US 10637968

I'll search for authoritative information on US patent 10637968, including any recent docket activity.

Let me search for the specific claim language and any 2026 appellate/docket activity.

Let me retrieve the actual claim language for the '968 patent.

Summary — US 10,637,968 B2

Important upfront caveat: The searchable public record (Google Patents, Espacenet, USPTO/PTAB documents, Unified Patents, Docket Alarm) gives me reliable bibliographic, litigation, and PTAB/reexam data for this patent. However, I was not able to retrieve the verbatim claim text of US 10,637,968 (the "968 patent") within this session. My independent-claim overview below is therefore reconstructed from the patent's own Summary of the Invention (which tracks the claim language) and from the PTAB record — and is flagged as such. I have not fabricated claim language. Also note: your prompt states the current date is April 26, 2026, while my system date is 2026-10-01; the docket findings below reflect what is indexed as of the later date.

1. Bibliographic data

Note on the cross-reference text: The specification text indexed for this document recites that "the present application is a continuation application of U.S. non-provisional patent application Ser. No. 14/025,109, filed Sep. 12, 2013, which is a divisional application of ... Ser. No. 12/836,059, filed Jul. 14, 2010 and issued as U.S. Pat. No. 8,560,604." Taken literally, this cross-reference does not reconcile cleanly with application No. 16/396,696 filed 2019-04-28. I am flagging this as an apparent inconsistency in the fetched text rather than correcting it.

2. Abstract (as published)

"A system designed for increasing network communication speed for users, while lowering network congestion for content owners and ISPs. The system employs network elements including an acceleration server, clients, agents, and peers, where communication requests generated by applications are intercepted by the client on the same machine. The IP address of the server in the communication request is transmitted to the acceleration server, which provides a list of agents to use for this IP address. The communication request is sent to the agents. One or more of the agents respond with a list of peers that have previously seen some or all of the content which is the response to this request (after checking whether this data is still valid). The client then downloads the data from these peers in parts and in parallel, thereby speeding up the Web transfer, releasing congestion from the Web by fetching the information from multiple sources, and relieving traffic from Web servers by offloading the data transfers from them to nearby peers."

3. Plain-language overview of the independent claims

Confidence: moderate — derived from the patent's Summary of the Invention, not from verbatim claim text. The '968 is a continuation in the Bright Data/Luminati "faster and more efficient data communication" family, issued with roughly 28+ claims (the PTAB petition challenged claims 1, 2, 8, 9, 11–18, 24, and 26–28, so claim 28 at least exists).

Two independent-claim families are described in the Summary and are the likely independent claims:

1. Network/system claim — A network for accelerating data communication, comprising:

   • a client communication device that originates a data request for data from a data server;
   • an agent communication device assigned to that data server, which receives the client's request and keeps track of which client communication devices have received responses to data requests from the assigned data server (i.e., the agent is the per-server index of who already holds the content);
   • a peer communication device that stores portions of data received in response to a request and can serve those portions back to a client on request; and
   • an acceleration server that decides which agent is assigned to which data server and tells the client that assignment.

2. Communication-device claim — A single device (memory + processor) configured to:

   • originate a data request for data from a data server;
   • be assigned to a data server;
   • receive requests from other devices and track which client devices have received responses from its assigned data server; and
   • store portions of the data it received, which portions can be transmitted to other devices on request.

There may also be method claims in the family (the related §101 briefing and the sibling patents use "A method for fetching …" language), but I do not have the '968's method-claim text to confirm.

The unifying idea: rather than a global content index (the proxy/BitTorrent problem described in the background), the system shards responsibility for web servers across agents by IP address, has agents maintain per-URL, per-chunk peer lists using checksums as keys, chunks content (e.g., 16 KB), and has clients fetch chunks in parallel from multiple peers, after validating cached data against the origin server (e.g., HTTP conditional requests).

4. Docket / contested-proceedings status (as indexed)

• PTAB — IPR2021-00249, Teso LT, UAB; Code200, UAB; Metacluster LT, UAB; Oxysales, UAB v. Luminati Networks Ltd., filed 2020-12-07, challenged claims 1, 2, 8, 9, 11–18, 24, 26–28 of the '968. Institution DENIED on 2021-06-02 under §314(a) as a discretionary denial (Google Patents labels it "Not Instituted – Procedural"). Source: Docket Alarm institution decision; IP Verse case record.
• Ex parte reexaminations 90/014,816 and 90/014,652 (both on US 10,637,968), filed 2021-07-30. A Final Office Action issued 2022-09-29 rejecting all claims asserted against Oxylabs. Docket Alarm shows 90/14,816 with status "Appeal Brief Filed (or Remand from Board) – Awaiting Examiner Action." This is the live invalidity risk vector on this patent. (Sources: Docket Alarm; Oxylabs legal timeline.)
• District court litigation (E.D. Tex., as reflected in the patent's own litigation metadata and the IPR papers): 2:19-cv-00396 (Luminati Networks Ltd. v. Code200, UAB, et al.); 2:19-cv-00414; 2:20-cv-00188 (Luminati Networks, Ltd. v. NetNut, Ltd.).
• CAFC 2026 — no docket found specifically naming U.S. 10,637,968. My searches of 2026 Federal Circuit materials did not return an appeal docketed under this patent number. What does exist at the Federal Circuit/SCOTUS level in 2026 is the Bright Data appeal concerning different patents — U.S. 10,257,319; 10,484,510; 11,044,342; and 11,044,344 — where the Federal Circuit (Aug. 1, 2025, nonprecedential) affirmed PTAB invalidation, and the Supreme Court denied certiorari on February 23, 2026 (Petition No. 25-779, Bright Data Ltd. v. Code200 UAB). The '968 is not among those four patents. (Sources: Supreme Court docket PDF; Law360; Oxylabs legal timeline; iipla.org.)
• Because the '968 reexamination is at the appeal-brief stage, a future appeal to the Federal Circuit (or to the PTAB under 35 U.S.C. §134(b)) remains possible — but as of my search, I found no such 2026 CAFC proceeding.

5. Explicit uncertainties

1. Verbatim independent-claim text for the '968 was not retrieved. Claim scope should be confirmed against the granted claim set (USPTO PatentCenter / the patent PDF), not this summary.
2. Total claim count is not confirmed; the record establishes at least claim 28.
3. The filing/cross-reference discrepancy noted in §1 is unresolved on the fetched record.
4. Litigation/reexam status is volatile — reexam outcomes and any appeal can change; treat the "Active" and "Awaiting Examiner Action" labels as search-result snapshots, not legal conclusions.
5. The '968 has been grouped by some aggregators alongside the four Bright Data patents the Supreme Court declined to review; that grouping is about the patent family/shared specification, not this patent number, and I have kept them separate deliberately.

Key sources: Google Patents (https://patents.google.com/patent/[US10637968B2](/patent/US10637968B2)/en); Espacenet (https://si.espacenet.com/publicationDetails/biblio?CC=US&NR=[10637968B2](/patent/10637968B2)); PTAB institution decision IPR2021-00249 (https://www.docketalarm.com/cases/PTAB/IPR2021-00249/...); Docket Alarm reexam docket (https://www.docketalarm.com/patentapps/US/90-14,816/...); Unified Patents patent page (https://portal.unifiedpatents.com/patents/patent/US-10637968-B2); Oxylabs legal timeline (https://oxylabs.io/legal-timeline); Supreme Court cert petition No. 25-779 (https://www.supremecourt.gov/DocketPDF/25/25-779/...).

I'll research litigation involving US patent 10637968. Let me start with the sources provided.

Let me search for the specific district court cases and the parties involved.

Let me get more details on the remaining cases and outcomes.

Litigation Involving U.S. Patent No. 10,637,968 ("the '968 Patent")

Patent identification (as given): U.S. Patent No. 10,637,968 B2, "System Providing Faster and More Efficient Data Communication," application no. 16/396,696, inventors Derry Shribman and Ofer Vilenski, priority date Oct. 8, 2009, granted April 28, 2020. Original assignee Luminati Networks Ltd.; now associated with Bright Data Ltd. (Web Spark Ltd./Luminati/Bright Data chain).

I located the following matters referencing this specific patent number ('968). I did not find any Court of Appeals for the Federal Circuit appeal decided on the merits of the '968 patent itself.

1. District Court — E.D. Texas

A. Luminati Networks Ltd. v. Code200, UAB, Oxysales, UAB, and Metacluster LT, UAB

• Case No.: 2:19-cv-00396-JRG (E.D. Tex., Marshall Division)
• Plaintiff: Luminati Networks Ltd. (later Bright Data Ltd.)
• Defendants: Code200, UAB; Oxysales, UAB; Metacluster LT, UAB
• Filed: December 6, 2019 (original complaint asserted the '511 patent); the '968 patent was added by Amended Complaint on June 9, 2020 (Dkt. 26)
• Jurisdiction: U.S. District Court, Eastern District of Texas (Judge Rodney Gilstrap)
• Status/Outcome: Litigated through claim construction — the court issued a claim construction order on Feb. 8, 2021 (Dkt. 97) construing terms of the '511 and '968 patents, and denied defendants' § 101 Rule 12(b)(6) motion. I could not confirm a final disposition (verdict, settlement, or dismissal) from the sources retrieved. A related later filing, Code200, UAB et al. v. Bright Data Ltd., appears on litigation trackers, consistent with subsequent declaratory/related proceedings.

B. Luminati Networks Ltd. v. NetNut Ltd.

• Case No.: 2:20-cv-00188-JRG-RSP (E.D. Tex., Marshall Division)
• Plaintiff: Luminati Networks Ltd. (later Bright Data Ltd.)
• Defendant: NetNut Ltd.
• Filed: June 11, 2020
• Jurisdiction: U.S. District Court, Eastern District of Texas (Judge Rodney Gilstrap; Magistrate Judge Roy S. Payne)
• Asserted: '968 patent (claims 1, 2, 8, 9, 11, 12, 15, 17, 18, 26, 27, 28) and '511 patent, plus trade-secret claims
• Status/Outcome: Terminated December 10, 2021. The AO-120 notice (Dkt. 234) reflects that all claims and causes of action asserted by Luminati against NetNut were dismissed without prejudice under Fed. R. Civ. P. 41(a)(2).
• Related: NetNut separately sought ex parte reexamination of the '968 patent (request Jan. 12, 2021; granted by the USPTO March 3, 2021).

C. Luminati Networks Ltd. v. Tefincom S.A. (d/b/a NordVPN)

• Case No.: 2:19-cv-00414-JRG (E.D. Tex.)
• Plaintiff: Luminati Networks Ltd. (later Bright Data Ltd.)
• Defendant: Tefincom S.A. d/b/a NordVPN
• Filed: December 31, 2019; the '968 patent was added via Amended Complaint on November 12, 2020 (Dkt. 22)
• Jurisdiction: U.S. District Court, Eastern District of Texas (Judge Rodney Gilstrap)
• Status/Outcome: Case docket indicates status "Closed" (a final civil docket dated Sept. 7, 2022 is listed). The precise terminating instrument (settlement vs. dismissal vs. judgment) was not confirmed in the retrieved records.

2. PTAB — Inter Partes Review

Teso LT, UAB, Code200, UAB, Metacluster LT, UAB, and Oxysales, UAB v. Luminati Networks Ltd.

• Proceeding No.: IPR2021-00249
• Patent challenged: U.S. Patent No. 10,637,968 (application 16/396,696)
• Petitioner(s): Teso LT, UAB et al. (with Code200 UAB, Metacluster UAB, Oxysales UAB as additional petitioners/POA filers)
• Patent Owner: Luminati Networks Ltd.
• Filed: December 7, 2020
• Outcome: Institution Denied — decision dated June 2, 2021; notice of refund issued July 21, 2021. (Panels included Judges Russell Cass, Sheila McShane, and Thomas Giannetti; Tech Center 2400, Art Unit 2459.)
• Source: https://ipverse.greyb.com/ptab-web/cases/case-details/IPR2021-00249 and https://www.docketalarm.com/cases/PTAB/IPR2021-00249/Teso_LT_UAB_v._Luminati_Networks_LTD/

3. Related matter (context, not a '968 patent suit)

• Teso LT, UAB v. Luminati Networks Ltd., No. 2:20-cv-00073 (E.D. Tex.) — a business-torts action (false advertising/unfair competition) brought by Teso against Luminati; referenced in PTAB papers as a related district court case. This is a counter-action to the patent suits above and does not assert the '968 patent.

Summary Table

Caveats / Confidence Notes

• I confined results to records referencing US 10,637,968 / application 16/396,696 specifically; I excluded filings referencing sibling patents (e.g., 10,257,319; 10,484,510; 10,484,511; 10,469,614; 8,560,604), which were asserted in separate cases (e.g., the Teso LT 2:19-cv-00395 action) but are different patent numbers.
• Outcomes for the Code200 and Tefincom cases are not fully confirmed from the sources retrieved; I am explicitly flagging this rather than inferring a result. The most authoritative confirmations I obtained were: the NetNut case dismissal (via the E.D. Tex. AO-120, Dkt. 234) and the IPR2021-00249 denial of institution.
• The Google Patents family record lists only three E.D. Tex. case links for this patent (2:20-cv-00188, 2:19-cv-00414, 2:19-cv-00396) plus the IPR2021-00249 proceeding — consistent with the matters above.
• Primary source pages consulted: https://patents.google.com/patent/US10637968/en ; https://portal.unifiedpatents.com/litigation/Texas%20Eastern%20District%20Court/case/2%3A19-cv-00414 ; https://www.courtlistener.com/docket/17241227/luminati-networks-ltd-v-netnut-ltd/ ; https://ipverse.greyb.com/ptab-web/cases/case-details/IPR2021-00249

If you'd like, I can attempt to pull the specific terminating docket entries for the Code200 and Tefincom cases to close the two open outcome questions.

I'll research the PTAB proceedings on US 10,637,968. Let me start with searches.

Let me dig deeper into the IPR2021-00249 details and check for related proceedings on the patent family.

Let me get the specifics of the institution decision for IPR2021-00249.

I have enough from the dockets, PTAB filings, and litigation records to answer, though the step limit stopped me short of pulling the full text of the institution decision. Here is the analysis, with the gaps flagged explicitly.

Proceedings overview

One (1) AIA trial proceeding has been filed against US 10,637,968: IPR2021-00249. Breakdown by status: 0 active / 0 claims invalidated / 0 claims sustained / 0 settled / 1 institution denied. (Note the discrepancy with the structured "PTAB proceedings on file" block you provided, which says the ODP API returns no AIA trials — that is a data gap in the ODP ingest, not an absence of activity. Web-sourced PTAB dockets, the Google Patents "Family has litigation" tab, and multiple PTAB mandatory notices independently confirm IPR2021-00249 on this patent.)

Bottom-line defensive posture: The '968 patent has never had a claim canceled, narrowed, or even adjudicated on the merits by the PTAB. Its single IPR was denied institution on 2021-06-02 before trial. That is not the same as "the patent survived two IPRs and is hardened" — the denial appears to have been discretionary (§ 314(a) / Fintiv), not a merits victory — but it does mean (a) every claim of the '968 is presumptively intact, and (b) no statutory estoppel attached, so the door is technically still open to a future IPR on different art. For a defendant today, the patent is un-narrowed but also un-hardened: there is no FWD, no Federal Circuit affirmance, and no § 315(e) estoppel to lean on.

IPR2021-00249 — Teso LT, UAB; Code200, UAB; Metacluster LT, UAB; and Oxysales, UAB v. Luminati Networks Ltd.

• Type: Inter Partes Review
• Filed: 2020-12-07
• Status: "Institution Denied" (verbatim from the PTAB docket). Plain English: the Board issued a decision on 2021-06-02 declining to institute trial; the patent never entered the trial phase, and the petitioners' post-institution fee was refunded on 2021-07-21.
• Judge panel: Administrative Patent Judges Thomas L. Giannetti, Sheila F. McShane, and Russell E. Cass (Tech Center 2400, Art Unit 2459). This is the same panel that denied the companion Luminati-family IPRs.
• Petition grounds: The petition challenged the '968 patent. Patent Owner's preliminary-response expert, Dr. V. Thomas Rhyne, analyzed two principal references: MorphMix — A Peer-to-Peer-based System for Anonymous Internet Access (Ex. 1013) and U.S. Pub. No. 2008/0196098 to Cottrell (Ex. 1017), supported by a battery of RFCs (RFC 1945, 1035, 1034, 1939, 937, 918, 821, etc.) filed as exhibits. The precise statutory basis (§ 102 vs. § 103) and the exact claim numbers challenged were not confirmed from the materials I retrieved — do not assume claim 1. This should be verified against the petition in PTAB E2E.
• Institution decision: Denied — 2021-06-02. I was unable to retrieve the four corners of the decision text before research was cut off. Based on the identical posture of the four companion petitions denied by this same panel — IPR2021-00122 ('511 patent), IPR2020-01358 ('510 patent), IPR2020-01266 ('319 patent), and IPR2020-01506 ('614 patent), all denied by the Board exercising its discretion under 35 U.S.C. § 314(a) (Apple v. Fintiv) in view of the advanced parallel E.D. Tex. litigation — the '968 denial was very likely a § 314(a) discretionary denial as well, not a merits ruling. Flag: I could not verify this specific reasoning; confirm by pulling Paper 12 (the institution decision) from PTAB E2E before relying on it.
• Final Written Decision: None issued — no trial was instituted, so no FWD, no claim-level verdict, and no claims canceled or sustained.
• Settlement / termination: Not a settlement. The proceeding ended by denial of institution; no termination-on-settlement entry appears on the docket.
• Appeal: None. Denial of institution is not appealable to the Federal Circuit under 35 U.S.C. § 314(d) (and no appeal was taken). The later consolidated Federal Circuit appeals involving this patent family — lead case Bright Data Ltd. v. Code200, UAB, No. 2023-2144 (Fed. Cir. judgment entered 2025-08-01) and the related appellate/Supreme Court activity — arise from different patents in the family (e.g., the '319, '510, and '614), not from the '968, because the '968 IPR never reached a FWD to appeal.
• Defensive value: Because institution was denied, no § 315(e)(2) estoppel attached — the Teso/Oxylabs entities are not barred from re-asserting MorphMix/Cottrell-type art in district court. Conversely, the patent owner cannot point to a merits win; the denial was procedural. The practical takeaway: IPR2021-00249 is not a shield for an accused infringer, but it is also not proof the '968 claims are valid.

Strategic summary

Claim status on the '968. Every claim is UNTESTED at the PTAB — no independent or dependent claim has been canceled, narrowed by disclaimer, or confirmed. There are zero CANCELED claims and zero SUSTAINED (FWD-confirmed) claims, because IPR2021-00249 died at the institution stage. Contrast this with sibling family patents: petitions against the '319 and '510 patents (IPR2021-01492 and IPR2021-01493, filed by NetNut) were instituted and reached final written decisions, and those FWDs and their claim-construction fallout were affirmed by the Federal Circuit on 2025-08-01 in the consolidated appeals (Bright Data v. Code200, 2023-2144 et al.). That appellate invalidation is on other patents, not the '968 — do not conflate them. The '968 thus stands as an untouched, full-scope asset.

Estoppel landscape. § 315(e)(2) estoppel is triggered only by a final written decision, 35 U.S.C. § 315(e)(2); a denial of institution creates no estoppel. So for a defendant being asserted against on the '968 today: the MorphMix/Cottrell and RFC-based grounds the Teso group raised in IPR2021-00249 remain fully available as invalidity grounds in litigation (subject only to the usual IPR-timing dynamics and the district court's discretion, not statutory estoppel). Conversely, because the patent owner won a denial, the patent owner is not estopped from anything either. The cleanest read: there is no estoppel on either side from this proceeding.

Pattern signals. Yes — there is a clear, coordinated, multi-front pattern:

• The same petitioner bloc — Teso LT, Code200, Metacluster LT, and Oxysales (all part of the Oxylabs / coretech corporate family) — filed five IPRs against five Luminati/Bright Data patents within roughly six months (2020-07-14 through 2020-12-07): IPR2020-01266 ('319), IPR2020-01358 ('510), IPR2020-01506 ('614), IPR2021-00122 ('511), and IPR2021-00249 ('968). All five were denied institution.
• A second challenger, BI Science (2009) Ltd., filed IPR2020-00166 and IPR2020-00167, both of which terminated on settlement (2020-04-28) — a § 315(b)-adjacent signal that the patent owner actively settles to avoid PTAB scrutiny.
• A genuine defensive-aggregator presence: Unified Patents is listed as the petitioner-side actor on IPR2021-00249's PTAB-adjacent record ("Petitioner: Unified Patents PTAB Data"), and the patent's public record carries the Unified Patents / PTAB and Darts-ip litigation tags. The patent owner (Bright Data, f/k/a Luminati) has litigated this family aggressively across the E.D. Tex. docket (2:19-cv-395, -396, -397, -414; 2:20-cv-00188; 2:22-cv-00011) and pursued PTAB and appellate review hard on the sibling patents, up to a U.S. Supreme Court petition (No. 25-779) on claim construction — signaling a well-funded, litigation-experienced owner that does not abandon fights.

The practical cross-proceeding picture for the '968: it is the least PTAB-exposed member of the family. Every other aggressively-asserted sibling has either been through a real IPR or been the subject of a settled IPR — the '968 alone escaped with a bare denial. That makes it both free of fatal PTAB holdings and free of helpful ones.

Recommended next steps

• Pull the actual record. Fetch the IPR2021-00249 institution decision (Paper 12, dated 2021-06-02) and the original petition from PTAB E2E (https://ptab.uspto.gov) or via the Docket Alarm docket (https://www.docketalarm.com/cases/PTAB/IPR2021-00249/Teso_LT_UAB_v._Luminati_Networks_LTD/). Confirming (i) the exact challenged claims, (ii) the § 102/§ 103 basis, and (iii) whether the denial was § 314(a)/Fintiv or merits-based is the single highest-value open item — my analysis flags the § 314(a) basis as the probable but unverified conclusion.
• If you are a defendant being asserted on the '968: there is no FWD to link to and no claim cancellation to quote. The absence is itself the signal: this patent has never been invalidated, but it also has never been defended through a full PTAB trial, and no estoppel bars you from filing your own IPR on art the Teso bloc did not use. Consider (a) whether the identical § 314(a)/Fintiv factors that killed the 2021 petition would now cut the other way if the parallel litigation posture has changed, and (b) art beyond MorphMix/Cottrell.
• If you are the patent owner / a licensee defending the '968: the favorable events are (1) the 2021-06-02 denial of institution and (2) the E.D. Tex. jury verdict in the Teso litigation finding related asserted patents valid and infringed (per the PTAB mandatory notice in IPR2022-00936). But be candid that the '968 itself has no PTAB merits win; if you cite the family's Federal Circuit story, cite it as to the '319/'510/'614 patents (2023-2144 et al., judgment 2025-08-01), not the '968.
• No live milestones to track. There is no pending institution-deadline, oral hearing, or statutory one-year FWD deadline on the '968, because no trial is pending. Any new IPR you file would start that clock fresh.
• Do not fabricate a second proceeding. I found exactly one AIA trial on US 10,637,968. The other Luminati/Bright Data IPRs surfaced in the same searches (IPR2020-00166/-00167, IPR2020-01266, IPR2020-01358, IPR2020-01506, IPR2021-00122, IPR2021-01492/-01493, the IPR2022-xxxx series, PGR2022-00061, etc.) are directed to other patents in the family and must not be attributed to the '968.

I'll research the assignment record for US 10637968, starting with the known legal events and then digging into USPTO records, litigation, and corporate history.

Let me dig into the corporate history (Hola Networks / Web Spark / Luminati) and try to find the actual assignment reel/frame records.

I have the corporate picture. Now let me find the specific reel/frame records and correspondents, and confirm the parent-patent ownership chain.

I was able to reconstruct the chain from the patent's own legal-events record plus PTAB, litigation, and trade-press sources, but I could not retrieve the actual USPTO reel/frame numbers or correspondent-of-record fields for this patent from the sources reachable to me — those fields are only exposed in the Assignment Center itself. I flag each such gap explicitly rather than guessing.

Inventors

Pattern notes: No unusual inventor-departure pattern. Both inventors co-founded Hola Networks (founded 2012) and are the only two named inventors on this entire family; they retained equity and operational roles through the 2017 EMK Capital majority-stake sale (Vilenski continued as Luminati CEO). This is the opposite of the "inventors gone within 12 months → portfolio fire-sale" tell. Note, however, that the invention's priority chain reaches back to a 2009 provisional (61/249,624), i.e., 3 years before Hola Networks is generally reported as founded — so attributing the 2009 conception to a Hola employment relationship is not supported.

Original assignee

There is a database conflict worth stating plainly:

• On the recorded assignment and sibling family members: Web Spark Ltd. (Netanya, IL) is the assignee/applicant of record on this family — e.g., US 10,523,788 and US 10,387,316 both name "WEB SPARK LTD., Netanya (IL)" as applicant (71) and assignee (73). The first recorded event on the '968 is the inventors assigning to Web Spark Ltd. on the filing date (2019-04-28).
• Google Patents auto-lists the "Original Assignee" for US 10,637,968 as Luminati Networks Ltd., and Unified Patents lists the original assignee as Hola Networks Ltd. (parent companies: Bright Data Ltd / Web Spark Ltd). These are aggregator-derived labels, not reel/frame-grounded, and I could not reconcile them against a primary document.

Business of the corporate group: Hola Networks Ltd. built a peer-to-peer VPN/browser extension; its enterprise proxy division was split out in 2014 as Luminati Networks Ltd., an IP-proxy / residential-proxy network for businesses (ad verification, brand protection, price comparison, data collection). UK private-equity firm EMK Capital LLP acquired a majority stake in Luminati in August 2017 at a reported ~$200M enterprise value; Hola retained a standalone VPN/CDN business. Luminati was renamed Bright Data Ltd. in March 2021.

Current status: Operating (private). Bright Data is an active commercial web-data/proxy platform with tens of thousands of paying customers, and it is a litigant in both directions — it has sued proxy competitors (Oxylabs/Teso, NetNut, BI Science, Code200) and it has been sued by Meta and X Corp. on scraping/contract theories. This is a live operating company, not a dissolved or bankrupt shell.

Assignment timeline

Field-availability caveat: The sources reachable to me (Google Patents "Legal Events," Unified Patents, Espacenet, PTAB papers) expose the assignment events below but not the reel/frame numbers or the correspondent-of-record. I did not fabricate those. Retrieve them at the USPTO Assignment Center by searching patent 10,637,968: https://assignmentcenter.uspto.gov/

2019-04-28 (executed on/at filing) / recorded 2019-04-28 — Reel/Frame not retrievable from reachable sources

• Conveyance: Assignment
• Assignor: Derry Shribman and Ofer Vilenski (the inventors)
• Assignee: Web Spark Ltd. (Netanya, Israel)
• Correspondent: not retrievable — no correspondent data in any source I could reach
• Context: Original inventor-to-company assignment, executed on the filing date (original acquisition).

2019-12-05 / recorded 2019-12-05 — Reel/Frame not retrievable

• Conveyance: Assignment
• Assignor: Web Spark Ltd.
• Assignee: Luminati Networks Ltd.
• Correspondent: not retrievable
• Context: Intra-group consolidation — IP moved from the Web Spark holding entity to the operating brand (Luminati). Note the timing: recorded one day before the Luminati v. Code200/Teso/Oxylabs complaints were filed on 2019-12-06 (E.D. Tex. 2:19-cv-395/-396/-397), consistent with housekeeping to tidy the standing record before assertion.

2021-03-31 / recorded 2021-03-31 — Reel/Frame not retrievable

• Conveyance: Change of Name
• Assignor: Luminati Networks Ltd.
• Assignee: Bright Data Ltd.
• Correspondent: not retrievable
• Context: Change of name only — the March 2021 corporate rebrand. No change in beneficial ownership.

Not found: no Security Agreement, License, Merger, Release, or Correction records surfaced; and no transfer of this patent outside the Hola/Web Spark/Luminati/Bright Data corporate group was found.

Family context (not reel/frame-verified): this patent is a continuation of 14/025,109 (filed 2013-09-12), which is a divisional of 12/836,059 (filed 2010-07-14, issued as US 8,560,604), claiming priority to provisional 61/249,624 (2009-10-08). US 8,560,604 was the subject of a 2018 pre-suit licensing demand to Oxylabs/Teso (referenced in Teso's June 20, 2018 letter) but was not asserted in that litigation.

Timeline diagram

timeline
    title Ownership of US 10637968
    2009 : Priority provisional filed
    2019 : Continuation filed
         : Assigned to Web Spark Ltd
         : Assigned to Luminati Networks Ltd
         : Code200 suit filed in December
    2020 : Patent issued
    2021 : Renamed Bright Data Ltd

NPE / troll-pattern signals

1. Shell-entity transfer — not present. The chain runs Web Spark Ltd. → Luminati Networks Ltd. → Bright Data Ltd. (change of name). None carries an "IP / Patents / Licensing / Holdings / Ventures" suffix; all are tied to a real operating business at a commercial address (Netanya/Tel Aviv, Israel). Web Spark Ltd. is a mild anomaly — a holding-name entity that is not the consumer brand — but there is no evidence (no registered-agent address, no single-member LLC, no licensing-only posture) to call it a shell. Not a finding on its own.

2. Known asserter in the chain — not present. No link in the chain matches the named NPE list (Acacia, Marathon, IV, IPNav, Wi-LAN, Mosaid/Conversant, Vringo, Pendrell, Round Rock, MPHJ, Lumen View, Spangenberg entities, etc.). The current owner, Bright Data Ltd., appears in Unified Patents and RPX dockets as an operating-company plaintiff and as a defendant (Meta, X Corp.), not as a high-frequency troll plaintiff. It does not appear on asserter watchlists.

3. Repeat correspondent across the chain — unclear / not retrievable. This is the strongest diagnostic the prompt asks for, and it is exactly the field I could not obtain. No correspondent-of-record name is exposed in any reachable source. I will not infer one from litigation counsel (e.g., the various E.D. Tex. plaintiff's counsel of record are not the same thing as assignment correspondents). Action item for the reviewer: pull the correspondent field for the three records at https://assignmentcenter.uspto.gov/ and check the 2019-12-05 and 2021-03-31 entries against each other.

4. Cascading transfers — not present. Three recorded events over ~2 years, all inside one corporate group; no chained single-purpose LLCs, no transfers across unrelated names, no shared registered-agent addresses evident.

5. Pre-litigation transfer — present but weak/ambiguous. The Web Spark → Luminati assignment was recorded 2019-12-05, one day before the 2019-12-06 complaints (2:19-cv-395/-396/-397). This looks like deliberate sequencing to clean up standing before filing. However, the assignee is the operating group's own entity, so it is not NPE-style "arrange the chain to enable assertion" behavior — it is portfolio hygiene inside a single company.

6. Bankruptcy fire-sale — not present. No bankruptcy, no Chapter 7/11, no proceedings-based sale found for any chain entity.

7. Privateering — not present. Bright Data asserts the family directly and in its own name against direct competitors (Teso/Oxylabs, NetNut, BI Science, Code200/Metacluster/Oxysales), and pairs the patent counts with non-patent claims (trade-secret, tortious interference) — the signature of a market competitor dispute, not of an operating company farming patents out to an NPE. No SEC filing or Patent Progress/EFF coverage indicates privateering.

8. Defensive aggregator — not present. The chain terminates at an operating company, Bright Data Ltd.; it does not end at RPX, AST, LOT, Unified Patents, or OIN. The patent is live offensive ammunition, not neutralized. (It has been challenged — IPR2021-00249 filed by Teso LT, UAB, not instituted (procedural); and it is in pending ex parte reexamination Control No. 90/014,816 — but that is third-party invalidity attack, not defensive acquisition.)

Verdict

Operating-company assertion.

The chain is an intra-group ownership sequence — Web Spark Ltd. → Luminati Networks Ltd. (2019-12-05) → Bright Data Ltd. (change of name, 2021-03-31) — terminating at a live operating company, not an NPE or defensive aggregator. Bright Data is a $200M-enterprise-value proxy/web-data business (EMK Capital majority stake, 2017) that asserts the family patents directly against competitors (the '968 was asserted against Code200/Metacluster/Oxysales in E.D. Tex. 2:19-cv-00396), and it litigates as a defendant as well (Meta, X Corp.) — the behavioral profile of a market participant, not a patent monetizer. Two cautions for the record: (i) the Board's Final Written Decision on sibling patent 11,044,344 (IPR2022-00353) found Bright Data failed to establish a nexus between its products and the claims, so "practices the claims" is not affirmatively established; and (ii) the reel/frame and correspondent fields could not be retrieved, so signal #3 (repeat correspondent) remains open.

Verify at: https://assignmentcenter.uspto.gov/ (search patent number 10,637,968) — retrieve the three records' reel/frame and correspondent entries to close out signal #3.

I'll research US patent 10637968, focusing on its prior art citations. Let me run several searches to gather the relevant information.

Let me dig deeper into the specific prior art citations and the IPR petition against this patent.

Let me look at the exact prior-art references cited against this patent and the IPR petition.

Prior-Art Analysis — US 10,637,968 B2

0. Retrieval limitation (important, stated up front)

I was able to confirm the identity and legal/procedural record of US 10,637,968 B2. However, I was not able to retrieve, from the sources available to me, the two documents that would give a complete, authoritative "each patent citation" list for this patent:

1. The examiner's PTO‑892 "Notice of References Cited" / IDS "References Cited" list on the face of US 10,637,968 B2 itself; and
2. The exhibit list / ground-by-ground prior-art mapping from the IPR2021‑00249 petition (Unified Patents) directed at this patent.

Under my operating rules I will not fabricate patent numbers, citations, or claim mappings. So below I give (A) what is verified, (B) the prior art that is documented in the record I could reach (the patent's own background section and the family-level third‑party‑submission references), and (C) an explicit statement of what I could not verify and therefore will not assert. Where I give a reference, I give its citation and dates as found, and I flag the § 102 caveat rather than asserting a definite anticipation.

1. Patent identification (verified)

Continuation chain (from the specification): US 16/396,696 is a continuation of US 14/025,109 (filed 2013‑09‑12), which is a divisional of US 12/836,059 (filed 2010‑07‑14, issued as US 8,560,604 on 2013‑10‑15), which claims priority to provisional 61/249,624 (filed 2009‑10‑08). All share the 2009‑10‑08 priority date. Source: https://patents.google.com/patent/US10637968/en

Procedural/legal record (verified from the patent page):

• PTAB case IPR2021‑00249 (Petitioner: Unified Patents) — status: Not Instituted (Procedural). https://portal.unifiedpatents.com/ptab/case/IPR2021-00249
• U.S. District Court litigation, E.D. Tex.: cases 2:19‑cv‑00396, 2:19‑cv‑00414, and 2:20‑cv‑00188.
• Reassignments: Web Spark Ltd. (2019‑04‑28) → Luminati Networks Ltd. (2019‑12‑05) → Bright Data Ltd. (2021‑03‑31, change of name).

2. A critical § 102 threshold: the effective priority date

Because US 10,637,968 B2 claims priority to 2009‑10‑08, the prior-art date that governs 35 U.S.C. § 102(a)/(b)/(e) for the claimed subject matter is the 2009‑10‑08 priority date (assuming the continuation chain is entitled to it). This has a decisive consequence:

• A reference is § 102 prior art only if it was published/patented/publicly available before 2009‑10‑08.
• Any reference dated after 2009‑10‑08 is not § 102 prior art for this patent unless the priority claim is successfully challenged (e.g., intervening prior art under § 102(a)(2) / AIA, or a priority break).

This matters because several of the references that appear in this patent family's prosecution record carry dates in 2010–2013, i.e., after the '968 priority date (see § 3.2 below). Those would be relevant to the related, later-filed family members, not to the '968's own priority-based prior art.

3. Prior art documented in the record I could reach

3.1 Prior art discussed inside the '968 specification (background/§ 102(b)-style disclosures)

The specification expressly frames the invention against two categories of prior art, with figures:

I note these are described and distinguished by the applicant — they are the closest admitted background art but are not, on their face, mapped by the applicant to specific claims as anticipating references. I could not verify that any of them appears on the '968's PTO‑892 list.

3.2 Family-level third-party-submission references (37 C.F.R. § 1.290)

The justia record for the related family member US 11,233,879 ("System providing faster and more efficient data communication," same Shribman/Vilenski family, Bright Data/Luminati) lists references that were the subject of multiple third-party submissions under 37 C.F.R. § 1.290 filed July 22–25, 2019 in related applications (e.g., US App. Nos. 16/140,749; 16/140,785; 16/214,433/451/476/496; 16/292,363/364/374/382; 16/365,250; 16/365,315). Source: https://patents.justia.com/patent/[11233879](/patent/11233879)

These references are family-level; I could not confirm they were specifically entered against US 16/396,696 (the '968 application) or mapped to specific '968 claims. Presenting them with that caveat:

Note: several of these span dates around the 2009‑10‑08 line, and the § 102 treatment of each depends on an evidentiary determination (public accessibility, exact date, and whether the priority chain holds) that I cannot make from the search results alone.

4. What I could NOT verify (and therefore do not assert)

• The '968's own backward patent-citation list (PTO‑892 / IDS "References Cited"): I did not retrieve it. I therefore cannot give a definitive "each patent citation" table with full citations, publication/filing dates, and claim-by-claim § 102 mapping for US 10,637,968 B2 as requested.
• The IPR2021‑00249 petition's prior-art grounds. My searches returned an unrelated PTAB document (about an unrelated "…329 Patent"), not the Unified Patents petition papers for the '968. I could not confirm the specific references (EX1001–EXxxxx) or their claim mappings.
• The '968's exact issued claim set. The full text I obtained ends before the claims; I have the Summary/abstract claim concepts but not the numbered claims. Without the numbered claims, any precise "which claim(s) it anticipates under § 102" statement would be speculative, so I decline to make one.

5. Honest bottom line

• Verified: US 10,637,968 B2 is real and correctly identified; 2009‑10‑08 priority; the family dates, IPR, and litigation record above are accurate.
• Documented prior-art context: The most directly relevant prior art in the record I could reach is (i) the proxy-caching art (generic proxy servers / Akamai) and (ii) the P2P file-transfer art (BitTorrent), both of which the applicant itself discusses and distinguishes in the specification, plus (iii) the family-level § 1.290 proxy/hide-IP references listed in § 3.2.
• Most of that art is better characterized as § 103 background than as § 102 anticipation, because the '968 claims (per its Summary) are directed to a specific combination — a client that intercepts requests, an acceleration server that assigns an agent by destination-server IP, an agent that tracks which clients received responses for its assigned server, and peers that store and serve data portions/chunks by checksum — a combination the generic proxy and BitTorrent references do not, on the material I saw, each disclose in a single reference.
• To complete the requested per-citation § 102 mapping, I would need the '968's PTO‑892/IDS citation list and the IPR2021‑00249 petition exhibits, which I could not retrieve here. I recommend pulling (a) the patent's "References Cited" from USPTO PatentCenter/Global Dossier, and (b) the IPR2021‑00249 papers, before finalizing any § 102 anticipation conclusions.

If you can supply the PTO‑892 citation list or the IPR petition exhibit list, I can complete the full reference-by-reference, claim-by-claim § 102 table precisely rather than in the provisional form above.

I'll start by trying to recover the prior-art references associated with this patent, since the page text supplied to me was truncated before any "Citations"/"Prior Art" listing.

The provided page text was truncated before any "Citations"/"Prior Art" section, so I reconstructed the prior-art set from the IPR exhibit list and will verify it. Let me get the actual claim language and the petition's grounds.

Obviousness Analysis — U.S. Patent No. 10,637,968 B2

"System providing faster and more efficient data communication"
Assignee of record: Web Spark Ltd. / Bright Data Ltd. (originally Luminati Networks Ltd.) · Inventors: Derry Shribman, Ofer Vilenski · Priority: 2009‑10‑08 · Filed: 2019‑04‑28 · Granted: 2020‑04‑28 · Anticipated expiration: 2030‑07‑14

0. Scope and evidence caveats (read this first)

Three verification problems materially affect this analysis, and I am flagging them rather than papering over them:

1. The "Prior Art" section you asked me to use was not present in the material supplied to me. The full text of the Google Patents page that was pasted into this task is truncated mid‑sentence at paragraph [0358] of the description ("In response to receiving the IP address of the server 152, the acceleration server"). The page as delivered therefore contains no Citations, Cited By, Similar Documents, or "Prior Art" listing. There are likewise no claims in the supplied text. I could not retrieve a Google Patents citation list in the searches available to me.

2. Consequently, I reconstructed the prior‑art set from the only authoritative § 103 challenge on record against this patent — the inter partes review petition exhibit list, retrieved at:

• https://www.docketalarm.com/cases/[PTAB](/ptab)/IPR2021-00249/Inter_Partes_Review_of_U.S._Pat._10637968/docs/12-07-2020-Petitioner/Petition-5-Petition_for_Inter_Partes_Review_of_US_Patent_10,637,968.pdf
• https://www.docketalarm.com/cases/PTAB/IPR2021-00249/Teso_LT_UAB_v._Luminati_Networks_LTD/
  That docket was filed 2020‑12‑07 and is captioned Teso LT, UAB v. Luminati Networks Ltd. Status: Institution Denied ("Not Instituted ‑ Procedural" per the Google Patents page). Judges: Cass, McShane, Giannetti.

1. Identifier discrepancy, reported literally and not corrected: the Google Patents page designates the PTAB case IPR2021‑00249, whereas the petition PDF's own caption text reads "Case IPR2020‑00249." Both strings exist in the record as retrieved. Similarly, the Google Patents page attributes the IPR data to petitioner "Unified Patents PTAB Data," while the petition itself names "Code200, UAB … and Oxysales, UAB" as Petitioners and the docket caption names Teso LT, UAB. I am not resolving any of these discrepancies.

2. I was unable to retrieve the substantive content of two asserted references — US 2006/0059091 (Wang) and US 7,007,228 (Carro) — before exhausting my search budget. I therefore treat them as documented‑but‑unverified and do not rely on them in the grounds below.

1. The asserted prior‑art set

From the IPR exhibit list (Ex. 1012–1037), the § 103 references are:

All of these predate the 2009‑10‑08 priority date, with the exception of RFC 2616 being pre‑priority as well (1999). Everything in the set qualifies as prior art.

Note also the specification's own admissions: FIG. 1 (proxy/cache servers, e.g. Akamai) and FIG. 2 (peer‑to‑peer file sharing, BitTorrent by name) are expressly described as prior art, as is the "Web2.0"/dynamic‑content problem. Under settled law, the patent's own background admissions are usable as prior art and as evidence of the state of the art and of the problem confronting the inventor.

2. Governing framework

• Pre‑AIA 35 U.S.C. § 103(a) applies. The '968 is a continuation of Ser. No. 14/025,109 (filed 2013‑09‑12), itself a divisional of Ser. No. 12/836,059 (filed 2010‑07‑14, issued as US 8,560,604), claiming priority to provisional 61/249,624 (2009‑10‑08). Because the family never contained a claim with an effective filing date on or after 16 March 2013, the pre‑AIA §§ 102/103 regime governs. Practically this matters mainly for § 102(e)/§ 103(c) treatment of the Cottrell publication.
• Graham v. John Deere, 383 U.S. 1 (1966): (1) scope/content of the prior art; (2) differences from the claims; (3) level of ordinary skill; (4) objective indicia.
• KSR Int'l v. Teleflex, 550 U.S. 398 (2007): the combination of familiar elements according to known methods is obvious when it does no more than yield predictable results; "if a technique has been used to improve one device, and a person of ordinary skill in the art would recognize that it would improve similar devices in the same way, using the technique is obvious."
• Level of ordinary skill (POSITA): a bachelor's degree in computer science or electrical engineering (or equivalent) plus two to three years of experience with Internet networking, HTTP/TCP‑IP, proxy caches, and distributed/peer‑to‑peer systems; or a master's degree plus one year. This is the level reflected in the '968 specification's own detail and is the level Petitioner's expert (Dr. Michael Freedman, Ex. 1011) was evidently retained at.

3. Reconstructed claim scope

The supplied text contains no claims, so I reconstruct from (a) the petition's own list of challenged claims — claims 1‑2, 8‑9, 11‑12, 15‑18, 24, 26‑28 in Ground 3 — and (b) the near‑identical sibling patents in the same family whose claim 1s were reproduced in the corresponding complaints (see the complaints retrieved from brightdata.com/static/patent/ and luminati.io/static/patent/):

• Sibling '319 (Ser. No. 15/957,945), claim 1 — method by a first client device: receiving, from the second server, the first content identifier; sending, to the first server over the Internet, an HTTP request that comprises the first content identifier; receiving the first content from the first server over the Internet in response to the sending of the first content identifier; and sending the first content by the first client device to the second server, in response to the receiving of the first content identifier.
• Sibling '510 (Ser. No. 16/278,107), claim 1 — same, plus establishing a TCP connection with the second server and sending the received content over that connection.

The '968 shares the same specification (it is the same disclosure, with FIGS. 1‑15 as reproduced in the supplied text). Its independent claim is therefore best characterized as a method/system in which a first client device is instructed by a second server (the "agent") which content identifier (URL) to fetch, fetches it over the Internet from an HTTP web server, and returns the fetched content to the second server — with dependent claims adding the proxy/agent‑selection, validity‑checking, chunking/checksum, peer‑list, and IP‑address‑assignment features described in the specification.

This breadth is the crux of the § 103 problem. The claim, so construed, is essentially "an intermediate device fetches a URL that a controlling server told it to fetch, and hands the bytes back."

4. Grounds of combination

Ground 1 — Cottrell in view of RFC 2616 and the general knowledge of a POSITA

(This is Ground 3 of the petition verbatim: "OBVIOUSNESS OF CLAIMS 1‑2, 8‑9, 11‑12, 15‑18, 24, 26‑28 BY COTTRELL + RFC 2616 + GENERAL KNOWLEDGE")

Cottrell (US 2008/0196098 A1 / US 8,375,434 B2) discloses a system "for protecting identity of network devices (102, 104, and 106) in a network environment" comprising "an apparatus having an interface to the network for completing connections to destination devices (152, 154, and 156) on the public side of the network," with "a masking element (140) for associating at least one masking identifier with a communication from the network device and masking the identifier of the network device from the destination device." Source: https://patentimages.storage.googleapis.com/59/18/89/d4dcf259dbff4f/US8375434.pdf

Element mapping:

Motivation articulated under KSR: Cottrell expressly frames the invention as concealing the identity of the requesting machine from the destination. Its architecture necessarily requires a device other than the requester to be the network‑visible interlocutor of the destination. Once that architecture is chosen, using HTTP — the dominant application protocol of the Web as of 2004‑2008 — as the protocol of the intermediary‑to‑destination exchange is the mere application of a known protocol to a known architecture, producing nothing more than the predictable result (KSR, 550 U.S. at 416, 421). There is no disclosed or inherent incompatibility between Cottrell and HTTP; Cottrell's destination devices are precisely "the public side of the network," i.e., web servers. The motivation is therefore not merely "these are both in networking" (an impermissible generic motivation) but a specific, stated design need — concealment of the requester + retrieval of web content — that HTTP directly serves.

Weakness to concede: Cottrell's claims and abstract emphasize masking identity, not the return of fetched content to a controlling server for the server's own subsequent use. Whether the "second server" is disclosed in the Cottrell‑as‑intermediary role depends on how the Board (or a court) construes "second server." The specification's own FIG. 10/11 flow — agent sends list of peers/checksums, client fetches from peers, client notifies agent which chunks it received, agent "adds the client to the list of peers for these chunks" (block 436) — is where the real bookkeeping invention lies, and Cottrell alone does not teach that bookkeeping.

Ground 2 — Rennhard (MorphMix) in view of RFC 2616 and the general knowledge of a POSITA

MorphMix (Rennhard 2004) discloses a peer‑to‑peer anonymity system in which a central entity coordinates which peer nodes a client should use, and the client's traffic is relayed through those peer nodes ("tunnels") toward a destination, so that the destination sees the peer, not the client. This maps onto the specification's "agent"/"peer" dichotomy and onto the acceleration server's role of telling the client which other devices to use.

Element mapping:

Motivation: MorphMix establishes that it was known by 2004 to use ordinary users' machines as relay/intermediary nodes under the direction of a central coordinator. A POSITA seeking to reduce origin‑server load and ISP congestion — the very problems the '968 specification recites in its Background — would have been motivated to overlay HTTP onto that known relay topology, because the load‑shedding benefit follows directly and predictably from having a third‑party machine, rather than the origin web server, serve the bytes. RFC 2616 supplies the conditional‑request validation mechanism (If‑Modified‑Since, ETag, max‑age, no‑cache) that the '968 specification recites at FIG. 12 as the way to test "whether a certain HTTP request is still valid" — i.e., RFC 2616 alone supplies an entire claimed dependent‑claim feature (the validity check of block 510).

The same MorphMix‑plus‑RFC 2616 combination was in fact asserted against the sibling US 10,257,319 in IPR2021‑01492 (NetNut Ltd. v. Bright Data Ltd.), whose grounds included "GROUND 5: ANTICIPATION OF CLAIMS 1, 17, 19, 21‑29 BY MORPHMIX" and "GROUND 6: OBVIOUSNESS … BY MORPHMIX + RFC 2616 + GENERAL KNOWLEDGE" (Ex. 1005, Teruya Declaration). This is important: the same family member was challenged with MorphMix as a primary reference, including as an anticipation reference, which signals that MorphMix was viewed by at least one expert as disclosing the core intermediary architecture.

Ground 3 — The specification's admitted BitTorrent‑type P2P art in view of RFC 2616 and a known proxy/cache (FIG. 1 art)

The '968 specification itself admits as prior art: (i) proxy servers that "request the data from the Web servers … on their behalf, and cache the responses … to provide to other client devices that make similar requests" (FIG. 1), and (ii) peer‑to‑peer file sharing in which "each consumer can serve up data to other consumers … providing the consumer multiple points from which to download the data … thus increasing the speed of the download" (FIG. 2), naming BitTorrent expressly.

Every architectural limitation of the independent claim and most dependent claims is therefore admitted art:

• splitting content into chunks with a checksum used as the key (BitTorrent's piece/hash model, which the '968 describes at [FIG. 7] as "the checksum is the 'key' to the data");
• a central index/tracker that tells the client which peers hold which chunks (the specification concedes the BitTorrent index exists, criticizing only its scale);
• parallel download from multiple sources (conceded as the point of FIG. 2);
• caching responses at an intermediary for reuse by others (conceded as the point of FIG. 1).

Motivation: the specification states the unmet need in its own words — that a proxy solution "cannot deal well with dynamic data," that a P2P index "cannot deal with dynamic HTTP data," and that a "heretofore unaddressed need exists in the industry." Where the inventor's own background identifies two known techniques and their respective deficiencies, combining them to obtain the complementary advantages each lacks is precisely the "predictable use of prior art elements according to their established functions" that KSR holds obvious (KSR, 550 U.S. at 417). RFC 2616 supplies the HTTP framing (methods, headers, conditional GET, Range) needed to make a P2P chunk distribution scheme operate over the Web — a step that would have been routine and that yields no more than the expected result of downloading web content in pieces from several sources.

Ground 4 — Cottrell in view of MorphMix and RFC 1122 / RFC 791 / RFC 2460 (for the "acceleration server assigns agents" and IP‑proximity limitations)

The specification's FIG. 7 and its corresponding description recite that "the acceleration server assigns a list of IP addresses to each communication device functioning as an agent. Each communication device will be the agent for any Web servers whose IP address is in the range 'owned' by that communication device," and FIG. 13 recites selecting "the next 5 active agents whose IP address is closest to the IP address of the server who is targeted in the request," with the example "192.166.3.103 is closer to 192.166.3.212 than to 192.167.3.104."

• RFC 791 / RFC 1122 / RFC 2460 supply the structure of IP addresses — that they are aggregatable, hierarchical, and usable as a proximity/grouping metric, and (per RFC 791) that routing is address‑based.
• MorphMix supplies the notion of a central coordinator that distributes node lists to clients.
• Cottrell supplies the intermediary device that actually performs the fetching on behalf of the requester.

Motivation: partitioning a finite address space among servers by numerical range, and selecting the nearest server by address comparison, is a known load‑balancing/hashing technique — the '968's own CPC classification places it squarely in H04L67/1004 (server selection for load balancing) and H04L67/1023 (server selection based on a hash applied to IP addresses or costs). A POSITA building the MorphMix/Cottrell‑style intermediary system at Internet scale would have been motivated, as a matter of ordinary engineering, to shard the globe among coordinator/agent machines by IP range to bound each machine's index size. That is the classic "design incentive" rationale recognized in KSR and In re Kahn, and the result is entirely predictable. Notably, the '968 specification itself describes the mechanism as an arbitrary administrative choice ("when a first ever communication device goes online … the acceleration server assigns all IP addresses in the world to this communication device"), which is strong evidence that the assignment scheme is a conventional implementation detail rather than an inventive contribution.

Ground 5 — RFC 2616 alone in view of the general knowledge of a POSITA (targeting the dependent validity‑check claims)

For any dependent claim limited to checking whether cached content is still valid (the FIG. 12 flow: applying "max age," "no cache," and HTTP conditional requests to the request in question, then "an HTTP conditional request is sent to the relevant Web server"), RFC 2616 is anticipatory on its face. The specification concedes as much: "It should be noted, that the abovementioned description with regard to FIG. 12 is an explanation of how to check if HTTP information is still valid." A patent specification's admission that a recited step is simply how the standard protocol works is a § 103 admission that the step adds nothing patentable. See In re Baxter Travenol, 952 F.2d 388 (Fed. Cir. 1991) (a reference is prior art for all it discloses/teaches), and the general principle that a claim limitation that is coextensive with a published standard is met by that standard.

5. Reasons a POSITA would have combined these teachings (KSR rationales, consolidated)

1. Predictable use of prior‑art elements according to their established functions. Chunking + checksum keys (BitTorrent), caching at an intermediary (FIG. 1 proxy art), and HTTP semantics (RFC 2616) each perform in the combination exactly the function each performs alone; the combination yields no more than the sum of expected results.
2. Known technique to improve similar devices in the same way. Central‑coordinator‑directed relaying through user machines (MorphMix) had already been used to serve one purpose (anonymity); applying it to content delivery is the same technique toward a different, equally predictable end.
3. Obvious to try, with a finite number of identified, predictable solutions. Given the stated need (faster delivery, less origin load, less ISP congestion, tolerance for dynamic content), the candidate techniques were the two the specification itself names — proxy caching and P2P distribution — and combining them was one of a small number of predictable approaches.
4. Market/design incentive. The specification recites enormous capital costs for a pure proxy solution ("in the range of billions of dollars") and index‑maintenance costs for pure P2P ("again in the billions"). Economic pressure to avoid capital expenditure while retaining the benefits is a legitimate, articulated motivation for the combination.
5. Common sense / ordinary engineering. Selecting a server by IP proximity, sharding an index by address range, and validating a cache entry with a conditional GET are, on this record, routine implementation choices of the type KSR expressly permits to be supplied by the artisan's ordinary creativity rather than by an express reference teaching.
6. Express, not generic, linkage. Each ground pairs a reference disclosing the architecture (Cottrell's intermediary; MorphMix's coordinator‑directed relays) with a reference disclosing the protocol or mechanism (RFC 2616's HTTP; the RFC 791/1122/2460 addressing) that the architecture requires. This avoids the "generic motivation" defect of In re Nuvasive / Personal Web because the motivation is derived from the primary reference's own stated purpose.

6. The patent owner's strongest non‑obviousness arguments (and how they fare)

A. "No reference teaches the agent keeping track of which client devices have received responses."
This is the inventor's genuine point of novelty and the one place the prior art is thinnest. The FIG. 11 flow (agent "adds the client to the list of peers for these chunks, specifically, since this client is now storing these chunks") is a dynamic peer‑inventory registration keyed to content identity — functionally BitTorrent's announce/peer‑list maintenance, and, in Ground 3, arguably fully met by the admitted BitTorrent art combined with RFC 2616. But Cottrell alone does not teach it, and MorphMix's coordinator does not maintain a content‑keyed inventory. A well‑pleaded Ground 1 that does not supplement Cottrell with a tracker‑type reference is vulnerable on this limitation.

B. "The prior art teaches a 'first server' that sends a list of peers rather than performing the selecting."
This is verbatim the argument that succeeded for the applicants during prosecution of the parent '511 application, where they overcame an obviousness rejection over Yu and Kageyama "by arguing non‑combinability and that the mapped prior art 'first server' sent a list of peers rather than performing the selecting" (from the PTAB exhibit retrieved at ptacts.uspto.gov). If the '968's "agent" is construed as itself performing the content fetching rather than returning a peer list, the Yu/Kageyama‑type art loses its footing — a distinction worth flagging as the single most likely reason a merits challenge would fail.

C. Proof of the combination from the primary reference's own field is required.
A § 103 challenge built on RFCs alone (Ground 5) is weak on any limitation beyond bare protocol semantics, because an RFC is a specification of how a protocol works, not a teaching of why to architect a system around it.

D. What the IPR denial does and does not prove.
The Google Patents page records IPR2021‑00249 as "Not Instituted ‑ Procedural"; the PTAB docket records "Institution Denied." A procedural denial — as opposed to an institution decision on the merits — carries no weight on obviousness. It does not establish that the claims are non‑obvious; it establishes only that the Board did not reach the grounds (commonly under § 325(d), General Plastic, Fintiv, or a petition defect). Any citation of the denial as evidence of validity would be a misuse of it. Conversely, the fact that a § 101 challenge under Alice was also pressed in E.D. Tex. (2:19‑cv‑00396, denied at Dkt. 85; re‑raised and ruled on in 2:19‑cv‑00395 at Dkt. 303, 2021‑02‑16) is relevant because a claim that reads on the generic intermediary architecture is both more likely abstract and more likely obvious — the two risks travel together.

7. Bottom line

• The strongest § 103 position is Cottrell + RFC 2616 + a BitTorrent‑type content‑keyed peer‑inventory reference, with RFC 1122/791/2460 for the address‑based agent assignment. Cottrell supplies the intermediary device that fetches on another's behalf and masks identity; RFC 2616 supplies the content identifier, the HTTP request/response, and the conditional‑request validity check; the specification's own admitted FIG. 1/FIG. 2 art supplies chunking with checksum keys, multi‑source parallel download, caching, and content‑indexed peer lists; the RFC addressing references supply IP‑range sharding and proximity‑based server selection.
• MorphMix + RFC 2616 is a viable alternative primary combination for the independent claims, and it was in fact asserted against sibling US 10,257,319 in IPR2021‑01492, including as an anticipation ground.
• The two defensible islands of non‑obviousness are (i) the agent's content‑keyed, dynamically maintained record of which client devices have received responses to requests for a given assigned server and (ii) the "agent performs the selecting/fetching rather than returning a peer list" distinction that carried the parent '511 application through prosecution over Yu and Kageyama. A challenge that does not independently meet both risks failing.
• Confidence levels: high on the identity of the asserted references and on the § 103 framework; high that RFC 2616 discloses the validity‑check dependent claims; moderate on the element‑by‑element mapping of the independent claims, because I did not have the verbatim claim text of US 10,637,968 and reconstructed it from the challenged‑claim list in the petition and the sibling patents '319/'510. Before filing or relying on this analysis, the actual claim 1 and its dependents of the '968 should be pulled from the granted patent and the mapping re‑verified line by line, and the content of Wang (US 2006/0059091) and Carro (US 7,007,228) should be independently reviewed — I could not verify either.