US 8756336

I'll search for authoritative information on this specific patent number.

US Patent 8,756,336 B2 — Summary

I searched for the specific number 8756336 (treated literally, no auto-correction). The authoritative full text (provided) is corroborated by Google Patents, the PTAB record, and litigation dockets. Note that this patent shares a specification with a large family, but its granted claims are different from the parent patents' claims (see the caveat at the end).

Bibliographic data (from the patent front matter / Google Patents)

Prosecution chain (as stated in the patent): This is a divisional of Ser. No. 13/922,227 (US 8,738,794), which was a continuation of Ser. No. 13/740,214 (US 8,700,790), which was a continuation of Ser. No. 12/333,303 (US 8,392,591), which claims benefit of provisional 61/017,202. It also incorporates Ser. No. 11/901,802 ("Online Publishing Of Multimedia Content").

Abstract (verbatim)

"Disclosed herein is a method and system for utilizing a digital data capture device in conjunction with a Bluetooth (BT) enabled mobile device for publishing data and multimedia content on one or more websites automatically or with minimal user intervention. A client application is provided on the BT enabled mobile device. In the absence of inbuilt BT capability, a BT communication device is provided on the digital data capture device. The BT communication device is paired with the BT enabled mobile device to establish a connection. The client application detects capture of data and multimedia content on the digital data capture device and initiates transfer of the captured data, multimedia content, and associated files. The digital data capture device transfers the captured data, multimedia content, and the associated files to the client application. The client application automatically publishes the transferred data and multimedia content on one or more websites."

Claims overview

The patent has 5 claims. Only claim 1 is independent; claims 2–5 depend on it.

Claim 1 (independent) — plain language. A method of segmented data transfer from a Bluetooth-enabled mobile device to one or more web services:

1. Determine the size of the data to be transferred.
2. Transfer the data using a segmented data transfer protocol, chosen based in part on the present data size and the device's data-transfer capabilities. The transfer is performed by the mobile device and involves:
   • Dividing ("segmenting") the data into segments at the application layer level;
   • Applying a segment identifier to each segment at the application layer level;
   • Applying a user identifier to each segment for each destination web service, at the application layer level, where the user identifier uniquely identifies the particular user of that web service;
   • Sending each segment (with its user identifier and segment identifier) to the web service;
   • Receiving a synchronization message from the web service confirming that segment transfer completed successfully;
   • Repeating the send-and-confirm steps until all segments are transferred successfully.
3. Interruption recovery: If the segmented transfer is interrupted, the mobile device (a) sends a synchronization request to the web service for that user identifier once the network connection is restored; (b) receives a synchronization reply identifying the segmented data the web service already received for that user identifier; and (c) resumes the transfer from the point immediately after the last successful receipt by the web service.

Dependent claims:

• Claim 2: The transferred data comprises one or more of audio, video, image, text, or digital data.
• Claim 3: The transferred data is multimedia data and data associated with multimedia data.
• Claim 4: Transferring further comprises configuring the data at the mobile device so it can be received and reassembled by the web service(s).
• Claim 5: The user identifier comprises one or more elements of user information — e.g., user-name, user-password, user-device-information, user settings, or user preferences.

Key takeaway: Although the specification frames the invention around Bluetooth camera-to-phone uploading, the granted claims of this divisional are narrowly directed to application-layer segmentation + per-segment user/segment identifiers + synchronization/acknowledgment + resume-after-interruption when transferring data from a Bluetooth-enabled mobile device to web services.

Litigation / PTAB status (as of the search results)

• PTAB: TikTok Inc. v. CellSpin Soft, Inc., IPR2024-00757, filed Apr. 1, 2024, against this patent (App. No. 14/109,972). Institution granted Sept. 30, 2024; Patent Owner's rehearing request denied Oct. 21, 2024; Director Review of the institution decision granted sua sponte (June 5, 2025); oral argument held (May 21, 2025); additional briefing authorized and supplemental briefs filed (Feb. 2, 2026). Search sources describe the case status as "Trial Instituted" / pending. (https://ipverse.greyb.com/ptab-web/cases/case-details/IPR2024-00757 ; https://www.docketalarm.com/search/?q=patent%3A%288756336%29)
• District court: CellSpin Soft, Inc. v. ByteDance Ltd. et al., No. 2:23-cv-00496 (E.D. Tex.) — this patent (8,756,336) is one of the asserted patents. The case was stayed pending IPRs (Jan. 27, 2025) and the stay was lifted (Jul. 16, 2026). A related case, 2:23-cv-00227 (E.D. Tex.), is also listed for this family.
• Recent family-level PTAB activity (2026): Law360 reports that in April 2026 the PTAB agreed to review whether three Cellspin patents are invalid after earlier TikTok challenges were rejected, and in June 2026 three challengers settled their PTAB disputes. I could not confirm from the results that these 2026 events specifically involve this patent (8,756,336) versus sibling patents — treat that link as uncertain.

Caveats / uncertainty

• I did not find an authoritative CAFC 2026 docket specifically for 8,756,336. The 2026 activity I found is at the PTAB (IPR2024-00757 and related proceedings), not the Federal Circuit. I cannot confirm any CAFC appeal for this exact patent.
• Some third-party records (e.g., the Unified Patents portal for a family member, US 10,425,806) list a priority date of 2007-12-27, whereas the patent front matter and Google Patents state 2007-12-28. I have reported the patent's own stated date.
• The live status of IPR2024-00757 (instituted vs. terminated) is ambiguous across the 2026 sources; the Director Review and later settlement activity suggest the record is still evolving. I flag this rather than assert a final outcome.

I'll research litigation involving US patent 8756336. Let me run several targeted searches.

Let me dig deeper into specific cases and confirm which assert the '336 patent.

Let me verify the 2025 cases and confirm whether the '336 patent is asserted there, plus check the Panasonic and ByteDance details.

Litigation Involving U.S. Patent No. 8,756,336

Patent: U.S. 8,756,336 B2 — "Automatic multimedia upload for publishing data and multimedia content"
Patent owner: CellSpinSoft Inc. (Cellspin Soft, Inc.), Mountain View, CA
Asserted family: The '336 patent is a member of the Cellspin family (family ID 40799912) that also includes 8,392,591; 8,737,794; 8,762,560; 8,798,539; 8,862,757; 8,892,752; 8,898,260; 8,904,030; 9,226,138; 9,258,698; 9,319,870; 9,749,847; 9,900,766; 10,425,806; 11,234,121; 11,659,381; 12,058,779.

I identified two district-court litigations in which the '336 patent was put in suit, plus one PTAB proceeding directly challenging it. Both district cases are the two E.D. Tex. suits flagged on the patent's Google Patents "litigation" record. Details below, with the caveat that I am relying on court-record aggregators (CourtListener/RECAP, Docket Alarm, Justia Dockets), the USPTO/PTAB docket, and Law360 — not on a direct PACER pull.

1. Cellspin Soft, Inc. v. Panasonic Corporation, et al.

Clarification (avoiding number confusion): This 2023 E.D. Tex. action is distinct from Cellspin Soft, Inc. v. Panasonic Corporation of North America, No. 4:17-cv-05941 (N.D. Cal.), a 2017 case that asserted U.S. 9,258,698 (not the '336 patent) and was dismissed under § 101.

2. Cellspin Soft, Inc. v. ByteDance Ltd., et al.

3. PTAB Proceeding Directly on the '336 Patent

Important Negative Findings (to avoid conflating similar numbers)

• The well-known 2017–2022 Cellspin campaign — Cellspin Soft, Inc. v. Fitbit, Moov, Nike, Fossil/Misfit, Garmin, Canon, GoPro, Panasonic, JK Imaging (N.D. Cal., coordinated before Judge Yvonne Gonzalez Rogers; Fed. Cir. Nos. 2018-1817; 2019-1588; 22-2025; 23-1526) — asserted U.S. 8,738,794; 8,892,752; 9,258,698; and 9,749,847. It did NOT assert the '336 patent. Do not attribute those decisions (the Cellspin Soft v. Fitbit, 927 F.3d 1306 § 101 rulings, the Garmin cert petition 19-400, or the 2024 non-infringement/summary-judgment appeals) to the '336 patent.
• The 2025 E.D. Tex. cases — Cellspin Soft v. SZ DJI Technology (2:25-cv-00949), Cellspin Soft v. Garmin Corporation (2:25-cv-00950), and Cellspin Soft v. Fossil Stores I/Fossil Group (2:25-cv-00683) — assert family patents generally described as "three patents related to Bluetooth data transfer." Based on the docket material I could retrieve, the complaints center on the '794 and '752 patents; I could not confirm that the '336 patent is asserted in any of these 2025 actions. Treat any '336 connection there as unverified.
• Similarly, the 2026 PTAB proceedings that other petitioners (including a group that "settled" per Law360, June 30, 2026) brought against "three Cellspin patents" are not confirmed to involve the '336 patent.

Caveats

• I was unable to pull PACER directly; case numbers, filing dates, and outcomes above are drawn from CourtListener/RECAP, Docket Alarm, Justia Dockets, the USPTO/PTAB docket, and Law360 (the latter two primarily for the PTAB proceeding).
• The Google Patents page for US 8,756,336 corroborates only two Texas district-court entries (2:23-cv-00496 and 2:23-cv-00227) and the IPR2024-00757 PTAB entry — which matches the district cases above.
• If others exist, they would most likely be additional 2023–2026 E.D. Tex. complaints by Cellspin that name the '336 patent alongside family patents; for those I could not verify the asserted patents, I have so stated rather than assume.

I'll research the PTAB proceedings on this patent. The structured block says ODP returns none, but the patent family data references IPR2024-00757 — let me verify.

Substantial PTAB activity exists despite the ODP gap. Let me dig into the FWD, appeals, and related proceedings.

Critical development found — the institution was vacated by Director Review. Let me confirm the specifics and check for any other proceedings or appeals.

PTAB Proceedings — US Patent 8,756,336 ("Automatic multimedia upload for publishing data and multimedia content")

Data-integrity flag (read first). The structured "PTAB proceedings on file" block states the USPTO ODP API returns no AIA trial proceedings for this patent. That is not accurate as of today. Both the Google Patents family record you supplied (which carries the annotation "PTAB case IPR2024-00757 filed (Pending - Instituted)") and voluminous public USPTO/PTAB documents confirm one AIA trial proceeding on the '336 patent: IPR2024-00757, TikTok Inc. v. Cellspin Soft, Inc. I have used web search to reconstruct it below. I could not independently confirm whether any other petitioner filed a later IPR against the '336 patent (see gaps at the end). Verify against PTAB E2E before relying on the count.

Proceedings overview

Total proceedings on file: 1 (IPR2024-00757). Breakdown: 0 claims invalidated, 0 claims sustained on the merits, 0 settled, 0 institution denials at the Board level — instead, a unique outcome: institution was granted, then vacated by the Director on 2026-03-30, and the petition was denied, terminating the trial before any Final Written Decision. Bottom-line defensive posture for a defendant: the '336 patent has NOT been hardened by IPR, but it also has NOT been narrowed — claims 1–5 are fully intact and UNTESTED at the FWD level. There is no IPR estoppel and no invalidity judgment to lean on. The one genuinely useful artifact is the Board's 2024-09-30 institution decision, which found a reasonable likelihood that claims 1–5 are obvious over three art combinations — admissible-flavored persuasive material for a district court, but with no preclusive or estoppel effect whatsoever.

The proceeding's headline is procedural, not substantive: TikTok's petitions were killed on a real-party-in-interest / "foreign government is not a 'person'" theory, not on the merits of the prior art.

IPR2024-00757 — TikTok Inc. v. Cellspin Soft, Inc.

• Type: Inter Partes Review (35 U.S.C. §§ 311–319). Part of a coordinated seven-petition campaign on the CellSpin patent family (see below).

• Filed: 2024-04-01 (notice according filing date 2024-04-25, Paper 3). § 315(b) one-year window ran from service of the E.D. Tex. complaint in Cellspin Soft, Inc. v. ByteDance Ltd. et al., No. 2:23-cv-00496.

• Status: Petition denied; institution vacated; IPR terminated (Director Decision, Paper 42, 2026-03-30). Gloss: the case is over; no trial, no FWD, no claims canceled. The Google Patents label "Pending - Instituted" is stale.

• Judge panel: Board panel — Gregg I. Anderson, Cynthia M. Hardman, Michael A. Valek. Director Review: initiated sua sponte by Acting Director Coke Morgan Stewart (Paper 34, 2025-06-05); decided by Director John A. Squires (Paper 42, 2026-03-30).

• Petition grounds: Challenged claims 1–5 (all claims). Three § 103 obviousness grounds:

  • Ground 1: Aaltonen in view of Drescher — claims 1–5;
  • Ground 2: Takahashi in view of Na — claims 1–5;
  • Ground 3: Takahashi in view of Na and Lind — claims 1–5.

  Petitioner's expert: Dr. Patrick Traynor. Patent Owner's expert: Dr. Gajendragad. Petitioner RPIs identified: TikTok Inc., ByteDance Inc., ByteDance Ltd., ByteDance PTE Ltd., TikTok PTE Ltd.

• Institution decision: Instituted 2024-09-30 (Paper 8) on all challenged claims and all asserted grounds. The Board rejected Patent Owner's Fintiv discretionary-denial arguments (Petitioner filed a Sotera stipulation) and its § 316(b) "geopolitical" arguments premised on the U.S. government's TikTok policy. Panel rehearing denied 2024-10-21 (Paper 12, opinion by Judge Valek); Patent Owner did not seek rehearing of the merits determination.

• Final Written Decision: NONE ISSUED. The trial was stayed on 2025-06-05 when the Acting Director ordered sua sponte Director Review, before the statutory FWD deadline (~2025-09-30). No claim was canceled; no claim was held patentable. The Board's merits record stops at the institution stage.

• Settlement / termination: No settlement. Termination came from Director Review. Timeline:

  • 2025-03-05/18 — Patent Owner moved to terminate, arguing the Chinese Communist Party ("CCP") is an unnamed RPI and that, under Return Mail Inc. v. USPS, 587 U.S. 618 (2019), a sovereign-controlled entity is not a "person" eligible to petition.
  • 2025-06-02 — Board denied the motion to terminate (Paper 32/33): motion waived (not raised in the POR filed 2025-01-03); an unidentified RPI does not divest jurisdiction; Return Mail concerned a U.S. federal agency, not a foreign government.
  • 2025-06-05 — Acting Director sua sponte Director Review (Paper 34); cases stayed.
  • 2026-01-15/16 — YMTC (Yangtze Memory Techs. v. Micron, IPR2025-00098/-00099, Paper 38) designated informative: petitioner's failure to provide clarity as to its identity is fatal to institution.
  • 2026-01-22 — Petitioner announced TikTok USDS Joint Venture LLC; Director authorized additional briefing (Paper 38) on the RPI question and the JV's effect.
  • 2026-03-18 — Tianma Microelectronics v. LG Display, IPR2025-01579, Paper 12, designated precedential: a foreign government is not a "person" under the AIA and cannot petition.
  • 2026-03-30 — Director Decision (Paper 42): "Vacating the Decisions Granting Institution and Denying Institution of Inter Partes Review." Because Patent Owner's evidence sufficiently put RPI into dispute and Petitioner elected not to introduce rebuttal evidence, the Director held Petitioner failed to carry its burden to show a foreign government was not an RPI; Petitioner is deemed ineligible to file an IPR; institution was vacated, the petition denied, no trial instituted. Joinders in the companion cases were also vacated.

• Appeal: No Federal Circuit appeal identified. A Director decision denying institution is generally not judicially reviewable (35 U.S.C. § 314(d); Cuozzo), so an appeal of Paper 42 is unlikely. I could not confirm any CAFC docket. Flagged as unverified.

• Defensive value: Double-edged. (1) Helpful: the Board already found a reasonable likelihood that every claim (1–5) is unpatentable over Aaltonen+Drescher and Takahashi+Na(+Lind) — strong persuasion value to a district court or a new petitioner. (2) Unhelpful: nothing is canceled, no estoppel attaches (see below), and a fresh IPR faces (a) the § 315(b) one-year bar if the defendant was served long ago, (b) § 325(d)/General Plastic friction given the earlier (albeit vacated) challenge, and (c) the Director's new line of authority cutting against sovereign-linked petitioners — which is irrelevant to a normal corporate defendant. Do not tell a court or client that any claim of the '336 patent has been invalidated; it has not.

Strategic summary

Claim status. All five claims — claim 1 (independent) and dependent claims 2–5 — are INTACT and UNTESTED. None was canceled (no FWD ever issued); none was adjudicated valid either. The institution decision's claim-by-claim "reasonable likelihood" findings are not a validity or invalidity holding and carry no collateral estoppel. Note the '336 patent is not among the patents that Unified Patents took to ex parte reexamination: the rehearing decision (Paper 12 n.9) records that "Patent Owner has not identified any reexamination involving the '336 patent" — Unified's reexam requests covered the '757 and '260 patents (CRU grants 2024-07-15), not the '336.

Estoppel landscape — the key takeaway. § 315(e)(2) estoppel applies only "after the petitioner requests or institution of an inter partes review… results in a final written decision." Here there was no final written decision — institution was vacated. Therefore:

• TikTok/ByteDance and their privies are NOT statutorily estopped from re-raising Aaltonen/Drescher or Takahashi/Na/Lind, e.g., in the stayed E.D. Tex. case or a new petition (subject to § 315(b) timing and the discretionary factors). But do not over-read this — the practical realities of General Plastic and the § 315(b) clock, plus the fact that TikTok is now characterized as sovereign-controlled for AIA standing purposes, make a TikTok refiling doubtful.
• For an unrelated defendant, the entire prior-art universe — including Aaltonen, Drescher, Takahashi, Na, Lind, and anything else — remains available. There is no IPR-based estoppel shadow on any ground. Because the earlier petition died on RPI/standing rather than the merits, there is no § 325(d) "previously presented" estoppel binding on the Office with respect to the merits, though a new panel may still weigh the earlier institution decision as a discretionary General Plastic factor.

Pattern signals.

• One petitioner, many patents. TikTok/ByteDance filed a coordinated seven-petition wave on the CellSpin family, all filed ~2024-04-01/04-05 and all instituted on 2024-09-30/10-01: IPR2024-00757 ('336), -00759 ('757), -00760 ('260), -00767 ('381), -00768 ('121), -00769 ('766), -00770 ('030). All seven were addressed by the Director's Paper 42 and all were vacated/denied on 2026-03-30 on the same RPI/foreign-government theory. No FWD issued in any of them.
• Joinder track. LifeScan, Senseonics, and Ascensia Diabetes Care were joined as petitioners to IPR2024-00768, -769, -770 via their own petitions IPR2025-00102, -00103, -00104; the Director vacated those joinders and left it to the Board to decide whether those petitions stand on their own. Not relevant to the '336 patent, but part of the picture.
• Defensive aggregator present. Unified Patents is in the chain, but on the ex parte reexam side for the '757/'260 patents (CRU granted 2024-07-15) — not on the '336. Unified surfaced in the '336 record only because Patent Owner complained about it in its Fintiv rehearing request (a footnote confirms no reexam on the '336).
• Patent Owner posture. CellSpin litigated aggressively and creatively — Fintiv discretionary denial, § 316(b) geopolitics, expert battles, and ultimately the CCP/RPI/Return Mail kill shot. It has not needed to defend any FWD on appeal on this patent. The related Cellspin Soft v. Fitbit (Fed. Cir. 2019) § 101 precedent is the family's other notable legacy.

Recommended next steps

1. Correct the record internally. Update any docket/database that shows "IPR2024-00757 — Instituted/Pending." The operative event is the Director's 2026-03-30 decision (Paper 42) vacating institution and denying the petition. Link it: https://www.uspto.gov/sites/default/files/documents/ipr2024-00757_etc_tiktok_v_cellspin_soft_03302026_paper42.pdf (and Paper 38, .../ipr2024-00757_tiktok_v_cellspin_order_paper38.pdf).
2. If you are a defendant, do NOT claim the '336 claims are invalid. Nothing was canceled — no FWD. Instead, use the 2024-09-30 institution decision as persuasive prior-art teaching: it found a reasonable likelihood that claims 1–5 are obvious over (i) Aaltonen + Drescher and (ii) Takahashi + Na (+ Lind). Those same references are fully available to you in district court.
3. Check your § 315(b) clock now. If you were served with a complaint asserting the '336 patent more than one year ago, a new IPR is time-barred unless you can join or find an exception. If you are within the window, a fresh IPR by a non-sovereign petitioner is procedurally viable (no FWD = no estoppel), but expect General Plastic/§ 325(d) briefing and consider piggybacking on the already-vetted Aaltonen/Takahashi grounds.
4. Watch the parallel litigation. Cellspin Soft, Inc. v. ByteDance Ltd. et al., No. 2:23-cv-00496 (E.D. Tex., Judge Gilstrap; Mag. J. Payne) was stayed 2025-04-11 pending the IPRs, with trial previously set for 2025-08-04. Because no FWD will now issue (the IPRs were terminated), the stay rationale has evaporated — expect a motion to lift the stay. That is your near-term event risk. A companion case, 2:23-cv-00227, also appears in the record.
5. Verify the gaps I could not close (do not treat as confirmed): (a) whether any CAFC appeal of Paper 42 or of any related Board action was filed; (b) whether any petitioner other than TikTok has since filed an IPR against the '336 patent specifically; (c) the exact disposition of IPR2025-00102/-00103/-00104 after the joinders were vacated. Confirm on PTAB E2E/Center (https://ptacts.uspto.gov/ptacts/) and the PTAB Decisions page, and check CourtListener for any CAFC docket.

Bottom line: The '336 patent survived, but only because the Board never got to decide it — the trial was extinguished on a foreign-sovereign-RPI theory unique to TikTok. For the rest of the world, the patent is fully intact, unobscured by estoppel, and open to attack, with a ready-made institution record pointing at claims 1–5.

I'll research the assignment record for US 8756336. Let me gather authoritative data from multiple sources.

I have substantial data. Let me gather more on Cellspin's corporate status, the original assignment correspondent, and litigation patterns.

Let me get the original assignment details and check for any further recorded assignments or releases.

I've reached the research step limit. Below is the assignment-chain reconstruction built on the authoritative patent text provided plus the assignment/litigation records I retrieved.

Assignment & Ownership Analysis — US 8,756,336 B2

Primary sources checked: USPTO Patent Assignment Search (reel/frame indices via assignmentcenter.uspto.gov and assignment.uspto.gov), the recorded assignment documents themselves (surfaced through Docket Alarm / IPR exhibit copies of the USPTO EPAS filings), Google Patents legal events, and litigation databases (RPX Empower, Docket Alarm, CourtListener).

⚠️ Critical framing note: The chain for this patent contains only one title-transfer assignment (inventors → Cellspin). Everything in 2018 is a security interest (a lien), not an assignment of title. Google Patents renders these as "Assigned to ," which is misleading — the assignee of record for title purposes never changes. I flag this because a careless reader could mistake the 2018 cluster for a shell-entity transfer cascade. It is not.

Inventors

Notable patterns:

• All three inventors were affiliated with the assignee, CellSpinSoft Inc. — this is a company-originated patent, not an individual-inventor filing.
• Two of the three inventors (Singh and Klein) later appear as secured lenders to their own company in the 2018 security-interest recordings (see timeline). Founders lending to their own distressed company alongside outside individual lenders is a notable distress signal.
• I found no evidence that the inventors departed the assignee within 12 months of filing. On the contrary, Singh is still identified as Cellspin's president years later, and the family kept filing continuations through 2023 (US 12,058,779).

Original assignee

• Entity: CellspinSoft Inc. (rendered "CellSpinSoft Inc." on the patent front matter; "Cellspin Soft, Inc., a California corporation" in the security agreements).
• Addresses of record (two, over time): 4423 Fortran Drive, Suite #116, San Jose, CA 95134 (original assignment record) → 1410 Mercy Street, Mountain View, CA 94041 (2018 security agreements and litigation filings).
• Line of business: Mobile-blogging / multimedia-upload software. Per Cellspin's own submissions, it shipped a commercial mobile app (ad-supported, later premium) for 300+ phone models enabling one-touch posting of photos/video/audio to Flickr, YouTube, Picasa, Facebook, eBay, etc. It won Silicon India's 2008 Top Mobile Application award. So an actual product was shipped.
• Current status: Operating as a patent-assertion entity; effectively wound down as a product company. Evidence: multiple money judgments against it (a Sept. 26, 2018 writ of execution for $65,942.15 in Cellspin v. Fitbit; a July 6, 2018 money judgment in Cellspin v. Fossil), the 2018 IP-collateral pledges, and the fact that third-party databases classify it as a high-frequency patent plaintiff (Ex Parte AI Lab: "Cellspin Soft Inc … Total Cases 39, Active Cases 10, Type NPE"). No record of a bankruptcy filing was found.

Assignment timeline

Documented recordings for this patent. The original assignment reel/frame is confirmed (32141/316). For the 2018 security-interest filings I recovered the execution dates, parties, and EPAS confirmation IDs but not the reel/frame numbers — I will not invent them.

• 2013-12-19 (recorded) — Reel 32141 / Frame 0316 (recorded Assignment of Assignors' Interest)

  • Conveyance: Assignment (initial, from inventors)
  • Assignor: Singh, Gurvinder; Klein, Marcos; Laviano, Vince
  • Assignee: CellSpinSoft Inc. (officially "Cellspinsoft Inc."), 4423 Fortran Drive, Suite #116, San Jose, CA 95134
  • Correspondent: Not confirmed on the index record; Ashok Tankha (Reg. No. 33,802) of Lipton, Weinberger & Husick, 36 Greenleigh Drive, Sewell, NJ 08080, is the long-time prosecution attorney for the entire family and is the most likely filer, but I could not verify this from the assignment record itself — treat as unconfirmed.
  • Context: Initial acquisition by the operating company (founders assigning to their own corporation). Execution-date anomaly: the index shows Singh executing Dec. 3, 2012, and Klein/Laviano in early 2014 — i.e., the assignment paperwork was executed across a spread of dates rather than at filing; recorded together at reel 32141/316.

• 2018-07-24 and 2018-07-27 (executed) — recorded 2018-07-27 — Reel/Frame: not recovered (EPAS IDs observed: PAT5071658, PAT5071256, PAT5071177)

  • Conveyance: Security Interest (11 separate short-form IP security agreements — a lien, not a transfer of title)
  • Assignor (conveying party): CellSpin Soft, Inc.
  • Assignees (secured parties), all natural persons: John W. Kastelic (Atlanta, GA); Gurvinder Singh (Santa Clara, CA); Rajeev Virmani (Hoboken, NJ); Donald A. Williams (San Ramon, CA); Marcos Klein; James Bland Kessinger; Neil Silverman; Amir Khan; Ram Akella; David C. Lin; Eamonn McSweeney
  • Correspondent: JOHN W KASTELIC (self-filed; johnwkastelic@gmail.com; 715 Flat Shoals Ave SE, Atlanta, GA 30316). He is simultaneously a secured party and the correspondent of record on the filings. ⚠️ Repeat correspondent flag: the same name appears across multiple links in this 2018 cluster.
  • Context: Securitization / collateral pledge — Cellspin pledges all IP (15 patents + applications in the family, including 8,756,336) as collateral for pre-existing loans. Notably, several underlying Security Agreements date back to 2007–2011 (e.g., Virmani's master SA dated May 10, 2007; Singh's dated Sept. 11, 2007), so these are late perfection filings of old debt dated to coincide with the 2018 litigation/financial pressure, not new financings.

• No release, merger, change-of-name, or title-transfer assignment appears in the record after 2018. Cellspin Soft, Inc. remains both assignee and the named plaintiff in every case asserting this patent.

Timeline diagram

timeline
    title Ownership of US 8756336
    2007 : Priority provisional filed
    2008 : Nonprovisional filed
    2013 : Divisional 14109972 filed
         : Reel 32141 frame 316 recorded
    2014 : Patent issued Jun 17
    2017 : First Cellspin suits filed
    2018 : Eleven IP security interests recorded
         : Money judgments against Cellspin
    2023 : ByteDance suit asserts this patent

NPE / troll-pattern signals

Caveat on the 2018 filings: my search results confirmed the execution dates, the conveying/receiving parties, and the EPAS submission IDs, but I could not confirm the reel/frame numbers for the security interests. I am not fabricating them.

Verdict

NPE — moderate confidence.

Justification: On the pure assignment-chain test, the classic troll pattern is absent — there is exactly one title transfer (inventors → Cellspin Soft, Inc., Reel 32141/0316, recorded 2013-12-19), and the current assignee of record is the same original operating company, which actually shipped a mobile app embodying the claims. The NPE call rests on two things: (1) a strong entity-level signal — Cellspin is classified as an NPE and has prosecuted a large multi-defendant campaign (Fitbit, Nike, Under Armour, Fossil, Garmin, Nikon, then ByteDance/TikTok) after ceasing to sell products; and (2) a cluster of 11 IP security interests recorded 2018-07-24/27 (correspondent John W. Kastelic, self-filed), several of which perfect debt dating to 2007–2011 — evidence the company had no operating revenue and was liquidating value through assertion. No shell-transfer signal (Signal 1) and no cascading-transfer signal (Signal 4) is present, which is why this is moderate, not high, confidence.

Verification: USPTO Assignment Center — https://assignmentcenter.uspto.gov/ (search patent 8756336); original assignment at Reel 32141 / Frame 0316. Alternate index: https://assignment.uspto.gov/patent/index.html

Flags for the earlier sections: (i) The previously generated summary listed the 2018 events as "SECURITY INTEREST" — good; I'd reinforce that these do not change the assignee of record, to prevent a shell-transfer misread. (ii) The previously generated section's "2007-12-28 vs. 2007-12-27" priority-date discrepancy does not affect the assignment chain, but note that two 2018 security agreements reference master Security Agreements dated May 10, 2007 and Sept. 11, 2007 — i.e., pre-dating even the priority filing, consistent with a company that was financed on the strength of its IP from the outset.

I'll research this systematically. Let me start by searching for the patent and its cited references.

Prior-Art Analysis — US 8,756,336 B2

Scope note and methodology

I searched for the specific number 8756336 (interpreted literally, no auto-correction). The authoritative full text you supplied lists the references on the face of the patent under "Citations (5)" — i.e., the five references cited (four published US applications and one US patent). I treat those five as "the patent citations for 8,756,336," and I corroborated each against Google Patents and related sources.

Important framing: the granted claims of 8,756,336 are narrowly directed to segmented data transfer at the application layer with per-segment segment identifiers and per-segment user identifiers, a synchronization message, and resume-after-interruption. Keeping that in mind is essential, because most of the five cited references are directed to different layers/aspects and therefore do not fully anticipate claim 1. I say so explicitly below rather than overstate.

Also note the statutory posture: the application was filed Dec 18, 2013, but is a divisional claiming benefit back to Dec 28, 2007 (provisional 61/017,202). Assuming every claim retains the 2007 priority date, the pre-AIA 35 U.S.C. §§ 101–103 framework applies (the AIA first-inventor-to-file regime governs only if a claim has an effective filing date on/after Mar. 16, 2013). I flag where § 102(e) versus § 102(b) matters.

The five cited references (as listed on the patent)

Sources: Google Patents entry for US8756336B2 (https://patents.google.com/patent/US8756336B2/en) and the corroborating pages noted per reference below.

Reference-by-reference analysis

1. US 2002/0167932 A1 — McGowan (Intel)

• Full citation: U.S. Patent Application Publication 2002/0167932 A1, "Method and apparatus for encapsulating universal serial bus messaging over link layer communication protocol," McGowan, Steve B.; filed Mar. 16, 2001; published Nov. 14, 2002.
• Brief description: Encapsulates USB HID messaging over a Bluetooth link. DisclosesSegmentation and Reassembly (SAR), L2CAP-based reliable delivery, per-report "Transaction Header" (THdr/DATA/DATC), and MTU-negotiated splitting of a payload into segments that the receiver concatenates back into "a single large data object." This is link-layer (Bluetooth HID/L2CAP) segmentation, not application-layer segmentation.
• Potentially anticipates: Claim 4 (data "configured… so that it can be received by the one or more web services and reassembled") is the only claim whose subject matter (segmentation/reassembly) is even arguably touched — but only at the link layer, and McGowan has no "web service," no user identifier, and no sync/resume. It does not anticipate independent claim 1. Best characterized as a § 102(b) reference usable for the concept of segment/reassembly (and § 103 fodder), not a full anticipation.
• Source: https://patents.google.com/patent/US20020167932A1 ; https://uspto.report/patent/app/20020167932

2. US 2003/0157960 A1 — Kennedy

• Full citation: U.S. Patent Application Publication 2003/0157960 A1, "Remote storage for electronics," Kennedy, Richard; filed Feb. 20, 2002; published Aug. 21, 2003.
• Brief description: The closest conceptual reference to the specification. A Bluetooth-enabled digital camera transfers images over a Bluetooth link to a cellular telephone (the intermediate device), which forwards them over 3G/Internet to a remote server. Discloses real-time / automatic / manual upload modes, threshold-triggered automatic upload (e.g., "memory nearly full"), multiple upload destinations, and local buffering for network delays. Photos are offloaded to a web-based server "through the user's cell phone."
• Potentially anticipates: This is the strongest of the five for the multimedia-to-web-service aspect. It is a § 102(b) reference (published 2003, well before the 2007 priority). It maps to claim 2 (data = image/audio/video) and claim 3 (multimedia data and associated data) conceptually, and to the preamble of claim 1. However, Kennedy does not disclose (i) segmentation at the application layer, (ii) per-segment segment identifiers, (iii) per-segment user identifiers keyed to each destination web service, (iv) a synchronization message confirming each segment, or (v) the sync-request/sync-reply/resume-after-interruption logic. Because every dependent claim incorporates claim 1's limitations, Kennedy cannot anticipate any claim as a whole; it is a § 103 / background reference.
• Sources: https://patents.google.com/patent/US20030157960 ; https://www.freepatentsonline.com/y2003/0157960.html

3. US 2008/0109317 A1 — Singh (same inventor)

• Full citation: U.S. Patent Application Publication 2008/0109317 A1, "Wireless dissemination of environment aware information," Singh, Gurvinder; filed Oct. 22, 2007 (priority Oct. 26, 2006); published May 8, 2008.
• Brief description: A client application on a mobile device detects nearby base-station/beacon devices over Bluetooth/Wi-Fi/WiMAX, receives vendor/service information, categorizes it, and correlates advertisements with environmental/time data. Uses OBEX push/pull, a client application on the mobile device, and active/passive pull modes.
• Potentially anticipates: Limited relevance. It shares the inventors' general architecture (client application on a mobile device; Bluetooth push/pull; OBEX), which could bear on claim 1's environment but does not disclose segmented transfer, identifiers, sync messages, or resume. It does not anticipate any claim.
• Disqualification caveat (important): Under pre-AIA § 102(e), a published application is prior art only if it is "by another." This reference names Gurvinder Singh, who is a common inventor on 8,756,336. Exercising the rule literally and per your instruction not to auto-correct, I flag that a common-inventor reference is generally not "by another" and therefore may not qualify as § 102(e) art at all. (It is also not § 102(b) art, because it published after the Dec. 28, 2007 priority date.) Treat this citation as likely non-prior-art / attorney-found art rather than substantive prior art.
• Source: https://patents.google.com/patent/US20080109317

4. US 2009/0023428 A1 — Behzad et al. (Broadcom)

• Full citation: U.S. Patent Application Publication 2009/0023428 A1, "Method and system for creating a personalized journal based on collecting links to information and annotating those links for later retrieval," Behzad, Arya; Buer, Mark; Karaoguz, Jeyhan; MacInnis, Alexander; Quigley, Thomas; Walley, John; filed Sep. 25, 2007 (provisional 60/950,947 filed Jul. 20, 2007); published Jan. 22, 2009. (Later granted as US 8,781,450 B2.)
• Brief description: Captures multimedia and pre-tags it with indexed information such as user ID, time, date, location, and environmental condition; uses context-data tags, biometric "data segment IDs," and data segments; the tagged data may be stored locally or transferred to a remote host/storage for post-processing, including over Bluetooth/WPAN to a wireless mobile communication device.
• Potentially anticipates: The most relevant of the five to the identifier concepts. The pre-tagging with a user ID and the notion of segment identifiers/tags bear on claim 1's per-segment user-identifier and segment-identifier limitations and on claim 5 (user identifier = user-name, password, device information, settings, preferences). Note the timing: filed/priority in 2007 (before the Dec. 28, 2007 priority date) but published Jan. 22, 2009 (after) — so it can only be § 102(e) art, not § 102(b). Behzad nevertheless does not disclose transfer to a "web service" with per-destination user identifiers, the synchronization message per segment, or the sync-request/sync-reply/resume logic. It therefore does not anticipate claim 1; it is best used for § 103 on the identifier-tagging limitations and possibly for claim 5.
• Sources: https://patents.google.com/patent/US20090023428A1 ; https://www.patents-review.com/a/20090023428-method-system-creating-personalized-journal-based-links.html

5. US 7,519,916 B1 — Microsoft

• Full citation: U.S. Patent 7,519,916 B1, "Methods for tailoring a bandwidth profile for an operating environment," Microsoft Corporation; filed Jun. 16, 2003; issued Apr. 14, 2009.
• Brief description: Relates to tailoring a bandwidth profile to an operating environment (adapting data-transfer parameters to the environment/device). From the patent's own citation list, this is the reference most plausibly cited against the "due in part to present data size and data transfer capabilities" language of claim 1.
• Potentially anticipates: Only the adaptive-transfer / capability-based limitation of claim 1—and even then only arguably. It is § 102(e) art (application filed 2003, granted 2009). It plainly does not disclose application-layer segmentation, per-segment user/segment identifiers, synchronization messages, or resume-after-interruption, and thus does not anticipate any claim.
• Evidence limitation: My tool budget was exhausted before I could retrieve the full text of US 7,519,916 B1. I am relying on the bibliographic data and title as cited on the face of 8,756,336 and in Google Patents. I have not independently verified its specification content, and I flag that rather than fabricate detail.

§ 102 conclusions (claim-by-claim)

• Claim 1 (independent): No single cited reference anticipates it. Its distinctive combination — determining data size; segmented transfer at the application layer; a segment identifier and a user identifier applied to each segment for each destination web service; a synchronization message confirming each segment; repetition to completion; and sync-request → sync-reply → resume-from-last-successful-point after an interruption — is not disclosed by any of the five. The references are best deployed under § 103 (e.g., Kennedy for the multimedia-to-web-service pipeline + Behzad for user-ID tagging + McGowan for segmentation/reassembly), but that is an obviousness theory, not § 102.
• Claim 2 and Claim 3: Kennedy (and, to a lesser extent, Behzad) discloses transferring image/audio/video/multimedia and associated data to a remote/server destination. But because both claims incorporate claim 1's full limitations, neither is anticipated by Kennedy or Behzad standing alone.
• Claim 4 (reassembly): McGowan discloses segmentation and reassembly (SAR) — but at the Bluetooth/HID-link layer, not for a web service. Relevant under § 102/§ 103 for the reassembly concept only; does not anticipate claim 4 as a whole (which requires a web service and the claim-1 framework).
• Claim 5 (user identifier elements): Behzad's "user ID" pre-tagging is the closest, but it does not disclose the enumerated elements (user-name, user-password, user-device-information, user settings, user preferences) in the per-destination-web-service context. Not anticipated; at most § 103 material.

Bottom line ranking of the five, by relevance to this divisional's claims

1. US 2003/0157960 A1 (Kennedy) — closest to the overall camera→phone→web upload concept (claims 2–3, claim 1 preamble); § 102(b).
2. US 2009/0023428 A1 (Behzad) — closest to the identifier-tagging limitations (claims 1, 5); § 102(e) only.
3. US 2002/0167932 A1 (McGowan) — segment/reassembly concept (claim 4); § 102(b), but wrong layer.
4. US 7,519,916 B1 (Microsoft) — capability/bandwidth-adaptive transfer (part of claim 1); § 102(e); text not fully verified.
5. US 2008/0109317 A1 (Singh) — same-inventor reference; likely not "by another" and thus likely not § 102(e) art; minimal substantive relevance.

Caveats and cross-references to the earlier section

• Consistency with the prior summary: This matches the earlier-generated summary's identification of the "Citations (5)". No contradiction found for the citation list; the earlier section's characterization of granted claim 1 as narrowly directed to application-layer segmentation + identifiers + sync/resume is confirmed and is precisely why none of these five references anticipates claim 1.
• Examiner-cited vs. IPR-cited art: The five above are examiner-cited references. Substantive validity challenges in this family have relied on different art — e.g., TikTok's IPR2024-00757 against this patent, and the earlier IPR2019-00127 against sibling US 9,258,698, which used Japanese references such as JP 2003-60953 (Hiroishi), JP 2005-303511 (Takahashi), and JP 2004-96166 (Nozaki). If you want the references actually being asserted against 8,756,336 (as opposed to its face citations), the IPR petitions — not the face of the patent — are the better source. I did not independently review the IPR2024-00757 petition's ground list here.
• Unverified items: Full text of US 7,519,916 B1 was not retrieved (tool budget exhausted); dates for references 4 and 5 reflect the patent's own citation list and Google Patents and were not independently re-verified against USPTO PatentCenter.
• Pre-AIA vs. AIA: I have assumed pre-AIA § 102 throughout because the claims are presumed to keep the Dec. 28, 2007 effective filing date. If any claim in this application ever had an effective filing date on/after Mar. 16, 2013, the AIA § 102(a)(2)/102(b) regime (and its "commonly owned / inventor-common" exceptions) would change the treatment of the applications, especially the common-inventor Singh reference.

I'll ground this analysis in the prior-art references listed on the patent page and verify against the live PTAB record before building the § 103 combinations.

Let me pull the specific IPR grounds and the remaining cited references to ground the combinations.

Obviousness Analysis Under 35 U.S.C. § 103 — U.S. Patent No. 8,756,336 B2

Building on the claim breakdown and bibliographic section already produced above — those are treated as given and not repeated here. This section adds the § 103 attack.

0. A threshold note on "the Prior Art section of this page"

The Google Patents page carries two distinct prior-art populations, and they are not the same set:

1. The examiner-cited references — "Citations (5)" on the page's face:

   • US 2002/0167932 A1 — McGowan — Method and apparatus for encapsulating universal serial bus messaging over link layer communication protocol
   • US 2003/0157960 A1 — Kennedy — Remote storage for electronics
   • US 2008/0109317 A1 — Singh — Wireless dissemination of environment aware information
   • US 2009/0023428 A1 — Behzad — …personalized journal based on collecting links to information and annotating those links for later retrieval
   • US 7,519,916 B1 — Microsoft — Methods for tailoring a bandwidth profile for an operating environment

2. The family-citation population ("Family Cites Families (115)" / "Families Citing this family (57)"), e.g. US 5,806,005 (Ricoh), US 6,209,048 (Ricoh), US 6,750,902 (Fotonation), US 7,246,147 (Canon).

Critically, the live challenge to this patent does not rest on that list. My search confirmed that in TikTok Inc. v. CellSpin Soft, Inc., IPR2024-00757, the Petitioner's grounds are built on Aaltonen (US 2005/0209927) + Drescher (WO 2008/028508) (Ground 1) and Takahashi + Na (US 2006/0129631) (Grounds 2A/2B), with Bosschaert (US 2007/0174393), Foo (US 2013/0315235), Gourley's HTTP: The Definitive Guide, and the '757 Singh patent as exhibits (docketalarm, Pet. Reply to POR; Pet. exhibits TTI‑1005–TTI‑1013). I flag this explicitly because the combinations the Board actually instituted on are drawn from a different reference set than the page's "Citations (5)". I give you both, clearly labeled, and mark where the two do and do not overlap.

Governing law / date. Per the institution record, the Board applied a date of invention of December 28, 2007 (Dr. Traynor's declaration, Ex. 1028 ¶ 5), i.e. pre-AIA § 103(a) governs. That means art must qualify under § 102(a)/(b)/(e) as of the 2007 critical date. I apply that test below and note where a reference's date is a problem.

POSITA. The Board adopted, without dispute, Petitioner's definition: "at least a bachelor's degree in electrical engineering, computer engineering, or computer science, and two years of experience in the field of consumer electronics, with exposure to digital camera technology and wireless communications" (Institution Decision, IPR2024-00770 applying the same record; Cellspin did not dispute it). I use that level throughout.

1. Availability of the candidate references as § 103 prior art (pre-AIA)

Caveat flagged: the examiner's use of Ihara (2012) and the statement in the sibling '757 file that the case was "examined under the first inventor to file provisions of the AIA" are in tension with a 2007 priority claim. The IPR applied pre‑AIA with a 2007 date. I use the IPR position (pre‑AIA, 2007) as ground truth and treat Ihara/Foo as non-prior-art background.

2. Claim 1 element decomposition (labels used below)

3. Reference-by-reference teaching map

3.1 Kennedy (US 2003/0157960) — the anchor

Directly on point for the preamble and the "why segment" rationale. From the disclosure:

• *"a Bluetooth-enabled camera that communicates to a cellular telephone via a Bluetooth wireless link. The telephone may then use a 3G wireless link and the Internet to communicate with a remote device such as a server, an application service provider ('ASP')…"* — covers [1pre] (the BT-enabled phone is the device transferring to a web service) and supplies the web service/ASP destination.
• Modes: "real-time upload, automatic upload or manual upload"; in real-time mode the device "transfers its data as the data is acquired and as quickly as the wireless connections will allow" — supports [1b]'s "data transfer capabilities" rationale and provides the automatic behavior.
• "multiple upload destinations can be supported due to the fact that the PC also has its own storage capability and can therefore buffer data" — motivation for buffering/segmentation.

3.2 McGowan (US 2002/0167932) — segmentation, identifiers, reassembly

• *"if a report … is larger than 672 bytes, then the HID device implements segmentation and reassembly (SAR)"* — size-triggered segmentation → [1a], [1b].
• "the HID SAR module divides data into segments less than or equal to the L2CAP layer's negotiated MTU … The first segment will contain a DATA header and all subsequent segments will contain a DATC header. The length defined by the DATA header will identify the total payload size" → [1c], [1d] (per-segment header identifier + size).
• Report IDs / report concatenation ([0057]–[0059]) — per-item identifiers → [1d].
• "it will concatenate all subsequent segments with DATC headers, until all of the expected payload is received. When all segments are assembled, the payload can then be passed to the upper software levels as a single large data object" → claim 4 (reassembly) and the [1h] "until all … successfully" notion.
• Baseband timeouts / re-request behavior ([0070]–[0072]) → points toward [1i]/retry.

Weakness (PO's likely attack): McGowan's SAR sits at the HID/L2CAP boundary, not the "application layer" the claim demands. Expect Cellspin to press this. It is why the live IPR does not lead with McGowan — it leads with Aaltonen + Drescher and Takahashi + Na, which operate at the HTTP/application level. See §4.

3.3 US 7,519,916 (Microsoft) — "due in part to data size and transfer capabilities"

• "providing a number of bandwidth profiles … allowing the user an option to select one of the bandwidth profiles", selected "depending on one or more characteristics of an operating system" (i.e., the operating environment) → [1b].
• Low-profile parameters include "selective part synchronization," "reduced data transfer frequency," buffer size parameter (cl. 5), check parameter / frequency of data checking (cl. 6), selective file synchronization (cl. 2). → [1a]/[1b] (deciding how to transfer based on size/capability) and, arguably, [1i]–[1k] (part-synchronization = resume semantics).

3.4 Behzad (US 2009/0023428) — destination/user association

A personalized journal assembled from collected links and annotations and published for later retrieval. Its value for this claim is modest: it supplies the notion of associating uploaded content with a particular user's destination (the journal), supporting [1e], and reinforces "one or more web services." It is not a strong teaching of per-segment user identifiers.

3.5 Family-cited camera-upload art (Ricoh / Fotonation / Canon)

• US 5,806,005 (Ricoh) — wireless image transfer from a digital still video camera to a networked computer.
• US 6,209,048 (Ricoh) — peripheral with integrated HTTP server for remote access using URL's → concretely supplies "web service" over HTTP.
• US 6,750,902 (Fotonation) — camera network communication device.
• US 7,246,147 (Canon) — upload and retrieval by an image device.

These establish that camera→network upload and HTTP-server-based retrieval were routine before 2007, shoring up [1pre] and undercutting any argument that the camera-to-web-service context was inventive.

3.6 The live-IPR references (Ground 1 / Ground 2A‑2B)

• Aaltonen (US 2005/0209927) — an upload agent receiving "upload content" sent as "block[s] of packets"; "each information packet can include an identifier that identifies placement of the respective data packet in relation to the other data packets" (or a table "uniquely identifying the block … and the correct order"), with CRCs; UDP/TCP and HTTP chunked transfer. → [1c], [1d], reassembly (claim 4).
• Drescher (WO 2008/028508) — segmented/streaming media upload with resume semantics → [1i]–[1k].
• Takahashi + Na (US 2006/0129631) — multimedia messaging; per the Pet. Reply, "MMS is compatible with and endorses HTTP"; supplies sequence/segment identifiers and a user identifier, and a "web service" → [1d], [1e], [1f].
• Bosschaert (US 2007/0174393) — application-layer/file-transfer background.
• Gourley, HTTP: The Definitive Guide — used to establish HTTP chunked transfer (application-layer segmentation) as well-known; directly aimed at the "application layer level" limitation that McGowan does not cleanly reach.

4. The § 103 combinations

Below, each combination is stated as a ground, with the elements it reaches and the motivation to combine.

Ground 1 (page-art lead ground): Kennedy + McGowan

Motivation. Both are in the same field (short-range wireless data movement from a portable device) and address the same problem: a file/large payload exceeds what the link can carry in one shot. McGowan expressly frames SAR as the answer for BT payloads exceeding MTU, and Kennedy expressly acknowledges that the intermediary "can buffer data" and that transfer proceeds "as quickly as the wireless connections will allow." A POSITA seeking to make Kennedy's BT camera→phone→server upload robust for large images/video would have had a reason and a reasonable expectation of success in adopting McGowan's SAR. This is a classic KSR "known technique to improve a similar device in the same way."

Ground 2: Kennedy + McGowan + US 7,519,916

Adds a clean teaching for [1a]–[1b]'s "due in part to present data size and data transfer capabilities." US 7,519,916 selects a transfer profile from measured operating-environment/bandwidth characteristics and includes part-synchronization and buffer-size parameters — supplying both the adaptation rationale and part of the resume rationale. Motivation: Kennedy's own premise is that links are constrained and variable; US 7,519,916 supplies the known technique of adapting transfer behavior to measured capability — a predictable, beneficial combination.

Ground 3: Kennedy + McGowan + US 7,519,916 + Behzad (or + Pryor)

Adds [1e]'s per-destination user identifier. Behzad supplies the "content published to a user's personalized journal" association; Pryor (US 2005/0273592), which the examiner actually used in the family, supplies user information carried in the HTTP request header — i.e., a user identifier transmitted with the upload. Motivation: publishing to the user's account requires identifying the user; both references make that explicit.

Ground 4: Kennedy + McGowan + Ricoh/Fotonation/Canon

Uses the family-cited camera-upload art to cement [1pre] — Ricoh's US 6,209,048 (peripheral-integrated HTTP server / URL-based remote access) directly supplies the "web service" and HTTP transport; US 5,806,005 and US 7,246,147 confirm camera→network upload was routine.

Ground 5 (live IPR Ground 1): Aaltonen + Drescher

Operates at the HTTP/application level and therefore directly targets the claim's "application layer level" qualifier that McGowan only awkwardly reaches:

• Aaltonen: segmented "upload content" as blocks of packets, with per-packet identifiers/tables identifying placement/order and CRCs → [1c], [1d]; HTTP chunked → application layer.
• Drescher: segmented/streaming upload with interruption-resume → [1i]–[1k].
  Motivation: both target the identical problem (large media upload over an unreliable link) using the identical known mechanism (segmentation + identifiers + resume); combining them is a KSR "predictable variation." Per the docket record, the Board instituted on all challenged claims (1–5).

Ground 6 (live IPR Ground 2A/2B): Takahashi + Na (± Lind)

Multimedia messaging over MMS/HTTP supplies [1c]–[1f] (segment/sequence identifier, user identifier, HTTP web service). Motivation: multimedia messaging systems already segment and sequence media for transport and carry subscriber/recipient identities; moving an MMS-style upload to an HTTP web service was a natural, predictable step, and Na "endorses HTTP."

Ground 7 (fallback superset)

Kennedy + McGowan + US 7,519,916 + Aaltonen + Drescher + Pryor — the "belt and suspenders" combination covering every element at the application layer while preserving Kennedy's automatic media-upload motivation. Expect a petitioner to fall back to this if the Board credits Cellspin's "McBowan is not application layer" argument.

5. Motivation to combine — consolidated

Applying KSR (550 U.S. 398, 418–21) and the Graham factors:

1. Same field of endeavor / same problem. Every reference addresses movement of digital media/data from a constrained portable device over a wireless/wired link to a remote store — "[a]ny need or problem known in the field … can provide a reason for combining." The shared problem is payload > link/MTU capacity and transient connectivity.
2. Known technique, predictable result. Segmentation + per-segment identifiers + acknowledgement + resume is a standard reliable-transfer technique (McGowan for BT payloads; Aaltonen/Drescher at HTTP). Applying it to Kennedy's upload is a "simple substitution of one known, equivalent element for another to obtain predictable results."
3. Design incentive in the references themselves. Kennedy: buffers data, supports multiple destinations, transfers "as quickly as the wireless connections will allow." McGowan: SAR is mandatory for payloads > MTU. US 7,519,916: adapt transfer profile to measured bandwidth. These are explicit prompts, not hindsight.
4. Finite, predictable solutions. For "reliably send a large file over a lossy mobile link," the set is small (segment-and-ack with resume; part-synchronization). KSR "obvious to try."
5. Interoperability / compatibility. Kennedy's phone already speaks HTTP/3G; McGowan already encapsulates higher-layer protocols over BT; Na/Aaltonen already use HTTP. Combining requires no incompatible redesign.
6. Multiple-destination support is a stated goal (Kennedy's PC "can therefore buffer data"; the '336 specification itself lists Flickr/Picasa/YouTube/eBay), which motivates [1e]'s "each destination web service" user identifier.

6. Dependent claims 2–5

• Claim 2 (audio/video/image/text/digital data) — Kennedy ("pictures"), Singh '317 (environment-aware multimedia), Na/Takahashi (MMS multimedia), Behzad (journal incl. mixed media). Obvious; no separate inventive weight.
• Claim 3 (multimedia data + associated data) — Aaltonen (content + controls), Kennedy (image + associated files), Behzad (links + annotations ≈ associated data). Obvious.
• Claim 4 (configure data so the web service can reassemble) — McGowan is nearly dispositive: it teaches reassembly from tagged segments into "a single large data object"; Aaltonen teaches the upload agent reassembling blocks; Gourley teaches HTTP chunked reassembly. Obvious.
• Claim 5 (user identifier = user-name, user-password, user-device-information, user settings, or user preferences) — Pryor (user info in HTTP header), US 7,519,916 (settings/preferences/buffer parameters), Behzad (user's journal identity). These are the enumerated ordinary credential/context fields; routine design choice. Obvious.

7. Anticipated Patent Owner (Cellspin) rebuttals — and the likely outcome

From the actual POR/reply record, Cellspin's arguments are:

1. "Aaltonen fails to teach applying a segment identifier…" and "Petitioner did not identify which field is the segment identifier" (Second Gajendragad Decl., Ex. 2022 ¶¶ 24–27). → Contestable, but a § 103 (not § 102) case survives via explicit "it would have been obvious" reasoning; Aaltonen ¶ 103 does recite identifiers/tables identifying packet order.
2. "HTTP headers are text-only, not binary; the claim requires data-content ≠ identifier." (Ex. 2022 ¶¶ 23–25). → A narrow-reading argument that depends on adopting Cellspin's proposed constructions of "segment identifier"/"user identifier"; the Board's preliminary construction posture favored the ordinary meaning.
3. Bodily-incorporation / "improper for bodily incorporation of Takahashi and Na" (Pet. Reply) → Belden/KSR answer: motivation to combine ≠ bodily incorporation.
4. No motivation to combine. → Undermined by the explicit shared problem and by US 7,519,916/Kennedy prompts (§ 5).
5. Different field (BT-HID vs. camera upload). → KSR: analogous art, common problem, predictable combination.
6. No RPI / termination motions (CCP theory). → Procedural, and the Petitioner's opposition (Paper 25) argues waiver; not an obviousness argument.

Secondary considerations (§ 103 "objective indicia"): I found no evidence of commercial success, long-felt need, or unexpected results tied to the '336 claims, and no copying evidence, in the available record. Absent a nexus, these do not rebut the prima facie case.

Balance: The strongest surviving Cellspin point is the "application layer level" limitation as applied to McGowan (which sits above baseband but arguably below "application layer"). That is precisely why the live ground leads with Aaltonen/Drescher (HTTP application layer) and Takahashi/Na, and why Gourley's HTTP chapter was cited — to lock down application-layer, chunked segmentation independent of McGowan. On the page-art combination alone, Grounds 1–4 are strong on the segmentation/transfer-capability elements but weakest on [1c]/[1d] at the application layer; adding Aaltonen/Drescher (Grounds 5/7) closes that gap.

8. Bottom line

• Under pre-AIA § 103(a) with a December 28, 2007 invention date and the Board's undisputed POSITA definition, claims 1–5 of US 8,756,336 are prima facie obvious.
• Lead combination on the page's own prior-art section: Kennedy + McGowan, optionally with US 7,519,916 (transfer adaptation), Behzad/Pryor (user identifier), and Ricoh/Fotonation/Canon (web-service/HTTP upload context). Claims 2–5 fall with routine modifications.
• The combination the live IPR actually uses and on which the Board instituted: Aaltonen + Drescher (Ground 1) and Takahashi + Na (Ground 2), which better reach the "application layer level" limitation.
• The single most vulnerable element for a petitioner is [1c]/[1d] (segmentation + identifiers at the application layer level); the strongest references for that element are Aaltonen, Drescher, and Gourley's HTTP chunked transfer, not McGowan.

9. Caveats / uncertainty (stated rather than smoothed over)

• The page's "Citations (5)" are not the IPR's grounds. I reported both sets; do not conflate them. The IPR references (Aaltonen, Na, Drescher, Takahashi, Bosschaert, Foo, Gourley) appear in the live record but not in the page's examiner-citation list.
• Drescher (WO 2008/028508) is available only under § 102(e) (published after the 2007 critical date) — its status depends on its PCT filing date, US designation, and English-language publication, which I could not fully verify from the snippets.
• Singh US 2008/0109317 shares an inventor (Gurvinder Singh) with the '336 patent and is likely disqualified for § 103 under pre-AIA § 103(c)/common ownership; I therefore do not rely on it as a primary ground.
• My content descriptions of Aaltonen, Na, Takahashi, and Drescher are drawn from partial snippets (the POR/reply excerpts and exhibit lists), not full-text review; the element mappings for those references should be verified against the full documents before being relied on in a filing.
• IPR status is unsettled: the record shows institution, a sua-sponte Director Review of the institution decision, additional briefing (Order, Jan. 23, 2026), and contested termination/RPI motions. I did not find a final written decision for IPR2024-00757, and I could not confirm whether the 2026 PTAB settlement/review events reported at the family level involve this patent specifically.
• Pre-AIA vs. AIA tension: a sibling application was examined under AIA first-to-file using 2012 art (Ihara), which is inconsistent with a 2007 priority claim; the IPR applied pre-AIA/2007. The two postures can yield different prior-art sets, and the correct one is a legal question I flag rather than resolve.

Sources: Google Patents US8756336B2 (authoritative full text, supplied); Google Patents/Justia US7519916; Freepatentsonline/US20030157960 and US20020167932; IPR2024-00757 docket (docketalarm) — Institution Decision, Pet. Reply to POR (Paper 20), Pet. Opposition (Paper 25), Ex. 2022 (Gajendragad Decl.), Ex. 1028 (Traynor Suppl. Decl.); IPR exhibit list TTI-1005–TTI-1013; Cellspin v. ByteDance, No. 2:23-cv-00496 (E.D. Tex.).