US 7047561

I'll search for authoritative information on US 7,047,561, including any 2026 CAFC activity.

Let me look specifically at the litigation history and any 2026 appellate activity.

Let me check the 2018 appeal and confirm whether there is any 2026 appellate activity.

Summary — US 7,047,561 B1

Note on sourcing/uncertainty: The full text of the patent (as published) was available to me and is authoritative for the bibliographic and claim content below. The USPTO/CAFC litigation searches I ran did not surface any 2026 Court of Appeals for the Federal Circuit docket activity involving this patent. The only Federal Circuit appeals tied to the '561 patent that I can verify are from 2017 and 2018 (detailed below). I state explicitly where I lack authoritative information rather than inferring.

Bibliographic data

Abstract (verbatim)

The present invention relates to a firewall for use in association with real-time Internet applications such as Voice over Internet Protocol (VoIP). The firewall applies an application proxy to the signaling and control channels and a packet filter to the bearer channels. One of the features of hybrid firewall is that the application proxy can instruct the packet filter as to which bearer channels to enable and disable for the duration of a real-time Internet application session. The hybrid firewall can also intelligently perform network address translation (NAT) on Internet protocol packets incoming and outgoing to the firewall.

Plain-language overview of the independent claims

• Claim 1 (firewall/apparatus). A firewall handles IP packets that carry data for a real-time Internet application, where each packet belongs to one of three channel types: a signaling channel, a control channel, or a bearer channel. The firewall contains two distinct mechanisms — an application proxy and a packet filter. Packets on the signaling and control channels are handed to the application proxy; packets on the bearer channel are handed to the packet filter. (In substance: deep, layer-7 inspection for the security-sensitive call-control traffic; fast, low-latency layer-3/4 filtering for the delay-sensitive media stream.)

• Claim 12 (method). A method of protecting a computer network that sends/receives IP packets formatted per a real-time Internet protocol, again with each packet belonging to a signaling, control, or bearer channel. Steps: (i) receive a stream of IP packets; (ii) apply the signaling- and control-channel packets to an application proxy; (iii) apply the bearer-channel packets to a packet filter.

• Claim 17 (computer-readable medium). A computer-readable medium holding computer-executable program code for the same three operations as claim 12 (receive a packet stream; send signaling/control packets to an application proxy; send bearer packets to a packet filter).

Dependent claims (context). Claims 2–5 specify the application type (VoIP, fax-over-Internet, video-over-Internet, voice messaging). Claim 6 covers the application proxy instructing the packet filter which bearer-channel packets to enable/disable for the duration of a session. Claim 7 covers NAT at layers 3 and 4 for signaling, control and bearer packets; claim 8 covers NAT at layer 7 (embedded addresses/ports) in signaling and control packets; claim 9 covers the proxy instructing NAT to stay active for the whole session independent of data traffic flow; claim 10 covers control logic; claim 11 covers housings (dual-homed workstation, general-purpose workstation, dedicated appliance, or ASIC). Claims 13–16 and 18–21 mirror these limitations in method and CRM form.

Literal-reading note (per the no-auto-correction rule): the published claim text contains several apparent typographical/OCR artifacts that I am reporting as they appear rather than correcting — e.g. claim 11 recites "said application proxy and said packet filter ane housed"; claim 15 recites "Internet Protocol (EP) addresses"; claim 17 recites packets "foxed in accordance with a real-time Internet protocol"; claim 19 recites "flier comprising computer program code." I cannot confirm from the sources retrieved whether these appear in the printed patent or are rendering errors.

Litigation history (for context)

• Genband US LLC v. Metaswitch Networks Ltd / Metaswitch Networks Corp., E.D. Tex. Case No. 2:14-cv-00033 (filed 2014-01-21). Asserted claims of the '561 patent were claims 6, 17, and 20, directed at Metaswitch's Perimeta Session Border Controller. A January 15, 2016 jury verdict found infringement of all asserted claims and validity, with $8,168,400 in damages. The September 29, 2016 bench-trial opinion (Judge Gilstrap) found the '561 asserted claims patent eligible under 35 U.S.C. § 101 and rejected Metaswitch's equitable defenses, but denied a permanent injunction. This ties to the Google Patents litigation entry for E.D. Tex. and to claim-construction briefing (Magistrate Judge Roy S. Payne's supplemental claim construction order of August 6, 2015 addressed the '561 patent).
• Federal Circuit No. 2017-1148 — Genband US LLC v. Metaswitch Networks Corp., 861 F.3d 1378 (Fed. Cir. July 10, 2017) (precedential). The court vacated the denial of a permanent injunction and remanded, holding it could not tell whether the district court applied too stringent a "drive demand" causal-nexus standard. The '561 patent appears in the list of asserted patents in that appeal, but the holding centered on injunction law, not the '561 claims.
• E.D. Tex. No. 2:18-cv-00155 — Sonus Networks, Inc. v. Metaswitch Networks Ltd et al. (Stanford NPE Litigation Database lists '561 among the patents in this case).
• Federal Circuit No. 2018-1869 — listed in the Google Patents litigation data for this family. I do not have an authoritative record of its specific holding or disposition, so I will not characterize it.

On the requested "CAFC 2026 dockets"

I found no authoritative evidence of any 2026 Federal Circuit proceeding involving US 7,047,561. Given the patent's status ("Expired – Lifetime," adjusted expiration April 17, 2023), new appellate activity in 2026 would be unusual. If you have a specific 2026 docket number or party name, I can run a targeted check — but I will not fabricate a case number, holding, or party list.

I'll research litigation involving US 7047561, starting with the litigation records already referenced in the patent data and then searching for additional cases.

Let me dig deeper into the specific cases and the CAFC appeals.

Litigation Involving U.S. Patent No. 7,047,561

Bottom line: Yes — U.S. 7,047,561 (the '561 patent, "Firewall for real-time internet applications") has been asserted in litigation. It was one of several patents in a pair of related patent‐infringement suits in the Eastern District of Texas in which the plaintiff (Nortel's successor, GENBAND/Sonus/Ribbon) accused Metaswitch Networks of infringement. The '561 patent also generated two Federal Circuit appeals. All district‐court cases were resolved (verdict, then final judgment and settlement).

Below are the known actions. I've kept strictly to the '561 patent (US 7,047,561 B1); similar numbers have been excluded.

---

1. GENBAND US LLC v. Metaswitch Networks Ltd. / Metaswitch Networks Corp. — E.D. Tex. (the primary case)

Procedural history (based on the district and appellate opinions):

• Jan. 15, 2016 — A jury found that Metaswitch infringed all asserted claims of the seven/eight patents (including the '561 patent) and that the claims were not invalid, and awarded $8,168,400 in damages. (The court's post‐trial opinion records the verdict and the assertion of '561 claims 6, 17, and 20 against Metaswitch's Perimeta product.)
• Sept. 29, 2016 — After a bench trial, the district court (211 F. Supp. 3d 858) rejected Metaswitch's equitable defenses (laches, implied waiver, etc.) but denied GENBAND's request for a permanent injunction.
• March 22, 2018 — The district court entered final judgment awarding GENBAND roughly $8.9 million in royalties (plus costs later set at ~$0.4 million).
• A companion counterclaim suit by Metaswitch (patents asserted against GENBAND) was filed as Metaswitch Networks Ltd. v. GENBAND US LLC, 2:14-cv-00744 (E.D. Tex.). A jury there (March 2016) found GENBAND did not infringe Metaswitch's patents and that three of them were invalid.

Sources: Genband US LLC v. Metaswitch Networks Corp., 861 F.3d 1378 (Fed. Cir. 2017); district‐court post‐trial opinion, E.D. Tex. 2:14-cv-00033, Dkt. 582; Stanford NPE Litigation Database (patent 7047561); Unified Patents litigation portal for case 2:14-cv-00033.

---

2. Federal Circuit Appeal No. 2017-1148 (GENBAND's appeal of the injunction denial)

Sources: CAFC 17-1148 opinion; 861 F.3d 1378; IPWatchdog and law-firm bulletins (Wiley, Brinks, National Law Review).

---

3. Federal Circuit Appeal No. 2018-1869 (Metaswitch's appeal of the final judgment)

Sources: Ribbon Communications 2019 Form 10-K; E.D. Tex. 2:18-cv-00155 docket (stay of "follow‐on" Federal Circuit appeal); Unified Patents litigation portal listing CAFC 18-1869.

---

4. Sonus Networks, Inc. v. Metaswitch Networks Ltd. et al. — E.D. Tex. (follow‐on case)

Sources: Justia Dockets, E.D. Tex. 2:18-cv-00155; Unified Patents litigation portal for 2:18-cv-00155 (status: closed; termination 2019-06-30); Stanford NPE Litigation Database (identifies 2:18-cv-00155 as case involving patent 7047561).

---

Additional context (assignments / expiration)

• Current owner: The patent issued to Nortel Networks Limited; ownership passed (via Nortel's 2010 Carrier VoIP sale and later transactions) to GENBAND US LLC, then to Ribbon Communications Operating Company, Inc. (f/k/a GENBAND US LLC and Sonus Networks, Inc.).
• Legal status: Expired – Lifetime, with an adjusted expiration date of April 17, 2023 (per the Google Patents legal-status data). Any litigation would therefore be concluded/settled; no active assertion of the '561 patent is currently indicated.

---

Caveats

• This summary is drawn from public litigation records (Stanford NPE Litigation Database, Unified Patents litigation portal, Justia/PACER dockets) and the reported opinions. Docket numbers and filing dates are reproduced literally from those sources; minor date discrepancies (e.g., the -155 filing date of April 17 vs. April 18, 2018) reflect differences among the source databases.
• I did not find any litigation in which the '561 patent was asserted by a non‐practicing entity, nor any case unrelated to the GENBAND/Sonus–Metaswitch dispute.
• Related, non‐'561 proceedings in the same commercial dispute include a Texas state‐court trade‐secret action (filed March 28, 2017) and a Metaswitch antitrust suit in the S.D.N.Y. (filed Nov. 9, 2018); these do not directly involve the '561 patent.
• I was unable to verify a merits disposition for Federal Circuit appeal 2018-1869 beyond the stay/settlement record; if a formal dismissal order is needed, it should be confirmed on the CAFC docket.

I'll research whether any AIA trial proceedings exist against US 7,047,561, starting with targeted PTAB searches.

The first pass surfaced only false positives (the "'561" hits there refer to different patents, e.g. US 8,078,561 and Genband's US 8,687,640). Let me run more targeted queries against the '561 patent specifically.

Let me run more targeted queries to confirm whether any petition was ever filed against this specific patent.

PTAB / AIA Trial Proceedings — US 7,047,561 ("Firewall for real-time internet applications")

Proceedings overview

Total AIA trial proceedings on file: 0. No IPR, PGR, or CBM petition has ever been instituted, denied, or terminated against US 7,047,561 — so the breakdown is 0 active / 0 claims invalidated / 0 claims sustained / 0 settled / 0 institution denials, and the bottom line for a defendant is the mirror image of the usual outcome: the claims are untouched by the PTAB, but so is your estoppel exposure to a prior petitioner — there is nobody to inherit a winning record from. This patent's validity was instead litigated to a jury verdict in district court (Genband US LLC v. Metaswitch Networks Ltd., E.D. Tex. No. 2:14-cv-00033), where asserted claims 6, 17, and 20 came out valid, infringed, and patent-eligible — and the court then denied Metaswitch's Rule 50(b) JMOL. That is a materially worse defensive posture than a patent with a mixed IPR record, because you cannot point the court at a PTAB-issued cancellation of anything.

There are no proceedings to enumerate

The canonical list — the structured "PTAB proceedings on file" block sourced from the USPTO Open Data Portal — is empty for this patent. I was unable to contradict it. My searches surfaced only false positives, which I flag below because a naive keyword search will lead you to the wrong patent and a wrong conclusion.

Important caveat on my verification. I ran targeted searches (patent number, title, patent owner/petitioner pairs, litigation-linked IPRs) and did not complete a live PTAB E2E / PatentCenter "Patent Number" docket query before exhausting my search budget. I therefore state the zero-count as the best-supported conclusion from the canonical ODP data plus every keyword search I could run, not as the output of a docket query I personally executed. If you need a certifiable negative for a brief, pull the PTAB E2E "Inter partes review / Post-grant review / Covered business method" result set for 7,047,561 yourself.

False positives you must not conflate with this patent

---

Strategic summary

Claim-by-claim status: nothing is CANCELED, nothing is PTAB-SUSTAINED, and most of the patent is UNTESTED. Every one of the 21 claims remains as issued — the patent simply expired on its adjusted expiration date of 2023-04-17 before any AIA petition matured. The only claims ever adjudicated on the merits were 6, 17, and 20, and the adjudication was favorable to the patent owner: a January 2016 jury verdict of infringement and validity over Metaswitch, followed by Judge Gilstrap's 2016-09-29 order denying Metaswitch's renewed JMOL on both infringement and invalidity, and by the court's separate conclusion that the "'561 Patent Asserted Claims are Patent Eligible" under § 101. Independent claims 1 and 12 — the apparatus firewall claim and the method claim — were never asserted and never construed on the merits of validity. So there is no cancelled claim to hand a court, and equally no PTAB precedent construing claim 1 narrowly.

The real defensive asset here is claim construction, not cancellation. Magistrate Judge Payne's supplemental claim construction order (E.D. Tex., 2015-08-06) construed "applying the Internet protocol packets associated with the signaling channel and control channel to [an/the] application proxy" to mean "applying, for the setup, duration, and take down of the real-time Internet application, all IP packets associated with the signaling channel and control channel to the application proxy." The court also construed "application proxy" as "a hardware and/or software component configured to operate at an upper level of the protocol stack and communicate with external services on behalf of a client," and "packet filter" as a component configured to examine incoming and/or outgoing packets and apply pre-defined filtering rules. Those constructions cut both ways; the "all IP packets … for setup, duration, and take down" gloss is a genuine non-infringement hook for an accused product that, e.g., hands only some signaling packets to a proxy, or that drops out of the call after setup. That argument survived a jury and a JMOL motion because of the accused product's specific conditional behavior (rate limiting / congestion control), so it is product-specific, not a general escape hatch.

Estoppel landscape: empty — and that is the point. Section 315(e)(2) estoppel attaches only to "a petitioner in an inter partes review … that results in a final written decision under section 318(a)." There is no such petitioner, no such FWD, and therefore no estoppel bar on any prior-art ground for any party — including Metaswitch and its privies, who are free to run in district court any § 102/§ 103 ground they never raised at the PTAB, because they never ran one there. But the reverse is also true: there is no estoppel working for you either. You cannot borrow a prior petitioner's grounds, because there is no prior petitioner. Conversely, because Metaswitch tried and failed to invalidate claims 6/17/20 before a jury, a new defendant inherits an adverse validity verdict on those three claims if it is in privity with Metaswitch — but it inherits nothing on claims 1–5, 7–16, 18, 19, and 21, which remain wholly untested.

Pattern signals. (1) No repeat-petitioner pattern exists, because there is no petitioner at all. (2) The patent owner has never had to defend a PTAB appeal on this patent — the CAFC activity it does have (Nos. 2017-1148 and 2018-1869) is litigation-driven, and per the previously generated sections the 2017-1148 appeal (Genband US LLC v. Metaswitch Networks Corp., 861 F.3d 1378) turned on injunction law, vacating the denial of a permanent injunction and remanding for clarification of the "drive demand" causal-nexus standard — not on the '561 claims. (3) No defensive aggregator (e.g., Unified Patents) appears anywhere in the chain for this patent; the Google Patents litigation entries are all private-party suits. (4) Politically, this is a manufacturing/telecom operating company patent (Nortel → Genband → Ribbon), not a troll patent — which is precisely why it drew district-court invalidity attacks and JMOL motions rather than a coordinated IPR campaign.

---

Recommended next steps

• State the negative plainly. If you are preparing a brief or an IPR-strategy memo: there are no AIA trial proceedings on US 7,047,561. The absence is itself a signal — this patent was asserted in 2014, 2016, and 2018 and still never attracted an instituted IPR, which tends to mean the invalidity case in district court was thin, not that challengers were asleep. Do not overstate the negative; just certify it. Confirm it independently at PTAB E2E: https://ptacts.uspto.gov/ (patent-number search across IPR/PGR/CBM).
• Verify the 2025 Smart RF papers are about 8,078,561 before you cite them. The 2026-10-14 and 2026-10-16 FWD deadlines in those records belong to a different patent: https://ptacts.uspto.gov/ptacts/public-informations/petitions/[1557431](/patent/1557431)/ and https://ptacts.uspto.gov/ptacts/public-informations/petitions/[1557562](/patent/1557562)/
• Anchor any non-infringement theory in the E.D. Tex. constructions, not in PTAB outcomes. The supplemental claim construction order is the dispositive document: Genband U.S. LLC v. Metaswitch Networks Ltd., No. 2:14-cv-33-JRG-RSP (E.D. Tex. Aug. 6, 2015). The operative limitation is the requirement that all signaling/control-channel IP packets go to the application proxy for setup, duration, and take down. Read your accused product's post-setup behavior against that gloss.
• Weigh the validity verdict you would be fighting. Claims 6, 17, and 20 carry a jury finding of validity and a denied JMOL (2016-09-29 order, E.D. Tex. 2:14-cv-33). If your proposed invalidity art is the same art Metaswitch lost on, expect the court to treat the verdict as persuasive on those claims. Different art on claims 1 and 12 (never asserted, never adjudicated) is a cleaner target — but note the patent is expired as of 2023-04-17, so any proceeding you contemplate is about past damages only.
• If you were served long ago, check § 315(b) before budgeting. One year from service of an infringement complaint is a hard bar for co-defendants and privies, and the Click-to-Call line makes even voluntarily dismissed suits count from the date of service. Given the 2014 and 2018 E.D. Tex. filings and the 2023 expiration, most realistic petitioners are time-barred regardless of merits.
• Do not rely on any 2026 Federal Circuit docket for this patent. Consistent with the earlier sections of this analysis, I found no authoritative evidence of 2026 CAFC activity involving 7,047,561, and I have not located a 2018-1869 disposition I can characterize. Cite only what you can retrieve: the CAFC opinion in Genband US LLC v. Metaswitch Networks Corp., 861 F.3d 1378 (Fed. Cir. 2017) via https://www.courtlistener.com/ and the Federal Circuit's own docket, and the E.D. Tex. filings via the Stanford NPE database for this patent (https://npe.law.stanford.edu/patent/7047561-0), which lists only Sonus Networks, Inc. v. Metaswitch Networks Ltd et al, No. 2:18-cv-00155.

Bottom line for a defendant today: there is no PTAB record to attack this patent with, no cancellation to point to, and no estoppel to exploit. Your leverage is (a) the narrowing "all packets, setup through take down" construction, (b) the small set of claims with any adjudicated validity finding (6, 17, 20) versus the large untested remainder, and (c) the fact that the patent expired in 2023, which caps exposure to historical damages if you were already practicing.

I'll research the assignment record for US 7,047,561, starting with the USPTO assignment data and the correspondent-of-record details the task emphasizes.

Excellent — I found a chain-of-title document. Let me pull the full record and the reel/frame details.

Ownership Chain — US 7,047,561 B1 ("Firewall for real-time internet applications")

Sourcing note. The full patent text supplied in this session (Google Patents legal-events extract) is authoritative and is the backbone of the timeline below; every reel/frame number cited here appears verbatim in that record or in a USPTO filing located in this session. One important limitation up front: the correspondent-of-record field (the attorney/agent who filed each recording) was not returned for any reel/frame in this chain by the sources I could reach. The Google Patents legal-events extract carries date, code, conveyance, owner name and reel/frame — but not the correspondent block that Assignment Center displays on the "Assignment Details" tab. I therefore state that explicitly rather than inferring a correspondent. Guidance for pulling the missing field is in the Verdict section.

Cross-check source located this session: a USPTO filing dated 2010‑11‑17 in application 09/735,427 (a §3.73(b) statement of chain of title) recites, for the Nortel patent family, the same GENBAND‑era reel/frame pairs — Reel 024468/0507 and Reel 024879/0475 and /0519 — corroborating the Google Patents legal events. (archived USPTO document)

---

Inventors

Unusual-pattern check: I found no evidence that the inventor departed the original assignee within 12 months of filing, and no evidence of a 102(c)/assignment-from-inventor irregularity. The inventor‑to‑Nortel assignment was executed one day before the filing date and recorded the same day — a routine, contemporaneous employee assignment, not a late-recorded cleanup. Not determinable: whether Lee remained at Nortel through the 2009 insolvency (declining-employer retention is common in this portfolio generally, but I have no record specific to this inventor, so I make no claim).

Literal-reading note: the printed record hyphenates/abbreviates the corporate name inconsistently ("Nortel Networks Ltd" in the bibliographic field vs. "NORTEL NETWORKS LIMITED" in the assignment record). These are the same entity; I have not auto-corrected either rendering.

---

Original assignee

Nortel Networks Limited (rendered "Nortel Networks Ltd" on the face of the patent).

• Did it ship a product embodying the claims? Yes — the patent is internal R&D from Nortel's voice/VoIP program. The specification itself is drafted for H.323/MGCP/Megaco/SIP call handling (claims 1, 7–9 recite signaling/control/bearer channel splitting plus NAT at layers 3, 4 and 7, and Fig. 4 describes a "dual homed commercial workstation" firewall). This maps onto Nortel's Carrier VoIP / softswitch product line (e.g., the Succession/Communication Server family). I can describe the technical fit confidently from the patent; I have not independently verified a specific SKU shipping this exact firewall implementation, so that narrower claim is not asserted here.
• Primary line of business: global telecommunications network equipment (wireline/wireless/optical/enterprise), headquartered in Brampton, Ontario; at its peak one of the largest telecom equipment vendors in the world.
• Current status: Insolvent and wound down. Nortel filed for creditor protection on 2009‑01‑14; the principal U.S. proceeding is In re Nortel Networks Inc., Bankr. D. Del. Case No. 09‑10138 (docket materials were retrieved this session), with parallel Canadian (CCAA) and EMEA administration proceedings. Nortel's businesses and patents were liquidated in 2009–2011. The entity no longer operates.

---

Assignment timeline

All dates below are executed / recorded as given in the Google Patents legal-events record for US 7,047,561 (which mirrors the underlying USPTO records). Conveyances marked "(lien)" are security interests — they are recorded against the patent but do not convey title; I include them because Assignment Center lists them and because they are diagnostically useful.

1. 2000‑09‑27 / 2000‑09‑28 — Reel 011147/0144

• Conveyance: Assignment of assignors' interest
• Assignor: Michael C. G. Lee (inventor)
• Assignee: Nortel Networks Limited
• Correspondent: not retrieved (see sourcing note)
• Context: Original inventor-to-employer assignment — executed one day pre-filing.

2. 2010‑05‑27 / 2010‑06‑02 — Reel 024468/0507

• Conveyance: Change of Name
• Assignor: GENBAND Inc.
• Assignee: GENBAND US LLC
• Correspondent: not retrieved
• Context: Change of name only — an intra-corporate relabeling, no transfer of ownership and no new consideration. (The same reel/frame pair is the standard GENBAND change-of-name record appearing across the entire GENBAND patent estate, e.g. US 7,965,627 and US 7,167,473.)

3. 2010‑05‑27 / 2010‑08‑25 — Reel 024879/0519

• Conveyance: Assignment of assignors' interest
• Assignor: Nortel Networks Corporation (as listed for this patent in the legal-events record)
• Assignee: GENBAND US LLC
• Correspondent: not retrieved
• Context: Acquisition — GENBAND's purchase of Nortel's Carrier VoIP and Application Solutions (CVAS) business out of Nortel's insolvency, closing 2010‑05‑27. This is the operative title transfer for the '561 patent.

4. 2010‑05‑28 / 2010‑06‑18 — Reel 024555/0809 (lien)

• Conveyance: Patent Security Agreement
• Assignor: GENBAND US LLC
• Assignee: One Equity Partners III, L.P., as Collateral Agent
• Correspondent: not retrieved
• Context: Securitization — acquisition financing; a lien, not a title transfer.

5. 2010‑10‑28 / 2010‑11‑09 — Reel 025333/0054 (lien)

• Conveyance: Security Agreement
• Assignor: GENBAND US LLC
• Assignee: Comerica Bank
• Correspondent: not retrieved
• Context: Securitization — senior secured credit facility lien.

6. 2010‑05‑27 / 2012‑03‑30 — Reel 027992/0443

• Conveyance: Corrective Assignment
• Assignors: Nortel Networks Limited and Nortel Networks Corporation
• Assignee: GENBAND US LLC
• Correspondent: not retrieved
• Context: Internal/cleanup — corrects the conveying-party data on Reel 024879/0519 to add Nortel Networks Limited as an assignor. This is a chain-of-title defect repair, and it is the single most analytically interesting record in the chain: the 2010 transfer named the U.S./parent entity while the patent was actually held by the Canadian IP-holder (NNL). Note the effective date is backdated to 2010‑05‑27, i.e. the correction relates nunc pro tunc to the original sale date.

7. 2012‑12‑19 / 2014‑01‑10 — Reel 031968/0955

• Conveyance: Release by Secured Party
• Assignor: One Equity Partners III, L.P., as Collateral Agent
• Assignee: GENBAND US LLC
• Correspondent: not retrieved
• Context: Lien release (clears record #4) — not a title transfer.

8. 2014‑12‑02 — Certificate of Correction (no reel/frame; not an assignment)

• Context: USPTO-issued certificate of correction. Flagged for completeness because it post-dates issuance and appears in the same legal-events stream; it is not an ownership record.

9. 2016‑07‑01 / 2016‑07‑06 — Reel 039269/0234 (lien)

• Conveyance: Patent Security Agreement
• Assignor: GENBAND US LLC
• Assignee: Silicon Valley Bank, as Administrative Agent
• Correspondent: not retrieved
• Context: Securitization — refinancing lien.

10. 2016‑07‑01 / 2016‑07‑07 — Reel 039280/0467

• Conveyance: Release and Reassignment of Patents
• Assignor: Comerica Bank, as Agent
• Assignee: GENBAND US LLC
• Correspondent: not retrieved
• Context: Lien release (clears record #5).

11. 2016‑07‑01 / 2017‑01‑03 — Reel 041422/0080 (lien correction)

• Conveyance: Corrective Assignment
• Assignor: GENBAND US LLC
• Assignee: Silicon Valley Bank, as Administrative Agent
• Correspondent: not retrieved
• Context: Corrects Reel 039269/0234 by removing patent no. 6,381,239 from that security agreement. Another record-accuracy repair — worth noting as a pattern of lien records being filed and then cleaned up.

12. 2017‑12‑21 / 2017‑12‑29 — Reel 044986/0303

• Conveyance: Termination and Release of Patent Security Agreement
• Assignor: Silicon Valley Bank, as Administrative Agent
• Assignee: GENBAND US LLC
• Correspondent: not retrieved
• Context: Lien release (clears record #9).

13. 2017‑12‑29 / 2018‑01‑02 — Reel 044978/0801 (lien)

• Conveyance: Security Interest
• Assignors: GENBAND US LLC and Sonus Networks, Inc.
• Assignee: Silicon Valley Bank, as Administrative Agent
• Correspondent: not retrieved
• Context: Securitization — post-merger combined-entity collateral package. Sonus Networks, Inc. appearing as a co-assignor reflects the Sonus/GENBAND combination that produced Ribbon Communications.

14. 2020‑03‑03 (recorded) — Reel/frame not shown in the retrieved record (lien)

• Conveyance: Security Interest
• Assignor: Ribbon Communications Operating Company, Inc.
• Assignee: Citizens Bank, N.A., as Administrative Agent
• Correspondent: not retrieved
• Context: Securitization — replacement credit facility.

15. 2021‑12‑06 (recorded)

• Conveyance: Termination and Release of Patent Security Agreement (at R/F 044978/0801)
• Assignor: Silicon Valley Bank, as Administrative Agent
• Assignee: Ribbon Communications Operating Company, Inc.
• Correspondent: not retrieved
• Context: Lien release (clears record #13).

16. 2024‑06‑24 (recorded)

• Conveyance: Release by Secured Party
• Assignor: Citizens Bank, N.A.
• Assignee: Ribbon Communications Operating Company, Inc.
• Correspondent: not retrieved
• Context: Lien release (clears record #14) — the last recorded event, leaving the patent unencumbered.

Chain-of-title observation on record #16 and the "current assignee" field. Google Patents lists the current assignee as Genband US LLC, while records #14–16 name "Ribbon Communications Operating Company, Inc. (f/k/a GENBAND US LLC and Sonus Networks, Inc.)". Because the parenthetical expressly identifies Ribbon as the same legal entity formerly known as GENBAND US LLC, this reads as a corporate renaming, not a new acquisition. Notably, no separate change-of-name assignment transferring GENBAND US LLC → Ribbon Communications Operating Company, Inc. appears in this patent's legal-events stream. That is a gap I flag rather than paper over; the name-change record may exist under a different reel or may simply not have been re-recorded per patent. Title outcome is unaffected (same entity), but a strict chain-of-title opinion should confirm it.

Also worth flagging: there is no assignment to any entity other than Nortel / GENBAND / Ribbon in this entire record. Every other entry is either a security interest or a lien release.

---

Timeline diagram

timeline
    title Ownership of US 7047561
    2000 : Filed by Nortel Networks
         : Inventor Lee assigns to Nortel
    2006 : Patent issues as US 7047561
    2009 : Nortel files for bankruptcy
    2010 : GENBAND buys Nortel Carrier VoIP
         : GENBAND name change recorded
         : One Equity Partners lien
         : Comerica Bank security pact
    2012 : Corrective assignment recorded
    2014 : One Equity lien released
         : Metaswitch suit filed in E D Tex
    2016 : Silicon Valley Bank security
         : Comerica release recorded
    2017 : SVB release and new lien
    2018 : SVB security for GENBAND and Sonus
    2020 : Citizens Bank security interest
    2021 : SVB lien terminated
    2024 : Citizens Bank lien released

---

NPE / troll-pattern signals

1. Shell-entity transfer — NOT PRESENT

No assignee in the chain carries an "IP / Patents / Licensing / Holdings / Ventures" suffix, and none is a purpose-built licensing vehicle. The title-holding entities are Nortel Networks Limited (record #1), GENBAND US LLC (records #2, #3, #6) and Ribbon Communications Operating Company, Inc. (records #14–16). GENBAND and Ribbon are operating telecom-equipment vendors with product lines and customers, not licensing shells. The only non-operating assignees are lienholders: One Equity Partners III, L.P. as Collateral Agent (024555/0809), Comerica Bank (025333/0054), Silicon Valley Bank (039269/0234, 041422/0080, 044978/0801), Citizens Bank, N.A. (2020‑03‑03). Secured-party records are not title transfers and are not shell-entity evidence.

2. Known asserter in the chain — NOT PRESENT

No assignee matches any published NPE register (Acacia, Marathon, Intellectual Ventures, IPNav, Wi‑LAN, Mosaid/Conversant, Vringo, Pendrell, Innovatio, MPHJ, Lumen View, Round Rock, Document Generation Corp, Spangenberg entities, or any Unified Patents / RPX high-frequency plaintiff). The asserter here is GENBAND US LLC, a practicing session-border-controller and softswitch vendor, suing a direct competitor (Metaswitch) — the inverse of the NPE pattern. Caveat, stated deliberately: the '561 patent's Nortel siblings were swept into Rockstar Consortium / Constellation / Spherix, which are classic privateering vehicles; the '561 patent was not in that residual-portfolio sale (see signal 6). I have found no evidence linking the '561 patent to Rockstar or its subsidiaries, and I do not assert one.

3. Repeat correspondent across the chain — UNCLEAR / INSUFFICIENT DATA

This is the signal the task weights most heavily, and it is the one I cannot close. The legal-events data I could reach supplies conveyance type, parties, dates and reel/frame — but not the correspondent block for any of the 16 records. I therefore cannot report an attorney name or firm, and I will not guess one. What I can say from the record shape: the 2010 Nortel→GENBAND acquisition records (024468/0507, 024879/0519, 027992/0443) were plainly prepared as a single estate-wide filing batch — identical execution date 2010‑05‑27 across the family, identical reel series 024879, and the same corrective-assignment mechanism applied repeatedly (the 09/735,427 §3.73(b) statement lists the same pairs for dozens of unrelated Nortel patents). That is consistent with one law firm handling the entire GENBAND recording program, which is exactly the repeat-correspondent pattern — but "consistent with" is not a finding. To convert this to a finding, pull the correspondent field for each of the 16 reel/frames at Assignment Center (see Verdict).

4. Cascading transfers — NOT PRESENT

There is no chain of LLCs transferring title in rapid succession. Title moved exactly twice in 24 years: inventor → Nortel (2000) and Nortel → GENBAND (2010). Records #2 and #6 are a name change and a corrective assignment, not transfers; records #4, #5, #9, #13, #14 are liens. The closest thing to a rapid sequence is 2010‑05‑27 → 2010‑05‑28 → 2010‑10‑28 (sale plus two financings within five months), which is ordinary acquisition-financing choreography, not NPE layering.

5. Pre-litigation transfer — NOT PRESENT

The first infringement suit naming this patent was filed 2014‑01‑21 (GENBAND US LLC v. Metaswitch Networks, E.D. Tex. 2:14‑cv‑00033). The last title-conveying assignment was recorded 2010‑08‑25 — approximately 3 years 5 months before suit, and roughly 3 years 9 months after its execution. The nearest later assignment-adjacent record, the 2012‑03‑30 corrective assignment (027992/0443), is a correction with a backdated 2010‑05‑27 effective date, not a new transfer, and still lands ~22 months before the complaint. Neither is within the 6‑month window that would suggest the chain was arranged to enable assertion, set venue, or manufacture standing. The GENBAND acquisition was a business purchase, and the '561 patent rode along in a multi-thousand-patent asset package — the lawsuit followed the acquisition, not the reverse.

6. Bankruptcy fire-sale — PRESENT, BUT QUALITATIVELY DISTINCT (this is the one signal I mark present, with a firm qualifier)

The original assignee did enter insolvency: Nortel filed 2009‑01‑14 (In re Nortel Networks Inc., Bankr. D. Del. 09‑10138; CCAA in Canada), and the '561 patent did pass through that estate — the 2010‑05‑27 transfer to GENBAND closed out of the Nortel insolvency process.

But the mechanism matters and cuts against the NPE read. Nortel sold its assets in two very different channels:

• the "Business Line Sales" — going-concern sales of operating divisions (GENBAND bought the Carrier VoIP / Application Solutions business, closing 2010‑05‑27, which is why the '561 execution date matches the close); and
• the "Residual Patent Sale" — the ~6,000-patent auction won by Rockstar Bidco for ~$4.5B, later distributed to Rockstar/Constellation/Bockstar/Spherix.

The '561 patent went out through the first channel. There is a second, independent corroboration in this session's sources: the 2010‑11‑17 USPTO §3.73(b) statement — filed five months after the GENBAND purchase — already recites the '561 patent (issued May 16, 2006) in the GENBAND chain of title via reels 024468/0507 and 024879/0475 & /0519. A patent heading to Rockstar could not simultaneously be sitting in GENBAND's recorded portfolio in November 2010. So the '561 patent was never part of the Rockstar residual auction; it was part of an operating-business acquisition. Characterization: Chapter 11 was the setting, not the transaction type. This is not Kodak-style patent-only fire-sale behavior.

7. Privateering — NOT PRESENT (for this patent)

No operating company transferred the '561 patent to a licensing vehicle to assert on its behalf against competitors. The operating company (GENBAND) asserted in its own name against a direct competitor (Metaswitch) and litigated the patent through to a jury verdict of infringement and validity (January 15, 2016; $8,168,400) and a $8.9M final judgment (2018‑03‑22), plus a Federal Circuit appeal (No. 2017‑1148, 861 F.3d 1378, vacating denial of a permanent injunction). Operating companies litigating their own products' patents against competitors is the defining non-privateering posture. Explicit contrast for the reader: Rockstar is the textbook privateering vehicle for other Nortel patents; it is not in this patent's chain.

8. Defensive aggregator (anti-NPE) — NOT PRESENT

The chain does not terminate at RPX, Allied Security Trust, LOT Network, Unified Patents or Open Invention Network. It terminates at Ribbon Communications Operating Company, Inc., a publicly traded operating telecom vendor, unencumbered as of the 2024‑06‑24 Citizens Bank release. No neutralization occurred.

---

Verdict

Operating-company assertion

Justification. The assignment record contains no shell-entity transfer, no known NPE/asserter in the chain, no cascading LLC layering, no pre-suit transfer within 6 months of the 2014‑01‑21 complaint, and no defensive-aggregator terminus. Title moved only twice — inventor→Nortel (Reel 011147/0144, 2000) and Nortel→GENBAND US LLC (Reel 024879/0519, executed 2010‑05‑27, corrected at Reel 027992/0443), a going-concern acquisition of Nortel's Carrier VoIP business rather than a patent fire-sale. Every other recorded event is a security interest or lien release (024555/0809, 025333/0054, 039269/0234, 041422/0080, 044978/0801, plus the 2020 Citizens Bank lien), and the chain ends unencumbered at Ribbon Communications Operating Company, Inc. after the 2024‑06‑24 release. The patent was asserted by the practicing owner in its own name against a direct competitor (Metaswitch), yielding a jury infringement verdict and an $8.9M judgment — the signature of operating-company assertion, not NPE behavior. The bankruptcy element (signal 6) is real but flows through Nortel's business-line sale, not the Rockstar residual auction, and is therefore not an NPE indicator.

Two honest caveats that keep this verdict from being higher-conviction:

1. Signal 3 (correspondent recurrence) is unresolved. I could not retrieve correspondent data, so the one signal most likely to reveal a concealed NPE architecture is blank — not negative. If the 16 reel/frames share a single recording firm whose name appears on Unified Patents/RPX high-frequency-plaintiff rosters, signal 3 would flip to present and the verdict would need revisiting. Given that all title holders are demonstrably operating entities with products, I judge this unlikely, but I will not call it dismissed.
2. No recorded change-of-name for GENBAND US LLC → Ribbon appears in this patent's legal-events stream. Consistent with a pure rename (same legal entity, stated expressly in the record), but worth confirming in a formal chain-of-title opinion.

Verification

• Assignment Center (primary): https://assignmentcenter.uspto.gov/ — search by patent number 7047561 (or by the reel/frame pairs above). Open the "Assignment Details" tab for each record to capture the Correspondent block (name, firm, street address) that this analysis could not retrieve; that single field resolves signal 3.
• Legacy interface (same underlying records): https://assignment.uspto.gov/patent/index.html
• Cross-check of the GENBAND chain of title: the 2010‑11‑17 §3.73(b) statement in application 09/735,427, archived at the Internet Archive item gov.uspto.patents.application.09735427, which lists the 024468/0507 and 024879/0475 & /0519 pairs for the Nortel family.
• Bankruptcy / sale context (signal 6): In re Nortel Networks Inc., Bankr. D. Del. No. 09‑10138, Dkt. 5202 (bidding procedures, 2011‑04‑04) and Dkt. 5935 (sale order, 2011‑07‑11).

No fabrication note: every reel/frame, date, party and conveyance above is reproduced from the authoritative patent record or the USPTO filing cited. Where a field was unavailable — the correspondent block, the Citizens Bank reel/frame, and the GENBAND→Ribbon name-change record — I have said so rather than filling the gap.

I'll research the specific prior art references cited on the face of US 7,047,561. Let me start with targeted searches.

Let me research the remaining cited references.

Prior-Art Analysis — U.S. Patent No. 7,047,561 B1

Patent under analysis: US 7,047,561 B1 ("Firewall for real-time internet applications"), App. No. 09/671,250, filed 2000-09-28, issued 2006-05-16, inventor Michael C. G. Lee, original assignee Nortel Networks Limited.

Note on the number: I confirmed the target is US 7,047,561 (rendered "US7047561B1"). I excluded same-family/similar numbers; one search incidentally surfaced a USPTO paper referencing "US-20110055916-A1" and an unrelated application (15/614,956) — those are not this patent and are disregarded.

Critical date: Because the '561 application was filed 2000-09-28 with no earlier priority, the pre-AIA §102 framework applies, and the presumptive invention date is on/before 2000-09-28. Every patent reference on the face of the '561 patent has a filing/priority date before 2000-09-28.

Important methodological caveat (read first): My task was to identify anticipatory art under 35 U.S.C. §102 — which requires a single reference to disclose every element of a claim. Based on the full text I retrieved, none of the ten cited references is a true §102 anticipation of independent claims 1, 12, or 17, because none discloses the core limitation — routing signaling- and control-channel packets to an application proxy while routing bearer-channel packets to a packet filter. What these references actually do is supply the building blocks that would be combined in a §103 obviousness challenge. I flag below, per reference, (a) the claims most "close" for a hypothetical single-reference §102 argument, and (b) the reality that the better characterization is §103. I do not overstate any reference as anticipatory where the disclosure does not support it.

---

The ten U.S. patent citations (front-page references)

The Google Patents record lists these twice — once as "Citations (9)" and once as "Patent Citations (10)" (the 10th being US6061798A, a continuation of US5826014A). The lists are otherwise identical. Sources: https://patents.google.com/patent/US7047561B1/en (Patent Citations table).

(References 8–10 issued after the '561 filing date but were filed before it, so they can only be prior art under pre-AIA §102(e), not §102(a)/(b).)

---

Reference-by-reference analysis

1. US 5,606,668 A — Shwed (Checkpoint Software Technologies Ltd.)

Title: "System for securing inbound and outbound data packet flow in a computer network." Filed 1993-12-15; issued 1997-02-25.
Description: The foundational Checkpoint packet-filter/inspection patent. It discloses a system administrator GUI for entering security rules (network objects, services), which are converted into a filter script used by the packet filter (see the summary in Baum, US 6,400,707, which cites it). It is essentially a rule-driven packet filter operating at the lower layers.
Potential §102 mapping: The reference is cited as background technology. It bears on the packet-filter half of claim 1 (the "packet filter" element) and on claim 10 (control logic specifying operating parameters). It discloses nothing about application proxies, signaling/control vs. bearer channel separation, or real-time Internet protocols. It does not anticipate any of claims 1, 12, or 17. At most it is §103 background for the packet-filter element.

2. US 5,793,763 A — Cisco Technology, Inc.

Title: "Security system for network address translation systems." Filed 1995-11-03; issued 1998-08-11.
Description: NAT combined with an "adaptive security algorithm" — i.e., stateful filtering of inbound packets before NAT translation (allow only traffic matching an established outbound session; drop/log source-routed packets; permit DNS/ICMP selectively). This is the Cisco PIX-style NAT+firewall. (Detailed in the Slashdot discussion of this patent: https://slashdot.org/story/00/09/20/[1415229](/patent/1415229)/.)
Potential §102 mapping: Relevant to claims 7 and 14 (NAT at layers 3/4 applied to signaling, control, and bearer packets) and the NAT concept in claim 1's overall firewall. It does not disclose channel-based proxy/filter splitting, so it does not anticipate claims 1/12/17; it is §103 material for the NAT dependent claims.

3. US 5,826,014 A — Coley & Wesinger (Network Engineering Software)

Title: "Firewall system for protecting network elements connected to a public network." Filed 1996-02-06; issued 1998-10-20.
Description: A stand-alone "firewall box" running a plurality of proxy agents assigned to incoming requests by service/port number; an assigned proxy verifies authority and then completes the connection on behalf of the source. It also notes combination with packet filtering for anti-spoofing (claims 17–18 recite OS/router packet filtering). (Text: https://patents.google.com/patent/US6061798A/en and the US5826014 claim set at uspto.report.)
Potential §102 mapping: Strong disclosure of the application-proxy element (proxy agents per service) and of an optional packet filter element. But nothing ties the proxy to signaling/control channels and the packet filter to bearer channels of a real-time Internet application. Not an anticipation of claims 1/12/17; usable as §103 art for the proxy element.

4. US 6,061,798 A — Wesinger & Coley (Network Engineering Software)

Title: "Firewall system for protecting network elements connected to a public network." Priority 1996-02-06; filed 1998-10-19; issued 2000-05-09.
Description: A continuation/refiling in the same family as US 5,826,014, with claims reciting per-port proxy agents, protocol-conformance checks, logging, and discarding packets with nested executable commands (see claim set at https://patexia.com/us/patent/06061798). Substantively the same disclosure as #3.
Potential §102 mapping: Same as #3 — proxy-agents/packet-filter combination, but no real-time channel architecture. Not anticipatory of claims 1/12/17. (Because it has the earliest priority of the pair — 1996-02-06 — it is the better §102(e) reference of the two.)

5. US 5,968,176 A — Nessett et al. (3Com Corporation)

Title: "Multilayer firewall system." Filed 1997-05-29 (priority 1997-05-28); issued 1999-10-19.
Description: A distributed/pervasive firewall spanning multiple protocol layers and multiple network devices (routers, switches, repeaters, NICs, end systems), coordinated by a centralized policy-definition component that translates security-policy statements into device-specific configuration data. It explicitly builds on US 5,606,668 for packet filtering and discusses TCP/application-layer filtering. (Text: https://patents.google.com/patent/[US5968176A](/patent/US5968176A).)
Potential §102 mapping: Relevant conceptually to the "firewall comprising different mechanisms at different layers" idea and to control logic (claim 10). But it discloses a distributed policy firewall, not a single firewall that sends signaling/control packets to an application proxy and bearer packets to a packet filter. Not an anticipation of claims 1/12/17. §103 background only.

6. US 6,119,171 A — Ip Dynamics, Inc.

Title: "Domain name routing." Filed 1998-01-29; issued 2000-09-12.
Description: Domain-name-based routing/address mapping (a name-to-route mapping scheme). This is the reference with the weakest nexus to the '561 claims — it concerns naming/routing, not channel-differentiated proxy/filter firewalling. (I could not retrieve a full-text description in this session, so I rely on the title/assignee and the patent's own classification of it as background; I flag this as lower-confidence.)
Potential §102 mapping: No anticipation of any claim. It may have been cited for generic identity/address-mapping context. Treat as non-material to anticipation.

7. US 6,006,272 A — Lucent Technologies

Title: "Method for network address translation." Filed 1998-02-23; issued 1999-12-21.
Description: A NAT method for a home network behind a router; notably, the specification discusses that the configured (Internet-unique) source address "can be included in the payload in accordance with H.323 and IP security can be applied as well." (Text: https://patents.google.com/patent/US6006272.)
Potential §102 mapping: Relevant to claims 7–9/14–16/19–21 (NAT at layers 3/4, and the H.323 payload-address problem). It is background for the NAT feature and touches the H.323-embedded-address issue, but it does not disclose reaching into layer-7 payloads to translate embedded addresses under control of an application proxy that also coordinates channel NAT synchronously. Not anticipatory; strong §103 support for the NAT-dependent claims.

8. US 6,400,707 B1 — Baum, Eggerl, Burton & Cloutier (Bell Atlantic Network Services)

Title: "Real time firewall security." Filed 1998-08-27; issued 2002-06-04. §102(e) reference (issued after the '561 filing but filed before).
Description: This is the single most relevant cited reference for the '561 patent. It discloses a firewall for packetized voice in a hybrid Internet/PSTN network, comprising:

• a static filter device positioned between the gateway and the originating terminal that filters the conversation (bearer) stream of packetized voice;
• a control processor connected to the packet switch and the filter device that receives a real-time copy of the call set-up signaling;
• the control processor derives filter parameters from the signaling dialog and generates a filter-device control signal that reconfigures the filter device with customized per-conversation filters that "disappear on termination of the conversation."
  It also expressly characterizes both packet filtering and application proxies as known, and states proxies are too slow for this application. (Text: https://patents.google.com/patent/US6400707 and https://www.freepatentsonline.com/[6400707](/patent/6400707).html.)

Potential §102 mapping (closest of all ten):

• Claim 1 / 12 / 17 (channel-based proxy/filter split): Close but not anticipatory. Baum separately handles signaling (control processor analyzes set-up signaling) and bearer (filter device filters the voice stream) — conceptually the same division of labor as the '561 patent. But Baum's signaling path uses a control processor that copies and analyzes signaling and reconfigures a filter; it does not route signaling/control through a layer-7 application proxy that terminates and regenerates the connection, and it does not use the three-way "signaling channel / control channel / bearer channel" framing. A §102 anticipation theory is colorable but vulnerable because the "application proxy" element is not clearly disclosed applied to the signaling/control channels.
• Claim 6 / 13 / 18 (application proxy instructs packet filter which bearer packets to enable/disable for the session duration): Baum's control processor dynamically reconfigures the filter per conversation and the filter disappears at termination — materially the same concept. This is the reference most likely to be used against claim 6 (one of the three claims actually asserted in the Genband v. Metaswitch litigation, along with claims 17 and 20).
• Claim 10 (control logic): disclosed (the control processor).

Bottom line on Baum: Its practical relevance is as the primary §103 reference (alone or combined with a proxy reference such as Wesinger/US 5,826,014), and as the strongest §102(e) candidate for the dynamic "enable/disable bearer channels for the session duration" limitation of claim 6.

9. US 6,728,885 B1 — Networks Associates Technology, Inc.

Title: "System and method for network access control using adaptive proxies." Filed 1998-10-09; issued 2004-04-27. §102(e) reference.
Description: Network access control using adaptive proxy techniques (a proxy whose behavior/configures itself based on context). (Full text not retrieved this session; I rely on title/assignee and its treatment as proxy-art background — flagged as lower-confidence.)
Potential §102 mapping: Bears on the application-proxy element (claims 1/12/17) and, at most, claim 10 (control logic). It does not disclose channel-based splitting for real-time Internet applications. Not anticipatory; §103 background for the proxy element.

10. US 6,650,641 B1 — Cisco Technology, Inc.

Title: "Network address translation using a forwarding agent." Filed 1999-07-02; issued 2003-11-18. §102(e) reference.
Description: NAT implemented via a forwarding agent (relevant to how translations are assigned/managed). (Full text not retrieved this session — lower-confidence description.)
Potential §102 mapping: Bears on the NAT dependent claims (7, 14, 19; and the NAT sub-elements of 8/9). No channel-based proxy/filter teaching. Not anticipatory; §103 background.

---

Non-patent citations (also on the face of the patent)

These five are listed under "Non-Patent Citations (5)" and were cited as general firewall/security background:

1. "General Firewall White Paper," www.ntresearch.com/firewall.htm.
2. "NAT and Networks," www.suse.de/~mha/linux-ip-nat/diplom/node2.html.
3. "White Paper Definitions," www.e2s.be/.../Surfcontrol/fdefns.htm.
4. "White Paper – The Firewall and Online Security," www.xpertdata.nl/service/papers/firewall.html.
5. Axent Technologies Inc., "Everything You Need to Know About Network Security," pp. 1–33, 1998.

Potential §102 mapping: None anticipates any claim. They are general-prior-art/background references establishing that packet filtering, application proxies, and NAT were well-known before 2000 (relevant only to the level of ordinary skill and to §103 motivation-to-combine arguments). I could not verify the exact content of the four web white papers from this session; treat their descriptions as generic.

---

Overall assessment

1. No cited reference is a clean §102 anticipation of independent claims 1, 12, or 17. The novel core of the '561 patent — "the firewall applying the IP packets associated with the signaling channel and the control channel to the application proxy, and the IP packets associated with the bearer channel to the packet filter" — is not disclosed as a channel-split architecture in any single cited reference.

2. The most relevant prior art is US 6,400,707 (Baum, "Real time firewall security"), which independently splits real-time voice handling into a signaling-analysis path and a bearer-filtering path with dynamically reconfigured per-conversation filters. It is the best candidate for (a) a §102(e) attack on the dynamic enable/disable limitation of claim 6 and (b) the lead §103 reference combined with a proxy reference.

3. The proxy art cluster (US 5,826,014 / US 6,061,798 / US 6,728,885) supplies the application-proxy element; the filter/NAT cluster (US 5,606,668, US 5,968,176, US 5,793,763, US 6,006,272, US 6,650,641) supplies packet filtering, multilayer policy, and NAT. A §103 combination of Baum + a proxy reference is the most plausible invalidity theory.

4. No cited reference mentions the "application proxy instructs the packet filter / synchronizes NAT for the full session independent of traffic flow" features (claims 6, 8, 9, 16, 21) except Baum's dynamic per-conversation filter reconfiguration (claim 6 analogue) — the layer-7 NAT synchronization with the proxy appears to be the strongest distinguishing feature over all cited art.

Confidence/limitations: Full text was retrieved and verified for US 6,400,707, US 6,061,798, US 5,826,014, US 5,968,176, and US 6,006,272, and for US 5,606,668 via secondary summaries. For US 6,119,171, US 6,728,885, and US 6,650,641 I could not retrieve full text within this session; their descriptions above are based on title/assignee/date and should be confirmed against the actual documents before being relied on for a formal invalidity opinion. I have not fabricated any disclosure I could not support, and I have not characterized any reference as anticipatory where only obviousness-type support exists.

I'll ground the analysis by checking the disclosures of the key cited references, since the Prior Art section of the patent page gives me titles/assignees but not full text.

Obviousness Analysis — U.S. Patent No. 7,047,561 under 35 U.S.C. § 103

Analyst note on scope and sources. This section builds on, and does not repeat, the earlier Patent Summary and Litigation Summary. The prior art analyzed below is taken from the Prior Art section of the patent page (the 9–10 examiner‑cited patents, the 5 non‑patent citations, and — where a pre‑2000‑09‑28 priority date makes them availabe — the earlier‑dated "Cited By" documents). Because the patent page supplied only titles/assignees for several references, I retrieved the disclosures of the most load‑bearing ones (US 5,968,176; US 6,400,707; US 6,728,885; US 6,006,272; and the GB 2 365 256 "similar document") and cite them specifically. Where I could not retrieve a full disclosure (e.g., US 5,793,763; US 5,826,014; US 6,110,171; US 6,650,641), I say so and reason only from the title/assignee as the page presents them. No claim text or identifier has been auto‑corrected; the literal artifacts noted in the earlier section (e.g., claim 11 "ane", claim 15 "EP", claim 17 "foxed", claim 19 "flier") stand.

---

1. Governing law and the person of ordinary skill

Applicable statute. The application was filed 2000‑09‑28, before the AIA first‑inventor‑to‑file date (2013‑03‑16). Pre‑AIA § 103(a) therefore controls, and the cited documents qualify as prior art under pre‑AIA § 102(e) (US patents granted on applications filed before 2000‑09‑28) and/or § 102(b) (publications more than one year before the critical date).

Standard. Graham v. John Deere Co., 383 U.S. 1 (1966) (scope/content of claims; differences over prior art; PHOSITA level; secondary considerations); KSR Int'l Co. v. Teleflex Inc., 550 U.S. 398 (2007) (a claimed combination is obvious where the elements were known, the combination does no more than yield predictable results, and the PHOSITA had "a reason to combine… with a reasonable expectation of success"; common sense and design incentives may supply the motivation, and a mere "teaching away" must be evaluated on the whole record).

Level of ordinary skill. A bachelor's degree in EE/CS (or equivalent) with ~2–5 years of experience in network security/firewalls, plus working familiarity with packet‑telephony protocols (H.323, SIP, MGCP, T.38) as of mid‑2000. This is the level the '561 specification itself presumes ("Persons skilled in the art will appreciate…").

Claim construction of the key limitations. The three independent claims (1, 12, 17) reduce to one concept: a single firewall that routes packets by logical channel type of a real‑time Internet application — signaling + control channels → application proxy; bearer channel → packet filter. Dependent claims add (i) application type (2–5); (ii) proxy‑driven enable/disable of bearer‑channel filter entries per session (6); (iii) NAT at L3/L4 across all channels (7) and at L7 for embedded addresses in signaling/control (8); (iv) NAT persistence for the whole session independent of traffic flow (9); (v) control logic (10); and (vi) housing options (11). Under Phillips, "signaling channel," "control channel," and "bearer channel" should be read against the specification's own definitions (col. describing FIG. 3: bearer = UDP/RTP stream; signaling/control = TCP stream).

---

2. The relevant prior‑art landscape

---

3. Combination 1 — the primary § 103 case

(a) Combination: US 6,728,885 + US 6,400,707, optionally with US 5,968,176

This is the strongest combination because the two references together supply every element of the independent claims as a matter of known, predictable design.

Claim 1 (and 12/17). US 6,728,885 already discloses a firewall containing both an application proxy (211) and a packet filter (DPF 207 / filters 209, 215), with a routing component that sends some packets to the proxy and others to the filter (Fig. 2; the DPF forwards connection‑establishing packets up the stack to the proxy and forwards the rest at the packet level). That is the literal architecture of claim 1. US 5,968,176 independently confirms that firewalls are known to perform "packet filtering at the network layer, the transport layer and the application layer, and application proxying" in one coordinated system.

What is missing from US 6,728,885 alone is the predicate for the split: routing by real‑time‑application channel type. That predicate comes from US 6,400,707, which is directed to the same problem (securing packetized voice / real‑time traffic without latency) and expressly separates the "call set‑up signaling" from the "conversation stream of packetized voice." US 6,400,707 teaches that the signaling is processed by an intelligent control element while the voice bearer stream is passed through a fast filter. Recasting that same partition as "signaling/control → proxy; bearer → filter" is a predictable application of a known two‑mechanism firewall (US 6,728,885) to the known channel structure of a real‑time protocol (US 6,400,707 + H.323v2 / SIP as cited in US 6,006,272).

Motivation / reasonable expectation of success. The '561 specification's own background supplies the design incentive and is an admission against interest: it concedes that an "H.323 application proxy… is very processor intensive… and such a firewall cannot handle many VoIP calls simultaneously," and that "[t]he use of a packet filter, while much faster than an application proxy, is very insecure." US 6,728,885's own summary makes the identical point (a proxy "requires large amounts of processing power and a corresponding loss of performance"; a packet filter "allows data to travel directly… creating potential security risks"). US 6,400,707 adds the real‑time constraint (latency). A PHOSITA confronting the admitted trade‑off — proxy is secure but slow, filter is fast but weak — has an explicit reason to put each mechanism where it does the most good: deep inspection on the low‑volume, security‑sensitive signaling/control channel; fast filtering on the high‑volume, delay‑sensitive bearer channel. KSR holds that where a technique is used to improve one element and the combination is "obvious to try" with predictable results, § 103 is met. No reference needs to be physically modified to defeat the claim — the '561 is precisely the arrangement of known elements each performing its known function.

(b) Claim 6 — proxy instructs the filter which bearer packets to enable/disable for the session

Claim 6 is, if anything, more clearly obvious than claim 1:

• US 6,400,707: the control processor derives filter parameters from call set‑up signaling and reconfigures the filter, and the filter "disappears on termination of the conversation." That is "instruct the packet filter as to which… bearer channel to enable and disable for the duration of a session."
• US 6,728,885: the proxy creates the connection rule, and "once the connection terminates, the connection rule is removed and the proxy is notified." Same teaching, in a proxy+filter firewall.

Two independent references teach session‑scoped, signaling‑driven, automatically torn‑down filter entries. Combining them yields claim 6 as a predictable result.

(c) Claims 7–9 — the NAT limitations

• Claim 7 (NAT at L3/L4 across all channels). US 6,728,885 places NAT module 205 inside the same firewall as the proxy and filter; US 5,793,763, US 6,006,272 and US 6,650,641 each disclose NAT for IP flows. Applying NAT to the L3 addresses and L4 ports of the packets transiting the firewall is the ordinary function of a NAT module.
• Claim 8 (L7 NAT for embedded addresses in signaling/control). This is the one limitation the examiner might regard as the closest to novel, but the motivation is squarely in the art: US 6,006,272 expressly recognizes that H.323 carries address information in the payload ("it can be included in the payload in accordance with H.323"), and it cites the H.323v2 draft. Because the claimed firewall already parses the signaling/control payload in an application proxy (US 6,728,885's proxy "performs the necessary application‑level filtering"), it would have been obvious to rewrite the embedded addresses with the same translation applied to the headers — the proxy is the only component with the knowledge and the opportunity to do so, and the problem (header/payload mismatch breaking H.323 through NAT) was documented. This is a classic KSR "known problem, known solution in the same field" situation.
• Claim 9 (NAT persists for the session independent of traffic flow). The NAT‑timeout problem is admitted in the '561 background and is general knowledge (NPL "NAT and Networks"). Given a proxy that knows when a session starts and ends (US 6,728,885's register/notify/connection‑rule‑removed cycle; US 6,400,707's session‑scoped filters), pinning the NAT binding for the session's duration — i.e., tying the NAT entry's lifetime to the connection rule already maintained by the proxy+filter — is an obvious, predictable design choice with no new mechanism required.

(d) Claims 10–11 — control logic and housing

• Claim 10 (control logic specifying operating parameters). US 5,968,176's policy definition component / configuration interface and US 5,606,668's GUI‑driven rule specification (as summarized in US 6,400,707) disclose exactly this.
• Claim 11 (dual‑homed workstation / general‑purpose workstation / appliance / ASIC). Dual‑homing two NICs into two networks is the textbook firewall deployment; US 6,728,885 and US 5,826,014 are host‑based firewalls. The '561 specification itself concedes the hardware is conventional ("a Sun Solaris or HP or high end Windows machine"). No patentable weight.

(e) Claims 2–5 — application type

Claims 2–5 (VoIP, fax‑over‑Internet, video‑over‑Internet, voice messaging) are field‑of‑use limitations on the same firewall. US 6,400,707 itself covers audio and video service; the record (US 6,006,272) shows fax/voice/video over IP were recognized real‑time protocols (H.323, T.38). Selecting among known real‑time applications for the same architecture is not an inventive contribution.

(f) Claims 12–16 and 17–21 — method and CRM mirrors

These recite the same steps as claim 1 plus the same dependent subject matter, and add only the computer‑readable‑medium vehicle, which is a routine implementation choice (a programmed general‑purpose computer). Their obviousness tracks claims 1–11; no separate inventive weight.

---

4. Combination 2 — an "all‑of‑record" alternative

For completeness, a second, more conservative route exists using only the examiner‑cited art:

US 5,968,176 + US 6,400,707 + US 6,006,272 / US 5,793,763

1. US 5,968,176: a firewall combining packet filtering and application proxying at multiple layers → claim 1's two mechanisms.
2. US 6,400,707: the real‑time‑telephony firewall with signaling‑driven, session‑scoped filter reconfiguration → claims 1 (real‑time channel context + filter for the voice stream) and 6.
3. US 6,006,272 (+ US 5,793,763): NAT, including the H.323 payload issue → claims 7–8.

This route is weaker than Combination 1 on the "proxy for signaling" allocation (US 5,968,176 splits by device/layer, not channel), but it is a viable fallback.

---

5. Anticipated patentee rebuttals and the honest counter‑analysis

A § 103 conclusion here is contested, not slam‑dunk. Three points deserve weight:

1. Teaching away in US 6,400,707. Bell Atlantic's patent expressly says "[a]pplication proxies would not be capable of handling the number of calls expected in this application except at a prohibitive cost" and that proxy latency "increases exponentially." A PHOSITA reading US 6,400,707 for a real‑time voice firewall is arguably steered away from including an application proxy at all. The patentee would argue, and the earlier Litigation Summary corroborates (jury verdict of validity; summary judgment of § 101 eligibility), that the '561's insight — use the proxy only on signaling/control, never the bearer stream — is what dissolves the trade‑off the references accepted. Whether this "teaching away" is enough is a factual question (the doctrine requires the prior art to criticize, discredit, or discourage the claimed solution; here US 6,400,707 criticizes proxies for the bearer path, which is not where the '561 uses the proxy — a court could read the reference as actually supporting the channel split). I flag this as the single most important battleground.

2. The "channel‑type" predicate is not in US 5,968,176 or US 6,728,885. US 6,728,885 splits by connection‑establishing packet / registered port; US 5,968,176 splits by network device/layer. Neither says "signaling/control → proxy, bearer → filter." The patentee will argue the art lacks this motivation, and that the references are not about real‑time Internet protocols. The counter is that H.323/SIP channel taxonomy was itself well known by mid‑2000 (US 6,006,272's citation of the H.323v2 draft shows the field had the specification), so applying it to a known two‑mechanism firewall is a predictable design choice — but this is a motivation‑to‑combine argument that turns on the evidentiary record.

3. Objective indicia. The patentee may assert unexpected results, industry praise, or copying. I have no authoritative, verified evidence of genuine unexpected results or commercial success causally linked to the '561 claims; I will not manufacture it. The E.D. Tex. jury's invalidity finding is not a legal obviousness determination and does not tell us whether these references were asserted in a § 103 defense (the earlier summary identifies the '561 validity challenge as centering on § 101 eligibility).

Honest caveat. I have not retrieved the full disclosures of US 5,793,763, US 5,826,014, US 6,061,798, US 6,110,171, or US 6,650,641, and my analysis of those is limited to title/assignee. My claim‑by‑claim conclusions rest primarily on US 5,968,176, US 6,728,885, US 6,400,707 and US 6,006,272, whose texts I verified. If a full § 103 challenge is being prepared, the disclosures of those four references should be charted claim‑element‑by‑claim‑element against each limitation (particularly the L7‑NAT limitation of claim 8).

---

6. Bottom line

Overall: A prima facie § 103 case can be made for all 21 claims, with US 6,728,885 (+ US 6,400,707 and US 5,968,176) as the core combination and US 6,006,272 / US 5,793,763 for the NAT limitations. The case is strongest on dependent claims 6–11 and most vulnerable on independent claims 1/12/17, where the outcome turns on (i) whether the channel‑type split is deemed a predictable design choice versus (ii) the express proxy‑disfavoring language in US 6,400,707 constituting a teaching away.

Sources (verified disclosures): US 5,968,176 — https://patents.google.com/patent/[US5968176A](/patent/US5968176A) and https://uspto.report/patent/grant/[5968176](/patent/5968176) · US 6,728,885 — https://patents.google.com/patent/US6728885 · US 6,400,707 — https://patents.google.com/patent/US6400707 · US 6,006,272 — https://patents.google.com/patent/US6006272 · GB 2 365 256 candidate — https://patents.google.com/patent/DE60113435T2/en