US 6826694

I'll verify the current record for US 6,826,694 against USPTO and CAFC sources before summarizing.

Let me check for any 2026 CAFC activity and confirm the current USPTO status.

Let me do one final check for any recent (2026) activity and the certificate text.

US 6,826,694 — Verification Summary

Search scope note: I queried USPTO-mirrored records (Google Patents, Justia, FPO, Unified Patents portal, PTAB/docket sources) and searched for 2026 Federal Circuit dockets naming this patent. I found no CAFC appeal involving 6,826,694 docketed in 2026. The two Federal Circuit numbers the patent page associates with the family (2014-1724 and 2017-2429) are historical, not 2026 dockets, and I could not resolve either to a specific proceeding — see "Uncertainties" below.

Bibliographic record

Abstract (as issued)

"A system and method for high resolution access control for packetized information. A packet is received at a firewall. A rule corresponding to header information in the packet prescribes referring the packet to an access control proxy. The access control proxy analyzes the contents of the packet, and identifies a rule based upon the contents. The rule is implemented at the firewall."

Plain-language overview of the sole independent claim

Claim 1 — A method for filtering a packet, comprising:

• (a) receiving a packet that has at least one header parameter and a payload;
• (b) selecting an access rule based on the contents of that packet's payload;
• (c) implementing the access rule for a packet, wherein the access rule is selected based on a combination of (i) the contents of the packet received in step (a) and (ii) the contents of at least one other packet.

In plain terms: rather than deciding PASS/DROP purely from header fields (source/destination address, port, protocol) as a conventional firewall does, the system looks inside the packet payload to choose a filtering rule. The distinguishing twist is element (c): the rule must be picked from information spread across two or more packets combined — i.e., multi-packet/cross-packet content correlation (e.g., reassembling a request that spans several packets before deciding).

Specification embodiments (not separately claimed) include: a firewall that REFERs matching packets to a protocol-specific access control proxy; rule selection by URL, filename, domain name, or "nth degree domain name"; rules stored locally or fetched from an external library node; and selection based on payload content alone or payload-plus-header (e.g., DROP secret.html from SOURCE A but PASS it from SOURCE B).

Record conflict resolved (flagging for the earlier sections)

The previously generated "Prior art" and "Strategic summary" sections flagged an unresolved discrepancy in whether IPR2014-00786 ended in invalidation. It did. The Final Written Decision (Paper 46, entered September 23, 2015; panel Moore, Petravick, Wood) ordered: "claim 1 of U.S. Patent No. 6,826,694 B1 is unpatentable," finding claim 1 anticipated under § 102(b) by Rubin, obvious under § 103 over Norman + Rubin, and anticipated under § 102(e) by Cunningham. This supersedes the earlier "termination Sept. 23, 2015" entry (that date was a merits FWD, not a procedural termination). An IPR certificate (kind code K1) was issued February 12, 2018.

2026 status (what I could and could not confirm)

• No live litigation, no live PTAB proceeding, no 2026 CAFC appeal involving 6,826,694 surfaced in my searches. The 2013 IV II bank campaign is long closed/stayed-out, and the patent expired October 22, 2019.
• One third-party compilation (usinventor.org, "Dual-Venue Patents," Nov. 2019) lists 6,826,694 as "INVALID" — consistent with the cancelled claim, though I treat that aggregator entry as secondary confirmation only.

Uncertainties (stated rather than guessed)

1. IPR certificate text not retrieved. I could not read the February 12, 2018 certificate's verbatim per-claim disposition. Cancellation of claim 1 is an inference from the affirmed FWD plus the K1 certificate event — verify in USPTO PatentCenter (Document Retrieval, "IPRC/2018-02-20") before quoting.
2. CAFC docket numbers 2014-1724 and 2017-2429 appear in the patent's litigation metadata. I could not map either to a proceeding, and neither is a 2026 docket. Do not represent them as such without a direct CAFC docket pull.
3. Live USPTO API/PAIR query not performed in this pass; bibliographic data above comes from the authoritative full text supplied earlier plus USPTO-mirrored public records. The ODP ingest for this patent has previously returned empty/stale results (flagged in the earlier PTAB section) — a known data-hygiene issue for this number.
4. Inventor name variance persists across records: "Michah Lerner" (front page/Google Patents) vs. "Micah Lerner" (assignment reel 010775/0769). Same person; spelling discrepancy only.
5. "Assignee" is ambiguous: the issued patent names AT&T Corp.; the recorded current owner is Intellectual Ventures II LLC. Both are accurate as to their respective points in time.

I'll search for litigation involving US patent 6826694 across patent litigation sources.

Let me dig into the specific cases identified and verify details.

Litigation Involving U.S. Patent No. 6,826,694

Patent identified: U.S. 6,826,694 B1, "High Resolution Access Control," application 09/422,952, filed Oct. 22, 1999, issued Nov. 30, 2004, priority to provisional 60/105,188 (Oct. 22, 1998). Original assignee AT&T; later Worcester Technologies LLC, then Intellectual Ventures II LLC (effective May 23, 2013). Status: Expired – Lifetime (anticipated expiration Oct. 22, 2019).

All district-court filings below are the 2013 Intellectual Ventures II LLC ("IV II") banking campaign, in which the '694 patent was asserted as one of five patents-in-suit (U.S. 5,745,574; 6,826,694; 6,715,084; 6,314,409; 7,634,666). The seven suits are enumerated in the "Notice of Related Matters" filed in PTAB IPR2014-00786 and are also reflected in the litigation links on the patent's Google Patents page.

A. District Court Litigation

Note on a related but distinct case: IV also sued PNC (Intellectual Ventures I LLC and II LLC v. PNC Financial Services Group, Inc. and PNC Bank, N.A., W.D. Pa. No. 2:13-cv-00740, filed May 29, 2013). The PNC complaints retrieved assert the '666, '574, '002, '298, and '409 patents — not the '694 patent. I therefore do not count PNC as '694 litigation, though contemporaneous press coverage grouped the PNC and FNBO suits together.

Caveat on outcomes: For cases 2–7, I could confirm the stay orders and initial filings but did not retrieve final termination documents. Because nearly all of these cases were stayed pending the IPRs, and the '694 IPRs terminated with a Final Written Decision (see below), it is plausible the cases were closed by dismissal or settlement after the PTAB proceedings — but I will not state a specific outcome for those six without a retrieved docket entry. Only case #1 (BBVA Compass) has a confirmed final outcome.

B. PTAB Proceedings (Inter Partes Review)

Three IPR petitions against the '694 patent are listed on the patent's Google Patents page; two produced Final Written Decisions, and IPR certificates were issued Feb. 12, 2018.

The PTAB's institution decision on claim 1 of the '694 patent (Sept. 24, 2014) found a reasonable likelihood that claim 1 was invalid under 35 U.S.C. §§ 102 and/or 103. Grounds asserted in IPR2014-00786 included obviousness over Abraham (U.S. 5,983,270) in view of Rubin ("Blocking Java Applets at the Firewall"), anticipation by Rubin, anticipation by Cunningham (U.S. 6,219,786), and obviousness over Norman in view of Rubin.

C. Federal Circuit Appeals

Two CAFC case numbers are associated with the '694 patent family per the Google Patents litigation list:

• Appeal No. 2014-1724 (Court of Appeals for the Federal Circuit)
• Appeal No. 2017-2429 (Court of Appeals for the Federal Circuit)

I was not able to retrieve the opinions or dispositions for either appeal within the scope of this search, and I will not guess at their outcomes. The 2017-2429 appeal is chronologically consistent with an appeal from the IPR Final Written Decisions (which led to the Feb. 12, 2018 IPR certificates), but I cannot confirm that link from the retrieved material.

D. Litigation Summary / Caveats

• Known infringement litigation asserting the '694 patent: seven district-court suits, all filed in 2013 by Intellectual Ventures II LLC against U.S. financial institutions, all in 2013 (May–July).
• Confirmed outcome: Only the BBVA Compass case has a confirmed final disposition in the material retrieved — dismissed with prejudice on Sept. 11, 2018. The other six were stayed pending IPR; their final dispositions are unconfirmed here.
• PTAB: The '694 patent was subject to three IPR petitions (two by IBM, one by the Compass/Commerce/FNBO group); IPR certificates were issued Feb. 12, 2018 for IPR2014-00587 and IPR2014-00786.
• Sources consulted: Google Patents page for US6826694B1 (litigation links and legal events); Unified Patents litigation portal (case 2:13-cv-01106, Alabama N.D.); Stanford NPE Litigation Database (patent 6826694); Justia/CourtListener/Casetext dockets and opinions (N.D. Ga., S.D. Ohio, W.D. Mo.); the PTAB Notice of Related Matters in IPR2014-00786 (USPTO PTACTS); GreyB/Ex Parte PTAB case pages; Jones Day experience page (Compass Bank); GeekWire and Open Legal Blog Archive (2013 filing coverage).

If you need docket-level final dispositions for cases #2–#7 (Nebraska, S.D.N.Y., W.D. Mo., N.D. Ga., D. Minn., S.D. Ohio) or the two Federal Circuit appeals, those would require direct PACER/CAFC docket pulls, which I could not complete here.

I'll start by verifying the PTAB record, since the patent's own metadata contradicts the ODP "no proceedings" default.

The structured block and the patent's own record conflict — let me resolve the appeals and the certificate.

Now let me pin down the Federal Circuit dispositions and the certificate.

⚠️ Record conflict — read first

The structured "PTAB proceedings on file" block says the USPTO ODP API returns no AIA trial proceedings for US 6,826,694. That is wrong, and I'm flagging it rather than repeating it: the patent's own Google Patents record (authoritative, fetched 2026-09-11) lists three AIA trials — IPR2014-00587, IPR2014-00786, IPR2014-01465 — plus a PTAB inter partes review certificate issued 2018-02-12. Live web sources confirm all four. The ODP ingest for this patent number appears stale/empty; the canonical list below is reconstructed from the patent record + PTAB/CAFC primary and secondary sources, each flagged as to confidence. Do not rely on the "no PTAB activity" default here.

Proceedings overview

Three AIA trials total — 1 claims-invalidated (IPR2014-00786, affirmed on appeal), 1 claims-sustained (IPR2014-00587, affirmed on appeal), 1 terminated pre-FWD on § 315(e)(1) estoppel (IPR2014-01465); no settlements, no institution denials — and the bottom line for a defendant is that the ’694 patent's only claim (claim 1) was held unpatentable in IPR2014-00786, that judgment was affirmed by the Federal Circuit on 2016-12-09, an IPR certificate effectuating the judgment issued 2018-02-12, and the patent expired 2019-10-22, so the patent is dead twice over (cancelled claim + lapsed term); any demand letter citing claim 1 is baseless.

IPR2014-00786 — Compass Bank, Commerce Bancshares, Inc., and First National Bank of Omaha v. Intellectual Ventures II LLC

• Type: Inter Partes Review
• Filed: 2014-05-20
• Status: Final Written Decision (Outcome: Unpatentable) — the Board ordered claim 1 unpatentable; affirmed on appeal, and an IPR certificate issued 2018-02-12.
• Judge panel: James T. Moore, Meredith C. Petravick, Benjamin D. M. Wood. FWD authored by Judge Petravick (per PTAB secondary-source records; the oral-hearing transcript confirms the three-judge panel).
• Petition grounds: sole claim (claim 1). § 102(b) anticipation by Rubin; § 103 over Norman in view of Rubin; § 102(e) by Cunningham. (Patent Owner's preliminary response also framed the grounds as including § 103 combinations built on Rubin, and argued Grounds 1–3 were duplicative of IBM's IPR2014-00587 and redundant.)
• Institution decision: instituted 2014-09-24. IV's preliminary response had argued the petition was statutorily defective for failing to name BBVA as a real party-in-interest (§ 312(a)(2)) and that the grounds were duplicative/redundant; the Board instituted notwithstanding. Key construction adopted: "contents of the packet" = "the contents of the header and/or payload of the packet" — the Board expressly rejected IV's narrower "header and payload" reading, reasoning that step c "does not preclude the information coming from the header or payload portion of the contents of the packets."
• Final Written Decision: issued 2015-09-23. Claim 1 — the sole claim — held unpatentable. Order: "ORDERED that claim 1 of U.S. Patent No. 6,826,694 B1 is unpatentable." Reasoning: "Petitioner has demonstrated by a preponderance of the evidence that claim 1 is anticipated under 35 U.S.C. § 102(b) by Rubin and is unpatentable [under] § 103 over Norman and Rubin [and under] § 102(e) by Cunningham." IV's attempt to antedate Rubin/Cunningham via inventor Dutta's testimony failed here — the Board dismissed IV's motion to exclude and found the Dutta corroboration exhibits "insufficient to corroborate Mr. Dutta's testimony of prior invention regardless of their admissibility." Both parties' motions to exclude were dismissed as moot.
• Settlement / termination: none. The record was contested post-FWD: on 2015-11-20 the Board granted a joint request (Paper 49) to preserve the sealed record pending appeal, noting that inventor Lerner's testimony "called into question whether the sealed information . . . was truly confidential"; on 2016-03-30 the Board ordered Exhibits 2011 and 2012 unsealed (Paper 51).
• Appeal: Yes — IV appealed, and lost. Intellectual Ventures II LLC v. Compass Bank, Commerce Bancshares, Inc., First National Bank of Omaha, Fed. Cir., argued and decided 2016-12-09, per curiam (Dyk, Plager, Reyna): "AFFIRMED. See Fed. Cir. R. 36" (nonprecedential one-line affirmance). CourtListener: https://www.courtlistener.com/opinion/[4328712](/patent/4328712)/intellectual-ventures-ii-llc-v-compass-bank/ . The appeal docket number is reported by Patexia as 2016-1416 (secondary source — treat the docket number as moderate confidence; the judgment itself is confirmed). Do not confuse this with the March 2017 Rule 36 dismissals at Nos. 2016-1519/1520/1528, which concerned a different IV patent (the ’084 patent) in Intellectual Ventures II LLC v. Commerce Bancshares, Inc.
• Defensive value: Claim 1 is dead. The FWD's unpatentability judgment was affirmed, and the certificate issued — any infringement theory built on the ’694 patent's only claim is now sanction-bait. This is the single most useful document for a defendant.

IPR2014-00587 — International Business Machines Corporation v. Intellectual Ventures II LLC

• Type: Inter Partes Review
• Filed: 2014-04-07
• Status: Final Written Decision (Outcome: Patentable) — claim 1 not shown unpatentable; affirmed on appeal.
• Judge panel: James T. Moore, Meredith C. Petravick, Benjamin D. M. Wood (FWD Paper 54). Same panel as IPR2014-00786.
• Petition grounds: claim 1, § 103 obviousness — the FWD turned on Hughes (US 5,842,040) in view of Abraham (US 5,983,270). (A third-party AI summary of the petition also lists § 103 grounds over Coss + RFC 1123/RFC 959 and over Rubin + RFC 793 — that inventory is lower confidence; I did not retrieve the petition itself. The Ground-1/§ 103-over-Hughes-and-Abraham characterization is confirmed by contemporaneous reporting.)
• Institution decision: instituted 2014-09-24 (same date as the ’786 institution; secondary-source date, moderate confidence). The Board applied the broadest reasonable interpretation and agreed with Petitioner that "contents of the packet" means "the header and/or payload of the packet" — the same construction that later doomed the patent in IPR2014-00786.
• Final Written Decision: issued 2015-09-23. Claim 1 not shown to be unpatentable. IV successfully disqualified the asserted art by proving prior reduction to practice before the references' effective dates (Hughes: 1996-06-18; Abraham: 1997-03-11). The Board found the "AT&T documents" alone did not sufficiently corroborate inventor Dutta's testimony, but was "persuaded that Dr. Vrsalovic's testimony provides sufficient corroboration for Mr. Dutta's testimony," concluding that "Petitioner [failed] to demonstrate that claim 1 is unpatentable under 35 U.S.C. § 103 because Petitioner [had] not demonstrated that Hughes and Abraham are prior art." This was a swearing-behind / § 102(g)-style priority win on a specific pair of references — not a merits holding that the claim is broadly valid.
• Settlement / termination: none. Record-preservation order pending appeal entered 2015-11-20 (joint with IPR2014-00786).
• Appeal: Yes — IBM appealed and lost. International Business Machines Corp. v. Intellectual Ventures II LLC, 636 F. App'x 1008 (Fed. Cir. 2016), per curiam: "AFFIRMED. See Fed. Cir. R. 36." CourtListener: https://www.courtlistener.com/opinion/[8696168](/patent/8696168)/international-business-machines-corp-v-intellectual-ventures-ii-llc/ . ⚠️ I could not retrieve the opinion text to confirm the underlying PTAB number; the caption/posture (IBM as appellant against IV as appellee, 2016) fit an appeal of the § 318(a) decision in IPR2014-00587, but verify on PTAB E2E/PAIR before citing it as the ’694 appeal.
• Defensive value: IV won this one, but the win is art-specific and now academic — the same claim was cancelled in the parallel ’786 trial whose judgment was affirmed, so the ’587 patentability finding has no residual offensive value. Expect a patent owner to cite the ’587 FWD as evidence of "survival"; the correct rebuttal is that it survived only by disqualifying two references as non-prior-art, and the claim has since been cancelled.

IPR2014-01465 — International Business Machines Corporation v. Intellectual Ventures II LLC

• Type: Inter Partes Review
• Filed: 2014-09-09
• Status: Procedural Termination (Board order terminating the review 2015-11-06).
• Judge panel: not confirmed — the source excerpt truncates at the panel listing. The Board's panel for the related trials was Moore, Petravick, Wood; do not assume it without verification.
• Petition grounds: claim 1, § 102 anticipation by Estrin 1987 (the "sole ground of review").
• Institution decision: instituted (date not confirmed in the sources retrieved) — the Board reached a contested motion to terminate before oral hearing, which presupposes institution. ⚠️ Verify the institution date on PTAB E2E.
• Final Written Decision: none — the trial was terminated before the oral hearing, so no claim was adjudicated in this proceeding.
• Settlement / termination: terminated 2015-11-06 on statutory estoppel (Paper 32). IV moved to terminate on the ground that the 2015-09-23 FWD in IPR2014-00587 triggered § 315(e)(1) estoppel, because IBM reasonably could have raised Estrin 1987 in the earlier trial. The Board agreed, holding that "could have been raised" is measured by what "a skilled searcher conducting a diligent search reasonably could have been expected to discover" — not by the petitioner's actual knowledge — and rejecting IBM's "opened the door" argument. The Board declined IBM's request to continue without a petitioner, finding termination the "best means of securing the just, speedy, and inexpensive resolution of the proceeding."
• Appeal: none identified.
• Defensive value: this order is the most useful estoppel precedent in the file: it confirms § 315(e)(1) can kill a follow-on IPR even where the second reference was unknown to the petitioner, and it shows the Board will terminate rather than issue an FWD with no petitioner left. Practical takeaway for a defendant: serial IPR attack on this family was foreclosed by estoppel — the second IBM petition never reached the merits.

Strategic summary

Claim status (claim-level, and there is only one claim). US 6,826,694 issued with exactly one claim: claim 1. There are no claims 2–5 to assert. Claim 1: CANCELED — held unpatentable in IPR2014-00587's companion trial IPR2014-00786 (FWD 2015-09-23), affirmed by the Federal Circuit on 2016-12-09 under Rule 36, with the IPR certificate (kind code K1) issued 2018-02-12 per the patent's legal-events record. SUSTAINED: none, in any operative sense — the ’587 FWD found claim 1 "not shown to be unpatentable," but that finding rested on two references being disqualified as prior art by a 1996 reduction-to-practice showing, and it is superseded in practical effect by the ’786 cancellation. UNTESTED: nothing — the whole claim set has been through trial. Separately, the patent's anticipated expiration was 2019-10-22 and its status is "Expired – Lifetime," so even the ’587-invoked shadow of validity cannot be asserted prospectively; only a closed pre-October-2019 damages window would remain, subject to the usual § 286 six-year lookback.

Estoppel landscape. § 315(e)(1) was applied against IBM in IPR2014-01465, terminating it outright. § 315(e)(2) now bars IBM and the three bank petitioners (Compass Bank, Commerce Bancshares, First National Bank of Omaha) and their privies from re-litigating in district court any ground they raised or reasonably could have raised in the ’587/’786/’465 trials — i.e., Rubin, Norman, Cunningham, Hughes, Abraham, and Estrin 1987 are off the table for those parties. A new defendant who is not a privy of IBM or the banks is under no estoppel, but has no need for it: the cancelled claim cannot be asserted against anyone. Practical caution — IV's preliminary response argued BBVA was an undisclosed real party-in-interest in the bank IPR; the Board instituted anyway, so BBVA's RPI/privity status was never definitively resolved. If you are litigating alongside BBVA, verify that before relying on estoppel against it.

Pattern signals. The patent drew three AIA petitions in 2014 — two by IBM and one by a three-bank coalition (Compass Bank, Commerce Bancshares, First National Bank of Omaha), filed in response to IV's 2013 multi-district assertion campaign against financial institutions (N.D. Ala. 2:13-cv-01106; W.D. Mo. 2:13-cv-04160; D. Neb. 8:13-cv-00167, among others named in the FWD). IV was an aggressive PTAB defender: it won IPR2014-00587 by proving prior reduction to practice, then used that win offensively to estop IBM's third petition in IPR2014-01465, and it appealed the adverse ’786 judgment to the Federal Circuit (losing on 2016-12-09). No defensive aggregator (e.g., Unified Patents) appears in the Petitioner chain — note that the separate PTAB proceeding at IPR2014-01465 involved IBM, and the Unified Patents PTAB entry listed in the patent metadata is unrelated docket noise; Unified Patents appears elsewhere in this patent's history as a litigation-data provider, not as petitioner. The parallel Rule 36 dismissals at Nos. 2016-1519/1520/1528 (March 2017, IV v. Commerce Bancshares) concern the ’084 patent, not the ’694 patent — keep those separate in any brief.

⚠️ Two CAFC docket numbers appear in the patent's litigation metadata that I could not resolve to a proceeding: 14-1724 (2014 — predates all three FWDs, most likely a district-court appeal in the underlying assertion campaign) and 17-2429 (2017 — likely a post-IPR district-court appeal). I am not asserting what either was. Pull the CAFC dockets before referencing them.

Recommended next steps

• If you are a defendant receiving a demand on the ’694 patent: you do not need an invalidity case. Cite the IPR2014-00786 Final Written Decision, whose order reads: "ORDERED that claim 1 of U.S. Patent No. 6,826,694 B1 is unpatentable" (https://www.finnegan.com/a/web/[114949](/patent/114949)/2G4qRC/ipr2014-00786-final-written-decision-35-usc-c2a7-318a-and-3.pdf — third-party hosted copy; pull the official version from PTAB E2E, https://ptacts.uspto.gov/, by entering patent number 6826694). Pair it with the Federal Circuit's 2016-12-09 Rule 36 affirmance, https://www.courtlistener.com/opinion/4328712/intellectual-ventures-ii-llc-v-compass-bank/, and the IPR certificate issued 2018-02-12 recorded in the patent's legal events (https://patents.google.com/patent/US6826694/en). Also note the 2019-10-22 expiration — the patent has no forward life.
• Retrieve the certificate text itself from USPTO PatentCenter/Document Retrieval (the "IPRC" event dated 2018-02-20 in the patent's legal events) and confirm its verbatim per-claim disposition before quoting it in a filing. I inferred cancellation of claim 1 from the K1 certificate plus the affirmed unpatentability judgment; I did not retrieve the certificate's text, so verify the exact wording.
• Confirm the ’587 appeal by checking whether International Business Machines Corp. v. Intellectual Ventures II LLC, 636 F. App'x 1008 (Fed. Cir. 2016) is docketed against IPR2014-00587 on PTAB E2E. Its Rule 36 affirmance is real; its linkage to the ’694 patent is my inference from caption and posture.
• No active proceedings are pending. All three trials are closed (FWDs 2015-09-23; termination order 2015-11-06), so there are no institution deadlines, oral hearings, or statutory § 316(a)(11) FWD due dates to track. The absence of active PTAB work is not a defensive gap here — it reflects that the patent was fully adjudicated, invalidated, and expired.
• As a data-hygiene matter: re-query ODP for this patent number. The empty ODP result contradicts the patent's own record and the PTAB/CAFC docket history; if your workflow relies on that API default, it will mis-flag this patent as "no PTAB activity," which is the opposite of the truth.

I have the core assignment data from the authoritative patent text (Google Patents legal events, including reel/frame numbers). Let me try to surface the correspondents of record and verify the chain.

The FPO result shows the prosecution attorney of record (Knobbe Martens). Let me verify that and hunt for the assignment recording correspondents specifically.

Ownership Reconstruction — US 6,826,694 B1 ("High Resolution Access Control")

Data source note: The reel/frame, conveyance, assignor, and assignee data below are taken verbatim from the patent's recorded legal events (mirrored on the Google Patents page and matching USPTO Assignment Center records). I was unable to retrieve the correspondent of record (the attorney/agent who filed each recording) — the Assignment Center's correspondent field was not exposed in any source I could reach, and I will not fabricate names. See the dedicated note under "Relationship to prosecution counsel" below.

Inventors

• All three inventors were AT&T employees — the 2000 assignment to AT&T CORP., NEW YORK (reel 010775/0769) recites them as assignors of their interest in the application. This is an ordinary employer-takes-invention record, not a fire-sale indicator.
• Name variance — flag: the front page / Google Patents inventor field spells the third inventor "Michah Lerner," while the recorded assignment (reel 010775/0769) and the reassignment entry spell it "Micah Lerner." This is a spelling discrepancy in the source records, not a different person.
• Timing: the assignment was executed 2000-03-22 to 2000-04-10 and recorded 2000-05-08 — i.e., ~5–6 months after the 1999-10-22 filing. Late execution is common and does not evidence inventor departures. I found no evidence of inventors departing AT&T within 12 months, nor of any inventor appearing on later NPE-family records.

Original assignee

AT&T Corp. (successor designations: AT&T Properties, LLC → AT&T Intellectual Property II, L.P.).

• Primary business: U.S. telecommunications long-distance, data networking, and Internet services — a large operating carrier, not a patent holding vehicle.
• Did it ship a product embodying the claims? The patent describes firewall / access-control-proxy filtering; AT&T operated Internet and managed-firewall service infrastructure, so AT&T plausibly practiced or could practice the claimed subject matter internally. I found no evidence of a discrete commercial product specifically branded to these claims; the invention reads as carrier network-security infrastructure.
• Current status of the original assignee: AT&T Corp. was acquired by SBC Communications in 2005, which adopted the "AT&T Inc." name (SBC + AT&T). The original AT&T Corp. entity was absorbed via the AT&T Properties, LLC → AT&T Intellectual Property II, L.P. reorganization in 2011. AT&T Inc. (the successor operating company) remains active; the original AT&T Corp. entity no longer exists as an independent company.

Assignment timeline

(Chronological. "Recorded" dates are as listed in the legal events; "effective/executed" dates are as stated in the record.)

• 2000-03-22 to 2000-04-10 (executed) / recorded 2000-05-08 — Reel 010775/0769

  • Conveyance: Assignment of assignors' interest (original employee assignment)
  • Assignor: Partha P. Dutta; Mahesh M. Kumar; Micah Lerner
  • Assignee: AT&T Corp., New York
  • Correspondent: Not retrieved (correspondent field unavailable in the sources reached).
  • Context: Initial employer assignment of the invention from AT&T inventors to AT&T Corp.

• 2011-08-15 (effective) / recorded 2011-08-25 — Reel 026804/0081

  • Conveyance: Nunc pro tunc assignment
  • Assignor: AT&T Corp.
  • Assignee: AT&T Properties, LLC, Nevada
  • Correspondent: Not retrieved.
  • Context: Internal corporate reorganization (first step of AT&T's 2011 IP-holding-company restructuring).

• 2011-08-15 (effective) / recorded 2011-08-25 — Reel 026807/0185

  • Conveyance: Nunc pro tunc assignment
  • Assignor: AT&T Properties, LLC
  • Assignee: AT&T Intellectual Property II, L.P., Georgia
  • Correspondent: Not retrieved.
  • Context: Internal corporate reorganization (second step — patent-pooling into the AT&T IP II holding entity). Runs in parallel with reel 026804/0081 on the same recording date.

• 2011-09-23 (effective) / recorded 2011-11-16 — Reel 027241/0681

  • Conveyance: Assignment of assignors' interest (sale/transfer)
  • Assignor: "AT&T Intellectual Property I, LP" (name as originally recorded — see correction below)
  • Assignee: Worcester Technologies LLC, Delaware
  • Correspondent: Not retrieved.
  • Context: Transfer out of the operating carrier to a licensing-only Delaware LLC — the pivotal link in the chain.

• 2011-09-23 (effective) / recorded 2013-05-15 — Reel 030427/0182

  • Conveyance: Corrective assignment
  • Assignor: AT&T Intellectual Property II, L.P. (corrected assignor)
  • Assignee: Worcester Technologies LLC, Delaware
  • Correspondent: Not retrieved.
  • Context: Correction of the previous recording — reel 027241/0681 erroneously named the assignor as "AT&T Intellectual Property I, LP"; the corrective record confirms the true assignor was AT&T Intellectual Property II, L.P. This makes the operative chain AT&T Corp. → AT&T Properties, LLC → AT&T IP II, L.P. → Worcester Technologies LLC.

• 2013-05-23 (effective) / recorded 2013-05-28 — Reel 030493/0972

  • Conveyance: Merger
  • Assignor: Worcester Technologies LLC
  • Assignee: Intellectual Ventures II LLC, Delaware
  • Correspondent: Not retrieved.
  • Context: Transfer-to-asserter — Worcester Technologies LLC merged into Intellectual Ventures II LLC, vesting the '694 patent in IV II days before the 2013 assertion campaign.

Relationship to prosecution counsel (not the assignment correspondent)

The patent's attorney of record during prosecution was Matthew Smithers, Knobbe Martens Olson & Bear LLP, Irvine, CA (per the patent's attorney/agent field on FreePatentsOnline, https://www.freepatentsonline.com/[6826694](/patent/6826694).html). I want to be precise: this is the prosecution firm of record, which is not the same field as the assignment correspondent the task asks for. I could not confirm whether Knobbe Martens (or any other firm) was the recording correspondent on any of the reel/frame entries above, and I found no recurrence of a recording correspondent to report. Treat the repeat-correspondent signal as unverified rather than absent.

Timeline diagram

timeline
    title Ownership of US 6826694
    1998 : Priority date Oct 22
    1999 : Application filed Oct 22
    2000 : Inventors assign to AT&T Corp
    2004 : Patent issues Nov 30
    2011 : AT&T reorg creates IP II LP
         : Transfer to Worcester Technologies
    2013 : Worcester merges into IV II
         : First infringement suits filed
    2019 : Anticipated expiration

NPE / troll-pattern signals

1. Shell-entity transfer — PRESENT. Reel 027241/0681 (effective 2011-09-23) moved the patent from the AT&T operating family to Worcester Technologies LLC, a Delaware LLC with no known products, and reel 030493/0972 (effective 2013-05-23) then vested it in Intellectual Ventures II LLC, a Delaware LLC. Both assignees are licensing/holding vehicles, not operating companies. The "Holdings/Ventures/Technologies LLC" naming plus the operating-company-to-LLC direction of transfer is the classic tell.

2. Known asserter in the chain — PRESENT. Intellectual Ventures II LLC is a well-known NPE (Intellectual Ventures is explicitly on the standard asserter lists). Reel 030493/0972 makes IV II the current owner. Worcester Technologies LLC is the intermediary holding entity in the same transfer chain. This is not an inference from naming — it is the recorded assignee on reel 030493/0972, and IV II then filed the 2013 infringement campaign described in the litigation summary.

3. Repeat correspondent across the chain — UNCLEAR / UNVERIFIED. The assignment correspondent of record was not retrievable from the sources available, so I cannot confirm or deny a recurring recording attorney across reels 010775/0769, 026804/0081, 026807/0185, 027241/0681, 030427/0182, and 030493/0972. The known attorney of record (Matthew Smithers, Knobbe Martens) is prosecution counsel and should not be reported as the assignment correspondent. No finding.

4. Cascading transfers — PRESENT. Between the 2011-08-15 effective date (AT&T Corp. → AT&T Properties, LLC → AT&T IP II, L.P.) and the 2011-09-23 effective transfer to Worcester Technologies LLC, the patent passed through three holders in roughly five weeks (reels 026804/0081, 026807/0185, 027241/0681), then to IV II via merger in 2013 (reel 030493/0972). Consecutive chained transfers well inside a 24-month window.

5. Pre-litigation transfer — PRESENT (strong). The final link — the Worcester Technologies LLC → Intellectual Ventures II LLC merger, effective 2013-05-23 (recorded 2013-05-28, reel 030493/0972) — completes days before the first assertion. IV II's first '694 suit (Intellectual Ventures II LLC v. First National Bank of Omaha, D. Neb. 8:13-cv-00167) was filed 2013-05-29, six days after the merger's effective date. The chain was plainly arranged to present a clean standing record for assertion; compare the wider 2013 campaign (May–July 2013) in the litigation summary.

6. Bankruptcy fire-sale — NOT PRESENT. AT&T Corp. did not file Chapter 7/11 and this patent was not sold through insolvency proceedings. The 2011 transfers were a voluntary corporate reorganization and portfolio monetization, not a bankruptcy disposition.

7. Privateering — UNCLEAR. AT&T transferred the patent out to Worcester/IV, and IV then asserted it (against banks, not obviously AT&T's telecom competitors). The direction of transfer is consistent with monetization-by-asserter, but I did not locate SEC filings or Patent Progress/EFF coverage showing AT&T orchestrated assertion against its own competitors on its behalf. The transfers were recorded as nunc-pro-tunc/merger conveyances without any disclosed assertion-back arrangement. No confirmation.

8. Defensive aggregator (anti-NPE) — NOT PRESENT. The chain terminates at Intellectual Ventures II LLC, an asserter, not at RPX, AST, LOT, Unified Patents, or OIN. The patent was not neutralized; it was weaponized. (The later PTAB proceedings and Feb. 12, 2018 IPR certificates on reels' related patents/claims are validity events, not ownership events, and do not change the assignee.)

Verdict

NPE — high confidence.

Two or more strong signals are documented on the face of the recorded chain: (i) a shell-entity transfer out of the AT&T operating family to Worcester Technologies LLC (reel 027241/0681, effective 2011-09-23), and (ii) a known asserter as current owner — Intellectual Ventures II LLC — by merger (reel 030493/0972, effective 2013-05-23), reinforced by (iii) cascading transfers through three AT&T-family entities within ~5 weeks in 2011 and (iv) a pre-litigation transfer completing six days before IV II's first '694 suit (D. Neb. 8:13-cv-00167, filed 2013-05-29). The only unresolved signal is the repeat-correspondent check, which could not be completed because the Assignment Center correspondent field was not retrievable — that omission does not weaken the verdict, since the shell-transfer and known-asserter signals rest on the reel/frame records themselves.

Verification link: USPTO Assignment Center search for the patent → https://assignmentcenter.uspto.gov/ (search patent number 6826694; equivalently https://assignment.uspto.gov/patent/index.html).

Caveats: (1) The two 2011-08-25 recordings (reels 026804/0081 and 026807/0185) are parallel steps of the same internal AT&T reorganization, not sequential thirds. (2) The 2011-11-16 recording (reel 027241/0681) contained an assignor-name error ("AT&T IP I, LP") later corrected to AT&T IP II, L.P. by reel 030427/0182 — cite the corrected chain. (3) I did not fabricate any correspondent names; where the field was unavailable, I have said so.

I'll research the prior art references cited on the face of US 6,826,694. Let me search for each reference.

Let me research the remaining cited references.

Let me research the remaining references and the IPR invalidity grounds.

Prior Art Analysis — US 6,826,694 B1 ("High Resolution Access Control")

Patent analyzed: US 6,826,694 B1, Application 09/422,952, filed Oct. 22, 1999, priority to provisional 60/105,188 (Oct. 22, 1998), issued Nov. 30, 2004. One claim only — independent claim 1. All §102 anticipation analysis below therefore runs against claim 1 exclusively.

Sources used: the authoritative full text you supplied (Google Patents mirror of the USPTO record), FreePatentsOnline, the issued patents' own front pages, and Unified Patents / docketalarm exhibits. I did not pull a live USPTO PatentCenter / PAIR API record; where dates come from the reference's own front page rather than primary USPTO bulk data, I say so. Per the strict rule, all identifiers below are reproduced literally as they appear; I have not "corrected" any number.

Claim 1 elements (the anticipation targets):

• (a) receiving a packet having at least one header parameter and a payload;
• (b) selecting an access rule based upon the contents of the payload of the packet received in step (a);
• (c) implementing the access rule for a packet, wherein the access rule is selected based upon a combination of the contents of the packet received in step (a) AND the contents of at least one other packet.

Element (c) — rule selection requiring a combination of contents across ≥2 packets — is the narrow, distinguishing limitation. None of the references cited on the face of the '694 patent discloses it.

A. Patent Citations on the Face of US 6,826,694

Google Patents labels five of the six references: three were third‑party (IDS) citations (US 5,473,607; WO 96/05549; EP 0 762 707) and three were examiner citations (US 5,983,270; US 6,219,706; US 6,584,508).

1. US 5,473,607 A — Hausman et al. / Grand Junction Networks, Inc.

Description: Packet filtering performed in a network controller using hashing. Relevant packet fields ("candidate fields," e.g., the destination address field) are hashed and compared against a target hash table to classify packets as rejects, exact matches, or partial matches. The specification states filtering criteria "can be expressed as simple Boolean functions of data fields within the packet," and explicitly that "any portion or combined portions of the packet … might constitute the candidate field." The thrust is header/address-field filtering for efficient multicast/groupcast reception.

§102 assessment vs. claim 1: Does not anticipate. It reads on element (a) only (packets have addressing fields). Its filtering decisions are keyed to header/addressing fields, not to payload contents (element (b)), and it discloses nothing resembling selecting a rule based on a combination of contents of two packets (element (c)). Properly characterized as background art on packet filtering efficiency, not anticipatory art.

2. WO 96/05549 A1 — Shiva Corporation

Description: A remote access server (the disclosure names the Shiva "LanRover") mediates a remote user's access to a local network. The user's ID string is used to index a database of access filters; the server applies those filters to control access to network services. It contrasts "per-user" filtering with "per-port" and "per-server" schemes, emphasizing that per-user filtering ties restriction to remote user identity. Access is controlled by identity/destination/service — not by packet payload content.

§102 assessment vs. claim 1: Does not anticipate. No payload-content-based rule selection (element (b)); no cross-packet content combination (element (c)). User-identity-based access control is orthogonal to claim 1.

3. EP 0 762 707 A2 — Telia AB

Description: An arrangement providing network access through the telecommunications network using a remotely controlled filter — i.e., an access filter whose configuration is controlled remotely (e.g., a subscription/connection-level gate in a dial‑up telecom access architecture). Filtering operates at the connection/address level.

§102 assessment vs. claim 1: Does not anticipate. Nothing indicates payload-content rule selection (element (b)) or multi-packet content correlation (element (c)). Confidence note: my retrieved detail on this reference's internal disclosure is thinner than for the others; the assessment rests on the reference's title/abstract scope and the fact that it is a telecom access-control filter, not a content-inspection engine. If you need a paragraph-level cite, the EP A2 specification would have to be pulled directly.

4. US 5,983,270 A — Abraham et al. / Sequel Technology Corporation ⭐ (most content-relevant face citation)

Description: A network management program for traffic between an intranet and an internetwork. A filter executive optimizes user/group policies into rules passed to a filter engine; the filter engine filters outbound and verifies inbound packets against those rules. Critically, the rule set includes content-characteristic rules — the patent's claim 15 recites "(a) a file extension rule … (b) an application protocol rule … and (c) a combined site and protocol rule." Thus filtering is not purely address/port based; it can key on a file extension or application protocol derived from the traffic.

§102 / §103 assessment vs. claim 1: This is the closest content-oriented reference among the face citations and was the primary asserted ground in the IPRs (obviousness over Abraham in view of Rubin). It plausibly reads on element (a) and supports element (b) insofar as rules may depend on content attributes (file extension, application protocol). It does not, however, disclose element (c)'s requirement that the rule be selected from a combination of the contents of the received packet and the contents of at least one other packet. So it is §103 material, not a clean §102 anticipation.

Important procedural fact from the prior sections (cross-referenced, not repeated): In IPR2014‑00587 the Board credited the patent owner's actual reduction to practice in early 1996 (no later than June 11, 1996) — corroborated by the AT&T "GeoPlex" prototype testimony — which disqualified the Abraham reference (and the Hughes reference) as prior art to the '694 invention. That is why the Final Written Decision (Sept. 23, 2015) held claim 1 not shown unpatentable. This is consistent with Abraham's 1997 filing date postdating the inventor's earlier reduction to practice.

5. US 6,219,706 B1 — Fan et al. / Cisco Technology, Inc. ⭐ (strongest technical reference on the face)

Description: A firewall/access-control system implemented on a dedicated network device (a router) between a local network and an external network. It dynamically allocates channels through the firewall based on the context of an application conversation, using access control lists (ACLs) for header-based decisions plus stateful inspection. Significantly, "the system may selectively examine packet payloads to determine when new channels are about to be opened" — e.g., it inspects the payload of an FTP control channel to identify the negotiated data-channel port number and then dynamically modifies the ACL to enable that channel. The background section also discusses Check Point "Stateful Inspection" firewalls that "inspect not only the packet header but also the packet payload."

§102 / §103 assessment vs. claim 1: This is the most technically dangerous face citation. Its filing date (Oct. 16, 1998) precedes the '694 provisional date (Oct. 22, 1998) by six days, so it is at least arguably §102(e) prior art relative to the '694 priority date (subject to the provisional's §112 support for claim 1). It plainly reads on element (a) and materially on element (b) — a firewall that examines payload content to drive access-control action (opening/closing channels, modifying ACLs). Its analysis of an application conversation (multiple packets/state) edges toward element (c)'s multi-packet notion. However, it does not expressly disclose selecting a PASS/DROP access rule "based upon a combination of the contents of the packet … and the contents of at least one other packet." Its payload inspection drives channel/ACL provisioning rather than the cross-packet rule-selection the claim requires. → Strong §103 combination candidate; not a clean §102 anticipation.

6. US 6,584,508 B1 — Epstein et al. / Networks Associates Technology, Inc.

Description: A "data guard" built as a multi-part proxy: a first proxy agent (inside network), a second proxy agent (outside network), and a content-based filter application that reviews information passed between the two agents. The proxy agents translate protocol operations into protocol-independent data analyzed by a protocol-independent content filter; software wrappers constrain component behavior.

§102 assessment vs. claim 1: Relevant in spirit (content-based filtering between proxies), but weak as prior art for the '694. Its earliest date (provisional 60/143,553, Jul. 13, 1999) is after the '694 priority date (Oct. 22, 1998), so it is not §102(e) prior art relative to the '694's priority. It is also a system/apparatus reference and does not disclose element (c)'s cross-packet rule-selection. Its appearance in the citation list is best explained as general background on content-filtering proxies rather than a §102/§103 thrust.

B. Non-Patent Literature Citation (examiner)

Bellovin, S.M. and Cheswick, W.R., "Network Firewalls," IEEE Communications Magazine, vol. 32, no. 9, Sept. 1, 1994, pp. 50–57.

• Description: The canonical survey of firewall technology — packet-filtering routers, application-level gateways/proxies, and hybrid architectures; discusses why packet-filtering gateways can only act on header information.
• §102 assessment vs. claim 1: Background/tutorial art. It frames the problem the '694 patent addresses (header-only filtering limitations) but discloses no payload-content rule selection (element (b)) and no multi-packet content combination (element (c)). Does not anticipate claim 1. (This same Bellovin/Cheswick article is also cited on the face of Abraham, US 5,983,270.)

C. Additional Prior Art Asserted in the IPRs (not on the '694 face — flagged for completeness)

Although the task is focused on the face citations, the actual invalidity grounds pressed in the PTAB litigations named other references. Per the "Notice of Related Matters" and petition papers:

Result: the IPR2014‑00587 Final Written Decision (Sept. 23, 2015) held claim 1 not shown unpatentable, the Board crediting an early‑1996 reduction to practice that antedated (and thus disqualified) the key references, including Abraham. (IPR2014‑00786 ran in parallel; the earlier sections flag an unresolved discrepancy in one aggregator's "Unpatentable" characterization of that proceeding — verify against the Feb. 12, 2018 IPR certificate, which I could not read substantively from the retrieved sources.)

D. Bottom Line — §102 Anticipation Scorecard (as to sole claim 1)

Conclusion: No reference cited on the face of US 6,826,694 anticipates claim 1 under §102. The two most relevant references — US 6,219,706 (Fan/Cisco) and US 5,983,270 (Abraham/Sequel) — are best characterized as §103 obviousness material, because each addresses payload/content-based access control but neither discloses the claim's controlling limitation: selecting the access rule based on a combination of the contents of the received packet and the contents of at least one other packet. The remaining references (Hausman, Shiva, Telia, Epstein, Bellovin) are header-based filtering, user/connection access control, or background firewall art that do not reach elements (b) or (c) at all.

Explicit uncertainties

1. EP 0 762 707 A2 — I did not retrieve a paragraph-level description of its disclosure; the "does not anticipate" conclusion rests on its scope as a telecom access filter. Pull the A2 spec for a firm cite.
2. US 6,219,706's §102(e) status — its Oct. 16, 1998 filing precedes the '694 provisional date by six days, but §102(e) treatment depends on the provisional's §112 support for claim 1; treat as arguably prior art, not settled.
3. US 6,584,508's citation basis — its July 13, 1999 date falls after the '694 priority date; why the examiner cited it is not explained by the retrieved record.
4. IPR2014‑00786 outcome — the "Unpatentable" label from one aggregator vs. the IBM FWD on the same sole claim remains unresolved without the Feb. 12, 2018 IPR certificate text.
5. I could not directly query the USPTO PatentCenter/PAIR or PTAB APIs; the citation set, citation types, and dates above are taken from the supplied authoritative full text and each reference's own front page.

I'll research the prior art references before building the §103 analysis.

I have the critical IPR record now. Let me pull details on the remaining cited references.

Obviousness Analysis — U.S. Patent 6,826,694 under 35 U.S.C. § 103

Builds on the previously generated patent summary and litigation summary. Claim text, bibliographic data, claim construction, and the district-court/PTAB case tables are not repeated here.

0. Contradiction flag — a previously "uncertain" item is now resolved

The earlier sections left two open items that this analysis can now close, and one that needs correcting:

Why the two panels diverged (same panel — APJs Moore, Petravick, Wood — in both cases) is the single most important fact for the § 103 analysis, and it is not a technical disagreement about the art. It is a date disagreement:

• IPR2014‑00587 (IBM): grounds were § 103 over Abraham + Hughes, § 103 over Coss + RFC 1123/959, and § 103 over Rubin + RFC 793. Patent Owner proved an earlier actual reduction to practice (early‑to‑June 1996, GeoPlex prototype), which antedated the § 102(e) references (Abraham was filed Apr. 2, 1997). Claim 1 survived.
• IPR2014‑00786: the dispositive reference was Rubin, and the Board held claim 1 "anticipated under 35 U.S.C. § 102(b) by Rubin." A § 102(b) printed publication cannot be sworn behind by a Rule 131‑type showing of prior invention. The 1996 reduction‑to‑practice evidence was therefore irrelevant to Rubin.

So the technical reach of the prior art was never really the battleground — the prior art reaches claim 1 in both cases. What saved the claim in 00587 was invention date.

1. Legal framework and the operative critical dates

Pre‑AIA § 103(a) governs (application 09/422,952 filed Oct. 22, 1999; priority to provisional 60/105,188 filed Oct. 22, 1998). The Graham/KSR framework applies: scope and content of the prior art, differences from the claim, level of ordinary skill, and any secondary considerations.

Critical dates:

Key structural observation: claim 1 is the sole claim, with no dependent-claim fallback. Therefore, if any one of the combinations below renders claim 1 obvious, the entire patent is invalid — there is no narrower claim to retreat to.

2. Person of ordinary skill in the art (POSITA)

A POSITA here would have a B.S. in computer science or electrical engineering (or equivalent) plus roughly two years of experience in TCP/IP networking and network security, or a master's degree with about one year of experience — i.e., someone who in 1996–1998 was fluent in packet structure (RFC 791), TCP stream reassembly, packet-filter rules and rule ordering, MTU/fragmentation behavior, and the then-standard two-tier firewall architecture (packet filter in front, application proxy behind).

3. Prior-art universe and status

3a. From the "Prior Art" section of the patent page (of record)

Point worth emphasizing: three of the six cited references — Abraham, Fan, and Epstein — were cited by the examiner (asterisked on the face of the patent). Abraham in particular was right in front of the examiner and discloses payload-content rule selection. The remaining delta over the of‑record art is narrow.

3b. Art relied on in the IPRs (separately labeled — not on the patent face)

4. Claim 1 limitation-by-limitation

Under the Board's construction (BRI): "contents of the packet" = "the header and/or payload of the packet"; and "selecting" does not require choosing from a set of rules. Both constructions widen the claim and make element (c) easier to meet.

5. Obviousness combinations

Combination A — Abraham + Rubin (§ 103(a)) — the adjudicated combination

Mapping. Abraham teaches [a] and [b] (filter engine comparing packets against per‑user rules including a file-extension rule). Abraham is silent on multi‑packet correlation — a point Petitioner itself conceded in IPR2014‑00786, framing the gap as "every element of claim 1 except for selecting an access rule based on the contents of multiple packets." Rubin supplies exactly that: an application proxy that reassembles IP packets into a TCP stream and searches it for a signature that "need not arrive in the same IP packet." Together, [a]+[b]+[c].

Motivation to combine (why a POSITA would have done it):

1. Express teaching of combinability. Rubin describes the combination as a workable strategy: "an application proxy for Web traffic can be added to an otherwise packet-filtering environment." Where the reference itself names the combination, motivation is not speculative.
2. The known problem of fragmentation / MTU. Application-layer objects that content filters must inspect — URLs in HTTP request lines, filenames, file-extension markers, Java class signatures — routinely exceed a single IP packet's payload or are segmented. A content rule that inspects only one packet produces false negatives (the very evasion Rubin highlights). Any POSITA implementing Abraham's content rules would necessarily add multi-packet reassembly to make the feature work — which makes [c] not merely obvious but arguably inherent in any operative content-filtering firewall.
3. Same field, same problem, same architecture. Both are perimeter access-control references addressing "which traffic may pass the boundary." Abraham is the rule engine; Rubin is the inspection point.
4. Predictable result / no change in principle of operation. The combination is a standard two-tier firewall (packet filter + application proxy). Every element performs its known function; the output is the sum of known outputs (KSR, "the combination of familiar elements according to known methods … does no more than yield predictable results").
5. No teaching away. Rubin does not disparage content inspection; it criticizes single-packet filtering and points toward stream-level inspection — i.e., toward the combination.

Expectation of success: high — Rubin reports the technique as implemented and effective.

Combination B — Norman + Rubin (§ 103(a)) — the combination the Board actually held to render claim 1 obvious

The FWD states: "Petitioner has demonstrated … that claim 1 is … unpatentable under § 103 over Norman and Rubin." Norman discloses a proxy firewall scanning packet content for viruses and hot words ([a], [b], and per Petitioner's expert, [c] because scanned files necessarily span multiple payloads); Rubin supplies the "detailed description of using a proxy to search the reassembled payloads of multiple IP packets for a signature string (CA FE BA BE)."

Motivation: Norman is a proxy-based content scanner; Rubin supplies the missing reassembly detail for exactly the same purpose (finding a signature that straddles packets). A POSITA improving Norman's scanner so that it detects viruses/signatures that are split across packets would look to the well-known technique of stream reassembly taught by Rubin. Predictable, low-risk, same field, same purpose.

Combination C — Cunningham alone, or Cunningham + Abraham/Bellovin

The Board held claim 1 "anticipated under 35 U.S.C. § 102(e) by Cunningham." Anticipation is the epitome of obviousness (In re Fracalossi; In re Pearson), so no separate motivation analysis is needed for this ground — a single anticipating reference establishes § 103.

If one insists on a combination framing: Cunningham + Bellovin (explicit statement that firewall types are commonly used together) or Cunningham + Abraham (rule structuring / per-user policy granularity) would each be motivated by the shared field and the shared objective of applying application-layer policy at a network boundary.

Combination D — Abraham + Bellovin (both of record; strongest ground for a reexam/invalidity challenge)

This is the cleanest combination built only from references already cited on the patent's own face plus the examiner-cited Abraham, and it is the ground most likely to survive without new-art discovery:

• Abraham → [a], [b] (file-extension rule; application-protocol rule; combined site+protocol rule).
• Bellovin → the application-gateway/proxy mechanism for [c], plus an express motivation: *"The authors classify firewalls into three main categories: packet filtering, circuit gateways, and application gateways. Commonly, more than one of these is used at the same time."* (https://dl.acm.org/doi/abs/10.1109/35.[312843](/patent/312843))
• Motivation: Bellovin affirmatively tells the artisan to use more than one firewall category simultaneously; Abraham tells the artisan to make pass/drop decisions from payload content; Bellovin's application gateway is the standard tool for operating on reassembled application data. The combination is the predictable union of the two, with each reference doing what it already does.

Combination E — Hausman and/or Shiva and/or Telia, plus a content-proxy reference

• Hausman (packet filtering for data networks) + Abraham supplies the filter datapath and rule engine ([a]).
• Shiva WO 96/05549 (restricting access to a local computer network) supplies the boundary gateway and authenticated access restriction.
• Telia EP 0 762 707 (remote-controlled filter for network access) supplies externally-supplied filter control, which is the specification's own "rule obtained from a node external to the firewall" — notably, that embodiment is not claimed, so Telia is more useful as evidence of the field's general knowledge than as an element-by-element reference.
• Motivation: all three are in the same field of endeavor (network boundary access control), address the same problem, and are combinable with a content-inspecting proxy without altering their principles of operation.

Caveat: I retrieved only front-page/title-level data for Hausman, Shiva, and Telia. Treat Combinations E as candidate grounds requiring full-text verification.

Combination F — Fan + Bellovin (weakest; lead with caution)

Fan (US 6,219,706, Cisco) has an Oct. 16, 1998 priority — only six days before the '694's provisional filing. It is § 102(e) art at best and is defeatable by the June 1996 reduction-to-practice showing that already succeeded in IPR2014‑00587. A challenge built on Fan would face the same antedating defense that defeated IBM's Abraham/Hughes ground. Use Fan as corroboration of the state of the art, not as the primary reference.

6. Motivation-to-combine synthesis (KSR rationales consolidated)

7. Secondary considerations

I found no evidence in the retrieved record of unexpected results, commercial success with nexus, long-felt-but-unsolved need, industry praise, or copying. The only substantive evidence Patent Owner produced in the IPRs was priority evidence (the GeoPlex prototype and Dr. Vrsalovic's corroborating testimony) — which is a § 102 antedating argument, not a Graham secondary consideration. That evidence won IPR2014‑00587 and was unavailing in IPR2014‑00786.

8. Bottom line

1. Claim 1 is obvious under § 103 over multiple independent combinations — most convincingly Abraham + Rubin, Norman + Rubin, and Abraham + Bellovin (the last built entirely from references on the patent's face plus the examiner-cited Abraham).
2. This is not a contested technical proposition. The PTAB held the sole claim unpatentable in IPR2014‑00786: anticipated by Rubin under § 102(b), obvious over Norman + Rubin under § 103, and anticipated by Cunningham under § 102(e). Only limitation [c] (multi-packet rule selection) was ever in dispute, and the Board found it met by both Rubin and Cunningham.
3. The patent's survival in IPR2014‑00587 was a date outcome, not a merits outcome. The same panel accepted the inventors' June 1996 reduction to practice, which antedates § 102(e)/(a) references (Abraham, Hughes, and by extension Fan), but which cannot remove a § 102(b) bar (Rubin, published Feb. 1997).
4. Practical consequence: any § 103 challenge to claim 1 should lead with § 102(b)‑qualified art (Rubin, Bellovin, Norman, Hausman, Shiva, Telia) or Cunningham as § 102(e)/§ 103 art if the antedating record is available, because a challenge built on § 102(e) art alone (Abraham, Fan, Epstein) is exposed to the same 1996 antedating showing that defeated IBM.

9. Caveats and limits of this analysis

1. Full-text verification outstanding for Hausman (US 5,473,607), Shiva (WO 96/05549), Telia (EP 0 762 707), Fan (US 6,219,706), and Epstein (US 6,584,508). Their characterizations above derive from titles, front-page data, and the examiner's citation, not from retrieved specifications. Combinations D, E, and F should be confirmed against full text before being asserted.
2. IPR certificate content (Feb. 12, 2018) remains unverified; my statement that claim 1 was likely cancelled is an inference from the 00786 FWD, not a retrieved fact.
3. "Hughes" (IBM's IPR2014‑00587 ground 1 partner to Abraham) was not identified in my retrieval and is not characterized.
4. The reconciliation of the 00587 vs. 00786 outcomes as a § 102(b)/§ 102(e) distinction is my analytical inference, consistent with the FWD language quoted (00786: "anticipated under 35 U.S.C. § 102(b) by Rubin"). The 00587 FWD's own reasoning on the antedating evidence was not re-retrieved in this pass.
5. This is a technical/legal analysis for analytical purposes, not a legal opinion; obviousness and validity determinations belong to the courts and the Office.