US 7330431

I'll search for current information on this patent, including any 2026 litigation activity.

Let me check the Federal Circuit appeal referenced in the patent record and any 2026 docket activity.

Let me verify the reexamination outcome and the status of the Federal Circuit appeal.

US Patent 7,330,431 — Analyst Summary

Scope note: All figures below are drawn from the authoritative full text of US 7,330,431 B2 provided to me, cross-checked against live search results. I did not substitute any other patent number; every identifier below refers to US 7,330,431 / application 10/933,572 only.

1. Bibliographic Data

Assignment chain (from Google Patents record): Orckit Corrigent Ltd. → Corrigent Systems Ltd. → Orckit-Corrigent Ltd. (change of name) → Orckit IP, LLC (2016-05-10) → NAHUM COMMUNICATION N.T.B. LTD. (2022-04-14) → Corrigent Corporation (2022-04-14). A 2013 security agreement to Hudson Bay IP Opportunities Master Fund, LP was released later that same year.

Family: A Japanese counterpart, JP 2006074796 A (JP application 2005255608), priority 2005-09-02, titled "Multipoint-to-multipoint communication method and apparatus for ring topology."

2. Abstract (verbatim)

"A method for assigning bandwidth in a network including nodes coupled by links arranged in a physical topology, the method including: defining between the nodes logical connections associated with a data transmission service to be provided over the network, the logical connections having a connection topology different from the physical topology, and determining respective bandwidth requirements for the logical connections based on parameters of the service. The method further includes mapping the connection topology to the physical topology, so that each of the logical connections is associated with one or more links of the physical topology, and allocating a bandwidth for the service on each of the links in response to the bandwidth requirements of the logical connections and to the mapping."

3. Plain-Language Overview of the Independent Claims

Claim 1 — Method: bandwidth assignment by mapping a logical topology onto a physical topology

This is the base method claim. Four steps:

1. Define logical connections between the network's nodes for a data transmission service, where those logical connections form a connection topology that is different from the network's actual physical topology.
2. Determine bandwidth requirements for each logical connection from the service's parameters.
3. Map that logical connection topology onto the physical topology, so each logical connection is tied to one or more physical links.
4. Allocate bandwidth on each physical link based on (a) the logical connections' bandwidth requirements and (b) the mapping.

In short: take the "who talks to whom" picture of a service, superimpose it on the ring's real links, and size each physical link accordingly — rather than reserving worst-case bandwidth on every link.

Claim 12 — Method: two services with two different logical topologies, summed per link

Extends the claim 1 approach to two concurrent services, each with its own connection topology (both different from the physical topology). The method:

• Determines first and second bandwidth requirements from the respective services' parameters.
• Generates a first mapping and allocates a first bandwidth per link.
• Generates a second mapping and allocates a second bandwidth per link.
• Sums the first and second bandwidths to produce a total allocation for each physical link.

In short: run the claim 1 mapping once per service, then add the per-link results to get the combined reservation.

Claim 24 — Method: ring network, hub-and-spoke + full mesh, with selective allocation

The most specifically limited independent method claim. It requires:

• A ring network (nodes coupled by links).
• A first service whose logical connections have a hub-and-spokes topology and a second service whose logical connections have a full mesh topology.
• Determining first and second bandwidth requirements; generating first and second mappings onto the ring.
• Determining a first and a second per-link bandwidth.
• Summing them into a total per-link bandwidth.
• Allocating one of {first bandwidth, second bandwidth, total bandwidth} to each link depending on whether the first service, the second service, or both services are actually provided over the network.

In short: same machinery as claim 12, but locked to a ring and to the hub-and-spoke/full-mesh pairing, and it adds the "allocate only what you actually turned on" selection step.

Claim 25 — Apparatus: controller implementing the claim 1 method

A controller (software/hardware element) adapted to: receive a definition of logical connections for a service whose connection topology differs from the physical topology; determine per-logical-connection bandwidth requirements from service parameters; map the connection topology to the physical topology; and allocate per-link bandwidth based on those requirements and the mapping. Dependent claims 26 and 27 specify that the controller sits inside one of the nodes or is external to the network.

Claim 28 — Apparatus: controller implementing the claim 12 method

A controller adapted to perform the two-service, two-topology version (receive both connection definitions, determine both sets of bandwidth requirements, generate both mappings, allocate both per-link bandwidths, and sum them into a total per-link allocation). Dependent claims 29 and 30 again place the controller in a node or external to the network.

Dependent-claim coverage worth noting: guaranteed-bandwidth service (cl. 4, 15); class of service defined by the network's protocol (cl. 5, 16); multiplying by a correction factor / "deviation parameter" (cl. 6, 17); node-bandwidth parameters per node (cl. 8, 19); traffic monitoring and adjustment (cl. 9, 20); reserved bandwidth for a traffic subclass (cl. 10, 21); and comparing a "mapping bandwidth" against a "full bandwidth" computed by assuming all possible logical connections are provisioned (cl. 11, 23).

Worked examples in the spec: FIG. 3 uses a hub-and-spoke example with node A at 3R and B/C/D at R each, DP = 10%; FIG. 4 uses a balanced full mesh with each node at R, DP = 10%. In both, the mapping bandwidth for certain links falls below "full" bandwidth, so those links are provisioned at mapping bandwidth × 1.1 (e.g., 1.1R, 1.1/3 R, 2.2/3 R), while links where mapping equals full are provisioned at the full value.

4. Post-Grant Challenges and Litigation (search results)

This patent has an unusually active post-grant and litigation history. The Google Patents record carries a "Family has litigation" flag with these entries:

PTAB — inter partes review

• IPR2022-01514, Unified Patents, LLC v. Corrigent Corporation, filed Sept. 29, 2022 — Terminated Apr. 4, 2023 (petitioner's unopposed motion to dismiss; pre-institution dismissal / "Procedural Termination"). (ipverse case page)
• IPR2023-00169, Cisco Systems, Inc. et al. v. Corrigent Corporation, filed Nov. 17, 2022, instituted Jul. 6, 2023, Final Written Decision Jul. 3, 2024. Petitioners also included Dell. (ipverse case page)
• IPR2023-00515, Arista Networks, Inc. v. Corrigent Corporation, filed Jan. 24, 2023, instituted Aug. 4, 2023, Final Written Decision Jul. 3, 2024. (ipverse case page)

Petitioned grounds in IPR2023-00169 were §103 obviousness combinations over Baker (US 2004/0095946), Wright (US 2004/0174884 / US 7,292,542), Holender (US 6,104,699 / US 6,069,894 depending on filing), and Zheng (US 6,757,247), with a Sotera stipulation filed by Cisco and Dell in May 2023.

Ex parte reexamination

• Control No. 90/015,224 — Unified Patents reported that on March 5, 2024 the Central Reexamination Unit entered a final rejection of the challenged claims of the '431 patent. (Unified Patents news item)

District court

• Corrigent Corporation v. Cisco Systems, Inc., 6:22-cv-00396 (W.D. Tex.) — the '431 patent was among the asserted patents.
• Corrigent Corporation v. [Dell Technologies Inc. et al](/litigations/by-defendant/Dell%20Technologies%20Inc.%20et%20al)., 1:22-cv-00496 (D. Del.)
• Corrigent Corporation v. Arista Networks, Inc., 1:22-cv-00497 (D. Del.)

Federal Circuit (the 2026-relevant docket)

• Appeal No. 25-2036, Corrigent Corp v. Cisco Systems Inc, filed 08/21/2025, appeal from W.D. Tex. 6:22-cv-00396. As of the search results retrieved, the docket shows no decision and no outcome, with "Pending Judge Assignment" and an empty appeal-outcome field. (CAFC docket 25-2036 summary)

5. Explicit Uncertainty / Caveats

• Federal Circuit: I found no 2026 Federal Circuit decision on this patent. Docket 25-2036 is a 2025-filed appeal that appears still pending as of the retrieved data. I could not confirm that 25-2036 specifically concerns the '431 patent — the originating district case, 6:22-cv-00396, asserted multiple Corrigent patents (the record in that case references at least the '431, '369, '485, and '602 patents), so the appeal may implicate only some of them. Treat any patent-specific attribution of 25-2036 as unverified.
• IPR outcomes: I could confirm the filing, institution, and Final Written Decision dates for IPR2023-00169 and IPR2023-00515 (both FWDs dated Jul. 3, 2024), but not the substantive holdings (i.e., which claims, if any, were cancelled or upheld). Do not assume the claims were invalidated by the FWDs.
• Reexamination: The March 5, 2024 "final rejection" report comes from Unified Patents' own news page describing its ex parte reexamination request in control 90/015,224. I could not independently verify whether a reexamination certificate has issued, or the final claim status of the '431 patent.
• Assignee: Google Patents lists "Current Assignee: Corrigent Corp" but also flags that listed assignees "may be inaccurate"; the recorded chain (Orckit IP, LLC → NAHUM COMMUNICATION N.T.B. LTD. → Corrigent Corporation, both on 2022-04-14) is from assignment records and should be treated as the best available evidence rather than a legal conclusion.
• Legal status: The "Active / expires 2026-08-04" designation on Google Patents is expressly a non-legal assumption; given the reexamination and IPR activity, actual enforceability should be confirmed against the USPTO PatentCenter file and the most recent CRU/PTAB papers.

I'll research litigation involving US patent 7330431 using patent litigation databases and the sources you specified.

Now let me search for the specific district court and Federal Circuit cases referenced.

Litigation Involving U.S. Patent No. 7,330,431

Patent: U.S. 7,330,431 ("Multipoint to multipoint communication over ring topologies"), inventor Leon Bruckman, application 10/933,572, filed Sept. 3, 2004, granted Feb. 12, 2008.

Current patent owner/assignee of record: Corrigent Corporation (rights acquired via the 2022 chain: Orckit IP, LLC → Nahum Communication N.T.B. Ltd. → Corrigent Corporation). Original assignee: Orckit Corrigent Ltd. / Corrigent Systems Ltd. The patent remains active (adjusted expiration listed as Aug. 4, 2026).

Google Patents flags this patent family as having litigation, listing the district court, PTAB, and Federal Circuit matters below. Below I separate what is documented from what I could not fully verify.

A. District Court Litigations

1. Corrigent Corporation v. Cisco Systems, Inc.

• Court/Jurisdiction: U.S. District Court for the Western District of Texas, Waco Division
• Case No.: 6:22-cv-00396 (docketed 6:22-cv-00396-ADA)
• Filed: 2022 (complaint; the IPR petitions cite this as Corrigent's Cisco action)
• Plaintiff: Corrigent Corporation
• Defendant: Cisco Systems, Inc.
• Asserted patents: The '431 patent was pleaded alongside U.S. Patent Nos. 6,957,369; 7,113,485; 7,593,400; and 9,118,602.
• Status/Outcome: Proceeded to pre-trial motion practice in mid-2024. Available rulings (Omnibus Order, W.D. Tex., after June 24/July 3, 2024 pretrial conferences) include: Cisco's Rule 12(c) motion granted without prejudice (leave to amend); Cisco's summary-judgment motion of no written description of the '400 patent granted as to non-infringement; Cisco's § 101 motion resolved (denied as to '369 and '485, granted as to '400); Cisco's standing challenge denied; Corrigent's ownership/unclean-hands motion granted. Note: the '431 patent does not appear among the later dispositive-motion rulings for this case, indicating it may have been dismissed or dropped before trial — I could not definitively confirm the '431 patent's disposition from the available records, so treat that as an open item.

2. Corrigent Corporation v. [Dell Technologies Inc. et al](/litigations/by-defendant/Dell%20Technologies%20Inc.%20et%20al).

• Court/Jurisdiction: U.S. District Court for the District of Delaware (Wilmington)
• Case No.: 1:22-cv-00496 (1:22-cv-00496-RGA)
• Filed: April 19, 2022
• Judge: Richard G. Andrews
• Plaintiff: Corrigent Corporation
• Defendant: Dell Technologies Inc., et al.
• Asserted patents: Report to the Commissioner lists 6,957,369; 7,330,431; 7,593,400; 9,118,602
• Status: Claim construction order issued May 29, 2024 (D.I. 103). Ongoing as of the last-documented filings; the IPR of the '431 patent was filed by Cisco/Dell in parallel.

3. Corrigent Corporation v. Arista Networks, Inc.

• Court/Jurisdiction: U.S. District Court for the District of Delaware (Wilmington)
• Case No.: 1:22-cv-00497 (1:22-cv-00497-RGA)
• Filed: April 19, 2022
• Judge: Richard G. Andrews (related case: 1:22-cv-00496-RGA)
• Plaintiff: Corrigent Corporation (counsel: Ashby & Geddes; Steptoe, incl. James R. Nuttall)
• Defendant: Arista Networks, Inc. (counsel: Morris, Nichols, Arsht & Tunnell; Baker Botts)
• Asserted patents: 6,957,369; 7,330,431; 7,593,400; 9,118,602
• Status: Complaint filed April 19, 2022; docketed as an active patent-infringement action (35:271).

B. PTAB Inter Partes Review Proceedings

All three IPRs challenged U.S. 7,330,431 (Application 10/933,572, Tech Center 2600), with James Nuttall as Patent Owner counsel.

A consolidation order appears in the Cisco proceeding papers, and the parallel Cisco and Arista IPRs were coordinated. I could not confirm from the retrieved records the substantive holding of the July 3, 2024 Final Written Decisions (i.e., whether claims 25–27 were upheld, cancelled, or partly cancelled). That should be verified directly in the FWD documents if it matters for your analysis.

C. Court of Appeals for the Federal Circuit

• Court: U.S. Court of Appeals for the Federal Circuit
• Case No.: 25-2036
• Filed: 2025 (per Google Patents litigation data)
• Status: Listed as an active appeal. The record available to me does not identify the appellant/appellee or the lower proceeding appealed with certainty; given the timing, it is likely tied to the '431 patent district-court or IPR outcome, but I cannot confirm the linkage. Mark as unverified.

Summary / Caveats

• Known litigation exists — this is not a "no known litigation" case. There are at least three district court actions (Cisco in W.D. Tex.; Dell and Arista in D. Del.), three PTAB IPRs (Unified Patents, Cisco et al., Arista), and one Federal Circuit appeal (25-2036).
• The '431 patent was one of several patents asserted by Corrigent in each of these campaigns, which focused on Corrigent's Ethernet-switching/router portfolio.
• Items I could not fully verify and would flag as requiring primary-source confirmation: (1) the exact disposition of the '431 patent in the W.D. Tex. Cisco case; (2) the substantive outcome of the two July 3, 2024 IPR Final Written Decisions; and (3) the parties/issues on appeal in CAFC No. 25-2036.
• Sources: Google Patents patent page (https://patents.google.com/patent/[US7330431](/patent/US7330431)/en); GreyB/IPverse PTAB dockets (https://ipverse.greyb.com/ptab-web/cases/case-details/IPR2023-00169, /IPR2023-00515, /IPR2022-01514); USPTO PTAB petition filings; CourtListener dockets for D. Del. 1:22-cv-00496 and W.D. Tex. 6:22-cv-00396.

If you need the IPR final holdings and the CAFC appeal particulars pinned down, those require pulling the actual FWD PDFs and the Federal Circuit docket for 25-2036 directly.

I'll research the PTAB proceedings flagged in the patent record before drafting.

Now let me pull the claim-level outcomes of the two Final Written Decisions.

Let me get the IPR2023-00169 FWD outcome and the IPR2022-01514 termination details.

PTAB / AIA Trial Proceedings — US 7,330,431 ("Multipoint to multipoint communication over ring topologies")

Source-data discrepancy — flagged before anything else.
The structured "PTAB proceedings on file" block in this prompt (the ODP-derived field) states: "The USPTO ODP API returns no AIA trial proceedings for this patent as of the most recent ingest." That field is incomplete. The patent's own public record at Google Patents carries the litigation/PTAB data layer with three proceeding links for this patent, and web search independently corroborates all three. I am reporting the three documented proceedings rather than the null ODP field, because telling a defendant "no PTAB activity" here would be materially wrong. Every proceeding number below appears verbatim in the patent record or in a PTAB/party filing — none are invented.

Proceedings overview

Three AIA trial proceedings are on file against US 7,330,431: 0 active, 1 terminated before institution (procedural termination on an unopposed motion to dismiss — functionally a walk-away by the petitioner), and 2 taken to Final Written Decision (both 2024-07-03, in consolidated proceedings, by the same three-APJ panel). Of the two FWDs, one is confirmed at claim level: in IPR2023-00515, the Board held claims 1-10, 12-22, and 24-30 unpatentable and sustained only claims 11 and 23.

Bottom-line defensive posture: this patent is effectively gutted. All five independent claims — 1, 12, 24, 25, and 28 — were canceled in IPR2023-00515, taking with them every broadly-worded claim a demand letter would plausibly cite. Any assertion today must rest on dependent claims 11 and/or 23 (which add subclass/reserved-bandwidth limitations), and the claim-level disposition of the companion FWD in IPR2023-00169 — the proceeding in which Cisco/Dell challenged those very two claims on different art — is the one fact a defendant must verify before responding to a demand. Two further headwinds for the patent owner: an ex parte reexamination with a final rejection of the challenged claims (2024-03-05), and an adjusted expiration of 2026-08-04, which if accurate means the term has already run as of today.

IPR2023-00515 — Arista Networks, Inc. v. Corrigent Corporation

• Type: Inter Partes Review
• Filed: 2023-01-24
• Status: Final Written Decision (caption: "JUDGMENT Final Written Decision Determining Some Challenged Claims Unpatentable 35 U.S.C. § 318(a)")
• Judge panel: Nathan A. Engels, Christopher L. Ogden (opinion author), Brent M. Dougal. Petitioner counsel: Eliot Williams et al.; Patent Owner counsel: James R. Nuttall et al. (Steptoe & Johnson LLP).
• Petition grounds: § 103 obviousness of claims 1-30, built on Baker (US 2004/0095946) in view of Wright (US 2004/0174884) and Holender '699 (US 6,104,699), with Holender '894 (US 6,069,894) added for the dependent-claim grounds (Ex. 1020). Supported by the Declaration of Tal Lavian, Ph.D. (Ex. 1021). Claim 1's limitations were addressed serially (limitations 1.1–1.5 in the institution decision).
• Institution decision: Instituted 2023-08-04 on all challenged claims. The panel found Arista had shown at the threshold stage that "Baker, in view of Wright and Holender '699, teaches limitation 1.3," that "Baker discloses limitation 1.4," and that "Baker, in view of Wright, teaches limitation 1.5" — rejecting Corrigent's argument that Baker's mapping "simply maps nodes to other nodes" and does not associate logical connections with specific physical links.
• Final Written Decision (2024-07-03, Paper 10): verdict at claim level — claims 1-10, 12-22, and 24-30 unpatentable; claims 11 and 23 not shown unpatentable. Verbatim: "For the reasons below, we conclude that Petitioner has shown that claims 1-10, 12-22, and 24-30 are unpatentable, but has not shown that claims 11 or 23 are unpatentable." This canceled all five independent claims (1, 12, 24, 25, 28). The panel noted the case was consolidated with IPR2023-00169, and that because the consolidated cases had "overlapping, but different, grounds and evidence," this FWD "is a final written decision ... solely in the IPR2023-00515 proceeding." The Board declined to adopt either of Corrigent's proposed constructions of "in response to," holding the challenged claims obvious "whether 'in response to' is interpreted under its plain meaning or either of PO's proposed constructions." Oral hearing held 2024-04-10.
• Settlement / termination: none — case went to FWD.
• Appeal: not confirmed for this proceeding from the sources retrieved (see the CAFC note under IPR2023-00169).
• Defensive value: This is the bellwether. All independent claims are dead, so an infringement theory built on claim 1, 12, 24, 25, or 28 is untenable — the PTO has already canceled them by final written decision. Your only remaining exposure is claims 11 and 23.

IPR2023-00169 — Cisco Systems, Inc., Dell Technologies Inc., and Dell Inc. v. Corrigent Corporation

• Type: Inter Partes Review
• Filed: 2022-11-17
• Status: Final Written Decision
• Judge panel: Nathan A. Engels, Christopher L. Ogden, Brent M. Dougal (same panel as IPR2023-00515; trial consolidated). Petitioner counsel: Theodore Foster et al.; Patent Owner counsel: James R. Nuttall et al.
• Petition grounds (as recited verbatim in Cisco's Sotera stipulation filed in the W.D. Tex. case, 2023-05-17): § 103 — (1) claims 1-4, 7-9, 11, 25, 26 over Baker + Wright; (2) claims 5, 10, 12-16, 18-24, 27-30 over Baker + Wright + Holender '699; (3) claim 6 over Baker + Wright + Zheng (US 6,757,247); (4) claim 17 over Baker + Wright + Holender '699 + Zheng. That set covers all 30 claims. Corrigent's Rule 42.70 request in the consolidated case lists eight grounds for argument, adding grounds that incorporate Holender '894 (Ex. 1020) against claims 12-16, 18-24, 27-30 and claim 17. (The exact set instituted on 2023-07-06 was not verified from the primary decision document in the searches performed.)
• Institution decision: Instituted 2023-07-06 (Decision Granting Institution, per the case record). Petitioners filed Sotera stipulations (Cisco in 6:22-cv-00396 (W.D. Tex.); Dell in 1:22-cv-00496 (D. Del.)) under which they would not pursue in the co-pending district courts the specific grounds asserted in the petition if the Board instituted.
• Final Written Decision (2024-07-03): issued. ⚠️ Claim-level disposition not verified. I could confirm the FWD's existence, date, panel, and that it issued the same day as the IPR2023-00515 FWD and flowed from the same 2024-04-10 consolidated oral hearing — but the retrieved sources did not reproduce the '169 FWD's claim-by-claim verdict. Do not assume claims 11 and 23 survived this proceeding: those are precisely the two claims the Cisco/Dell grounds targeted on different art (Holender '699, Holender '894, Zheng) than Arista's. Verify Paper at PTAB E2E before relying on them.
• Settlement / termination: none — case went to FWD.
• Appeal: A Court of Appeals for the Federal Circuit filing is listed for this patent's family at docket No. 25-2036. The appellant, issues, and disposition are not established by the sources retrieved. Note the timing: a notice of appeal from the 2024-07-03 FWDs would ordinarily have been due ~2024-09-04, so a 2025-docketed appeal may arise from a different underlying decision — treat this as an open item, not a conclusion.
• Defensive value: Cisco and Dell's participation gives three more petitioners (with Arista) subject to § 315(e)(2) estoppel once FWDs issued. Combined with IPR2023-00515, the practical question for a defendant is narrow: does claim 11 and/or claim 23 still stand?

IPR2022-01514 — Unified Patents, LLC v. Corrigent Corporation

• Type: Inter Partes Review
• Filed: 2022-09-29
• Status: Procedural Termination (case record: Termination Decision 2023-04-04; Notice 2023-04-21)
• Judge panel: not public in the sources retrieved (no institution decision issued).
• Petition grounds: § 103 against claims 25-27 (apparatus claims). Ground 1: claims 25 and 27 obvious over Holender (US 6,069,894) in view of Wright (US 7,292,542) and the general knowledge of a POSITA. Ground 2: claims 25 and 26 obvious over Baker (US 2004/0095946) in view of Wright, Lemieux (WO 02/19616), and general knowledge. Supported by the Declaration of James T. Geier (Ex. 1003). Unified's petition also attached the three parallel district-court complaints (Cisco, Dell, Arista) as exhibits.
• Institution decision: none — the proceeding was terminated before institution. Sequence: Patent Owner Preliminary Response 2023-01-25 → Unified's Unopposed Motion to Dismiss and Terminate 2023-03-24 → Termination Decision 2023-04-04 → Unified's Request for Refund of Post-Institution Fees 2023-04-17. Because no FWD issued, no § 315(e)(2) estoppel attached to Unified from this petition.
• Final Written Decision: none issued.
• Settlement / termination: terminated by unopposed motion — i.e., the patent owner did not oppose dismissal and no FWD ever issued. Whether any agreement underlay the unopposed dismissal is not in the public record; if terms exist, they are not disclosed. Note the strategic pivot: Unified then pursued the patent through ex parte reexamination, which is not subject to Fintiv and generates no IPR estoppel.
• Appeal: none from this proceeding.
• Defensive value: low as a direct estoppel source, but the proceeding is a defensive aggregator signal — Unified (which certified it was the sole real party-in-interest, with a Jakel declaration per Worlds v. Bungie) entered and then exited in favor of a reexam, which has since produced a final rejection.

Strategic summary

Canceled vs. sustained vs. untested. In IPR2023-00515, the Board canceled claims 1-10, 12-22, and 24-30 — including every independent claim: 1, 12, 24, 25, and 28. Only dependent claims 11 and 23 were sustained under Arista's grounds. Claim 11 (depending from claim 1) and claim 23 (depending from claim 12) add, at most, the "plurality of subclasses of traffic / reserved bandwidth for one of the subclasses" limitations (see claim 10/21 subject matter and the specification's subclass-reservation disclosure). Untested/unknown is the key gap: IPR2023-00169's FWD addressed the same claim set on Cisco/Dell's alternative art, and its claim-level verdict is unverified here. If claims 11 and 23 also fell in IPR2023-00169, the patent has no surviving claims at all and a demand letter citing any claim is baseless.

Estoppel landscape. Three petitioners (Arista; Cisco/Dell) are now subject to § 315(e)(2) estoppel on any ground they raised or reasonably could have raised in the IPRs — Baker/Wright/Holender '699/Holender '894/Zheng and the RFC-based obviousness combinations are effectively off the table for them and their privies in the parallel litigations. Unified escaped estoppel because IPR2022-01514 died pre-institution; but Unified's reexamination (90/015,224) puts the art it relied on (Holender, Wright, Baker, Lemieux) in front of the examiner, and the CRU entered a final rejection of the challenged claims on 2024-03-05. For a new defendant not in privity with any petitioner, the prior-art field remains open — the named references above are public, and a second, differently-configured combination has not been litigated to FWD.

Pattern signals. Unified Patents acted as a pure defensive aggregator (RPI certification, Jakel declaration, no district-court involvement) and then converted to ex parte reexam. The three operating-company defendants, sued in W.D. Tex. (6:22-cv-00396) and D. Del. (1:22-cv-00496, 1:22-cv-00497), filed their own IPRs with Sotera stipulations to neutralize Fintiv. Corrigent is the asserting entity, having taken assignment on 2022-04-14 via Orckit IP, LLC → Nahum Communication N.T.B. Ltd. → Corrigent Corporation — five days before the complaints were filed. The same petitioners also took down the sibling US 7,113,485 in IPR2023-00168 (FWD 2024-07-03, "Determining All Challenged Claims Unpatentable"), showing a coordinated family-wide attack.

Recommended next steps

1. Verify the IPR2023-00169 FWD claim-by-claim before anything else. Pull the 2024-07-03 FWD from PTAB E2E (https://ptacts.uspto.gov/ptacts) or the Docket Alarm case page (https://www.docketalarm.com/cases/PTAB/IPR2023-00169/Cisco_Systems_Inc._v._Corrigent_Corporation/). This single document determines whether any claim of the '431 patent survives.
2. Pin down the '431 FWD source of record. The IPR2023-00515 FWD is available at https://www.docketalarm.com/cases/PTAB/IPR2023-00515/Arista_Networks_Inc._v._Corrigent_Corporation/docs/07-03-2024-Board/Final_Written_Decision__original-10-Final_Written_Decision__original.pdf — cite it and quote the disposition: "claims 1-10, 12-22, and 24-30 are unpatentable, but has not shown that claims 11 or 23 are unpatentable."
3. Track the CAFC appeal — CAFC No. 25-2036 (family litigation record: https://portal.unifiedpatents.com/litigation/Court%20of%20Appeals%20for%20the%20Federal%20Circuit/case/25-2036). Confirm the appellant, whether the notice of appeal targets the 2024-07-03 FWDs or the reexam, and check CourtListener for the Rule 30(b) docketing. Any reversal on claims 11/23 would reopen a narrow theory against you.
4. Confirm the term. The record lists an adjusted expiration of 2026-08-04 (Google Patents legal-status event dated 2026-08-04; the patent's base 20-year term from 2004-09-03 ran in 2024). If accurate, the patent expired roughly two months ago — no prospective injunction, no post-expiry damages, and past damages capped by the § 286 six-year lookback. Verify the PTA calculation in USPTO PatentCenter before relying on this.
5. Check the reexamination certificate. Ex parte reexam 90/015,224 terminated in a final rejection on 2024-03-05 (https://portal.unifiedpatents.com/exparte/90015224). Confirm which claims were rejected, whether Corrigent appealed to the PTAB, and whether a reexam certificate has issued — a certificate canceling claims 11/23 would close the file entirely.
6. If a demand letter arrives: respond that all five independent claims are canceled by FWD, that only dependent claims 11 and 23 are (so far) sustained, that the patent appears expired, and that the examiner has finally rejected the challenged claims in reexam. Demand that the patent owner identify which surviving claim, if any, the letter relies on — and note the § 315(e)(2) estoppel wall binding Arista, Cisco, and Dell.

No active PTAB trial is pending (subject to the CAFC item in step 3). Statutory milestones are moot: both FWDs issued within the one-year § 316(a)(11) window measured from institution dates of 2023-07-06 and 2023-08-04.

I'll ground this with searches before writing anything, since the record must be evidence-based.

Good context. Now let me find the specific reel/frame and correspondent data for US 7330431.

Let me pin down the reel/frame records and the IPR/litigation specifics.

US 7,330,431 — Assignment & Ownership Reconstruction

Patent: US 7,330,431 B2 — "Multipoint to multipoint communication over ring topologies"
App: 10/933,572 · filed 2004-09-03 · issued 2008-02-12 · priority 2004-09-03
Assignee of record (Google Patents): Original — Orckit Corrigent Ltd; Current — Corrigent Corp

Sourcing note up front. Google Patents' legal-events block for this patent surfaces the parties, conveyance types and event dates but not reel/frame numbers or correspondence data. I recovered reel/frame and correspondent strings from USPTO PTAB filings that reproduce the chain-of-title tables for other patents recorded in the same Orckit recordings (e.g., US 6,680,904; the 2016 13-property assignment). I have not independently verified that each reel/frame cited below names US 7,330,431 specifically, and I flag every such instance. I did not fabricate any reel/frame. Where the record is silent, I say so.

Inventors

• The application was filed by Orckit Corrigent Ltd., and the inventor's rights were assigned to Corrigent Systems Ltd. on 2004-11-17 — i.e., a standard employee-inventor assignment into the operating company, not a third-party acquisition. Bruckman is a recurring Corrigent inventor (also named on related Corrigent filings such as EP2057796, alongside Ronen Solomon, Rafi Ram, David Zelig and Shachar Katz).
• No unusual inventor pattern is determinable from the record. I found no evidence of inventor departures, and I will not infer one. Note only that the '431 patent is a single-inventor filing, which makes it a clean, self-contained assertion vehicle — relevant context for the transfers below, but not itself a finding.

Original assignee

Corrigent Systems Ltd. (renamed Orckit-Corrigent Ltd. effective 2014-04-01 by change of name), Tel Aviv, Israel.

• Line of business: packet-transport networking for carrier/service-provider backhaul — Ethernet switching and routing, RPR/SONET/SDH transport, VPLS and residential/business/mobile backhaul. The '431 patent's subject matter (Connection Admission Controller bandwidth allocation across physical ring links for multipoint services) sits squarely in that product line.
• Product embodying the claims: Corrigent's CM-4000 family of packet transport switches (CM-4140, CM-4206, CM-4314/4314T) and the CM-401x series. Corrigent represented in its 2022 Cisco complaint that this portfolio was "practiced by its products," citing a KDDI Japan deployment of 2,000+ Corrigent Ethernet switches as early as 2005. Caveat: that product-practice statement is a litigation position by the successor plaintiff, not independently verified here.
• Current status: Insolvent / dissolved. Following the dot-com bust the company accumulated debt and in 2012 entered Israeli debt-restructuring / insolvency proceedings. Its patents were sold out of those proceedings in 2016. Counsel for Cisco stated in a 2024 trial record that "I don't think Orckit-Corrigent exists." The 2013 Hudson Bay security agreement and its prompt release (below) are consistent with the restructuring.

Assignment timeline

Chronological. Recorded events as surfaced on Google Patents for US 7,330,431; reel/frame and correspondent fields are reproduced from PTAB chain-of-title exhibits for the same recordings and are marked with confidence.

2004-11-17 (executed) / recorded 2004-11-17 — Reel not captured on the '431 legal-events record

• Conveyance: Assignment of assignors' interest
• Assignor: Leon Bruckman
• Assignee: Corrigent Systems Ltd.
• Correspondent: not recovered
• Context: routine employee-inventor assignment into the operating company.

2013-03-18 (executed) / recorded 2013-03-18 — Reel not recovered

• Conveyance: Security Agreement
• Assignor: Orckit-Corrigent Ltd.
• Assignee: Hudson Bay IP Opportunities Master Fund, LP
• Correspondent: not recovered
• Context: securitization / financing — a lender taking the patent estate as collateral during the company's financial distress. Notably, the fund is named "IP Opportunities," i.e., a distressed-IP lender.

2013-07-26 (executed) / recorded 2013-07-26 — Reel not recovered

• Conveyance: Release by Secured Party
• Assignor: Hudson Bay IP Opportunities Master Fund LP
• Assignee: Orckit-Corrigent Ltd.
• Correspondent: not recovered
• Context: release of the security interest ~4 months after recording — collateral returned, ownership unencumbered ahead of the insolvency sale.

2014-04-01 (executed) / recorded 2014-04-01 — Reel not recovered

• Conveyance: Change of Name
• Assignor: Corrigent Systems Ltd.
• Assignee: Orckit-Corrigent Ltd.
• Correspondent: not recovered
• Context: internal reorg / name change only — same legal entity, no change in beneficial ownership.

2016-05-10 (recorded; executed 2016-04-25, effective date 2016-04-25) — Reel 038453/0840 and Reel 038507/0587 ⚠️ family-level, not verified for '431 specifically

• Conveyance: Assignment of assignors' interest
• Assignors: Orckit Communication Ltd. and Orckit-Corrigent Ltd. (reel 038453/0840); Corrigent Systems Ltd., Orckit Communication Ltd. and Orckit-Corrigent Ltd. (reel 038507/0587), each acting through Lior Dagan, Adv. (Trustee), 1 Azrielli Center, Tel Aviv
• Assignee: Orckit IP, LLC — 874 Walker Road, Suite C, Dover, Delaware 19904-2778; mailing address 831 Beacon Street #307, Newton, MA 02459
• Correspondent: not stated in the copy I recovered — the Correspondence Data field on the 2016 assignment submission was blank in the reproduced document. This is a gap; do not read it as an absence.
• Context: bankruptcy fire-sale / transfer-to-asserter. The agreement itself recites a Purchase Agreement dated 2015-11-30 between the insolvent assignors and Orckit IP, LLC, consummated through the Israeli insolvency proceedings.
• ⚠️ Coverage caveat: the 2016 transaction was recorded across at least two reel/frames covering different subsets of the portfolio. The Appendix A list I recovered from the 2016 assignment (13 properties) names 7,768,928 / 8,830,821 / 7,593,400 / 9,185,151 / 7,545,740 / 8,014,394 / 7,463,580 / 8,196,637 / 7,916,636 / 7,515,536 plus three applications — 7,330,431 is not on that list, so the '431 patent almost certainly moved under one of the sister recordings (e.g., 038507/0587, which adds Corrigent Systems Ltd. as assignor). Unconfirmed which reel/frame names '431.

2022-04-14 (event date on the '431 record) — Reel 060539/0804 ⚠️ reel cited from the 6,680,904 chain-of-title exhibit; not verified for '431

• Conveyance: Assignment
• Assignor: Orckit IP, LLC
• Assignee: Nahum Communication N.T.B. Ltd. — Ramat Gan, Israel
• Correspondent: not recovered for this entry
• Context: transfer to asserter conduit. Execution date 2021-12-31, recorded 2022-07-18 — i.e., recorded roughly three months after the Cisco complaint was filed, which is why the recorded date diverges from the 2022-04-14 event date on the '431 legal-events record.

2022-04-14 (event date on the '431 record) — Reel 060544/0799 ⚠️ reel cited from the 6,680,904 chain-of-title exhibit; not verified for '431

• Conveyance: Assignment of assignors' interest
• Assignor: Nahum Communication N.T.B. Ltd.
• Assignee: Orckit IP, LLC (listed at West Newbury, MA in this record)
• Correspondent: RICHARD A. BAKER, JR., 291 Main Street, West Newbury, MA 01985. ⚠️ Single confirmed appearance in the material I recovered — see Signal 3.
• Context: directionally inconsistent with the '431 event record. ⚠️ On the '431 Google Patents record the second 2022-04-14 hop runs Nahum → Corrigent Corporation, whereas reel 060544/0799 (as reproduced for US 6,680,904) runs Nahum → Orckit IP, LLC. I cannot reconcile these two directions from the sources available, and I am not going to guess which is correct for '431. Both readings point to the same conclusion (Nahum is a conduit and Corrigent Corporation is the terminus), but the intermediate leg is genuinely ambiguous.

2022-04-14 (event date on the '431 record) — Reel not recovered

• Conveyance: Assignment
• Assignor: Nahum Communication N.T.B. Ltd.
• Assignee: Corrigent Corporation (current assignee of record)
• Correspondent: not recovered
• Context: transfer-to-asserter / pre-litigation standing. The '431 patent is expressly listed as an asserted patent owned by Corrigent Corporation on the AO-120 form for the suit filed five days later (see Signal 5).

Litigation / administrative events tied to this patent (non-assignment, for context):

• 2022-04-19 — Corrigent Corporation v. Cisco Systems, Inc., W.D. Tex. 6:22-cv-00396 (Waco). The AO-120 notice lists asserted patents 6,957,369 · 7,113,485 · 7,330,431 · 7,593,400 · 9,118,602, all held by Corrigent Corporation. The '431 patent was mapped to Cisco's "Collaboration Platform" compatible products.
• 2022-04-14 — Corrigent Corp. v. Dell Technologies, Inc., D. Del. 1:22-cv-00496; a companion 1:22-cv-00497 also appears in the litigation data; E.D. Tex. 2:22-cv-00276 (Orckit Corporation v. Cisco) also appears.
• 2022-09-29 — IPR2022-01514, Unified Patents, LLC v. Corrigent Corporation, on US 7,330,431 / App. 10/933,572. Terminated 2023-04-04 (procedural termination). Respondent counsel: James Nuttall et al. (Corrigent's litigation counsel — not the assignment correspondent; do not conflate the two).
• IPR2023-00169 and IPR2023-00515 (Google Patents lists both as filed and reaching Final Written Decision in this family); IPR2023-00554 (pleaded against the '111 patent). I did not verify which of -00169/-00515 is directed to the '431 patent specifically.
• 2025 — appeal docketed at the Federal Circuit, 25-2036.
• 2026-08-04 — adjusted expiration; status active.

Timeline diagram

timeline
    title Ownership of US 7330431
    2004 : Filed by Orckit Corrigent Ltd
         : Inventor assigns to Corrigent Systems Ltd
    2008 : Patent issued
    2012 : Orckit Corrigent enters Israeli insolvency
    2013 : Security agreement to Hudson Bay IP
         : Hudson Bay releases security interest
    2014 : Corrigent Systems renamed Orckit Corrigent Ltd
    2016 : Insolvency sale to Orckit IP LLC
    2022 : Transfer to Nahum Communication
         : Transfer to Corrigent Corporation
         : Cisco suit filed in W D Texas
         : Unified Patents files IPR on 7330431

NPE / troll-pattern signals

1. Shell-entity transfer — PRESENT.
The 2016-05-10 recording (executed 2016-04-25; reels 038453/0840 and 038507/0587) moved the patent out of the operating group into Orckit IP, LLC, a Delaware LLC whose recorded address is 874 Walker Road, Suite C, Dover, Delaware — a suite-style service address, with a separate operating/postal address at 831 Beacon Street #307, Newton, MA. There is no evidence Orckit IP, LLC made or sold anything; its name carries the "IP" suffix and its function in the record is acquisition-plus-assertion (its CEO/owner, Binder, was described in the E.D. Tex. record as the "front man for Orckit IP"). Note the transfer to Orckit IP, LLC is not dated to this patent's record date though — see the coverage caveat above.

2. Known asserter in the chain — PRESENT.

• Corrigent Corporation (current assignee) is the plaintiff in 6:22-cv-00396, 1:22-cv-00496 and 1:22-cv-00497, and Unified Patents, LLC — a defensive aggregator — chose to file IPR2022-01514 against Corrigent on this very patent (filed 2022-09-29), plus IPR2023-00169/-00515 in the family. Unified does not generally expend IPR budget on operating-company competitors; it targets assertion entities.
• Orckit IP, LLC is separately documented in the E.D. Tex. trial record as partnering with an outside attorney (Pitcock) on a contingency — 15% of proceeds to file suits, 42% to Orckit IP, with an alleged side arrangement to principal Tamir — which is an assertion-funding structure, not a product business.
• ⚠️ I did not confirm a listing on a named RPX or Unified "high-frequency plaintiff" directory page; the evidence above is the IPR posture and the litigation/contingency record, which is strong but is not the same thing as a directory entry.

3. Repeat correspondent across the chain — UNCLEAR (thin).
The only correspondent string I recovered anywhere in this chain is RICHARD A. BAKER, JR., 291 Main Street, West Newbury, MA 01985, on reel 060544/0799. Two observations, kept separate from the finding:

• It is a single confirmed appearance, not the recurrence the signal requires, so per your own rule I am not scoring it as present.
• However, the assignee in that same record was listed at "West Newbury, MA," the correspondent's town — an address coincidence worth a targeted re-check.
• I could not recover the correspondent for the 2016 Orckit IP recording (field blank in the copy I obtained) nor for reels 060539/0804 or the final Corrigent Corporation leg. This is the single most valuable open item for your purposes, because a recurring Baker (or a recurring firm) across the 2016, 2021-12-31 and 2022-04-14 recordings would convert this to PRESENT.

4. Cascading transfers — PRESENT.
The chain runs through five distinct steps in ten years, and within 2022 specifically: Orckit IP, LLC → Nahum Communication N.T.B. Ltd. (executed 2021-12-31) → Corrigent Corporation, with two hops carrying the same 2022-04-14 event date on the '431 record. Nahum Communication N.T.B. Ltd. (Ramat Gan, Israel) appears only as an intermediate holder with no evident business of its own, i.e., a conduit. The 2013 security/release pair (03-18 → 07-26, ~4 months) is a second short-cycle hop. Flag the unresolved directionality of reel 060544/0799 (Signal 6/2022 timeline entry) rather than smoothing it over.

5. Pre-litigation transfer — PRESENT.
The final hop to Corrigent Corporation is dated 2022-04-14, and the Cisco complaint was filed 2022-04-19 — five days. This is about as clean an alignment of "clean standing record → assertion" as the data supports.

6. Bankruptcy fire-sale — PRESENT.
Orckit-Corrigent entered Israeli debt-restructuring/insolvency proceedings in 2012. The 2016 acquisition by Orckit IP, LLC was made under a Purchase Agreement dated 2015-11-30 and executed 2016-04-25 with Lior Dagan, Adv. (Trustee) signing for the assignors, recorded 2016-05-10. This is a court-supervised sale of an insolvent operating company's patent estate.

7. Privateering — NOT PRESENT.
This is not privateering. The transferor (Orckit-Corrigent) was insolvent and ceased operating; it did not hand patents to an NPE while continuing to compete against the defendants. The asset flow is a bankruptcy liquidation, not a covert proxy war — even though the E.D. Tex. record contains allegations that the buyer's principals had undisclosed involvement in the insolvency itself.

8. Defensive aggregator — NOT PRESENT.
The chain terminates at Corrigent Corporation, an active plaintiff, not at RPX, AST, LOT, Unified Patents or OIN. Unified Patents appears here only as IPR petitioner, which challenges the patent but does not neutralize ownership — the patent is still shown Active with an adjusted expiration of 2026-08-04.

Verdict

NPE — high confidence.

Four strong signals stack up on the face of the record: (i) the 2016-04-25/2016-05-10 insolvency sale out of Orckit-Corrigent into Orckit IP, LLC, a Delaware licensing vehicle at 874 Walker Road, Suite C, Dover (reels 038453/0840, 038507/0587); (ii) the 2022 cascade Orckit IP, LLC → Nahum Communication N.T.B. Ltd. → Corrigent Corporation (reels 060539/0804, 060544/0799), with Nahum as a pure conduit; (iii) the final hop landing five days before the 2022-04-19 W.D. Tex. complaint in which Corrigent Corporation is named as holder of the '431 patent on the AO-120 form; and (iv) Unified Patents, LLC targeting this specific patent in IPR2022-01514, an assertion-entity-directed filing. The transferor was insolvent and the current owner makes no products, so no operating-company-assertion or defensive reading is available.

Two items a prosecution analyst should close before filing this as final: (a) pull the correspondent of record on the 2016 Orckit IP recording and on reels 060539/0804 / the final Corrigent Corporation leg to test the recurring-attorney hypothesis (the only correspondent I have, Richard A. Baker, Jr., shares West Newbury, MA with the assignee on his one record); and (b) resolve the directional conflict on reel 060544/0799 (Nahum → Orckit IP per the 6,680,904 chain exhibit, versus Nahum → Corrigent Corporation per the '431 event record).

Verify at: USPTO Assignment Center — https://assignmentcenter.uspto.gov/ (legacy interface: https://assignment.uspto.gov/patent/index.html) — search patent number 7330431 or application 10/933,572. Patent text and legal events: https://patents.google.com/patent/US7330431/en. Litigation data: Unified Patents litigation portal — D. Del. 1:22-cv-00497; IPR2022-01514.

I'll research this patent and its cited prior art. Let me run several searches in parallel.

I have strong grounding on the patent and key references. Let me research a few more of the most §102-relevant citations.

Prior-Art Analysis — US 7,330,431 B2

Patent under analysis: US 7,330,431 B2, "Multipoint to multipoint communication over ring topologies," inventor Leon Bruckman, original assignee Orckit Corrigent Ltd. / Corrigent Systems Ltd., app. US 10/933,572, filed 2004-09-03, granted 2008-02-12. 30 claims (independent claims 1, 12, 24, 25, 28). Live source: https://patents.google.com/patent/US7330431/en (cited via USPTO PTAB petition documents at ptacts.uspto.gov, petition 1550567).

Important framing before the reference-by-reference table:

1. Statutory standard. Anticipation under 35 U.S.C. § 102 requires that a single reference disclose every element of a claim, arranged as in the claim. The 64 references below were cited by the examiner; collectively they were almost certainly relied on under § 103 (obviousness), not § 102. My § 102 column therefore states whether any single reference by itself appears to disclose all elements of a given claim — and in every case the answer is "no clean anticipation," for reasons given. Where a reference is strong on some but not all elements, I identify it as an obviousness (combination) reference, not an anticipatory one.

2. Sourcing. Publication/priority dates, assignees and titles are taken verbatim from the citation record of US 7,330,431 itself. Descriptions for references I retrieved full text or abstracts for are verified; for the remainder I describe based on title/assignee and flag where I am inferring. I did not obtain full text of all 64 references, so the § 102 calls are reasoned preliminary assessments, not a substitute for element-by-element charting.

3. What claim 1 actually requires (the key independent claim):

   • (a) defining logical connections between nodes for a service, the logical connections having a connection topology different from the physical topology;
   • (b) determining bandwidth requirements for the logical connections based on service parameters;
   • (c) mapping the connection topology to the physical topology so each logical connection is associated with one or more physical links; and
   • (d) allocating bandwidth on each link responsive to (b) and (c).

   Independent claims 12/24 add two different services (hub-and-spoke + full mesh) summed per link; claims 25/28 are the apparatus (controller) counterparts; 6, 11, 17, 23 add the correction-factor and mapping-vs-"full"-bandwidth comparison features.

Tier 1 — Closest art (ring bandwidth allocation / logical-to-physical design)

Tier 2 — Bandwidth / admission-control / resource-allocation art

Tier 3 — Ring-network topology, routing-direction and transport art

Tier 4 — General network optimization, QoS, MPLS/ATM, and miscellaneous (remainder of the 64)

These were cited but are further afield. None anticipates an independent claim; individually they map at most to isolated limitations.

Non-patent citations (14)

Also cited: RFC 3036 (LDP), RFC 2702 (TE over MPLS), RFC 2457 (DiffServ), RFC 2205 (RSVP), RFC 2892 (SRP, Tsiang — foundational to the ring/bandwidth background), OSPF (RFC 2328), Dziong et al. "A Framework for Bandwidth Management in ATM Networks — Aggregate Equivalent Estimation Approach" (IEEE/ACM ToN, Feb. 1997), RFC 2814 (RSVP-based admission control), RFC 1990 (PPP Multilink), IMA (Inverse Multiplexing over ATM), and two prosecution-related Office Actions from related Corrigent applications. These supply the § 103 "general knowledge" backdrop (RPR/SRP rings, CAC, effective bandwidth) rather than anticipatory disclosures.

Bottom line

• No cited reference anticipates any independent claim of US 7,330,431 under § 102. Each of claims 1, 12, 24, 25, and 28 requires the specific combination of (a) a logical connection topology different from the physical topology, (b) per-service bandwidth requirements, (c) mapping that logical topology onto physical links, and (d) link-by-link allocation — and no single reference discloses all four (least of all for a ring physical topology).
• The closest § 102-type disclosures:
  • US 5,854,903 A (Lucent) — best on the logical-network-over-physical-network mapping (elements a–c), but no ring, and capacity-sizing rather than the claim's allocation/comparison.
  • US 7,035,279 B2 (Corrigent) — best on ring link resource allocation with measured use (elements c–d), but a flow-allocation (not logical-topology-mapping) disclosure, and it is same-assignee/same-inventor art.
• Realistically, the examiner's citations support obviousness (§ 103) when combined — e.g., US 5,854,903 (logical→physical mapping) + US 7,035,279 (ring resource allocation) + US 6,625,155 (measured-traffic CAC adjustment) + US 6,339,488 (full-mesh) + US 6,826,158 (tree/ring).
• Litigation context (from the patent record): the patent has been the subject of Delaware and W.D. Tex. district-court cases and PTAB proceedings — IPR2023-00515, IPR2023-00169, and IPR2022-01514 (Unified Patents petitions; links at portal.unifiedpatents.com), plus a Federal Circuit appeal No. 25-2036. Any § 102/§ 103 positions actually advanced are in those petitions, which I did not fully retrieve.

Caveat: items marked "peripheral," and the § 102 columns for references whose full text I did not retrieve, are preliminary judgments based on the citation record, titles/assignees and partial text. A definitive § 102 chart requires element-by-element comparison against each reference's complete specification and claims.

I'll ground this analysis in the prior-art references actually listed on the patent's page, verifying their subject matter with targeted searches.

Now let me verify the remaining key references and the IPR-ground references.

Obviousness Analysis — US 7,330,431 B2 under 35 U.S.C. § 103

Scope and provenance. This analysis builds on the previously generated sections (bibliographic data, claim summary, post-grant/litigation history). I do not repeat that material. The prior art used here is the prior art listed in the "Prior Art" section of the US 7,330,431 page — i.e., the 64 examiner-cited U.S. patent references and the 14 non-patent citations — plus, secondarily, the references identified in the third-party IPR record already summarized. I verified the subject matter of the four load-bearing references by live search (see source table in §10). Where I could only rely on the title/assignee listed in the '431 record, I say so.

Minor date inconsistency to flag: the environment states today's date as 2026-09-28, while the task statement says April 26, 2026. Nothing in this analysis turns on the difference; I treat the art as of the '431 priority date, 2004-09-03.

1. Analytical framework and level of ordinary skill

Graham v. John Deere factors govern: (a) scope and content of the prior art; (b) differences between the prior art and the claims; (c) PHOSITA level; (d) secondary considerations. KSR Int'l v. Teleflex supplies the legal overlay: a combination is obvious if it is "the product of ordinary creativity," involves "known techniques" combined "to yield predictable results," or is a "predictable variation" of prior-art elements with a recognized need and an articulated reason to combine.

PHOSITA (proposed): a telecommunications network engineer with a B.S. in EE/CS and ~3–5 years' experience in packet-ring / SONET-SDH transport, TCP/IP layering, and network provisioning, familiar with RPR (IEEE 802.17), SRP (RFC 2892), OSPF (RFC 2328), RSVP-based admission control, and standard link-budgeting arithmetic.

Claim construction notes that matter for §103:

• "logical connections" / "connection topology" — the "who-talks-to-whom" graph of a service, expressly different from the physical ring. The specification treats this as a graph-theoretic overlay (hub-and-spoke, full mesh).
• "mapping" — the specification's own description is arithmetic routing: "CAC 24 typically selects from the available single links a minimum hop path between any two nodes, and typically also balances the allocation of the links between the two networks." This is conventional shortest-path-with-load-balancing.
• "allocating a bandwidth … in response to the bandwidth requirements … and to the mapping" — the worked examples (Tables I–VII) show this is summation of per-connection Mbps values onto physical links, plus a 1.1× multiplier where the sum falls below a worst-case figure. No claim element recites novel hardware, a new protocol, or a physical transformation.
• "full bandwidth" (cl. 11, 23) — a counterfactual worst-case reservation computed by assuming all possible logical connections are provisioned.
• "correction factor" (cl. 6, 17) = the DP (deviation parameter), 10% in both examples.

This construction matters: because the claimed steps are computational/provisioning steps over a known ring geometry, the non-obviousness inquiry concentrates almost entirely on (i) whether the overlay-to-physical mapping and (ii) the full-bandwidth-vs-mapping-bandwidth comparison with selective DP application were known or predictable.

2. Scope and content of the prior art (from this page's Prior Art section)

The cumulative picture: as of 2004-09-03 the art already contained (i) the concept of designing a logical/virtual topology over a physical network with guaranteed-bandwidth QoS commitments ('903); (ii) the specific instantiation of a full-mesh logical topology over a dual-ring core with a controller computing per-node-pair rates and summing them across ring links ('488); (iii) per-link resource accounting and path selection on both ringlets ('279, '912, RFC 2892); (iv) admission control for data networks ('155, RFC 2814, RFC 2205); (v) aggregate bandwidth estimation with a safety margin (Dziong); and (vi) sharing of reserved bandwidth among multiple connections ('008).

3. Claim 1 — element-by-element, and the primary § 103 combination

Primary combination (all non-commonly-owned art):
US 5,854,903 (Lucent) + US 6,339,488 (Nortel) + US 6,680,912 (Luminous) + US 6,625,155 (Lucent), optionally with US 6,647,008 (IBM) and RFC 2892.

Why a PHOSITA would have combined these (motivation, articulated):

1. Same field, same problem. All four address guaranteed/high-priority bandwidth in carrier transport (SONET/SDH, ATM, packet rings) with the shared objective the '431 specification itself states: point-to-multipoint and multipoint-to-multipoint traffic make link budgeting "difficult," and the naive answer ("allocate the full required bandwidth for all possible paths") "wast[es] large guaranteed bandwidth." That is an express design need in the art, supplying the "reason to combine."
2. Complementary, not overlapping, disclosures. '903 supplies the logical-topology-over-physical-network design methodology but is topology-agnostic; '488 supplies the ring-specific instantiation of the same idea (full mesh → ring, per-node-pair service rate, controller); '912 supplies the ring direction/hop-cost selection needed to turn a graph into specific span allocations on the two ringlets; '155 supplies the admission-control adjustment loop the '431 calls the CAC. Combining them yields nothing more than the '431's own flow-chart (FIG. 2: map → sum → compare → allocate).
3. Predictable results (KSR). The combination produces only arithmetic: the specification's Tables II and VI are sums of R and ⅓R terms; Tables IV and VII are the same sums with a 1.1 multiplier. There is no unpredictable interaction and no new hardware — a textbook "predictable variation … of known elements."
4. No teaching away. Every reference endorses economizing on guaranteed bandwidth; none disparages logical-overlay mapping, per-link accounting, or hop-balancing.
5. Commercial impetus. Bandwidth exhaustion and the cost of over-provisioning are recited in '903 and '488 as the driving problem. KSR recognizes market pressure plus finite design options as an obviousness indicium.

Independent, weak-secondary support: the '431 record itself lists US 7,035,279 B2 (Corrigent, "Flow allocation in a ring topology") among the examiner-cited references. '279's dispatcher-on-a-node, per-link resource accounting, quantum allocation, and per-link "sufficiency" check supply every remaining functional detail of claim 1's step (d) and steps (c)/(d) together — but see §9, because common ownership may disqualify it as §103 art.

4. Dependent claims 2–11 — combinations and where each falls

5. The genuinely contestable limitations: claims 11, 23, and the "DP-only-where-mapping<full" rule

The distinctive feature of the '431 is not "map a logical topology," nor "sum per-link bandwidth" — both are squarely in '903/'488 — but the comparison step:

"If the mapping bandwidth is equal to the full bandwidth, then CAC 24 uses the mapping bandwidth … If the mapping bandwidth is less than the full bandwidth, CAC 24 multiplies the mapping bandwidth by the deviation parameter to determine the actual bandwidth." (FIG. 2, steps 56/58)

Combined with the "R* not summed" node-budget rule of Table III (a node contributes its finite bandwidth R to one connection, so a second connection drawing on the same node bandwidth is not added), this is a specific, if simple, accounting algorithm.

Best available § 103 attack on claim 11/23:

• '903 expressly performs capacity sizing against guaranteed-bandwidth QoS commitments using effective bandwidth, which is the same conceptual move: reserve against an estimate, with a margin for variability. Comparing an "estimated/effective" aggregate against a "worst-case/full" aggregate is the ordinary way one decides whether a margin is needed.
• Dziong (NPL) teaches computing an aggregate equivalent bandwidth for a set of streams and applying a multiplicative safety factor — functionally the DP.
• '488 teaches distributing a node's finite service rate across its source-destination pairs, which is the same constrained-allocation problem as the R* rule; distributing a finite per-node budget among competing connections is elementary.
• KSR: routing/link-budgeting "comparison" and "multiply-if-below-threshold" logic is no more than the predictable application of a known mathematical relation. The '431 admits the arithmetic by presenting it as a table.

Best counter-argument for the patentee (and why it is narrow): no single reference of record is shown to disclose comparing a mapping bandwidth to a full bandwidth and applying a tolerance factor only on links where the mapping falls short, in the context of a ring with two counter-rotating ringlets and a per-node bandwidth budget. That is a real, though modest, gap; it is the type of limitation that is (i) easily designed around (apply the factor to all links, or use a different margin rule), and (ii) highly vulnerable to a teaching/suggestion/motivation argument from '903 + Dziong + '488, since the result (a conservative reservation bounded between the mapped estimate and the worst case) is exactly what all three aim at.

6. Claim 12 (and apparatus claim 28) — two services, two topologies, summed per link

Combination: US 5,854,903 (Lucent — expressly "multi-service networks") + US 6,339,488 (Nortel) + US 6,647,008 (IBM).

• '903 is directed to multiple services sharing a physical network, with logical subnetworks re-dimensioned over time and per-stream capacity requirements; adding the requirements of different services onto a common physical link is the premise of the reference.
• '488 supplies concurrency of multiple logical channels (per node pair / per destination buffer / per wavelength) multiplexed onto the shared ring, with a controller summing their rates onto each span.
• '008 supplies the specific step of sharing reserved bandwidth among several dependent connections.

Motivation: the patent's own premise ("different services operate in the network, each having a respective logical connection topology") is the standard multi-service transport scenario; '903's title and disclosure are directed precisely at multi-service logical network design. Summing per-service per-link reservations is a predictable arithmetic step (KSR), and the claim adds no new structure. Claim 22 ("first topology different from second") is met by '903's multiple logical partitions / '488's mesh-over-ring combined with any hub-and-spoke service.

Apparatus claims 25/28 (controller; in a node or external): met by '488 ("a centralized or distributed controller"); '279 ("a dispatcher, preferably in the form of a process running on one of the nodes"); '903 (a design/routing processor). Dependent claims 26/27 and 29/30 (controller inside a node vs. external) are, at most, a binary siting choice with no asserted advantage — the '431 specification itself lists both alternatives with no stated benefit, making them obvious design choices.

7. Claim 24 — ring + hub-and-spoke service + full-mesh service + selective allocation

Claim 24 is the most narrowly limited independent claim, and therefore the hardest, but still attackable:

Combination: US 6,339,488 (full mesh over a ring) + US 5,854,903 (service-specific logical topology; multi-service) + RFC 2892 / US 6,680,912 (ring direction & hop-cost path selection) + US 6,625,155 ('155), with the hub-and-spoke service topology supplied by the art's well-known "hubbed"/point-to-multipoint (e.g., video distribution) ring traffic pattern and by RPR/SRP class-based reservation messaging.

• The hub-and-spoke + full-mesh pairing is expressly juxtaposed in the ring-capacity literature cited on the '431 page: the Kalman/'912 family material contrasts ring capacity requirements for "a single 'hubbed' pattern" versus "a uniform 'meshed' pattern," i.e., the two topologies of claim 24 as the two canonical ring demand models. Selecting the model that matches the service (VODS vs. VPLS, per the '431 background) is a design choice with a clear rationale.
• The "sum first + second bandwidths" step is claim 12 machinery (§6).
• The selective allocation — allocate first, second, or total depending on whether the first, second, or both services are actually provided — is a per-service activation decision. '279's per-flow/per-request allocation and '903's "re-dimensioned from time-to-time" both teach provisioning according to which services are active; doing nothing for an inactive service is the trivially obvious baseline (KSR: "a step of doing something is obvious if the doing is trivial and the omission is the default").

Residual strength of claim 24: its specificity (ring + the exact hub/full-mesh pair + the three-way selective allocation) gives the patentee a narrower target and makes a § 103 showing more labor-intensive; but its narrowness also means trivial design-arounds (swap the topology pair; always allocate the total), which cuts against non-obviousness under KSR's "identified, predictable solutions" reasoning.

8. Third-party obviousness evidence (from the IPR record summarized earlier)

The IPR record supplies independent, adverse evidence that the claims were considered obvious — this is probative even though PTAB outcomes are not binding on a later § 103 analysis:

• IPR2023-00169 (Cisco/Dell v. Corrigent; filed 2022-11-17, instituted 2023-07-06, FWD 2024-07-03) was petitioned on § 103 grounds over Baker (US 2004/0095946), Wright (US 2004/0174884 / US 7,292,542), Holender (US 6,104,699 / US 6,069,894), and Zheng (US 6,757,247), with a Sotera stipulation.
• IPR2023-00515 (Arista v. Corrigent; FWD 2024-07-03) and IPR2022-01514 (Unified Patents; terminated pre-institution 2023-04-04) round out the challenge history; Unified also reported a March 5, 2024 final rejection in ex parte reexamination 90/015,224.

These references are not in the '431's own Prior Art section, so I flag them as secondary. Two of them (Baker US 2004/0095946 and Wright US 2004/0174884/US 7,292,542) are 2004-era ring/packet bandwidth-management publications coextensive in time with the '431, and Holender (US 6,104,699 / US 6,069,894) and Zheng (US 6,757,247) are earlier. The petitioned theory is therefore a different combination from the one I build in §§3–7, which is itself an obviousness signal: the claims sit at a junction where multiple independent combinations of known art converge. However, I could not verify the substantive holdings of the FWDs (which claims, if any, were cancelled/survived) — see §10 caveats. Do not read the FWD dates as proof of invalidation.

9. Critical caveat: pre-AIA § 103(c) common ownership may disqualify the Corrigent references

This is a methodological trap in using the '431's own citation list against it:

• US 7,035,279 B2 and US 6,952,397 B2 are Corrigent Systems references, and '279 names Leon Bruckman (the '431's sole inventor) among its inventors.
• Under pre-AIA 35 U.S.C. § 103(c)(1), subject matter that qualifies as prior art only under § 102(e)/(f)/(g) is disqualified in a § 103 analysis where, at the time the '431 invention was made, the subject matter and the claimed invention were commonly owned or subject to an obligation of assignment to the same person. Both patents were assigned to Corrigent Systems Ltd. (the '431 recorded an assignment to Corrigent Systems Ltd. on 2004-11-17).
• Consequence: '279 and '397 cannot be relied on as the primary § 103 art against the '431 merely because the examiner cited them. They are useful only as corroboration of the state of the art (what a PHOSITA knew) and, potentially, as § 102(a)/(b) art if publicly available more than one year before filing — not as the core § 103 combination.

For that reason, every combination in §§3–7 is built on non-commonly-owned art (Lucent '903, Nortel '488, Luminous '912, Lucent '155, IBM '008, and the IETF NPL), with Corrigent '279/'397 used only as corroboration. This is the single most important structural point in this analysis, and it is a point that a prior analyst summary — which leaned on the cited-art list generally — would not have surfaced.

(Practice note: the pre-AIA/AIA distinction matters here because of the 2004 filing date. Post-AIA § 102(b)(2)(C) is the analogue but is narrower/worded differently. Because the '431 is pre-AIA, I apply the pre-AIA framework.)

10. Verification status, confidence, and explicit uncertainty

11. Bottom line

Claim 1, 12, 25, 28, and most dependent claims are, in my assessment, vulnerable to a strong § 103 challenge built from non-commonly-owned art: US 5,854,903 + US 6,339,488 + US 6,680,912 + US 6,625,155 (with US 6,647,008 and RFC 2892 supplying the reserved-bandwidth-sharing, class-of-service, and ring-topology facets). Each element of the independent claims appears in that art, the references occupy the same field and attack the same well-articulated problem (wasteful full-bandwidth reservation in multipoint ring traffic), and the claimed steps are predictable arithmetic over known ring geometries — squarely KSR territory.

The narrow, defensible core is the pairing of (i) allocation-time comparison of a per-link "mapping bandwidth" against a computed "full bandwidth" with (ii) selective application of a single deviation parameter only to links where the mapping falls short, particularly under the per-node bandwidth budget rule of Table III (the "R* not summed" constraint) and its use in claim 24's selective first/second/total allocation. These limitations are not squarely disclosed by the verified non-commonly-owned references and are the strongest candidates for surviving a § 103 challenge — though they remain exposed to a Dziong-plus-'903-plus-'488 motivation-to-combine argument, and their narrowness invites easy design-arounds.

Methodological caution for any further work on this patent: do not use US 7,035,279 or US 6,952,397 as primary § 103 art — pre-AIA § 103(c) common ownership (both Corrigent Systems Ltd., with inventor overlap) likely disqualifies them, notwithstanding their presence on the '431 face.