routing information is obtained based on one or more probe packets sent hy the device to a server that is connectable to the device via the Internet (see Fig. 1 ¶ [0021] ¶ [0029] as described for the rejection of claim 1 and is incorporated herein) , whereby a series if packet hops was implemented to route the one or more probe packets to the server (see ¶ [0022] as described for the rejection of claim 1 and is incorporated herein), the routing .information includes a series of Internet Protocol (IP) addresses of the series of packet hops along a path to the server (see ¶ [0004] as described for the rejection of claim 1 and is incorporated herein);
create, based on the routing information, a fingerprint (e.g. flow record) describing an architecture of the path of the device to the server (see ¶ [0007] as described for the rejection of claim 1 and is incorporated herein) , the path comprising at least the series of packet hops (see ¶ [0007] as described for the rejection of claim 1 and is incorporated herein) ;
Dave fails to explicitly teach.
However Gentleman teaches
utilize a prediction n1odeJ to determine u label for the fingerprint (see ¶ [0181] as described for the rejection of claim 1 and is incorporated herein) , wherein the label is indicative of metadata of a user of the device(see ¶ [0084] ¶ [0244] as described for the rejection of claim 1 and is incorporated herein) , wherein the prediction model is trained using a training dataset (see ¶¶ [0251-0252] as described for the rejection of claim 1 and is incorporated herein) that includes pairs of fingerprints and labels using edge devices having known labels (see ¶ [0025] as described for the rejection of claim 1 and is incorporated herein) ,
The motivation to combine Gentleman with Dave is described for the rejection of claim 1 and is incorporated herein.
The combination of Dave and Gentleman fails to explicitly teach,
However Farooq teaches
the fingerprints of the training dataset (see ¶ [0037] as described for the rejection of claim 1 and is incorporated herein) are indicative of a routing information of an edge device to the Internet. (see ¶ [0047] Fig. 7D ¶ [0079] as described for the rejection of claim 1 and is incorporated herein)
The motivation to combine Farooq with the combination of Dave and Gentleman is described for the rejection of claim 1 and is incorporated herein.
In regard to claim 15, Dave teaches A computer program product comprising a non-transitory computer readable medium retaining program instruction, which program instructions when read by a processor (see ¶¶ [0089-0091] “ . . . If implemented in hardware, this disclosure may be directed to an apparatus such a processor or an integrated circuit device, such as an integrated circuit chip or chipset. Alternatively, or additionally, if implemented in software or firmware, the techniques may be realized at least in part by a computer-readable data storage medium comprising instructions that, when executed, cause a processor to perform one or more of the methods described above. For example, the computer-readable data storage medium may store such instructions for execution by a processor. A computer-readable medium may form part of a computer program product, which may include packaging materials. A computer-readable medium may comprise a computer data storage medium such as random access memory (RAM), read-only memory (ROM), non-volatile random access memory (NVRAM), electrically erasable programmable read-only memory (EEPROM), Flash memory, magnetic or optical data storage media, and the like. In some examples, an article of manufacture may comprise one or more computer-readable storage media. In some examples, the computer-readable storage media may comprise non-transitory media. The term “non-transitory” may indicate that the storage medium is not embodied in a carrier wave or a propagated signal. In certain examples, a non-transitory storage medium may store data that can, over time, change (e.g., in RAM or cache . . . “), cause the processor to:
obtain routing information of a device (see ¶ [0003] as described for the rejection of claim 1 and is incorporated herein) , wherein the routing information is obtained based on one or more probe packets sent hy the device to a server that is connectable to the device via the Internet(see Fig. 1 ¶ [0021] ¶ [0029] as described for the rejection of claim 1 and is incorporated herein) , whereby a series if packet hops was implemented to route the one or more probe packets to the server (see ¶ [0022] as described for the rejection of claim 1 and is incorporated herein), the routing .information includes a series of Internet Protocol (IP) addresses of the series of packet hops along a path to the server(see ¶ [0004] as described for the rejection of claim 1 and is incorporated herein) ;
create, based on the routing information, a fingerprint (e.g. flow record) describing an architecture of the path of the device to the server (see ¶ [0007] as described for the rejection of claim 1 and is incorporated herein), the path comprising at least the series of packet hops (see ¶ [0007] as described for the rejection of claim 1 and is incorporated herein);
Dave fails to explicitly teach.
However Gentleman teaches
utilize a prediction n1odeJ to determine u label for the fingerprint (see ¶ [0181] as described for the rejection of claim 1 and is incorporated herein) , wherein the label is indicative of metadata of a user of the device (see ¶ [0084] ¶ [0244] as described for the rejection of claim 1 and is incorporated herein), wherein the prediction model is trained using a training dataset (see ¶¶ [0251-0252] as described for the rejection of claim 1 and is incorporated herein) that includes pairs of fingerprints and labels using edge devices having known labels (see ¶ [0025] as described for the rejection of claim 1 and is incorporated herein) ,
The motivation to combine Gentleman with Dave is described for the rejection of claim 1 and is incorporated herein.
The combination of Dave and Gentleman fails to explicitly teach,
However Farooq teaches
the fingerprints of the training dataset(see ¶ [0037] as described for the rejection of claim 1 and is incorporated herein) are indicative of a routing information of an edge device to the Internet (see ¶ [0047] Fig. 7D ¶ [0079] as described for the rejection of claim 1 and is incorporated herein).
The motivation to combine Farooq with the combination of Dave and Gentleman is described for the rejection of claim 1 and is incorporated herein.
In regard to claim 16, the combination of Dave, Gentleman, and Farooq teaches wherein the fingerprint comprises the series of IP addresses of the series of packet hops or an encoding thereof (see Dave ¶ [0027] as described for the rejection of claim 2 and is incorporated herein)
In regard to claim 17, the combination of Dave, Gentleman, and Farooq teaches wherein the fingerprint comprises IP addresses associated with N consecutive packet hops from the device in accordance with the routing information (see Dave ¶ [0022], ¶ [0045] as described for the rejection of claim 3 and is incorporated herein)
In regard to claim 18, the combination of Dave, Gentleman, and Farooq teaches wherein the prediction model is configured to predict the labels based on a similarity between the fingerprints wherein a similarity between two fingerprints is determined based on a size of an identical subset of consecutive packet hops (see Gentleman ¶ [0135] as described for the rejection of claim 4 and is incorporated herein).
The motivation to combine the references is described for the rejection of claim 4 and is incorporated herein.
In regard to claim 19, the combination of Dave, Gentleman, and Farooq teaches wherein said utilizing the prediction model is performed on the device, to predict the label for the device without exposing the fingerprint to an external device (see Farooq ¶ [0055] as described for the rejection of claim 5 and is incorporated herein).
The motivation to combine the references is described for the rejection of claim 5 and is incorporated herein.
Claims 6 and 20 are rejected under 35 U.S.C. 103 as being un-patentable over Dave et al. (U.S. 2022/0166722 A1; herein referred to as Dave) in view of Gentleman et al. (U.S. 2022/0207163 A1; herein referred to as Gentleman) in further view of Farooq et al. (U.S. 2023/0162089 A1; herein referred to as Farooq) as applied to claims 1 – 5 and 14 – 19 in further view of Alexander et al. (U.S. 2019/0045241 A1; herein referred to as Alexander).
In regard to claim 6, the combination of Dave, Gentleman, and Farooq fails to explicitly teach,
However Alexander teaches wherein the metadata comprises at least one of: a workplace identity, demographic information, a geographic proximity, an income level, an education level, or user interests (see Alexander ¶ [0011] “ . . . The method may further include routing the replicated content stream over a path through the telecommunications network. For example, the replicated content stream may contain multiple packets, each of which may have a header populated by the network device with a routing label corresponding to the path the packet is to follow. . . “; see Alexander ¶ [0031] “ . . . In addition, content providers 204,206 may provide a content stream to more than one replicator 214-218 of the network 202. For example, content provider 204 may provide a content stream to replicator A 214 and replicator B 216. Further still, although only illustrated in FIG. 2 as providing a received content stream to two receivers 222,224, the replicator 214 may provide a received content stream to any number of receivers, with each receiver communicating with the replicator 214 through a particular output port of the replicator device. Generally and discussed in more detail below, the receivers 228-232 connect to and communicate with the network 202 in a similar manner as the content providers 204,206. In other words, the receivers 228-232 establish a communication session with one or more edge devices 222-226 (which may generally be referred to as “egress” edge devices) of the network 202 to begin receiving packets of information transmitted through the network 202. In many cases, the edge device 222-226 utilized by a receiver is the edge device that is the nearest geographically to the receiver, although any edge device may be utilized. Routing of the content stream to the egress edge devices 222-226 may be based at least on internal routing techniques utilized by the network 202 to reach the intended receiver 228-232. . . .”; see Alexander ¶ [0044] “ . . . Another advantage realized through the system described herein is the ability to pass particular types of metadata within the content stream. For example, many satellite-based distribution systems remove certain types of metadata from the content stream, such as Society of Cable and Telecommunications Engineers (SCTE)-35 metadata, which is used to signal local advertising insertion into a content stream. This SCTE-35 metadata may be removed when compressed into a satellite-based signal but may be retained in the content stream of the system described herein, enabling the dynamic insertion of advertisements corresponding to particular geographic locations or viewer demographics. Other metadata of the signal may also be retained that other distribution systems remove for transmission across a system . . .”)
It would have been obvious to one skilled in the art, before the effective filing date of the applicant’s claimed invention to incorporate a method and system for providing content through a terrestrial fiber network wherein a replicated content stream contains multiple packets, each of which may have a header populated by the network device with a routing label corresponding to the path the packet is to follow, as taught by Alexander, into a method and system for managing a plurality of network devices, a network device includes processing circuitry configured to: receive packet data corresponding to a network flow originating at a first device, the packet data destined to a second device; generate an entropy label to add to a label stack of the packet data, wherein the entropy label is generated from one or more attributes corresponding to the network flow as measured by packet hops that originated at the first device and is destined to the second device; generate a flow record including the entropy label, wherein the entropy label identifies the network flow amongst a plurality of network flows in the network; and send, to a controller of the network, the flow record, wherein the controller identifies the flow record based on the entropy label corresponding to the network flow originating at the first device and is destined to the second device, for providing for the dynamic data classification of the devices, including prediction of candidate data classification labels for devices associated with one or more applications and services, or edge nodes such that the assignment of one or more data classification labels for transmission to one or more computing devices. Examples enable the interactive and progressive application of machine learning techniques to data classification systems to assign data classification labels with probable certainty, and creating datasets with labels for the devices and stored routing information for the devices as taught by the combination of Dave, Gentleman, and Farooq. Such incorporation enables metadata from the label being used to enhance the transmission of the packets across the system.
In regard to claim 20, the combination of Dave, Gentleman, Farooq, and Alexander teaches wherein the metadata comprises at least one of: a workplace identity, demographic information, a geographic proximity, an income level, an education level, or user interests (see Alexander ¶ [0011], ¶ [0031], ¶ [0044] as described for the rejection of claim 6 and is incorporated herein).
The motivation to combine the references is described for the rejection of claim 6 and is incorporated herein,
Claims 7, 8, and 13 are rejected under 35 U.S.C. 103 as being un-patentable over Dave et al. (U.S. 2022/0166722 A1; herein referred to as Dave) in view of Gentleman et al. (U.S. 2022/0207163 A1; herein referred to as Gentleman) in further view of Farooq et al. (U.S. 2023/0162089 A1; herein referred to as Farooq) as applied to claims 1 – 5 and 14 – 19 in further view of Wang et al. (U.S. 2024/0356817 A1; herein referred to as Wang).
In regard to claim 7, the combination of Dave, Gentleman, and Farooq fails to explicitly teach,
However Wang teaches wherein the prediction model is generated using centralized learning performed on a central server (see Wang ¶ [0118] “ . . . Federated Learning is a machine learning setting where the goal is to train a high-quality centralized model while training data remains distributed over a number of clients. One or more clients may have an unreliable or relatively slow network connection. On each round, each client may independently compute an update to the current model based on the client's local data, and communicates this update to a central server, where the client-side updates are aggregated to compute a new global update. The typical clients in this setting may include mobile phones. As such, it may be advantageous to achieve efficient communications. Federated Learning may enable mobile phones to collaboratively learn a shared prediction model while keeping the training data on the mobile phone, decoupling the ability to do machine learning from the need to store the data in the cloud. The training data may be kept locally on users' mobile devices, and the devices may be used as nodes performing computation on their local data in order to update a global model. . . .”), wherein the training dataset comprises multiple training data, each of which are obtained from a different edge device (e.g. STA – station) (see Wang ¶ [0193] “ . . . The parameters of the Trigger Dependent Common Info subfield 800 may include the available dataset 806, which may request the STA(s) to indicate if it has any new training data being collected in the device, e.g., during a certain period. The parameters of the Trigger Dependent Common Info subfield 800 may include the processing capability 808, which may request the STA(s) to indicate the processing capability, e.g., how many GPUs are available in the STA. For example, if the value of the Processing Capability subfield is 1, it may indicate that the AP requests the STA to indicate the processing capability (e.g., the number of GPU available in the STA). If the value of the Processing Capability subfield is 0, it may indicate that the AP does not request the STA to indicate the processing capability. The parameters of the Trigger Dependent Common Info subfield 800 may include the updated training model 810, which may request the STA(s) to feedback the updated training model or updated training model parameters. . . .”).
It would have been obvious to one skilled in the art, before the effective filing date of the applicant’s claimed invention to incorporate a method and system for implementing a distributed machine learning model to train data collected in a wireless network, as taught by Wang into a method and system for managing a plurality of network devices, a network device includes processing circuitry configured to: receive packet data corresponding to a network flow originating at a first device, the packet data destined to a second device; generate an entropy label to add to a label stack of the packet data, wherein the entropy label is generated from one or more attributes corresponding to the network flow as measured by packet hops that originated at the first device and is destined to the second device; generate a flow record including the entropy label, wherein the entropy label identifies the network flow amongst a plurality of network flows in the network; and send, to a controller of the network, the flow record, wherein the controller identifies the flow record based on the entropy label corresponding to the network flow originating at the first device and is destined to the second device, for providing for the dynamic data classification of the devices, including prediction of candidate data classification labels for devices associated with one or more applications and services, or edge nodes such that the assignment of one or more data classification labels for transmission to one or more computing devices. Examples enable the interactive and progressive application of machine learning techniques to data classification systems to assign data classification labels with probable certainty, and creating datasets with labels for the devices and stored routing information for the devices as taught by the combination of Dave, Gentleman, and Farooq. Such incorporation enables generating prediction models using Federated systems.
In regard to claim 8, the combination of Dave, Gentleman, Farooq, and Wang teaches wherein each training data obtained from a respective edge device is processed to replace a permanent identifier (e.g. device address) of the respective edge device with a transient identifier (e.g. broadcast address) prior to being sent to the central server, whereby preserving privacy of data of the respective edge device (see Wang ¶¶ [0131-0133] “ . . . The example embodiments discussed below address federated learning or machine learning model sharing. The Enhanced Broadcast Services (EBCS) (e.g., 802.11bc) procedures provide both uplink (UL) and downlink (DL) broadcasting protocols. Such UL and DL broadcasting procedures may be suitable for the AP and the STAs to share their machine learning and federated learning models, weights, gradients, and/or other parameters. FIG. 4A depicts an example design of a ML/FL Model sharing announcement frame 402ln. ML/FL Model sharing announcement frame 402 may be referred to herein as a FL Announcement frame, an announcement frame, a sharing announcement frame, a ML announcement frame, FL sharing announcement frame, a ML sharing announcement frame, or a ML/FL sharing announcement frame. In one example application, a machine learning and federated learning model-sharing broadcasting procedure may be implemented in a circumstance in which an EBCS STA, such as an EBCS AP or an EBCS non-AP STA, is connected to a machine learning (ML) or federated learning server (LS) and the LS is configured to initiate or execute machine learning or federated learning algorithms in cooperation with one or more of STAs within the area or the network. The EBCS AP or EBCS STA may broadcast one or more FL announcement frames to announce that machine learning model or federated learning model sharing is currently in progress. Announcement frame 402 may contain one or more fields for communicating parameters related to an ML/FL model. For example, the ML/FL Support Indication frame may include one or more of the following fields: MAC header field 404, learning server (LS) Server Address field 406, ML/FL Model ID field 408, Number of Layers field 410, Weights Lists field 412, Gradient Lists field 414, Update schedule field 416, Model Updates 418, and frame check sequence (FCS) field 420. FCS field 420 may be used to check the integrity of the frame. The MAC header field 404 may contain a receiver address (RA). The RA may be set to a broadcast address, or an address such as multicast address that is associated with one or more ML/FL model sharing, or to another address that is agreed beforehand. MAC header field 404 may contain a transmitter address (TA). The TA may be set to the MAC address of the transmitter, such as the MAC address of the EBCS AP or STA, or to a multicast address that is associated with one or more ML/FL model sharing. . . .”).
The motivation to combine Wang with the combination of Dave, Gentleman and Farooq is described for the rejection of claim 7 and is incorporated herein, . Additionally, Wang modifies the network addresses of the devices for the training data to enable privacy of the device.
In regard to claim 13, the combination of Dave, Gentleman, Farooq, and Wang teaches wherein the prediction model is generated using federated learning performed on a central server (see Wang ¶ [0118] “ . . . Federated Learning is a machine learning setting where the goal is to train a high-quality centralized model while training data remains distributed over a number of clients. One or more clients may have an unreliable or relatively slow network connection. On each round, each client may independently compute an update to the current model based on the client's local data, and communicates this update to a central server, where the client-side updates are aggregated to compute a new global update. The typical clients in this setting may include mobile phones. As such, it may be advantageous to achieve efficient communications. Federated Learning may enable mobile phones to collaboratively learn a shared prediction model while keeping the training data on the mobile phone, decoupling the ability to do machine learning from the need to store the data in the cloud. The training data may be kept locally on users' mobile devices, and the devices may be used as nodes performing computation on their local data in order to update a global mode. . . “) , wherein each edge device provides a model update to the predictive model based on pairs of fingerprints and labels available to the edge device, whereby obfuscating training data generated by the respective edge device (see Wang ¶ [0119] “ . . . In accordance with a naive implementation of Federated Learning, each client may send a full model (or a full model update) back to the server in each round. For large models, this step may be the bottleneck of Federated Learning due to multiple factors. One factor is the asymmetric property of internet connection speeds: the uplink may be slower than the downlink. One way to reduce uplink communication (from the client to the server) cost in Federated Learning is to implement structured updates, where an update from a restricted pace (e.g., speed) can be learned and it can be parametrized using a smaller number of variables. Another is Sketched updates, where a full model is updated. Then it is compressed before sending to the server. . . . “).
The motivation to combine Wang with the combination of Dave, Gentleman and Farooq is described for the rejection of claim 7 and is incorporated herein, Additionally, Wang provides a federated solution to enable multiple devices to aggregate the data to be used for the training.
Claims 9 – 10 are rejected under 35 U.S.C. 103 as being un-patentable over Dave et al. (U.S. 2022/0166722 A1; herein referred to as Dave) in view of Gentleman et al. (U.S. 2022/0207163 A1; herein referred to as Gentleman) in further view of Farooq et al. (U.S. 2023/0162089 A1; herein referred to as Farooq) as applied to claims 1 – 5 and 14 – 19 in further view of McCourt Jr. (U.S. 2020/0019884 A1; herein referred to as McCourt).
In regard to claim 9, the combination of Dave, Gentleman, and Farooq fails to explicitly teach,
However McCourt teaches wherein the training dataset includes a partly fabricated training data that was reported by a training edge device (see McCourt ¶ [0159] “ . . . The technologies described herein provide various technological improvements to computer performance and efficiently. For example, the tolerator 1500 and/or the processes herein are technological improvements over those of the past because they provide a much more accurate and efficient device and/or process for determining whether training data is incorrect for training a binary signal classifier, tolerating such incorrect data, removing such incorrect data, and determining an accuracy score for the classifier. For example, the models described for equations (7)-(13) may be or be used as an error tolerant model for training of machine learning binary classifiers with training data that includes incorrect binary output labels. These models may ignore or compensate for the incorrect labeled training data by including a probability that each training data has an incorrectly identified output label. This keeps the model from overcommitting to the incorrectly identified output labels which is what happens for a model that only considers a correctly identified output label. Also, the models described for equations (12)-(13) may be or be used as an approximated error tolerant model for training of machine learning binary classifiers with training data that includes incorrect binary output labels that can be used by a computer because they add, not multiply, certain terms (they are quadratic; not cubic) to avoid computer based numeric overflow and underflow. Next, the models described for equations (7)-(13) may be or be used as an error tolerant model for determining the accuracy of any trained machine learning binary classifier by cleaning a set of real training data having outputs that are fabricated by or confirmed by a human, but are generated by customers or a computer . . . “) ,
the training edge device having a known correct label (see McCourt Fig. 5 ¶ [0011] “ . . . FIG. 5 is a plot showing predicted probabilities that a training data output label is correct according to the logistic regression error-tolerant model for a label of true and various choices of the priori odds ratios . . . “).,
the partly fabricated training data comprises a fingerprint of the training edge device that is paired with several labels (see McCourt ¶ [0122] “ . . . a machine learning binary classifier begins to be trained using a set of training data entries with a subset of training data entries that have incorrectly labeled known. Training at 1010 may include a tolerator or a machine learning binary classifier beginning to be trained by a person and/or computing device, with a set of training data entries that each have known inputs and a known output label and have the subset with incorrectly labeled known output labels. . . . “),
the several labels include the known correct label and at least one incorrect label, whereby preserving the privacy of data of the training edge device during the training process (see McCourt ¶ [0123] “ . . . training continues by determining a correct likelihood ratio that each training data entry of the set of training data entries has a correctly labeled output label and an incorrect likelihood ratio that each training data entry of the set of training data entries has an incorrectly labeled output label. Determining at 1020 may include a computing device determining the correct and incorrect likelihood ratios for each training data entry using a computer model of the tolerator or of the machine learning binary classifier. . . “).
It would have been obvious to one skilled in the art, before the effective filing date of the applicant’s claimed invention to incorporate a method and system for a machine learning binary classifier automatically tolerating training data that is incorrect by determining a correct and an incorrect likelihood ratio that each training data entry has a correctly and an incorrectly labeled output, so that the correct and an incorrect likelihood ratio are combined with a correct and an incorrect priori odds ratio that the set of training data entries have correctly and incorrect labeled output labels, as taught by McCourt, into a method and system for managing a plurality of network devices, a network device includes processing circuitry configured to: receive packet data corresponding to a network flow originating at a first device, the packet data destined to a second device; generate an entropy label to add to a label stack of the packet data, wherein the entropy label is generated from one or more attributes corresponding to the network flow as measured by packet hops that originated at the first device and is destined to the second device; generate a flow record including the entropy label, wherein the entropy label identifies the network flow amongst a plurality of network flows in the network; and send, to a controller of the network, the flow record, wherein the controller identifies the flow record based on the entropy label corresponding to the network flow originating at the first device and is destined to the second device, for providing for the dynamic data classification of the devices, including prediction of candidate data classification labels for devices associated with one or more applications and services, or edge nodes such that the assignment of one or more data classification labels for transmission to one or more computing devices. Examples enable the interactive and progressive application of machine learning techniques to data classification systems to assign data classification labels with probable certainty, and creating datasets with labels for the devices and stored routing information for the devices as taught by the combination of Dave, Gentleman, and Farooq. Such incorporation enables certain data to be protected so that the privacy can be maintained.
In regard to claim 10, the combination of Dave, Gentleman, Farooq, and McCourt teaches wherein the training dataset includes a partly fabricated training data that was reported by a training edge device (see McCourt ¶ [0159] as described for the rejection of claim 9 and is incorporated herein) ,
the training edge device having a known correct label and a known correct fingerprint (see McCourt ¶ ¶ [0122-0123] “ . . . a machine learning binary classifier begins to be trained using a set of training data entries with a subset of training data entries that have incorrectly labeled known. Training at 1010 may include a tolerator or a machine learning binary classifier beginning to be trained by a person and/or computing device, with a set of training data entries that each have known inputs and a known output label and have the subset with incorrectly labeled known output labels. After 1010, at 1020 training continues by determining a correct likelihood ratio that each training data entry of the set of training data entries has a correctly labeled output label and an incorrect likelihood ratio that each training data entry of the set of training data entries has an incorrectly labeled output label. Determining at 1020 may include a computing device determining the correct and incorrect likelihood ratios for each training data entry using a computer model of the tolerator or of the machine learning binary classifier . . .”).,
the partly fabricated training data comprises at least a first pair and a second pair (e.g. regression models) (see McCourt ¶ ¶ [0124-0125] “ . . . Determining the correct likelihood ratio may include fitting a first logistic regression model to each entry of the training data and determining the incorrect likelihood ratio comprises fitting a second logistic regression model to each entry of the training data; and fitting comprises minimizing a likelihood function that each entry of the training data fits a logistic regression model to estimate model parameters of a logistic regression model. Training at 1020 may include the tolerator determining a correct likelihood ratio Q (x.sub.i|l.sub.i, α) that each training data entry of the set of training data entries has a correctly labeled output label and an incorrect likelihood ratio Q (x.sub.i|¬l.sub.i, α) that each training data entry of the set of training data entries has an incorrectly labeled output label. . . .”),
the first pair comprising the known correct fingerprint and the known correct label (see McCourt, see ¶ [0126] “ . . . Determining the correct likelihood ratio at 1020 may be fitting a first logistic regression model to each entry of the training data . . . “)
the second pair comprising a fabricated fingerprint and the known correct label, whereby preserving a privacy of data of the training edge device during the training process (see McCourt, see ¶ [0126] “ . . . determining the incorrect likelihood ratio at 1020 is fitting a second logistic regression model to each entry of the training data. Here, fitting is minimizing a likelihood function that each entry of the training data fits a logistic regression model to estimate model parameters of a logistic regression model. The correct likelihood ratio may be a first sigmoid term for a label l.sub.i being correct, and the incorrect likelihood ratio may be a second sigmoid term for a label l.sub.i being incorrect that is a mirror image of the first sigmoid term . . . “).
The motivation to combine McCourt with the combination of Dave, Gentleman, and Farooq is described for the rejection of claim 9 and is incorporated herein. Additionally, McCourt performs label corrections.
Claims 11 – 12 are rejected under 35 U.S.C. 103 as being un-patentable over Dave et al. (U.S. 2022/0166722 A1; herein referred to as Dave) in view of Gentleman et al. (U.S. 2022/0207163 A1; herein referred to as Gentleman) in further view of Farooq et al. (U.S. 2023/0162089 A1; herein referred to as Farooq) as applied to claims 1 – 5 and 14 – 19 in further view of Alanazi (U.S. 2021/0152523 A1; herein referred to as Alanazi)
In regard to claim 11, the combination of Dave, Gentleman, and Farooq fails to explicitly teach,
However Alanazi teaches wherein the training dataset comprises pairs of fabricated fingerprints and labels (see Alanazi ¶ [0082] “ . . . The machine learning model(s) can be trained using multiple datasets, as is known by those skilled in the art, including, for example, a training dataset, a test dataset and a validation dataset. The model(s) can be trained and periodically or continuously updated to accurately predict data traffic, traffic characteristics and traffic patterns for each communication device 20, as well as the entire LAN 10. Model parameters can be updated on an ongoing basis based on analysis results from the activity monitor 145. . . .”), wherein a fabricated fingerprint is generated by modifying an IP address of at least one packet hop in the path (see Alanazi ¶ [0090] “ . . . The S/D synthesizer 160, which can be formed as a single device or module with the traffic synthesizer 150 or provided separately, can fabricate a source IP address or a destination URL. For instance, the S/D synthesizer 160 can identify a commonly used URL, such as, for example, a search engine URL, a news website URL, or any other website commonly accessed by the general public. The S/D synthesizer 160 can learn network activity for non-IoT communication devices 20 in the LAN 10 over time and identify destination addresses or URLs used by such devices. The S/D synthesizer 160 can generate a fabricated destination address by, for example, crawling the WWW and determining popular URLs. The fabricated destination address (or URL, such as, for example, <www.yahoo.com>) can be forwarded to the traffic synthesizer 150 to be injected into the packet header of fabricated data traffic, which can be sent by the entropy appliance 25 to the ISP server 40 (shown in FIG. 1). The S/D synthesizer 160 can change the fabricated destination URL for added obscurity or concealment from passive analyzers. . . .”).
It would have been obvious to one skilled in the art, before the effective filing date of the applicant’s claimed invention to incorporate a method and system for protecting data traffic from a communication device against fingerprinting or privacy leakage, as taught by Atanazi, into a method and system for managing a plurality of network devices, a network device includes processing circuitry configured to: receive packet data corresponding to a network flow originating at a first device, the packet data destined to a second device; generate an entropy label to add to a label stack of the packet data, wherein the entropy label is generated from one or more attributes corresponding to the network flow as measured by packet hops that originated at the first device and is destined to the second device; generate a flow record including the entropy label, wherein the entropy label identifies the network flow amongst a plurality of network flows in the network; and send, to a controller of the network, the flow record, wherein the controller identifies the flow record based on the entropy label corresponding to the network flow originating at the first device and is destined to the second device, for providing for the dynamic data classification of the devices, including prediction of candidate data classification labels for devices associated with one or more applications and services, or edge nodes such that the assignment of one or more data classification labels for transmission to one or more computing devices. Examples enable the interactive and progressive application of machine learning techniques to data classification systems to assign data classification labels with probable certainty, and creating datasets with labels for the devices and stored routing information for the devices as taught by the combination of Dave, Gentleman, and Farooq. Such incorporation protects data privacy by fabricating information that protects the device for being identified.
In regard to claim 12, combination of Dave, Gentleman, Farooq, and Atanazi, teaches wherein the training dataset includes a partly fabricated training data (see Alanazi ¶ [0052] “ . . . the entropy appliance 25 can be arranged to forge data traffic for each communication device 20 to generate an entropy factor to a portion or all data traffic emanating from the entropy appliance 25. The forged data traffic can include synthetic data that can be injected as filler data in the forged data traffic or packetized and transmitted as forged data traffic from a fabricated communication device. In the latter instance, a passive analyzer would likely identify the forged data traffic as emanating from the fabricated communication device, which does not exist in reality but, instead, is fabricated to confuse passive analysis technologies. . . .”) , wherein fabrication of training data is performed below a predetermined threshold (e.g. entropy factor), thereby enabling the prediction model to predict correct labels despite fabricated and incorrect information (see Alanazi ¶ [0052] “ . . . the forged data traffic can provide an entropy factor that reduces or eliminates any chance of the communication device 20 or the LAN 10 being identified or monitored through technologies such as fingerprinting or other passive analysis technologies. The forged data traffic can be generated or shaped by the entropy appliance 25 based on learned network activities and operational characteristics for each communication device 20. The forged data traffic can be generated by the entropy appliance 25 carrying out a data traffic forging process 200 (shown in FIG. 5) and/or traffic randomization process 300 (shown in FIG. 6). . . . “).
The motivation to combine Alanazi with the combination of Dave, Gentleman, and Farooq is described for the rejection of claim 11 and is incorporated herein. Additionally, Alanazi provides techniques to vary the fabrication so that accurate models can be created.
Conclusion
There are prior art made of record which are not relied upon but are considered pertinent to applicant’s disclosure. They are listed on the PTO-892 accompanying this action.
Any inquiry concerning this communication or earlier communications from the examiner should be directed to JAMES N FIORILLO whose telephone number is (571)272-9909. The examiner can normally be reached on 7:30 - 5 PM Mon - Fri..
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, John A. Follansbee can be reached on 571-272-3964. The fax phone number for the organization where this application or proceeding is assigned is 571-273-8300.
Information regarding the status of an application may be obtained from the Patent Application Information Retrieval (PAIR) system. Status information for published applications may be obtained from either Private PAIR or Public PAIR. Status information for unpublished applications is available through Private PAIR only. For more information about the PAIR system, see http://pair-direct.uspto.gov. Should you have questions on access to the Private PAIR system, contact the Electronic Business Center (EBC) at 866-217-9197 (toll-free). If you would like assistance from a USPTO Customer Service Representative or access to the automated information system, call 800-786-9199 (IN USA OR CANADA) or 571-272-1000.
/JAMES N FIORILLO/Primary Examiner, Art Unit 2444
Prosecution Insights