US 8595778

Search Results Summary

USPTO/Google Patents (US8595778B2): Confirmed via Google Patents and related records. The patent is active, with an adjusted expiration of 2032-05-29. PTAB records (IPR2025-00846) confirm the application number 12/604,678 and patent number 8595778.

PTAB/IPR activity (relevant to "dockets"):

• IPR2025-00846 — Google LLC v. Sandpiper CDN, LLC (filed May 7, 2025; institution decision Oct 10, 2025): Discretionary Denial (matches the Google Patents metadata noting "Not Instituted – Procedural"). The petition (Fransdonk/Norris/Carle grounds) attacked claims including claims 1, 2, 3, 11, and 13.
• IPR2026-00095 — Microsoft Corporation v. Sandpiper CDN, LLC (filed Nov 3, 2025; status Pending): Patent Owner's Preliminary Response filed Feb 13, 2026; Board Director discretionary-denial referral on Mar 17, 2026. Petitioner's exhibits cite Sandpiper's district-court complaints (e.g., Sandpiper CDN, LLC v. [Microsoft Corp.](/litigations/by-plaintiff/Microsoft%20Corp.), No. 2:25-cv-00664, E.D. Tex.) and the Level 3 → Sandpiper assignment (EX1021).

CAFC 2026 dockets: I found no CAFC appeal docket specific to 8595778 in my searches. The docket-alarm and PTAB results are the only 2026 litigation-adjacent records located. I cannot confirm whether any appeal from the IPR or district-court cases has been filed at the Federal Circuit; that remains uncertain.

Patent Summary — US8595778B2

Abstract / gist: The formal abstract text was not reproduced in the source materials I have, so I paraphrase with moderate confidence: the patent discloses computer-implemented methods for authorizing delivery of a video stream to an end user in a content delivery network by having an authorization manager "dip into" a content publisher's pre-existing subscription database (e.g., satellite/cable subscriber records) to determine whether the end user is authorized, and either notifying the user of non-authorization or initiating delivery. The specification also describes a companion content-delivery manager that detects trigger signals in a stream (e.g., temporal marks) to modify delivery — for blackouts, geo-filtering, and targeted ad insertion.

Independent claims (plain-language overview): The claims text was not included in the patent text provided to me, so the following is reconstructed from the specification (FIGS. 16–22) and the IPR2025-00846 petition excerpts. Treat claim numbering and scope as approximate:

• Claim 1 (method): A computer-implemented method for authorizing delivery of a video stream to an end user, where an authorization manager (i) receives the end user's request for the stream, (ii) queries a subscription database associated with the content publisher (not an Internet-specific database), (iii) processes the reply to determine authorization, (iv) if not authorized, transmits a notification to the end user, and (v) if authorized, initiates delivery of the video stream. (IPR excerpts confirm the "receiving a request from the end user" and subscription-database query elements.)

• Claim 2 (dependent on claim 1): Adds that the authorization manager receives metadata from the end user (e.g., a token and/or cookie) that identifies the content publisher associated with the end user, allowing authorization without a separate subscription-database query.

• Claim 3 (dependent on claim 2): Recites that the metadata is at least one of a token and a cookie (confirmed verbatim in the IPR petition).

• Claim 11 (method, later in the chain): Appears directed to processing a second request from a second end user for delivery of the video stream (multi-user scenario), with similar authorization processing.

• Claim 13 (dependent on claim 11): Recites that processing the second request to determine authorization includes analyzing at least one of a token and a cookie associated with the second request (confirmed verbatim in the IPR petition).

Additional independent-claim scope suggested by the specification (FIGS. 19–22) likely includes: a claim using proximity parameters (e.g., IP address → geographic region) to determine geo-restrictions/blackout applicability, and possibly system / computer-readable-medium counterparts — but I could not verify these without the claims text.

Uncertainty Notes

• The formal abstract and full claims set were absent from the patent text provided; claim descriptions above are inferred from the specification and IPR petition, not quoted from the granted claims.
• No CAFC 2026 appeal docket for 8595778 was found; only PTAB (IPR2025-00846, IPR2026-00095) and district-court activity (C.D. Cal., E.D. Tex.) were identified. If you need claim-by-claim verbatim text, the authoritative source is the USPTO full-text at patft.uspto.gov or the Patent Center record for US 12/604,678.

Based on my searches (prioritizing live search results over training data), here is what I found regarding litigation involving US Patent No. 8,595,778 ("User authentication in a content delivery network," currently assigned to Sandpiper CDN, LLC, formerly Level 3 Communications LLC):

Known Litigation Involving US 8,595,778

1. Sandpiper CDN, LLC v. Google LLC — District Court Case

• Plaintiff: Sandpiper CDN, LLC
• Defendant: Google LLC
• Jurisdiction: U.S. District Court, Central District of California (Western Division)
• Case Number: 2:24-cv-03951 (later styled 2:24-cv-03951-AB (RAO) after reassignment; originally assigned to Judge Philip S. Gutierrez/Magistrate Rozella A. Oliver, later before Judge Andre Birotte Jr.)
• Filing Date: May 10, 2024
• Patents-in-suit: The '778 patent is one of six asserted patents (8,595,778; 8,478,903; 8,645,517; 8,719,886; 9,021,112; 10,924,573), targeting Google's Cloud CDN and Media CDN (including YouTube/YouTube TV). The complaint's Exhibit B is the '778 patent.
• Status / Outcome (as of the current date):
  • Open/active.
  • Sept. 16, 2024: The court granted Google's motion to dismiss Counts II and IV with prejudice (Alice/§101 subject-matter-eligibility grounds), and the court's order addressed the '778 patent's claims as directed to the abstract idea of authorizing/restricting access (see Dkt. 28, C.D. Cal.).
  • Jan. 22, 2026: The case was stayed by stipulation/order pending final resolution of Google's related IPR proceedings (IPR2025-00806, IPR2025-00826, IPR2025-00860, IPR2025-00969, IPR2025-01010), with periodic joint status reports required and trial dates vacated.
  • Sources: UniCourt, DocketAlarm (exhibits filed in IPR2026-00095), RPX/Mondaq coverage.

2. Google LLC v. Sandpiper CDN, LLC — IPR Proceeding (PTAB)

• Petitioner: Google LLC
• Patent Owner/Respondent: Sandpiper CDN, LLC
• Jurisdiction: USPTO Patent Trial and Appeal Board (PTAB)
• Case Number: IPR2025-00846
• Filing Date: May 7, 2025
• Patent Challenged: US 8,595,778 (application no. 12/604,678)
• Status / Outcome: Not instituted — discretionary denial ("Procedural" per Unified Patents' PTAB data). Institution decision issued October 10, 2025; the proceeding was denied institution and terminated.
  • Sources: ipverse.greyb.com PTAB case page; Unified Patents portal (referenced on the Google Patents file wrapper); PTACTS exhibit list (dated Sept. 18, 2025) identifying U.S. Patent No. 8,595,778 and the related district court case.

Related Portfolio Litigation (NOT involving the '778 patent — listed for completeness)

• Sandpiper CDN, LLC v. Microsoft Corporation, No. 2:25-cv-00664 (E.D. Tex., filed June 26, 2025): The amended complaint (Oct. 30, 2025) asserts patents 8,478,903; 8,924,466; 9,456,053; 9,228,883; 9,762,692; and 10,701,173 — the '778 patent does not appear to be asserted in this case.
• Sandpiper CDN, LLC v. Comcast Cable Communications LLC, No. 2:24-cv-00886 (E.D. Tex.): Based on available summaries, this case asserts patents including 9,628,347 and 9,660,876 — not the '778 patent.
• Microsoft Corporation v. Sandpiper CDN, LLC, IPR2026-00095 (PTAB, filed ~2026): This IPR cites the Google case docket, but the available record does not indicate that it challenges the '778 patent (the patents in the underlying Microsoft case do not include '778).

Summary

The '778 patent is directly involved in two known proceedings: (1) the district court case Sandpiper CDN, LLC v. Google LLC, 2:24-cv-03951 (C.D. Cal.), filed May 10, 2024, which is currently stayed pending Google's IPRs; and (2) the PTAB proceeding IPR2025-00846, Google's IPR petition filed May 7, 2025, which was denied institution (discretionary) on October 10, 2025.

Note on confidence: I compiled this from live search results (UniCourt, DocketAlarm, RPX/Mondaq, PTACTS, ipverse.greyb.com, CourtListener/Justia) rather than from PACER directly. The Google district court docket is the most authoritative source for the precise current status of the stay and the dismissed counts; if you need exact docket-level confirmation, checking PACER for 2:24-cv-03951 would be the next step.

Proceedings overview

There is one AIA trial proceeding on file for US 8,595,778 — IPR2025-00846 (Google LLC v. Sandpiper CDN, LLC) — which was denied institution on discretionary/procedural grounds on 2025-10-10; no proceeding has been instituted, so no claims have been canceled or sustained by the PTAB, none have settled, and none are currently active. Defensive posture: the patent is PTAB-untested (all claims remain fully intact for PTAB purposes), so an IPR-based defense will not get you a cancellation — but the far bigger weapon is already on the record: the C.D. Cal. district court held the '778 claims ineligible under 35 U.S.C. § 101 (Sandpiper CDN, LLC v. Google LLC, No. 2:24-cv-03951, ECF No. 28 at 15 (Sept. 16, 2024)), and Sandpiper dropped the patent from its amended complaint (reserving appeal). A defendant facing the '778 today should lead with § 101, not IPR.

⚠️ Flag — structured data vs. live sources: The "PTAB proceedings on file" block (USPTO ODP ingest) reports no AIA trials, but that ingest is stale. Live sources (PTACTS, GreyB/IPVerse, Docket Alarm, Google Patents metadata) confirm IPR2025-00846 exists and was denied institution on 2025-10-10. Per the operating rules, I treat the live web sources as authoritative and flag the contradiction.

⚠️ Flag — correction to the prior summary: the prior section listed IPR2026-00095 (Microsoft v. Sandpiper) as relevant to this patent. It is not a proceeding on the '778. IPR2026-00095 challenges U.S. 8,478,903 ("Content Delivery System") — a different patent in the same Level 3 → Sandpiper portfolio (Docket Alarm and AI-Lab case pages both confirm Patent No. 8,478,903). Do not cite it as activity on 8595778.

IPR2025-00846 — Google LLC v. Sandpiper CDN, LLC

• Type: Inter Partes Review (35 U.S.C. § 311)
• Filed: 2025-05-07
• Status: Discretionary Denial — i.e., "Not Instituted – Procedural" (Google Patents metadata; GreyB/IPVerse). The proceeding terminated with the denial; no trial was ever instituted.
• Judge panel: Not identified in any source I could retrieve. Because the denial post-dates the USPTO's 2025-03-26 Interim Process for PTAB Workload Management, a discretionary denial of this kind is decided at the Director level (in consultation with PTAB judges) before any merits panel assignment — but I cannot name the deciding officials and will not guess.
• Petition grounds (from the petition itself, retrieved via PTACTS; exhibits EX1005–EX1011, EX1013):
  • Ground 1 — Anticipation under § 102 by Fransdonk (US 2003/0165241 A1): challenged the independent method/system claims and dependents, including claims 1, 6, 8, 10–12, 14–15, 17–18 as excerpted in the petition (the petition's claim-by-claim section shows claim 6 [blackout-rule limitation], claim 8 [non-subscriber notification], claim 11 [second end-user request], claim 12 [same-end-user determination], claim 14 [system claim], and claims 15, 17, 18).
  • Ground 2 — Obviousness under § 103 over Fransdonk + Norris (US 6,718,328 B2): claims 2, 3, and 13. Claim 3 (verbatim): "The computer-implemented method as in claim 2, wherein the metadata is at least one of a token and a cookie." Claim 13 (verbatim): "The computer-implemented method of claim 11, wherein processing the second request to determine that the second end user is authorized to receive delivery of the video stream includes analyzing at least one of a token and a cookie, wherein the at least one of a token and a cookie is associated with the second request received from the second end user."
  • Ground 3 — Obviousness under § 103 over Fransdonk + Carle (US 2007/0198839 A1): claims 4–8.
  • Additional references were listed as exhibits (Risan, Chatani, Bi, Bacso, Foti, Krishnamurthy, Mao, Biliris, Trimper), but I could not confirm they were asserted as separate grounds; treat the ground list above as the confirmed core.
• Institution decision: Denied — 2025-10-10 (discretionary/procedural; no merits decision). I was unable to retrieve the actual decision paper, so I will not quote its reasoning. What the record does show: (i) the petition itself litigated the Fintiv factors, arguing factor 2 favored institution because the '778 had been dropped from the parallel C.D. Cal. case; (ii) the '778 claims had already been adjudicated ineligible under § 101 in Sandpiper v. Google, C.D. Cal., ECF No. 28 (2024-09-16), which the new discretionary framework expressly permits the Director to weigh ("whether the PTAB or another forum has already adjudicated the validity or patentability of the challenged patent claims"); and (iii) the metadata label is "Not Instituted – Procedural." The specific basis (Fintiv vs. § 325(d) vs. prior adjudication) is unconfirmed — do not represent it as decided.
• Final Written Decision: None — a FWD issues only after institution, and institution was denied. No claim of the '778 has ever been held unpatentable (or patentable) by the PTAB.
• Settlement / termination: No settlement. The case terminated administratively upon the institution denial (GreyB lists a termination date contemporaneous with the 2025-10-10 decision).
• Appeal: None identified. Institution denials are not appealable as of right, and I found no Federal Circuit docket for IPR2025-00846 or for the related § 101 dismissal of the '778 counts (the C.D. Cal. docket, 2:24-cv-03951, is stayed pending instituted IPRs on other Sandpiper patents; whether Sandpiper ever pursues its reserved appeal of the § 101 order is unconfirmed).
• Defensive value: Modest but real. The denial means no § 315(e)(2) estoppel attaches — Google (and everyone else) remains free to raise Fransdonk, Norris, Carle, and any other art in district court. But it also means the PTAB has given you nothing on the merits. The decisive fact for a defendant is not this IPR — it is the district court's § 101 holding, which is now part of the public record in Google's IPR exhibits (EX1023, cited in Microsoft's IPR2026-00095 papers) and can be deployed in any venue.

Strategic summary

Claims of 8595778: CANCELED vs. SUSTAINED vs. UNTESTED. At the PTAB: zero claims canceled, zero claims sustained — the entire patent is UNTESTED (institution was denied before any merits review). Outside the PTAB, the C.D. Cal. court dismissed the asserted '778 counts as ineligible under § 101 (ECF No. 28 at 15, 2024-09-16), and Sandpiper's amended complaint dropped the patent while "reserv[ing] its right to appeal" (First Amended Complaint, ECF No. 57 at 1, 28). So the accurate picture is: all claims formally alive and enforceable on their face, but judicially tagged as ineligible subject matter in a published order — a tag any defendant should put in front of a judge on day one.

Estoppel landscape. Because IPR2025-00846 was denied, not instituted, no § 315(e)(2) estoppel bars Google or anyone else from re-asserting any ground — including the very grounds in the denied petition — in district court. For a defendant currently facing the '778: every § 102/§ 103 ground remains available (Fransdonk; Fransdonk+Norris; Fransdonk+Carle; anything else you find), and the § 101 ineligibility holding is an independent, already-litigated path. Nothing about the IPR closes off prior art for you.

Pattern signals. This patent sits in the middle of a coordinated 2024–2026 campaign by Sandpiper CDN, LLC — a Delaware entity formed in March 2024 that acquired the Level 3 CDN portfolio in April 2024 (recorded 2024-04-26) and immediately sued Google (C.D. Cal., 2:24-cv-03951), Comcast (E.D. Tex., 2:24-cv-00886), and Microsoft (E.D. Tex., 2:25-cv-00664). Google responded with a wave of IPRs across the portfolio — IPR2025-00846 (the '778), plus IPR2025-00860 (10,924,573), IPR2025-00969 (8,478,903), IPR2025-01010 (10,057,322) — and Microsoft followed with IPR2026-00095 ('903, now instituted), IPR2026-00174 ('692), and IPR2026-00190 ('053, denied). Note: the "Unified Patents" label in the Google Patents PTAB metadata is a data-attribution credit, not a petitioner — Unified Patents is not a party. The '778 has attracted exactly one IPR, and it was turned away on discretion before the merits — consistent with the fact that the parallel § 101 dismissal already neutralized the patent in the lead litigation. Sandpiper has shown no appetite for PTAB appeals on this patent (there is nothing to appeal).

Recommended next steps

1. Lead with § 101, not IPR. The single highest-value document in the file is the C.D. Cal. order: Sandpiper CDN, LLC v. Google LLC, No. 2:24-cv-03951, Order Granting Defendant's Motion to Dismiss Counts II and IV, ECF No. 28 at 15 (C.D. Cal. 2024-09-16) — holding the '778 claims ineligible under 35 U.S.C. § 101. It is already of record in the PTAB world (Google's EX1023 in IPR2025-00846; cited in Microsoft's IPR2026-00095 papers as "The district court determined two of those patents—U.S. 8,595,778 ('778 patent) and 8,719,886 ('886 patent)—are invalid under § 101"). File a § 101 motion to dismiss on the Alice framework and cite that order as persuasive authority; the claims (authorization by "dipping" into a pre-existing subscription database, token/cookie verification, geo-blackout) are classic abstract-idea candidates.
2. If you still want an IPR: nothing is estopped and the patent is un-instituted, so a fresh petition is procedurally available — but expect the same discretionary headwinds, and note the PTAB's 1-year trial clock only starts if the Director declines to deny. Realistically, a second petition on this patent is low-ROI while the § 101 order stands.
3. Monitor, don't confuse: IPR2026-00095 (instituted 2026-04-22) and the other portfolio IPRs concern different patents ('903, '573, '322, '692, '053) — useful for reading Sandpiper's litigation posture, but they have no direct effect on the '778.
4. Verify the appeal path: confirm on the C.D. Cal. docket (2:24-cv-03951) whether Sandpiper has appealed the § 101 dismissal to the Federal Circuit — I found no CAFC docket as of 2026-08-26, but that status should be re-checked before you rely on the order's finality.

Sources: PTACTS petition record for IPR2025-00846 (exhibit list and ground excerpts: ptacts.uspto.gov petition no. 1557719); GreyB/IPVerse case page (https://ipverse.greyb.com/ptab-web/cases/case-details/IPR2025-00846); Docket Alarm IPR2026-00095 pages (EX1031, EX1045, Microsoft Reply to Joinder Opposition); Google Patents metadata for US8595778B2 (status, assignments, "Not Instituted – Procedural"). No FWD exists, so no PTAB decision link is possible — do not cite one.

Inventors

• Ian Maloney — sole named inventor. Employer at filing: Level 3 Communications, LLC (the original assignment from Maloney to Level 3 was recorded on the filing date, 2009-10-23, per the Google Patents legal-event record: "ASSIGNMENT OF ASSIGNORS INTEREST (SEE DOCUMENT FOR DETAILS). Assignors: MALONEY, IAN").

No unusual inventor pattern: a single inventor assigned to the applicant at filing; there is no evidence of inventor departures preceding a portfolio sale, and this patent is a Level 3-originated application (filed 2009), not one of the legacy Sandpiper Networks/Savvis patents (RPX notes Farber and Swart are inventors on only one of the six Sandpiper-sued patents — 8,478,903 — not on 8,595,778).

Original assignee

• Level 3 Communications, LLC (named assignee on the issued patent).
• Product status: Yes — Level 3 was a major operating CDN/telecom provider and shipped products embodying the claimed subject matter (the patent covers user authentication in a CDN, squarely within Level 3's CDN business, which it operated from the mid-2000s until exiting the market in 2023).
• Current status: Operating, but no longer as Level 3 — acquired by CenturyLink in 2017 and rebranded Lumen Technologies. Level 3 exited the CDN business in late 2023 and sold off CDN assets (per Sandpiper's own complaint and PTACTS filing).

Assignment timeline

The Assignment Center does have records for this patent — two recorded conveyances:

• 2009-10-23 (executed same day as filing; exact execution date not separately confirmed) / recorded 2009-10-23 — reel/frame not confirmed from available sources (Google Patents legal events show the recording on 2009-10-23; the reel/frame string was not retrievable in my searches)

  • Conveyance: Assignment of Assignor's Interest
  • Assignor: Ian Maloney
  • Assignee: Level 3 Communications, LLC
  • Correspondent: not confirmed from available sources
  • Context: employee-inventor assignment to the applicant at filing — routine, not a red flag.

• 2024-04-24 (assignment date per RPX reporting; Sandpiper's complaint separately pleads the sale agreement was signed 2024-03-29 — likely purchase-agreement date vs. assignment-execution date) / recorded 2024-04-26 — Reel 068256 / Frames 0091–0115

  • Conveyance: Patent Assignment Cover Sheet (assignment of patent rights)
  • Assignor: Level 3 Communications, LLC
  • Assignee: Sandpiper CDN, LLC (Delaware LLC, formed 2024-03-21)
  • Correspondent: not confirmed from available sources
  • Context: bulk transfer of 80+ US patents from an operating company (Level 3/Lumen) to a newly formed Delaware LLC, which filed its first infringement suits within weeks (RPX; IPR2026-00190 Exhibit EX1036, "Patent Assignment Cover Sheet from Level 3 Communications, LLC to Sandpiper CDN, LLC, Reel No. 068256, Frames 0091-115").

No other recorded assignments were identified (no intervening LLC chain, no security agreement, no merger, no license recorded for this patent number). The chain is: Maloney → Level 3 (2009) → Sandpiper CDN (2024).

Timeline diagram

timeline
    title Ownership of US 8595778
    2009 : Filed by Level 3 Communications
    2013 : Patent issued
    2024 : Assigned to Sandpiper CDN LLC
         : First infringement suit filed

NPE / troll-pattern signals

1. Shell-entity transfer — Present. Sandpiper CDN, LLC was formed in Delaware on 2024-03-21 (RPX), received 80+ patents from operating company Level 3 via the Reel 068256/0091–0115 recording (2024-04-26), and has no products in commerce — its own public filing describes the portfolio as "assets... now managed by Sandpiper CDN" and solicits "partnership and licensing opportunities" (USPTO PTACTS record). This is a single-purpose, licensing-only Delaware LLC — the "IP holding LLC formed immediately before litigation" pattern, not a name-based inference.

2. Known asserter in the chain — Present. Sandpiper CDN is tracked as a high-frequency plaintiff by RPX Insight ("Recently formed Delaware plaintiff Sandpiper CDN, LLC has launched litigation, suing Alphabet (Google) (2:24-cv-03951) in the Central District of California") and is the listed assignee in Unified Patents' portal. It has sued Google (C.D. Cal., May 2024), Microsoft (E.D. Tex., 2:25-cv-00664), and faces IPR2025-00846 (Google), IPR2026-00095 and IPR2026-00190 (Microsoft). Also relevant: Level 3's prior bulk transfer of ~110 patents to Optic153 LLC (subsidiary of Equitable IP Corporation, a known monetization/NPE house) beginning April 2017 (RPX).

3. Repeat correspondent across the chain — Unclear. I could not retrieve the correspondent-of-record attorney names for either recording (2009-10-23 or Reel 068256). No recurrence can be established from available data; this signal is unsupported rather than disproven.

4. Cascading transfers — Not present for this patent. Only two recorded conveyances, 15 years apart. (Level 3's broader NPE-transfer pattern via Optic153 exists, but it involves different assets and does not touch this patent's chain.)

5. Pre-litigation transfer — Present. Assignment recorded 2024-04-26 (Reel 068256/0091–0115); Sandpiper filed its first suit against Google in C.D. Cal. in early May 2024 (RPX article dated May 12, 2024) — roughly two weeks after recording, well inside the 6-month window. The LLC itself was formed only ~5 weeks before the recording (2024-03-21). The chain was arranged to enable immediate assertion with clean standing.

6. Bankruptcy fire-sale — Not present. Level 3/Lumen is a solvent operating company; the transfer was a voluntary market-exit sale (per the complaint, Level 3 "decided to exit the CDN market in 2023 and began selling off its CDN assets"), not a Chapter 7/11 proceeding.

7. Privateering — Present, with a caveat. Level 3 — the operating CDN company (now Lumen) — transferred the patents to an NPE that immediately asserted them against CDN competitors (Google Cloud CDN/Media CDN, Microsoft Azure CDN). The complaint itself frames the sale as monetizing infringement that "depressed its revenue and profit." Caveat: Level 3 had exited the CDN market before the transfer, and I found no SEC or other evidence of a retained financial interest or revenue share, so the classic "operating company funds NPE to attack rivals while staying in the market" element is not confirmed — but the operating-company-to-asserting-NPE pattern (including Level 3's prior Optic153/Equitable IP transfers) is documented.

8. Defensive aggregator (anti-NPE) — Not present. The chain terminates at Sandpiper CDN, an asserting entity — not RPX, AST, LOT, Unified Patents, or OIN.

Verdict

NPE — high confidence.

Three strong signals drive this: (1) the transfer of 80+ patents from operating company Level 3 to a newly formed, product-less Delaware LLC (Sandpiper CDN, formed 2024-03-21) recorded at Reel 068256/Frames 0091–0115 on 2024-04-26; (2) Sandpiper's immediate, RPX- and Unified-Patents-tracked assertion campaign against Google and Microsoft filed within weeks of the recording; and (3) Level 3's established pattern of monetization transfers to NPEs (Optic153 LLC / Equitable IP Corporation, from 2017). The only weak link is the correspondent-identity signal, which could not be verified from available sources. Verification: https://assignmentcenter.uspto.gov/ (search by patent number 8595778); corroborating exhibit at IPR2026-00190, EX1036 (Reel 068256, Frames 0091–0115).

Prior Art Analysis — U.S. Patent 8,595,778 B2 ("User authentication in a content delivery network")

0. Patent identity verification (no look-alike numbers)

I searched for 8595778 specifically. Confirmed record:

• Patent number: US 8,595,778 B2 (publication US20100122303A1)
• Application: 12/604,678, filed Oct 23, 2009; priority date Nov 12, 2008
• Grant date: Nov 26, 2013; Inventor: Ian Maloney; Assignee: Level 3 Communications, LLC (now Sandpiper CDN, LLC)
• Status: Active, adjusted expiration May 29, 2032

I did not confuse this with similar numbers (e.g., 8,595,777/779, 8,597,778, etc.); no such near-identical records were returned. Note that the patent has a sibling patent, US 8,719,886 B2 ("Dynamic processing of streamed content"), filed the same day with the same priority date, which shares a large portion of the specification (FIGS. 9–15). Several of the citations discussed below map to the trigger/stream-modification subject matter of the sibling rather than to the granted authentication claims of 8,595,778 — I flag this where relevant.

The granted claims (verified from the Justia record and IPR2025-00846 petition)

• Claim 1 — method: (i) receive request from end user for delivery of a video stream associated with a content publisher; (ii) query a subscription database associated with the content publisher; (iii) process the reply to determine authorization; (iv) perform at least one of: transmit a non-authorization notification, or initiate delivery when authorized.
• Claim 2 — receiving metadata from the end user, processing it to identify the content publisher associated with the end user, and matching it to the publisher of the stream.
• Claim 3 — metadata is at least one of a token and a cookie (note: the Justia text has a typo "as in claim 3"; the IPR petition confirms claim 3 depends on claim 2).
• Claims 4–8 — proximity parameters (e.g., IP → geographic location), geo-based non-authorization, blackout rules with time/geographic restrictions, group subscription parameters, termination of an already-initiated delivery.
• Claims 9–10 — subscriber verification table entries (non-subscriber / subscriber).
• Claims 11–13 — second request from a second end user; claim 13 adds analysis of a token and/or cookie associated with the second request.

1. Citation lists located

Two overlapping but distinct citation sets were found:

Set A — Google Patents "Patent Citations" for the 8,595,778 family (identical to the EPO search report of family member EP2351298A4): four U.S. published applications + one NPL (a Wikipedia article). These are the four references that the EPO report associated with this family.

Set B — Justia's "References Cited" record for 8,595,778 (mirrors the USPTO face-of-patent list): a longer list of U.S. patents and published applications. My snapshot of this list was truncated (it cut off mid-way through the published-application section), so the list below is likely incomplete — in particular, the patent may also cite foreign-patent documents and "Other Publications" that I could not retrieve.

⚠️ Confidence flag: I could not load the full text of the cited references. Citation data (number, date, inventor) below is verified from search-result snippets; descriptions and § 102 claim mappings are my analysis from the verified titles/subject matter, not from full-text reads, except where noted. A rigorous anticipation determination requires the complete text of each reference.

2. Set A — Family-level patent citations (Google Patents / EP search report)

A1. US20020131511A1 — "Video tags and markers" — Ian Zenoni

• Priority: Aug 25, 2000; Published: Sep 19, 2002
• Description: Systems for associating "tags" and "markers" with video content (e.g., in a stream) to signal events such as advertising or other program points.
• § 102 relevance to the granted claims of 8,595,778: Low. It teaches video-stream markers analogous to the specification's "trigger signals" (FIGS. 9–15), which are not claimed in the granted authentication claims. It does not disclose the subscription-database authorization "dip" of claim 1, the token/cookie metadata of claims 2–3/13, the proximity/blackout logic of claims 4–8, or the subscriber-verification table of claims 9–10. Strongest relevance is to the sibling patent's trigger claims.

A2. US20080104624A1 — "Method and system for selection and scheduling of content outliers" — Motorola, Inc.

• Published: May 1, 2008 (application priority from 2006)
• Description: Selection and scheduling of "outlier" content items within a content line-up.
• § 102 relevance: Low/none for the granted claims — it does not address end-user authorization against a publisher subscription database.

A3. US20080201736A1 — "Using Triggers with Video for Interactive Content Identification" — ICTV, Inc.

• Published: Aug 21, 2008
• Description: Using trigger data embedded in video to identify and launch interactive content.
• § 102 relevance: Low for the granted claims; again aimed at the trigger-processing features of the shared specification (sibling patent's claims), not the authentication claims.

A4. US20080263056A1 — "Electronic content asset publication system" — Youbiquity, LLC

• Published: Oct 23, 2008
• Description: Publication and management of electronic content assets.
• § 102 relevance: Low/none for the granted claims on the available information.

A5. NPL — ANONYMOUS, "Geolocation software," Wikipedia (version of Dec 19, 2007; XP055375609)

• Date: Dec 19, 2007 (publication date of the cited Wikipedia revision)
• Description: Encyclopedia description of geolocation software — determining the physical/geographic location of an Internet user from data such as IP address.
• § 102 relevance: Moderate for claims 4–8, which require "processing proximity parameters … specify[ing] a geographic location of the end user" and deriving geo-based restrictions. As an NPL reference dated Dec 19, 2007 (pre-priority date Nov 12, 2008), it is prior art under § 102. It plausibly teaches the IP-address-to-geographic-region element (claim 4's proximity-parameter processing and the claim 5 geo/time blackout premise), but it does not by itself teach the claimed blackout rules, subscription parameters, or delivery-restriction steps — so standalone § 102 anticipation of claims 4–8 is unlikely; it is better used in a § 103 combination.

3. Set B — USPTO "References Cited" (per Justia record of 8,595,778)

Verified data from the snippet (number / inventor / date). I group them by subject so the § 102 relevance can be assessed coherently.

B1. Level 3 / Farber content-delivery-network patents (self-citations)

• 6,654,807 — Farber — Nov 25, 2003

• 6,928,442 — Farber — Aug 9, 2005

• 7,054,935 — Farber — May 30, 2006

• 2004/0139097 — Farber — Jul 15, 2004

• 2005/0114296 — Farber — May 26, 2005

• 2005/0198334 — Farber — Sep 8, 2005

• 2006/0218265 — Farber — Sep 28, 2006

• 2007/0233705 — Farber — Oct 4, 2007

• 2007/0233706 — Farber — Oct 4, 2007

• 2007/0233846 — Farber — Oct 4, 2007

• Description: Level 3's own CDN/edge-server patents (the specification's tiered CDN, director-agent redirection, and "cookies and data from other databases" authentication concepts derive directly from this family).

• § 102 relevance: These are the closest family-internal art to claim 1's environment (request for stream delivery over a CDN). They may disclose request handling, user redirection, and cookie-based identification, which is relevant to claims 2–3 (token/cookie metadata). However, I could not verify that any single Farber reference discloses the full claim-1 combination (explicit query of a subscription database associated with the content publisher + authorization notification/initiation). Treat as candidate anticipatory references requiring full-text element-by-element mapping; more likely they support § 103 obviousness arguments. (Note: in IPR2025-00846 the petitioner instead used Fransdonk/Norris/Carle — a different art set.)

B2. Interactive TV program guide / EPG references

• 6,668,133 — Yuen et al. — Dec 23, 2003

• 6,687,906 — Yuen et al. — Feb 3, 2004

• 6,732,369 — Schein et al. — May 4, 2004

• 6,760,537 — Mankovitz — Jul 6, 2004

• 6,799,326 — Boylan, III et al. — Sep 28, 2004

• 6,799,327 — Reynolds et al. — Sep 28, 2004

• 6,850,693 — Young et al. — Feb 1, 2005

• 6,859,799 — Yuen — Feb 22, 2005

• 7,039,935 — Knudson et al. — May 2, 2006

• 7,069,576 — Knudson et al. — Jun 27, 2006

• 2003/0188311 — Yuen et al. — Oct 2, 2003

• 2004/0010806 — Yuen et al. — Jan 15, 2004

• Description: Classic interactive-television-program-guide/EPG patents (schedules identifying when content is available to authorized users — the specification's "data repository (e.g., Electronic Programming Guide)" concept).

• § 102 relevance: Marginal to the granted claims. These teach schedule/EPG data, not the publisher subscription-database authorization dip of claim 1 or the token/cookie/geo-limitations of claims 2–8. They map to specification features (EPG querying) that were not carried into the granted claims of 8,595,778 (they are in the sibling's claims).

B3. Blackout / conditional-access-adjacent references (relevant to claims 4–8)

• 6,742,183 — Reynolds et al. — May 25, 2004

• 6,745,391 — Macrae et al. — Jun 1, 2004

• 6,756,997 — Ward, III et al. — Jun 29, 2004

• 6,762,337 — Belknap — Jul 13, 2004

• 6,963,910 — Belknap — Nov 8, 2005

• 6,965,890 — Dey — Nov 15, 2005

• 7,039,633 — Dey — May 2, 2006

• 7,206,748 — Gruse — Apr 17, 2007

• 7,404,010 — Gardell et al. — Jul 22, 2008

• 7,487,529 — Orlick — Feb 3, 2009

• Description: I have not verified the titles/texts of these. Based on subject context (TV distribution, rights/conditional access), several are plausibly relevant to geo/time-based delivery restrictions (claims 4–8), but I cannot responsibly assert specifics.

• § 102 relevance: Uncertain — requires full-text review. Do not rely on my characterizations here.

B4. Other cited U.S. patents and published applications (data verified; content not verified)

• 6,567,606 — Milnes et al. — May 20, 2003

• 6,587,837 — Spagna — Jul 1, 2003

• 6,588,013 — Lumley et al. — Jul 1, 2003

• 6,862,264 — Moura et al. — Mar 1, 2005

• 6,859,791 — Spagna — Feb 22, 2005

• 7,103,564 — Ehnebuske — Sep 5, 2006

• 7,110,984 — Spagna — Sep 19, 2006

• 7,117,259 — Rohwer — Oct 3, 2006

• 7,162,468 — Schwartz — Jan 9, 2007 (verified title: "Information retrieval system")

• 7,188,085 — Pelletier — Mar 6, 2007

• 7,647,418 — Ash et al. — Jan 12, 2010

• 7,720,432 — Colby et al. — May 18, 2010

• 2001/0029610 — Corvin et al. — Oct 11, 2001

• 2001/0047298 — (inventor name truncated in my snapshot) — Nov 29, 2001

• § 102 relevance: Not determinable from available data. I explicitly decline to guess titles or content for these; any anticipation claim against them would need their full texts. The Spagna entries (6,587,837; 6,859,791; 7,110,984) may relate to content-distribution/streaming architectures, but I could not verify.

4. Bottom-line § 102 assessment

Key caveats:

1. The full citation list on the face of the patent is larger than what I could snapshot (the Justia list was truncated mid-way; foreign documents and "Other Publications" were not retrieved). For a complete enumeration, pull the "References Cited" page from USPTO Patent Center (application 12/604,678) or the Google Patents "Citations" tab for US8595778B2.
2. My claim mapping is analytical inference from verified titles/assignee context, not from full-text reads. Several references (EPG set, trigger set) map to specification features that are not in the granted claims, so they are poor § 102 candidates against 8,595,778 itself.
3. The live IPR record (IPR2025-00846, Google v. Sandpiper; and IPR2026-00095, Microsoft v. Sandpiper) used a different prior-art set (Fransdonk, Norris, Carle) than the examiner-cited references above — if your goal is litigation-focused prior art, those IPR references are the ones currently being tested, but they are not the patent's own backward citations.

Obviousness Analysis — U.S. Patent 8,595,778 B2 ("User authentication in a content delivery network")

A. Scope note and source limitations

Before the merits, three caveats about the record I am working from:

1. The Google Patents page text provided to me does not reproduce the patent's formal "Citations" (prior-art) section, the abstract, or the granted claim set. I therefore reconstruct (a) the cited art from a Justia mirror of the patent and (b) the claims and the invalidity grounds from the IPR2025-00846 petition (Google LLC v. Sandpiper CDN, LLC), portions of which are quoted in the search results. Claim text below is reconstructed and approximate — for verbatim claim language, the authoritative source is the USPTO Patent Center record for US 12/604,678.
2. IPR2025-00846 was discretionarily denied on procedural grounds ("Not Instituted – Procedural," per the Google Patents metadata; institution decision Oct. 10, 2025). The PTAB therefore never reached a merits decision on obviousness. The analysis below is a merits-style assessment, not a statement of what the Board held.
3. One reference used in the petition, Carle, could not be fully identified (patent/publication number unverified in the available search results). I flag it explicitly where it appears.

B. Legal framework applied

Under 35 U.S.C. § 103, a claim is unpatentable if the differences between the claimed subject matter and the prior art are such that the subject matter as a whole would have been obvious at the time of invention to a person having ordinary skill in the art (POSITA). The Graham factors frame the inquiry: (1) scope and content of the prior art; (2) differences between the prior art and the claims; (3) level of ordinary skill; (4) secondary considerations. Under KSR Int'l Co. v. Teleflex Inc. (2007), obviousness can be shown by combining prior-art elements "according to known methods" to yield "predictable results," by a "design need" or "market pressure" to solve a known problem, and by "obvious to try" where a finite number of identified, predictable solutions exists. A motivation to combine may be found in the references themselves, in the nature of the problem, or in the general knowledge of the POSITA.

POSITA profile: a person with a bachelor's degree in computer science, computer engineering, or a related field and roughly 2–4 years of experience in networked content delivery, streaming media (RTSP/RTP), conditional access/DRM, and web-based authentication (tokens, cookies, subscriber databases), or equivalent experience.

C. Prior-art landscape

1. Art of record (Examiner-cited during prosecution)

The Justia mirror of US 8,595,778 lists the references of record, which cluster in the TV/EPG, conditional-access, and CDN fields: Milnes (US 6,567,606), Spagna (US 6,587,837; 6,859,791; 7,110,984), Lumley (US 6,588,013), Farber (US 6,654,807; 6,928,442; 7,054,935 — CDN streaming), Yuen (US 6,668,133; 6,687,906; 6,859,799), Schein (US 6,732,369), Reynolds (US 6,742,183; 6,799,327), Macrae (US 6,745,391), Ward (US 6,756,997), Mankovitz (US 6,760,537), Belknap (US 6,763,377; 6,963,910), Boylan (US 6,799,326), Young (US 6,850,693), Moura (US 6,862,264), Dey (US 6,965,890; 7,039,633), Knudson (US 7,039,935; 7,069,576), Ehnebuske (US 7,103,564), Rohwer (US 7,117,259), Schwartz (US 7,162,468), Pelletier (US 7,188,085), Gruse (US 7,206,748), Gardell (US 7,404,010), Orlick (US 7,487,529), Ash (US 7,647,418), Colby (US 7,720,432), Corvin (US 2001/0029610), and others. These show that geo-filtering/blackout, EPG-driven programming control, CDN streaming, and subscription/conditional-access authentication were all well-developed by the 2008 priority date.

2. The IPR2025-00846 art (the core of this analysis)

The petition attacks the patent with Fransdonk, Norris, and Carle — per the petition's own argument, art "the Office never applied" and presenting "different invalidity grounds" than prosecution.

(a) Fransdonk — a family of US applications by Robert W. Fransdonk (content security / conditional access), including Method and system to securely distribute content via a network (US 2003/0161473 A1 and US 2003/0163684 A1), Method and system to secure content for distribution via a network (US 2003/0167392 A1), Method and system to digitally sign and deliver content in a geographically controlled manner via a network (US 2003/0165241 A1; US 2012/0102547 A1), Content distribution using set of session keys (US 2006/0193474 A1), and Method and system to dynamically present a payment gateway for content distributed via a network (US 2007/0180496 A1). The petition cites paragraph numbers up to ¶[0376], most consistent with the large "securely distribute content" application. Key teachings attributed to Fransdonk in the petition:

• A conditional access agent 28 that "validate[s] subscriber content requests against, for example, content access criteria, local date and time, and subscriber credentials" (¶[0062]) and authenticates a content destination 22 (¶[0060]).
• The agent "interfaces with the conditional access server 36 to query subscriptions" and acts as a "broker" to send requested content to the correct destination based on the subscription query response (¶¶[0079], [0136], [0146]) — i.e., querying a subscription database associated with the content provider and processing the reply to determine authorization.
• Upon credential satisfaction, content is forwarded to a media client 49 using Real Time Streaming Protocol — "initiating delivery of the video stream" (¶[0149]).
• Geographic access criteria: Fig. 24 (¶¶[0372]–[0376]) — retrieving access criteria from the conditional access server, including geographic criteria (countries, states, provinces, counties, towns, municipal areas); determining the physical delivery address (block 554); determining the source IP address and mapping it to a geographic location (¶[0374]); and releasing content for delivery only if the delivery address complies (block 560).
• Blackout/time rules: "access control on the basis of region and date/time" (¶[0070]); ACProfileRegionBlackout and ACProfileCountryBlackout tables (¶¶[0128]–[0129]); time/date windows including TimeWindowFlag, DateWindowFlag, TimeWindowStart, TimeWindowEnd (¶[0130]); and the stated goal of blocking, e.g., "a certain country's users from viewing a live game broadcast based on exclusive broadcasting rights" (¶[0019]).

(b) Norris — US 6,718,328 B1, System and method for providing controlled and secured access to network resources (issued Apr. 6, 2004; the petition's shorthand "Norris" is the inventor name). Key teachings (per the petition and the patent text located in search results):

• A publisher/owner controls access to content hosted even on a distributed network — "multiple computers storing and providing access to the content" — exactly the CDN scenario of the '778 patent.
• A user receives a token from the publisher (included in an authenticated URL); a content server receives the token, verifies it, delivers the file (e.g., a movie) if valid, and denies access if invalid.
• Authentication via tokens and/or cookies (Norris 2:23–31, 3:56–63, 4:15–32, 5:20–25, 6:9–13, 7:49–8:8).
• Token bit-field structure (Norris 7:1–13 table): IP Address (bit 0), File path/URL (bit 1), Password (bit 2), Referer ID (bit 3), Time Window (bit 4) — with the IP address "allow[ing] the publisher to restrict access to content to specific users authorized by the publisher" (7:49–53), and publisher passwords preventing one publisher from using another's content. This is both user authentication and publisher identification via metadata.

(c) Carle (number unverified) — per the petition, discloses "method and systems for providing programmatic substitutions into video streams," including making content substitutions based on geographic information where content "may or may not be able to be viewed like, e.g., a 'blacked out' region for certain events," capturing geographic information such as ZIP code, postal address, and geographic [location] (Carle ¶¶[0001], [0011]). This is a geo-targeted ad-insertion/blackout substitution reference squarely in the video-streaming field.

D. Reconstructed claim set (approximate, from IPR petition excerpts)

E. Obviousness combinations and motivation analysis

Ground 1: Fransdonk + Norris → claims 1–3, 11, 13 (the petition's primary ground)

Claim 1. Fransdonk alone supplies nearly every element: the end user's request for content (content requester → conditional access agent 28); the "dip" into the publisher's subscription database (agent 28 "interfaces with the conditional access server 36 to query subscriptions," ¶[0079]); processing the reply to determine authorization ("verification … of content destination (e.g., subscriber) requests for secure content against access criteria defined by a content provider," ¶[0136]); and both branches — denying (releasing only when credentials/geographic criteria are met, ¶[0376]) and initiating delivery (forwarding to media client 49 via RTSP, ¶[0149]). The '778 specification's own description of "dipping" into a publisher's pre-existing subscriber database (FIGS. 16–22) is a functional description of what Fransdonk's conditional-access query already does.

Claims 2–3. Fransdonk's credentialing is heavy (certificates, keys, conditional-access servers). Norris supplies a lightweight, known alternative: publisher-issued tokens and/or cookies that (i) identify the authorizing publisher (publisher passwords in the token "prevent a valid publisher from using another publisher's content," Norris 7:49–53) and (ii) authenticate the user (IP address, time window). The petition correctly notes a POSITA would understand tokens and cookies as "metadata." Motivation: Fransdonk needs user-side credentials to present to the conditional access agent; Norris provides a standard, off-the-shelf mechanism for exactly that — publisher-issued, self-verifying tokens on distributed content servers. Substituting Norris's token/cookie authentication into Fransdonk's conditional-access architecture is a textbook combination of known elements to yield the predictable result of publisher-defined, edge-enforced access control.

Claims 11, 13. Norris expressly contemplates a distributed network with "multiple computers storing and providing access to the content" and content servers receiving requests "from users"; its token bit-field encodes each user's IP address, permitting individual requests to be associated with individual users — i.e., a "second request from a second end user" with its own token/cookie. Fransdonk likewise supports multiple subscribers/content destinations. Combining them renders the multi-user claims obvious for the same reasons as claims 2–3.

Ground 2: Fransdonk alone → claims 1, 4–8 (and portions of 9, 14, 15)

Fransdonk's Fig. 24 process is a nearly verbatim anticipation map for the geographic claims:

• Claim 4 (proximity parameters): Fransdonk "performs a lookup to determine the physical delivery address" of the user and "determines the source IP address … and attempts to map the source IP address to a geographic location" (¶¶[0373]–[0374]). Both are "proximity parameters specifying a geographic location of the end user," as the petition argues.
• Claim 5 (relative time + blackout): Fransdonk evaluates requests "based on access criteria specified by a content provider, local date and time information, and user credentials" (¶[0060]) and provides "access control on the basis of region and date/time" (¶[0070]).
• Claim 6 (blackout rules with time and geographic restrictions): the ACProfileRegionBlackout/ACProfileCountryBlackout tables (¶¶[0128]–[0129]) and TimeWindowFlag/DateWindowFlag/TimeWindowStart/TimeWindowEnd (¶[0130]) are literally blackout rules with associated time and geographic restrictions prescribed by the content provider (whose criteria the conditional access agent enforces, ¶[0062]).
• Claim 7 (group subscription parameters): geographic access criteria stored "alongside other subscription data" and applied to classes of users by region/country profiles are subscription parameters for a group of end users.
• Claim 8 (termination): conditional access systems inherently terminate/deny an in-progress delivery when criteria fail; this is an obvious, conventional implementation detail.

Motivation: none beyond applying Fransdonk's own disclosure — these are express teachings, not combinations.

Ground 3: Fransdonk + Carle → claims 4–8

Even if a challenger wanted a second, independent path to the geographic/blackout claims, Carle teaches programmatic substitution into video streams based on geographic information (ZIP code, postal address) and blackout regions. Motivation: both references solve the same problem — regional availability of broadcast content (blackouts, exclusive territorial rights). Fransdonk's conditional-access tables and Carle's stream-substitution engine are complementary: Carle supplies the stream-level substitution mechanics; Fransdonk supplies the publisher-controlled authorization criteria. A POSITA seeking to enforce regional blackouts in an Internet video stream would find the combination an obvious "substitution of one known element for another" (KSR), with a reasonable expectation of success because both operate on the same inputs (geographic location of the viewer) and the same output (restrict/alter stream delivery).

Extended combination: Fransdonk + Norris + Carle → claims 9, 12, 14, 15, 16

• Claim 9 (subscriber verification table; non-subscriber notification): maintaining a table of verification outcomes (authorized/not) keyed to users is a routine data-management step. Fransdonk's conditional access server stores subscriber/access data; Norris's server-side token verification inherently generates per-user pass/fail records. The "subscriber verification table" is a conventional data structure a POSITA would implement to cache Fransdonk/Norris verification results — particularly given the '778 specification itself describes the local end-user repository as "progressively updated with subscription data" (FIG. 17), which is the same idea.
• Claim 12 (second end user is the same as the first): Norris's IP-address-encoded tokens let a server recognize that a later request comes from the same user (same IP, same token); session/re-authentication handling of this kind is standard.
• Claim 14 and claim 15 (session information stored per relative time): logging a delivery session with its start/relative time is a trivially obvious bookkeeping step; Fransdonk's real-time, date/time-based evaluation (¶¶[0060], [0070]) and Norris's Time Window token component (bit 4) both already tie access decisions to time.
• Claim 16 (system claim): a system claim tracking the method claim maps one-to-one onto Fransdonk's hardware (conditional access agent + conditional access server + subscription database) and/or the combination with Norris's content server; system claims reciting "a content server configured to …" are obvious counterparts of the obvious method, under the ordinary rule that method and apparatus claims of the same scope rise or fall together.

Supplementing art (prosecution references)

The CDN context (edge servers, web server + streaming server, RTSP/RTP delivery) is supplied by the Farber CDN patents of record (US 6,654,807; 6,928,442; 7,054,935) and by the '778 patent's own admitted background (the Level 3 tiered CDN of FIGS. 1–8). The EPG/geo-filtering context (blackouts by region/time, targeted ads) is supplied by the Yuen/Schein/Knudson/Reynolds/Macrae EPG references of record. Thus even a ground built purely on art of record + Fransdonk would be strong.

F. Why a POSITA would have been motivated to combine (synthesis)

1. Same field of endeavor / analogous art. Fransdonk (conditional access for network content distribution), Norris (publisher-controlled access to content on third-party distributed servers), and Carle (geo-based stream substitution) are all directed to authorized delivery of media over networks — the identical problem the '778 patent addresses. The '778 patent itself cites to the same technological milieu (CDNs, RTSP/RTP, subscription databases, blackouts).
2. Complementary, non-competing teachings. Fransdonk is strong on the server-side authorization/subscription-query/blackout machinery but its user authentication is heavyweight (certificates, secure devices). Norris supplies exactly the missing piece — lightweight, publisher-issued token/cookie metadata that both authenticates the user and identifies the publisher — and was designed for content hosted on third-party distributed networks, which is the CDN setting. Combining them yields a complete system with predictable, improved results: edge servers that can enforce publisher-defined access criteria using standard web credentials. This is the paradigm KSR describes as obvious — "the combination of familiar elements according to known methods … [that] does no more than yield predictable results."
3. Design need / market pressure. By 2008, broadcasters moving live content (sports, premium channels) to the Internet faced a known, pressing problem: honoring territorial broadcast rights and blackout obligations online while reusing existing cable/satellite subscriber records. The '778 patent's "dipping" concept is a direct response to that market pressure — and it is the same response Fransdonk (query subscriptions), Norris (publisher-issued tokens), and Carle (geo-substitution) already gave individually. A POSITA would have been motivated to assemble those known components to meet the known demand.
4. Reasonable expectation of success. Every element (subscription database query, token/cookie authentication, IP-to-geo mapping, blackout tables, RTSP delivery, stream substitution) was known and commercially used before the priority date. Nothing in the combination requires new technology or produces an unexpected result; the claims read as an aggregation of known conditional-access, authentication, and geo-filtering functions.

G. Counterarguments and where the claims are strongest

An honest analysis should flag the patent owner's likely rebuttals:

1. "Pre-existing non-Internet subscription database." The '778 specification emphasizes reusing a content publisher's already-established satellite/cable subscriber database so the publisher need not "create a new subscriber database for Internet end users." Fransdonk's subscription database is part of its conditional-access system — not necessarily a legacy cable/satellite subscriber record set. A challenger would need to argue (as the petition does) that "subscription database associated with the content publisher" is not limited to legacy pay-TV records and that Fransdonk's access criteria comprising subscription information satisfies it. That is a reasonable claim-construction position, but it is the crux of the fight.
2. Claim 15's subscriber verification table with time-stamped session entries is the most detailed affirmative limitation; while a routine data structure, its combination of elements (table + authorized-subscriber entry + session info + relative time) may require relying on more than one reference and a slightly more aggressive "inherency + common knowledge" argument.
3. Secondary considerations (unavailable in my record) — e.g., licensing, long-felt need, copying — could weigh against obviousness; none were identified in the materials reviewed.
4. Procedural posture. The Board's discretionary denial of IPR2025-00846 means no merits finding of obviousness exists; the denial (procedural, per the metadata) is not evidence of patentability. The companion IPR2026-00095 (Microsoft) remains pending, and district-court litigation (e.g., Sandpiper CDN, LLC v. Microsoft Corp., No. 2:25-cv-00664, E.D. Tex.) is ongoing, so these same grounds may be re-litigated.

H. Bottom line

On the available record, the claims of US 8,595,778 B2 are strong candidates for obviousness under 35 U.S.C. § 103:

• Fransdonk alone renders claims 1 and 4–8 (and substantially 9, 14, 15) obvious — it discloses subscription-database querying, authorization processing, notification/initiation branches, IP/address-based proximity processing, region/country blackout tables, and time/date windows.
• Fransdonk + Norris renders claims 2–3 and 11–13 obvious — Norris supplies publisher-issued token/cookie metadata identifying both user and publisher on distributed content servers, exactly the missing authentication mechanism.
• Fransdonk + Carle provides an independent path to claims 4–8 via geo-based programmatic stream substitution/blackout.
• Fransdonk + Norris + Carle, supplemented by the CDN/EPG art of record (Farber; Yuen; Schein; Knudson), renders the full claim set — including the system claim (16) and the verification-table claims (9, 14, 15) — obvious under the KSR framework, with clear motivations: same field, complementary teachings, known market pressure to enforce territorial broadcast rights online, and a predictable, conventional combination with a reasonable expectation of success.

The primary genuine vulnerability in the petition's theory is claim construction of "subscription database associated with the content publisher" — whether it must be a legacy pay-TV subscriber database. If that construction battle is won by the petitioner (as the petition argues), the remaining limitations are largely express disclosures of Fransdonk, Norris, and Carle.

Uncertainty notes: (1) claim text is reconstructed from IPR petition excerpts, not the granted patent; (2) the specific Fransdonk publication used in the petition (family member and paragraph numbering) and the Carle patent number could not be fully verified from my search results — the Fransdonk family members and the petition's ¶¶[0001]/[0011] citations for Carle are identified above; (3) no CAFC appeal docket specific to this patent was located.