US 7760664

I'll run targeted searches for this specific patent number and its 2026 appellate activity.

Let me narrow the search to the patent number itself and the most recent district court case.

Search results and reliability caveat

I ran targeted searches for the literal identifier 7760664 / US 7,760,664 / US7760664B2 against publicly indexed sources (Google Patents, Unified Patents litigation data, and court-complaint aggregators). I did not find any Court of Appeals for the Federal Circuit 2026 docket, opinion, or Rule 36 judgment referencing patent 7,760,664. I also do not have direct PACER/CAFC docket access through my tools, so I cannot rule out a pending 2026 appeal — I can only say I found no public trace of one. Treat the "no CAFC 2026 activity found" statement as a search result, not an authoritative negative.

Similarly, searches for "US7760664" and "US 7,760,664" returned no CAFC-specific hits; the live results I did get are consistent with the patent text supplied.

---

Bibliographic data (grounded in the authoritative patent text)

Assignee caveat: the recorded assignments include a license subsequently corrected to an assignment, plus a second corrective reassignment (2022-02-16). The Google Patents "Current Assignee" field carries the disclaimer that listed assignees may be inaccurate. I would not treat Unwired Broadband, Inc. as a definitively confirmed current owner without checking the USPTO Assignment record directly.

Abstract (verbatim)

"A network provisioning system for establishing a path between two networks is disclosed wherein a common network device between those networks is modeled as a link between a first network element in one network and a second network element in a second network. A network routing graph is created by an inventory subsystem in a routing manager by inventorying the physical network elements and links in the network. The inventory subsystem then models those elements/links as a plurality of nodes and links between the nodes. At least one common network device, such as a digital cross connect connecting the two networks, is modeled as a link instead of a node. A routing engine then uses the network routing graph, including the link modeled from the common network device, to provision a path between the networks."

Independent claims in plain language

• Claim 1 — System. A network provisioning system for a path between a first and second network element that are interconnected through a digital cross connect system (DCS). It requires a routing-manager computer with (a) an inventory subsystem that models each network element as one or more routing nodes in a graph of nodes and interconnecting links, and (b) a routing engine that uses the graph to find the path. The distinguishing limitation: the inventory subsystem models the DCS as a link between the two network elements (not as a node). It also requires a cross-connection status database storing the status of each interconnection, indicating whether a cross-connection using the DCS was successfully provisioned.

• Claim 4 — Routing manager (means-plus-function). A routing manager with: means for creating a graph of routing nodes (representing network elements) and links; means for modeling a first DCS as a link between the routing nodes representing the first and second network elements; and means for storing a status of each interconnection indicating whether a cross-connection using the DCS was successfully provisioned.

• Claim 7 — Method (routing traffic between two networks). Determine, with a network configuration management system, the interconnections created by the DCS between at least two network elements; represent each of those interconnections as a link between the two network elements; and store a status of each interconnection in a cross-connection status database indicating whether the cross-connection using the DCS was successfully provisioned. (Claim 8 adds actually provisioning a path.)

• Claim 9 — Method (determining a path). Model the network elements of both networks as routing nodes; model the network links as routing links interconnecting those nodes; model the DCS as a routing link connecting a first routing node in network one to a second routing node in network two; and store a status of each node interconnection indicating whether the DCS cross-connection was successfully provisioned. (Dependent claims add storing routing links in a routing link table, determining the path using the routing link, and provisioning the path.)

• Claim 14 — Method (generic common network device). Same structure as claim 9 but the intermediate device is recited more broadly as a "common network device" rather than a DCS (so it is not literally limited to a digital cross connect), with the step of maintaining a status of the routing links indicating whether a cross-connection using at least one routing link was successfully provisioned.

Common inventive thread: instead of collapsing network elements into single nodes only at Layer 2 (the approach of the incorporated '187 application, issued as US 7,289,456 B2), this patent collapses the inter-network junction device (DCS/optical cross-connect) out of the node set entirely and represents it as a link, with provisioning status tracked separately in a cross-connection status database.

Litigation identified for this patent (from Unified Patents data in the source record)

NY S.D. 1:20-cv-07529; TX E.D. 2:20-cv-00158, 2:20-cv-00159, 2:20-cv-00160; NY S.D. 1:20-cv-07534; DE D. 1:21-cv-01323; TX W.D. 6:22-cv-00156, 6:22-cv-01268; TX W.D. 6:23-cv-00835; TX E.D. 2:25-cv-00716. Web search corroborates CommWorks Solutions LLC v. Earthlink LLC, No. 2:25-cv-00716 (E.D. Tex.) and a CommWorks Solutions, LLC v. Comcast Cable Communications, LLC action asserting 7,760,664 alongside 6,832,249, 7,027,465, 7,177,285, 8,923,846, and RE42,883.

Explicit uncertainties

1. No CAFC 2026 docket found for 7,760,664 — my tools cannot query PACER/CAFC directly, so this is an absence of evidence, not proof of no appeal.
2. Current ownership is ambiguous from the assignment records (license vs. assignment, two corrective documents); "Unwired Broadband Inc" is Google's automated listing.
3. Litigation list comes from Unified Patents' aggregated data presented in the source record, not from PACER verification, and should be re-verified before reliance.

I'll research litigation involving US 7760664 specifically, starting with the litigation data already present in the patent record and then verifying/expanding via web search.

The search surfaced a Stanford NPE Litigation Database entry and a case list. Let me pull the detailed litigation data.

Good progress. Let me now identify the remaining cases from the litigation list (E.D. Tex., D. Del., W.D. Tex.) and confirm outcomes.

Litigation Involving U.S. Patent No. 7,760,664 (the "'664 patent")

Patent identity (as verified): U.S. 7,760,664 B2, "Determining and Provisioning Paths in a Network," inventor Sanyogita Gupta, filed Apr. 7, 2005 (priority Sept. 30, 2004, provisional 60/614,609), granted July 20, 2010. Original assignee Telcordia Technologies; the patent later traveled through Telcordia Licensing → TTI Inventions A LLC → Nytell Software LLC → Intellectual Ventures Assets 130 LLC → CommWorks Solutions, LLC, with a license to Unwired Solutions, Inc. / Unwired Broadband, Inc. Google Patents lists the current assignee as Unwired Broadband, Inc. and a legal status of "Active, expires 2028-01-07." (US7760664B2, Google Patents)

Important scope note: All known infringement litigation over this patent was brought by CommWorks Solutions, LLC. Every case below is a district-court action; I found no CAFC appeal, no ITC action, and no IPR/PGR involving the '664 patent. (A search surfaced unrelated patents with similar-looking numbers — e.g., 7,778,664, 6,775,664 — which are not this patent and are excluded.)

---

District Court Litigation

The '664 patent carried ten district-court case links on the patent record (Unified Patents Litigation Data / Darts-ip), consistent with reporting that the patent "has been asserted against eight companies." Below I list each case, with parties confirmed where I could verify them.

Key corroborating details

• Mediacom (1:20-cv-07529) and RCN (1:20-cv-07534) were companion complaints filed Sept. 14, 2020 in the S.D.N.Y. asserting a nine-patent portfolio that expressly included U.S. 7,760,664 (listed as Exhibit F in the RCN complaint). Reported by Law Street Media and The IP Crew. The RCN complaint is viewable at Docket Alarm. Mediacom docket: Justia.
• EarthLink (2:25-cv-00716) — the complaint attached "Exhibit D. Evidence of Use Regarding U.S. Patent No. 7,760,664," and asserted at least claim 7 of the '664 patent. Docket history (stay, voluntary dismissal, order of dismissal) is at CourtListener.
• Charter (6:22-cv-00156), Extreme Networks (6:23-cv-00835), Consolidated Communications (2:20-cv-00159) — identified via the Stanford NPE Litigation Database, patent 7760664.
• Related discovery proceeding (not a separate infringement merits case): CommWorks Solutions, LLC v. RCN Telecom Services, LLC, No. A-22-MC-00133-RP (W.D. Tex. Mar. 22, 2022) — a subpoena-enforcement motion against non-party MediaTek USA, Inc. over chip source code, transferred to the S.D.N.Y. because the underlying RCN suit was pending there. Casetext.

---

Parallel PTAB/USPTO Proceedings

• Ex parte reexamination 90/015,057 — filed by Unified Patents on June 10, 2022; granted July 1, 2022 (substantial new question of patentability on the challenged claims). Unified Patents.
• Reexamination certificate C1 issued Nov. 8, 2022 — "Reexamination decision confirms claims" (effective Nov. 2, 2022), i.e., the challenged claims survived. (Google Patents legal events; see also Unified Patents portal 90/015,057.)

---

Caveats (stated per your strict rules)

1. Party names for E.D. Tex. 2:20-cv-00158, 2:20-cv-00160, D. Del. 1:21-cv-01323, and W.D. Tex. 6:22-cv-01268 could not be confirmed from the sources available to me. The patent record lists those case numbers and jurisdictions but not the defendants; I did not guess. These should be verified against PACER/Unified Patents' case list.
2. Outcomes for most 2020–2023 cases are unconfirmed. The strongest outcome data points are (a) the RCN case terminated by consent judgment, and (b) the EarthLink case stayed and then voluntarily dismissed in 2025. I flag the RCN "consent judgment" characterization as coming from a secondary commercial source (PatSnap blog), not from a primary docket.
3. The Google Patents "Family has litigation" block is populated from Unified Patents Litigation Data and Darts-ip; I relied on it as the authoritative case list per your instruction, and cross-checked party names only where independent sources were found.

If you want, I can next attempt to pull the specific defendant names and termination entries for the four unconfirmed case numbers directly from the Unified Patents litigation case list and PACER.

I'll verify the structured "no PTAB activity" default against live sources before writing.

Let me dig specifically for any AIA trial numbers tied to this patent or the CommWorks campaign.

One confirmatory pass for any IPR/CBM numbers and any appeal activity.

PTAB / AIA Trial Activity — U.S. Patent No. 7,760,664

"Determining and provisioning paths in a network" · Issued 2010-07-20 · Owner of record (asserting entity): CommWorks Solutions, LLC

Sourced from the structured "PTAB proceedings on file" block (USPTO ODP ingest) and verified against open-web sources. Where I could not confirm something, I say so.

---

Proceedings overview

Zero (0) AIA trial proceedings are on file against 7,760,664 — 0 active, 0 claims invalidated, 0 claims sustained on the merits, 0 settled, 0 institution denials — which means there is no PTAB Final Written Decision, no judge panel, no institution decision and no § 315(e)(2) estoppel to work with; all 14 claims (1–14) are live and wholly untested at the PTAB, and the only post-grant attack of any kind was a Unified Patents ex parte reexamination (90/015,057, requested 2022-06-10) that ended in confirmation of the claims, leaving the patent "unhardened but also unbroken."

⚠️ Important framing: the "absence of PTAB activity" is the headline here, not a footnote. CommWorks has asserted the '664 patent against at least eight companies since 2020 (RCN, Charter/Spectrum, Comcast, Mediacom, Extreme Networks, EarthLink, and others — see Google Patents family/litigation record), and not one of them has taken the patent to the Board. That is an unusual profile and it materially changes the defensive posture analysis below (see Estoppel landscape).

---

AIA trial proceedings on file

None — no IPR, PGR, or CBM proceeding identified for U.S. 7,760,664

• Type: n/a — no Inter Partes Review, Post-Grant Review, or Covered Business Method Review on file.
• Filed: n/a.
• Status: n/a (per the ODP structured data: no AIA trial proceedings returned).
• Judge panel: n/a — do not infer a panel; none exists.
• Petition grounds: n/a.
• Institution decision: n/a.
• Final Written Decision: n/a.
• Settlement / termination: n/a.
• Appeal: n/a — with no FWD there is no PTAB-to-Federal Circuit appeal on this patent.
• Defensive value: Neutral-to-favorable as a toolkit question: because no IPR ever ran, there is zero § 315(e)(2) estoppel against any petitioner, and no adverse PTAB precedent construing the claims (e.g., no Board reading of "modeling said digital cross connect system as a routing link," claims 1 and 9). But there is also no canceled claim to lean on — every one of claims 1–14 must be attacked on its own merits.

Caveat on completeness: I searched independently for IPR/PGR/CBM filings naming the '664 patent (including through the CommWorks campaign defendants) and surfaced none. The ODP structured block is authoritative for this task and returns none. If a petition were filed after the ODP ingest window it would not appear; before relying on the zero count, run a live patent-number search on PTAB E2E (https://ptab.uspto.gov) and on the Unified Patents litigation portal.

---

Related post-grant challenge (NOT an AIA trial) — for completeness, because it drives the defensive picture

Ex parte reexamination 90/015,057 — Requester: Unified Patents, LLC (defensive aggregator) v. Patent Owner: CommWorks Solutions, LLC

This is a reexamination under 35 U.S.C. §§ 301–307, not an AIA trial. No PTAB judges, no institution decision, no FWD, no appeal. But a defendant needs to know about it.

• Type: Ex parte reexamination (central reexamination unit, examiner-driven; no adversarial PTAB phase).
• Filed: 2022-06-10 (per Unified Patents' own announcement; Google Patents records the reexamination request event as 2022-08-23 with effective filing date 2022-06-10).
• Status: Concluded — claims confirmed. Google Patents legal events show "Reexamination decision confirms claims" (CONR, kind code C1, effective 2022-11-02, entered 2022-11-08). Request had been granted 2022-07-01, with the USPTO finding "substantial new questions of patentability on the challenged claims."
• Judge panel: n/a — reexamination is not a panel proceeding.
• Grounds: Patents/printed publications only (§ 302), the sole statutory basis available in reexam. Unified has not publicly disclosed the specific references; the SNQ finding means at least one reference or combination was new to the examiner. Pull the file history for the actual art.
• Institution decision (analog): Granted 2022-07-01. Requester's own announcement: "the USPTO granted Unified's request, finding substantial new questions of patentability on the challenged claims of U.S. Patent 7,760,664." (Unified Patents, 2022-07-05)
• Final outcome: Certificate confirming the claims. On the record as indexed, no claim of 7,760,664 was canceled — i.e., claims 1–14 came out of reexam intact. Verify against the face of the C1 certificate whether any dependent claim was amended rather than purely confirmed; the "confirms claims" entry is consistent with no amendment, but I have not inspected the certificate text myself and I am not going to represent otherwise.
• Settlement / termination: n/a — no settlement mechanism in ex parte reexam.
• Appeal: None. The requester has no right to participate after the order and no adverse decision to appeal where the claims are upheld; the patent owner had nothing to appeal. No Federal Circuit appeal exists.
• Defensive value: Mixed, and nuanced. Negative: a defensive aggregator (Unified Patents) already tried the cheapest post-grant route and the examiner maintained all claims, which is a modest "hardening" signal and gives CommWorks a talking point. Positive: (i) the claims still have never faced an adversarial PTAB validity trial; (ii) ex parte reexam creates no IPR-style estoppel and no party-status estoppel against the requester; (iii) a follow-on ex parte reexam on the same art is now dead on arrival under § 303(a) — the same references cannot supply a substantial new question — so that door is closed, which pushes any future challenge to new art or to an IPR.

Sources: Unified Patents — '664 challenged (2022-06-13) · Unified Patents — CommWorks reexamination granted (2022-07-05) · Reexam file, Unified portal 90015057 · Google Patents US7760664B2

---

Strategic summary

Claim status across all 14 claims — CANCELED: none. SUSTAINED: none (on the merits, at the PTAB). UNTESTED: all of claims 1–14. The patent emerged from reexamination 90/015,057 with its claims confirmed, so the full set — independent claims 1 (network provisioning system), 4 (routing manager, means-plus-function), 7 (method for routing between two networks), 9 (method of determining a path across a DCS), and 14 (method using a "common network device"), plus dependents 2, 3, 5, 6, 8, 10, 11, 12, 13 — remains enforceable. Independent claims 7, 9, and 14 are the commercially important ones: they are the broad method claims covering modeling the cross-connect "as a routing link," and claim 7 is the claim actually charted in current assertion (the 2025 EarthLink complaint identifies "at least independent claim 7" of the '664 — see the E.D. Tex. 2:25-cv-00716 docket). Note the earlier Comcast case pleaded induced infringement of the '664 (§ 271(b)), suggesting CommWorks has trouble with direct-infringement theories against the accused Cisco Nexus-class switching gear — that is a scope signal worth exploring, not a validity holding. Also treat third-party litigation databases with suspicion here: at least one (PatSnap) mischaracterizes the '664 as "contention-free traffic detection," which is actually the '465/'904 patents — do not import that patent's history into your '664 analysis.

Estoppel landscape — you have the cleanest slate available in patent litigation. Because no IPR or PGR was ever instituted, § 315(e)(2) estoppel is a non-issue: no petitioner, no privy, no RPI is barred from any ground. Every prior-art combination — § 102 and § 103 — remains available in district court and, for a petitioner with an open window, at the PTAB. The countervailing constraint is § 315(b): the one-year clock runs from service of a complaint alleging infringement of the '664, and the defendants served in the 2020–2023 wave (RCN, Mediacom, Comcast, Charter/Spectrum, Extreme, et al.) are long since time-barred, along with their privies and RPIs. The most recent wave (EarthLink, complaint filed 2025-07-15) would have had a bar date landing around mid-July 2026, i.e., effectively expired as of today (2026-09-14) assuming service on or about the filing date — confirm the actual service date before relying on that. Practical consequence: any new IPR on this patent realistically has to come from a party that has never been served, i.e., a defensive aggregator like Unified Patents or RPX, or a newly-sued defendant. That is precisely the role Unified already played in 2022 — via reexam, not IPR.

Pattern signals. (1) Same challenger, different vehicle: Unified Patents is the only entity that has attacked this patent at the Office, and it chose ex parte reexam — the anonymous, no-estoppel, no-§ 315(b) tool — over an IPR. With that reexam now confirmed and the same-art reexam route exhausted under § 303(a), Unified or another aggregator would have to bring new art or an IPR to do anything further. (2) No repeat-petitioner dynamic exists — there is no multi-IPR campaign to model, because there is no IPR campaign. (3) The patent owner is a serial assertor, not a PTAB appellant: CommWorks (parent: Brainbox Innovations, LLC, per the 2025 EarthLink Rule 7.1 disclosure) has run a wide MSO/networking-vendor licensing campaign with mostly early, no-precedent resolutions — e.g., the 6:23-cv-00835 action against Extreme Networks ended in a with-prejudice joint dismissal with no claim construction and no validity finding — so there is no Federal Circuit precedent on the '664 and no narrowing construction to exploit. (4) Ownership records are worth a look: the Google Patents "current assignee" field shows Unwired Broadband, Inc. based on a 2020–2022 license/corrective-assignment chain out of CommWorks Solutions, yet CommWorks sued as the owner in 2025. That discrepancy should be run down for standing/real-party-in-interest purposes; I am flagging it as a records inconsistency, not asserting a conclusion. (5) The clock: status Active, adjusted expiration 2028-01-07, all three maintenance fees paid (7.5-year and 11.5-year/12th-year payments on record), so there is roughly 16 months of enforceable term remaining — a defendant's leverage point for a nuisance-value resolution.

One live policy wildcard: under the USPTO's 2025–2026 discretionary-denial posture, patent age has become a stated basis for denying institution ("settled expectations" doctrine, disputed and the subject of an amicus filing in No. 25-1230), and a 2010-issued patent asserted only after years of dormancy is squarely in the category the Board has been declining to review. Treat the IPR route as unreliable for this patent notwithstanding the clean estoppel slate. (See Unified Patents amicus brief, 2026-05-29; RPX on the 2025 reexam surge.)

---

Recommended next steps

1. Do not plan a defense around a canceled claim — there isn't one. There is no FWD to cite, no claim to point to as dead. The '664 must be met on § 102/§ 103 in district court, on claim construction (nothing has ever been construed — that cuts both ways), and on non-infringement. The induced-infringement posture of the earlier Comcast pleading is a promising line of non-infringement attack for equipment-based accusations.
2. Pull the reexam file immediately — Unified portal 90/015,057 and the C1 certificate via USPTO Patent Center. Two things to extract: (a) the exact references the examiner considered and whether any were applied against the claims (if art was distinguished on the record, those statements are prosecution-history estoppel material and potential § 112/§ 103 ammunition); (b) confirmation of whether the certificate purely confirmed or amended any dependent claim.
3. Determine your own § 315(b) date before anything else. If you (or a privy/RPI) were served with a complaint alleging the '664 more than one year ago, IPR is unavailable to you personally — in which case a third-party IPR by a never-served aggregator is the only Board route, and coordination with such an entity should be evaluated now rather than after the bar date.
4. If an IPR window remains open, file early and build the petition against claims 7, 9, and 14 jointly (a claim-by-claim, limitation-by-limitation mapping of "modeling said digital cross connect system as a routing link" against the admitted prior art, including the patent's own incorporated '187 application lineage, U.S. Pat. App. 10/118,187). Anticipate a § 314(a)/Fintiv and age-based discretionary-denial fight, and pre-empt it with a Sotera-style stipulation and evidence that trial timing does not moot the Board's work.
5. Do not re-file an ex parte reexam on the 2022 art — § 303(a) will defeat it for failure to present a substantial new question. Any reexam needs genuinely new references or a new combination. Note for your budget: ex parte reexam imposes no § 315(e) estoppel and no § 315(b) bar, and remains the cheapest route in the current discretionary-denial environment.
6. Monitor for new filings against never-served defendants. CommWorks continues to file (2025-07-15 EarthLink in E.D. Tex. before Judge Gilstrap; note the earlier W.D. Tex./Judge Albright cases and the 2023 Extreme Networks matter). Each new complaint starts a fresh § 315(b) clock for that defendant and creates a new potential petitioner — worth watching as both a threat map and a coordination opportunity.
7. Run a live confirmation search on PTAB E2E by patent number before filing anything, to close the ingest-lag gap on the zero-count finding.

Bottom line for a defendant today: you are facing a 14-claim patent that is unbroken but unpurified — no IPR has ever touched it, so no claim is canceled and no estoppel exists, but also no PTAB panel has ever blessed it. The realistic levers are district-court invalidity and non-infringement, an induced-infringement/scope attack on the '664's path-modeling claims (especially asserted claim 7), and a settlement calibrated to a patent term ending 2028-01-07.

I'll verify the assignment records and correspondent data against live sources.

Ownership & Assignment Reconstruction — U.S. Patent No. 7,760,664

Scope/verification note up front. I built this section from the USPTO assignment data as ingested into the Google Patents legal-events record for US7760664B2 (Google Patents), which reproduces the recorded reel/frame, conveyance, assignor/assignee and execution/recording dates for each document. I could not retrieve the correspondent-of-record fields themselves — my tooling could not reach assignmentcenter.uspto.gov or assignment.uspto.gov directly, and the two targeted searches I ran for the correspondent data returned nothing usable. Everything below is therefore grounded in reel/frame, dates, conveyance and parties, and I mark correspondent data as a retrieval gap rather than inventing names. I flagged this explicitly in Signal 3.

Cross-checks performed: Google Patents legal events (primary), the litigation/PTAB sections generated earlier in this analysis, and one live search confirming CommWorks Solutions, LLC's corporate parent (Brainbox Innovations, LLC) via its Rule 7.1(a)(1) disclosure filed in CommWorks Solutions, LLC v. EarthLink, LLC, No. 2:25-cv-00716 (E.D. Tex.) (CourtListener).

---

Inventors

Pattern notes:

• Pre-filing assignment (unusual salience for a solo inventor). The instrument at reel 016218/0727 was executed 2005-03-23, before the non-provisional was filed on 2005-04-07, and the provisional (60/614,609) had been filed 2004-09-30. A pre-filing, pre-dating assignment of a sole-inventor case is the signature of a pre-existing employment/invention agreement, not a negotiated sale.
• The "all inventors departed within 12 months" fire-sale precursor is not present — but only because there is one inventor. With a single inventor and a same-day-as-employment assignment, there is no observable departure cluster to read. I found no evidence of Gupta leaving Telcordia within 12 months of filing, and I am not going to assert one.
• Repeat inventor across the same family lineage. Gupta is also the named inventor of the '187 application incorporated by reference into this patent (published as US 2003/0189919 A1, filed 2002-04-08, "Determining and provisioning paths within a network of communication elements," which issued as US 7,289,456 B2). Both appear in the "Patent Citations" table of the source record. This is a single-inventor continuity line — one Telcordia routing-architecture engineer generating the whole node-collapsing family, not a large lab team. That matters for § 103 obviousness-typology arguments (same inventor, same assignee, express incorporation by reference), which I flag for the invalidity workstream rather than for the NPE analysis.

---

Original assignee

Telcordia Technologies, Inc. — not "Individual."

Correction to a field in the previously generated bibliographic table: an earlier section reproduced Google Patents' "Original Assignee ⇒ Individual." That field is an artifact of the underlying application/bibliographic data, not the operative assignee of record. The executed assignment (Reel 016218/0727, effective 2005-03-23, recorded 2005-05-16) conveys the invention to Telcordia Technologies, Inc. (New Jersey). For ownership purposes, Telcordia was the assignee from before the filing date onward.

---

Assignment timeline

Title-affecting documents in chronological (execution) order. Reel/frame, conveyance, parties, execution and recording dates are as recorded; the correspondent field is not available to me and is noted as a gap on every entry.

1. Inventor → Telcordia (the originating assignment)

• 2005-03-23 (executed) / recorded 2005-05-16 — Reel 016218/0727
  • Conveyance: Assignment of Assignors' Interest
  • Assignor: GUPTA, SANYOGITA
  • Assignee: TELCORDIA TECHNOLOGIES, INC., New Jersey
  • Correspondent: ⚠️ not retrieved — pull from Assignment Center. (Business-significance note: the assignee address and correspondent on this 2005 document are almost certainly Telcordia's in-house IP docket, not an outside NPE shop; but I have not verified it and will not name it.)
  • Context: Pre-filing employment assignment. Execution predates the 2005-04-07 filing; this is the only "real" inventor-to-company transfer in the chain.

2. Security-interest release (encumbrance, NOT a title transfer)

• 2009-02-20 (executed) / recorded 2009-03-17 — Reel 022408/0410
  • Conveyance: Release of Security Interest
  • Assignor / releasing party: WILMINGTON TRUST COMPANY (as collateral agent)
  • Assignee / beneficiary: TELCORDIA TECHNOLOGIES, INC., New Jersey
  • Correspondent: ⚠️ not retrieved
  • Context: Securitization unwind. Releases the lien Wilmington Trust held over the portfolio, freeing the patents for the 2009 transfer that follows 99 days later. This document does not move title; it clears it. Note the sequence: security release → asset transfer within ~4 months is the classic pre-divestiture step.

3. Telcordia → Telcordia Licensing Company, LLC

• 2009-06-16 (executed) / recorded 2009-06-24 — Reel 022871/0920
  • Conveyance: Assignment of Assignors' Interest
  • Assignor: TELCORDIA TECHNOLOGIES, INC.
  • Assignee: TELCORDIA LICENSING COMPANY, LLC, New Jersey
  • Correspondent: ⚠️ not retrieved — this is the highest-value entry to obtain. The recording agent on the first carve-out to a licensing vehicle is the earliest point at which an NPE-side recording firm can be identified.
  • Context: Internal reorg into a licensing-only vehicle. Telcordia separates its patents from its operating business — "Licensing" suffix, dedicated LLC. First structural step away from an operating company.

4. Telcordia Licensing → TTI Inventions A LLC

• 2011-11-02 (executed) / recorded 2012-03-09 — Reel 027830/0088
  • Conveyance: Assignment of Assignors' Interest
  • Assignor: TELCORDIA LICENSING COMPANY LLC
  • Assignee: TTI INVENTIONS A LLC, Delaware
  • Correspondent: ⚠️ not retrieved
  • Context: Transfer to an asset-holding LLC, timed to the Ericsson acquisition. Executed Nov. 2011, recorded Mar. 2012. "TTI" plausibly abbreviates Telcordia Technologies, Inc. and "Inventions A" implies a lettered series of parallel asset LLCs (A, B, C…). That series inference is mine and is flagged as inference, not a record finding. The A-series naming plus the Ericsson-closing timing is consistent with the carve-out being settled as a condition of, or consequence of, the sale.

5. TTI Inventions A LLC → Nytell Software LLC (by merger)

• 2015-08-26 (executed) / recorded 2015-12-31 — Reel 037407/0912
  • Conveyance: MERGER
  • Assignor / disappearing entity: TTI INVENTIONS A LLC
  • Assignee / surviving entity: NYTELL SOFTWARE LLC, Delaware
  • Correspondent: ⚠️ not retrieved
  • Context: Internal corporate reorganization by merger — title passes by operation of law, not by bargain. Delaware LLC. Note the four-year dormancy (2011→2015) with no assertion of the '664 during that window.

6. Nytell Software LLC → Intellectual Ventures Assets 130 LLC

• 2019-10-30 (executed) / recorded 2019-11-01 — Reel 050886/0640
  • Conveyance: Assignment of Assignors' Interest
  • Assignor: NYTELL SOFTWARE LLC
  • Assignee: INTELLECTUAL VENTURES ASSETS 130 LLC, Delaware
  • Correspondent: ⚠️ not retrieved — second highest-value entry. The "Assets 130 LLC" series is IV's industrial-scale asset-holding nomenclature; the recording agent on this document is the IV-side filer.
  • Context: Acquisition by a known aggregator. This is the single most probative link in the chain for NPE classification: it moves the patent into the Intellectual Ventures asset family. "Assets 130" indicates a numbered shell series — 130+ single-purpose vehicles is itself the definition of an aggregation factory.

7. Intellectual Ventures Assets 130 LLC → CommWorks Solutions, LLC

• 2019-11-15 (executed) / recorded 2020-01-03 — Reel 051463/0026
  • Conveyance: Assignment of Assignors' Interest
  • Assignor: INTELLECTUAL VENTURES ASSETS 130 LLC
  • Assignee: COMMWORKS SOLUTIONS, LLC, Georgia
  • Correspondent: ⚠️ not retrieved
  • Context: Transfer to the asserting entity. 16 days after the IV acquisition executed (2019-10-30 → 2019-11-15). CommWorks is the plaintiff in all ten district-court actions on this patent. Its disclosed corporate parent is Brainbox Innovations, LLC (Rule 7.1 disclosure in 2:25-cv-00716).

8. CommWorks → Unwired Solutions, Inc. (LICENSE)

• 2020-09-18 (executed) / recorded 2020-11-05 — Reel 054443/0958
  • Conveyance: LICENSE
  • Assignor: COMMWORKS SOLUTIONS, LLC
  • Assignee: UNWIRED SOLUTIONS, INC., California
  • Correspondent: ⚠️ not retrieved
  • Context: Post-filing licensing arrangement — executed four days after the first two infringement complaints. The S.D.N.Y. complaints against Mediacom (1:20-cv-07529) and RCN (1:20-cv-07534) were filed 2020-09-14. This instrument was executed 2020-09-18. Whether that timing is coincidence or reflects a financing/licensing arrangement tied to funding the campaign is unknown to me — but the four-day gap is a fact worth running down.

9. Corrective assignment — name of assignee

• 2020-09-18 (executed, retroactive to the same date) / recorded 2021-02-04 — Reel 056981/0631
  • Conveyance: CORRECTIVE ASSIGNMENT — "to correct the assignee name previously recorded at reel 054443, frame 0958"; the instrument's own text "confirms the LICENSE" and separately "confirms the ASSIGNMENT."
  • Assignor: COMMWORKS SOLUTIONS, LLC
  • Assignee: UNWIRED BROADBAND, INC., California
  • Correspondent: ⚠️ not retrieved
  • Context: Change of name only as to the assignee (Solutions, Inc. → Broadband, Inc.) — plus a pleading/records inconsistency: the same reel contains both a "LICENSE" confirmation and an "ASSIGNMENT" confirmation. Which conveyance actually occurred is ambiguous on the face of the record.

10. Corrective assignment — nature of conveyance

• 2020-09-18 (executed, retroactive) / recorded 2022-02-16 — Reel 059907/0563
  • Conveyance: CORRECTIVE ASSIGNMENT — "to correct the nature of conveyance previously recorded at reel 056981, frame 0631"; the instrument now "confirms the ASSIGNMENT."
  • Assignor: COMMWORKS SOLUTIONS, LLC
  • Assignee: UNWIRED BROADBAND, INC., California
  • Correspondent: ⚠️ not retrieved
  • Context: Correction converting/settling the conveyance as an assignment. Two successive corrections (2021-02-04 and 2022-02-16) to a single 2020-09-18 instrument is a drafting-quality tell — the transaction was closed loosely and repaired ex post.

Non-assignment entries observed on the same record (not title transfers, listed so the timeline reconciles): fee-payment/entity-status events 2009-12-08, 2013-12-30, 2017-12-15, 2022-03-07, 2022-07-18; the ex parte reexamination request 2022-08-23 (effective 2022-06-10) and confirmation certificate C1 entered 2022-11-08.

⚠️ Standing / real-party-in-interest inconsistency to run down. The recorded chain terminates with Unwired Broadband, Inc. (Reels 056981/0631 and 059907/0563), and Google Patents lists Unwired Broadband as current assignee — yet CommWorks Solutions, LLC sued as patent owner in 2025 (EarthLink, 2:25-cv-00716, filed 2025-07-15). Either (a) the Unwired instruments conveyed something narrower than the full right, title and interest the "ASSIGNMENT" caption implies; (b) the chain of 2020–2022 corrections was itself erroneous; or (c) an unrecorded re-transfer exists. I cannot resolve this from the available records and I am not going to guess. It is a genuine § 281 standing / Rule 17 RPI question and should be raised for the file.

---

Timeline diagram

timeline
    title Ownership of US 7760664
    2004 : Provisional filed 30 Sep 2004
    2005 : Inventor assigns to Telcordia
         : Application filed 7 Apr 2005
    2009 : Security interest released
         : Telcordia Licensing Company LLC formed
    2010 : Patent issued 20 Jul 2010
    2011 : TTI Inventions A LLC acquired portfolio
    2012 : Telcordia acquired by Ericsson
         : Assignment recorded at USPTO
    2015 : TTI Inventions merged into Nytell Software LLC
    2019 : Nytell sold to Intellectual Ventures Assets 130 LLC
         : 16 days later sold to CommWorks Solutions LLC
    2020 : First infringement suits filed
         : Unwired license recorded after suit
    2021 : Corrective assignment changes assignee name
    2022 : Corrective assignment confirms conveyance
         : Reexam confirms all claims
    2025 : CommWorks sues EarthLink

---

NPE / troll-pattern signals

---

Verdict

NPE — high confidence.

Justification. Four independent strong signals fire on the recorded documents. (1) Shell-entity transfer: title passed from operating company Telcordia Technologies to four consecutive non-operating holders — Telcordia Licensing Company, LLC (Reel 022871/0920, exec. 2009-06-16), TTI Inventions A LLC (Reel 027830/0088, exec. 2011-11-02), Nytell Software LLC (Reel 037407/0912, merger, exec. 2015-08-26), Intellectual Ventures Assets 130 LLC (Reel 050886/0640, exec. 2019-10-30) — and never returned to an operating entity. (2) Known asserter: the chain runs through Intellectual Ventures (Reel 050886/0640) and terminates at CommWorks Solutions, LLC (Reel 051463/0026), the plaintiff in all ten district-court actions on this patent (2020–2025), whose disclosed parent is Brainbox Innovations, LLC. (3) Cascading transfers: the final two hops executed 16 days apart (2019-10-30 → 2019-11-15) and were asserted against eight-plus defendants beginning Sept. 14, 2020. (4) The litigation posture confirms the classification rather than the naming: ten suits across five districts, ten-plus years of dormancy broken only after the asset reached a plaintiff with no products, and a second infringement case filed as recently as 2025-07-15.

I am not resting the verdict on any single name: the Delaware LLC suffixes and the "Inventions A" / "Assets 130" series numbering are corroborative, but the load-bearing evidence is the reel/frame-documented structural fact that four successive non-operating entities held this patent for sixteen years. The one signal I explicitly could not establish is the repeat correspondent (Signal 3) — a retrieval gap, not a negative — and the one open records question is the two successive corrective assignments (Reels 056981/0631 and 059907/0563) that leave the current owner ambiguous between CommWorks Solutions, LLC and Unwired Broadband, Inc., notwithstanding CommWorks's 2025 assertion as owner. Both should be resolved against primary records before this section is relied on for standing or RPI purposes.

Verify at: USPTO Assignment Center — search patent 7760664 (mirror index: assignment.uspto.gov).

---

Contradictions with previously generated sections (flagged per instruction)

1. "Original Assignee: Individual." The previously generated bibliographic table lists the original assignee as "Individual; assigned to Telcordia Technologies, Inc." The operative assignment of record (Reel 016218/0727, effective 2005-03-23 — before the filing date) makes Telcordia Technologies, Inc. the assignee ab initio. Treat "Individual" as Google Patents metadata, not an ownership fact.
2. Current owner. The prior sections noted the license-vs-assignment ambiguity and the 2020/2021 corrections; this section adds the second corrective assignment (Reel 059907/0563, recorded 2022-02-16, correcting the nature of conveyance to an assignment). The inconsistency with CommWorks suing as owner in 2025 is therefore not resolved — it is reinforced, and it should be treated as an open standing question rather than a settled assignment chain.

I'll verify the USPTO record for 7760664 and each cited reference before analyzing.

Confirming the key cited references' content and dates.

I hit my tool-call budget mid-verification, so this section combines (a) the authoritative citation list from the US7760664 record itself and (b) the references I was able to verify live this session. Where I could not independently verify a reference in this session, I say so rather than describing it from memory as if confirmed.

---

Prior Art Analysis — U.S. Patent No. 7,760,664

Search confirmation. I searched for the literal string 7760664 and returned hits only for US 7,760,664 B2, "Determining and provisioning paths in a network" (Gupta). No similarly-numbered patents (e.g., 7,778,664; 6,775,664) were included in this analysis. The record's own citation table ("Patent Citations (17)," "Non-Patent Citations (3)") is the authoritative list and is what I analyze below.

Threshold legal frame (important — this is a pre-AIA patent)

US 7,760,664 was filed 2005-04-07, claiming priority to provisional 60/614,609 (2004-09-30). It is therefore governed by pre-AIA 35 U.S.C. § 102. Three consequences drive everything below:

• References published before 2004-09-30 are prior art under § 102(a)/(b) (and, if the publication is more than one year before the priority date, § 102(b) is a statutory bar not defeasible by a swear-behind).
• U.S. patents and U.S. application publications whose applications were filed before 2004-09-30 are prior art under § 102(e) even if they published after that date. This is the operative category for the Oracle, Ericsson, Castro, Fujitsu, and Nortel references.
• Because the reexamination certificate (C1, effective 2022-11-02) confirmed the claims, there are no amended claims to re-scope — the analysis is against the issued claim text as summarized earlier in this file.

Bottom line before the detail

No reference cited in US 7,760,664 discloses the limitation that distinguishes this patent — modeling a digital cross connect system as a link rather than as one or more nodes. Every independent claim (1, 4, 7, 9, and 14) contains that limitation. Therefore there is no § 102 anticipation of any independent claim, and no § 102 anticipation of any dependent claim, on the face of the cited art. The citation list is best understood as (i) background art for the graph-routing motivation and (ii) § 103 combination fodder for the individual elements. I state this as my own analysis of the cited disclosures — I do not have the reexamination file history and cannot report which references the examiner actually applied or how they were distinguished.

Claim-element key used below

Master table — the 17 cited patent documents

Non-patent citations of record (3): European Search Report for EP 05857725.5 (2009-08-25); International Search Report for PCT/US2005/034418 (mailed 2006-12-27); Notice of Rejection for JP 2007-534687 (mailed 2009-07-31, with English translation).

---

Per-reference analysis

Tier 1 — Closest art in the record: the same inventor's own '187 disclosure

US 2003/0189919 A1 — Sanyogita Gupta, "Determining and provisioning paths within a network of communication elements," filed 2002-04-08, published 2003-10-09; issued as US 7,289,456 B2 (2007-10-30), later reissued as RE43,704. This is the '187 application incorporated by reference into the '664 specification, and it is the same inventive entity.

Description (verified on Google Patents/RPX): models network elements as one or more routing nodes where any edge ports capable of interconnection can be interconnected; models network links and virtual trunks as routing links; keeps the graph in three tables (routing link table, routing node table, NMS/EMS table); and maintains "each cross-connection comprising a provisioned virtual trunk/circuit … which status indicates whether a cross-connection has been successfully provisioned."

§ 102 assessment: discloses E1, E2, E4, E5, E6 and much of E7. It does not disclose E3 — the DCS-as-link modeling — because the DCS concept is the very gap the '664 specification describes as unaddressed. No § 102 anticipation of claims 1, 4, 7, 9, or 14. It is simultaneously (i) the strongest § 103 starting point, (ii) an admission of prior art under the '664's own "Background"/"Detailed Description" narrative, and (iii) potentially double-edged for the patentee because the '664's claims are a narrow delta over it.

Tier 2 — Node-aggregation / "virtual node" art (the conceptual ancestor of collapsing nodes)

US 5,526,414 A — Northern Telecom (Bell-Northern Research), filed 1994-10-26, issued 1996-06-11. Verified. Dynamically Controlled Routing network in which a switching element's destination list "comprises a virtual destination node representing a group of two or more components, each component being a distinct physical network element," with each link to that virtual node being a set of circuit groups. This is genuine node aggregation (§ 102(b), long pre-dating the '664). § 102: touches E1 only; it is telephony call routing at the switch layer, not transport-path provisioning, and it says nothing about representing an interconnection device as a link. Claims 1, 4, 7, 9, 14 — E1 only; no anticipation.

US 6,091,720 A — Nortel Networks, filed 1994-10-26, issued 2000-07-18. Verified. Continuation-in-part of the '414 family (also EP 0789974), adding weighted-proportion overflow across the multiple circuit groups comprising the link to a virtual destination node. This is the single most conceptually adjacent piece of telephony art in the list, because it treats one graph link as a set of circuit groups spanning two or more physical components. § 102: touches E1 and arguably E2; still no DCS-as-link and no cross-connection provisioning-status database as claimed. No anticipation of any claim.

US 2005/0169179 A1 — Telefonaktiebolaget LM Ericsson (Publ), "Cluster-based network provisioning," priority 2004-02-04, published 2005-08-04. § 102(e) (application filed before the '664 priority date). Clusters of network elements provisioned as an abstract unit — structurally the same "collapse many elements into one modeled entity" idea that underlies E1. § 102: E1/E2 only; no E3, and it post-dates the '187 disclosure it resembles. Relevant to claims 1, 4, 9, 14.

US 5,764,740 A — Ericsson, filed 1995-07-14, issued 1998-06-09. Optimal logical-network capacity dimensioning with broadband traffic; models logical/virtual topology. § 102: E1/E2 for the broad "model the network as nodes and links" concept; nothing on cross-connects. Claims 1, 4, 7, 9, 14 — background.

Tier 3 — Topology modeling and transport provisioning

US 7,173,912 B2 — Fujitsu Limited (Jaber, Harbin, Yaseen, Mo, Bucklin), filed 2001-05-04 (priority 2000-05-05), issued 2007-02-06. Verified from the patent PDF. Determines asymmetric intranode connectivity between traffic-bearing components of a transport node, models the node, advertises the model (opaque LSAs) to a disparate node, and uses it there for path determination. § 102(e). Relevant to E1, E2, E4 — but note it runs in the opposite direction from the '664: it expands a node into modeled internal components, whereas the '664's invention removes a device (the DCS) from the node set entirely and treats it as a link. It is useful § 103 art only for the generic "nodes and links are modeled and used for path selection" proposition. Claims 1, 4, 9, 11, 14.

US 6,981,065 B1 — Nortel Networks, filed 2000-05-18, issued 2005-12-27. Verified. Provisioning a connection through a SONET/SDH transport network formed of multiple sub-networks, using a connection-oriented, OSI-layered framework. § 102(e). Relevant to E2, E5, E6 — provisioning a connection across sub-networks and tracking the connection's establishment. It does not model cross-connect junctions as links. Claims 7, 8, 9, 11, 12, 14.

US 2004/0107277 A1 — Fujitsu Network Communications, "Element management system with tree-structured representations," priority 1999-05-26, published 2004-06-03. § 102(b). EMS representing managed elements in tree structure — relevant to E1 (representing network elements as managed entities) and weakly to E5 (representing element/connection state). Not independently re-verified this session — description is from the citation record plus general knowledge of the family. Claims 1, 4, 7, 9, 14.

US 2003/0071840 A1 — Ye Huang, GUI for building and/or monitoring a telecommunication network, priority 1998-12-18, published 2003-04-17. § 102(b). A GUI layer over a telecom network model — relevant only to E1/E4 at the "there is a graph of the network" level. Not independently re-verified this session. Claims 1, 4, 9, 11.

Tier 4 — Data modeling, GUI, and service-provisioning peripherals

US 2005/0097108 A1 — Oracle International Corporation, "Network data model for relational database management system," priority 2003-10-29, published 2005-05-05. § 102(e). A relational network data model (nodes, links, paths persisted in an RDBMS). Strategically this is the most useful § 103 partner in the list for the '664's database-structure limitations, because it maps directly onto the routing link table / routing node table / cross-connection status database architecture. § 102 alone: no E3, so no anticipation. Claims 1, 4, 7, 9, 14.

US 2006/0015617 A1 — Fabian C. Castro, "Management system and method for service subscription provisioning," priority 2002-07-18, published 2006-01-19. § 102(e). Service-subscription provisioning management; relevant to E5/E6 (provisioning workflow and status). Not independently re-verified this session. Claims 1, 4, 7, 8, 9, 12, 14.

US 2002/0029298 A1 — Magnus Wilson, "Arrangement, a system and a method relating to management communication," priority 1997-02-24, published 2002-03-07. § 102(b). Management-communication arrangement; weak, generic relevance to E1 only. Not independently re-verified this session — I flag this as the reference in the list about which I know least and would not rely on my description of it.

Tier 5 — Background routing art and one clearly non-analogous citation

US 5,377,262 A (AT&T, 1991-12-30 / 1994-12-27) and US 5,297,137 A (IBM, 1991-01-30 / 1994-03-22) — adaptive-routing and multi-node packet-routing references. § 102(b). These establish only the ancient proposition that networks are modeled and routed as nodes and links (E1/E4 background). No anticipation of anything.

US 4,669,113 A (AT&T, 1985-04-26 / 1987-05-26) and US 4,284,852 A (Northern Telecom, 1979-06-22 / 1981-08-18) — dynamic nonhierarchical routing controller and alternate telephone routing, respectively. § 102(b). Background only; indeed US 4,284,852 is cited within US 5,526,414's own specification as "an early version of dynamically controlled routing," confirming its role as lineage background. No claim relevance.

US 4,788,421 A — Mitsubishi Denki, "Apparatus for controlling relative movement of an optical head to an optical disk with velocity detection," 1985-05-15 / 1988-11-29. I flag this one explicitly: it is an optical disk drive servo/velocity-detection patent and is not analogous art to network path modeling. Its presence in the citation list is almost certainly examiner-citation over-inclusion or an automated-association artifact. It does not anticipate any claim of US 7,760,664, and I would not build any argument — validity or validity-challenge — on it.

---

Data-quality warning worth carrying forward

The record's "Family Cites Families" block lists US 11,818,702 B2 (Asustek Computer Inc., 2022-03-25) as a family citation. That is not prior art to the '664 and is almost certainly a string-matching artifact of the '187 application's Ser. No. 10/118,187 (the "118,187" sequence matching patent number "11818702"). Do not treat it as a citation of record. The other four entries in that block (US 5,864,666; JP 3974705 B2; JP 3653660 B2; JP 4110671 B2) are foreign/family citations from the JP and EP prosecutions and are outside the US citation list analyzed here.

I also note the source record's citation table carries asterisk annotations whose legend ("* Cited by examiner, † Cited by third party") appears under the "Cited By" tables; I could not confirm with certainty which of the 17 were examiner-cited versus third-party-cited, so I have deliberately not sorted the list by that distinction. Seven of the seventeen carry the annotation in the record.

What this means for a § 102/§ 103 posture

1. There is no single-reference § 102 knockout in this record. The DCS-as-link limitation (E3) appears nowhere in the cited art — not in the telephony virtual-node art (which aggregates switches into nodes), not in the Fujitsu intranode-modeling art (which expands nodes), and not in the Oracle data-model art (which is agnostic about what a link represents).
2. The strongest invalidity theory from this record is § 103, and it runs through the patentee's own '187 disclosure (US 2003/0189919 → US 7,289,456 / RE43,704) as the primary reference, combined with (a) a virtual-node aggregation reference (US 5,526,414 or US 6,091,720) for the "collapse modeled entities" rationale, (b) US 2005/0169179 (Ericsson cluster-based) for grouping elements, and (c) US 2005/0097108 (Oracle) for the database-persistence limitations. The '664 specification's own admissions that DCS configurations are "relatively static" and that DCSs function as "common nodes between the respective networks" supply the articulated reasoning for why one of ordinary skill would model the DCS as a link instead of a node.
3. The obviousness attack must be built on the file history, not the patent face. The reexamination requester (Unified Patents) obtained a substantial-new-question finding and still lost — so whatever art was considered in 90/015,057 is the art that failed. Pull the reexam file before investing in any of the references above, because § 303(a) means those same references cannot sustain a second reexam, and any district-court § 103 theory that merely re-runs them starts from a known loss.

Caveats I will not paper over

• I did not retrieve the USPTO file wrapper for 11/101,136 or for reexamination 90/015,057; consequently I cannot state which of these 17 references the examiner applied under § 102 or § 103, or how they were overcome. The claim mapping above is my analysis, not a report of the prosecution record.
• Six references (US 2004/0107277, US 2003/0071840, US 2006/0015617, US 2005/0097108, US 2005/0169179, US 2002/0029298) were described from the citation record and general knowledge rather than verified live in this session because I exhausted my search budget. Treat those descriptions as provisional and verify before citing them.
• The claim-element key is derived from the claim text already summarized in the prior section of this file; if that text is later re-read against the certificate, confirm E3's exact wording ("models said digital cross connect system as a link between said at least a first network element and said at least a second network element") — claims 1, 4, 7, 9, and 14 all turn on it.

I'll verify what the cited prior-art references actually disclose before building the § 103 combinations, since the record gives me only titles for most of them.

Two key confirmations so far — the '187 application's own summary discloses cross-connection status tracking. Let me verify whether it also discusses digital cross-connects, and check the Nortel provisioning reference.

Obviousness Analysis — U.S. Patent No. 7,760,664 Under 35 U.S.C. § 103

Analysis date context — flagged inconsistency (per your "flag contradictions" rule): the task header states the current date is April 26, 2026, while the system date for this session is 2026-09-14, and the previously generated sections were written against the September date. Nothing in this § 103 analysis turns on the difference, but I am flagging it rather than silently reconciling it.

Method note. Per your instruction I worked from the prior-art citations listed in the record for US7760664B2 (the "Patent Citations," "Non-Patent Citations," and the "Prior art keywords" blocks). Because the record supplies mostly titles, I ran confirmatory searches on the most load-bearing references and quote what I found with URLs. Where I am inferring from a title only, I say so. I do not have full-text access to every reference, and the conclusions below that depend on unmatched full text are marked accordingly.

---

1. Two record-level contradictions that materially affect the § 103 analysis

(A) The record attributes the DCS-as-link invention to the '187 application. The "Definitions" block of the source record states, in connection with U.S. application Ser. No. 10/118,187: "At least one common network device, such as a digital cross connect located at a junction between the two networks, is modeled as a link instead of a node." If that sentence describes the '187 application, then the '187 discloses the single limitation that the '664 specification identifies as its point of novelty, and claim 1 is anticipated under § 102, not merely obvious.

My retrieval of the '187 publication's own text does not support that attribution. The '187's abstract and summary (US 2003/0189919 A1, published 2003-10-09; granted as US 7,289,456 B2 on 2007-10-30) teach the opposite emphasis: network elements are modeled as one or more routing nodes — "[t]he network links and virtual trunks are both modeled as routing links, wherein routing links interconnect the routing nodes to create the graph of the network." The "common network device … modeled as a link" sentence appears to be lifted from the '664's own Summary/Abstract and placed adjacent to the '187 discussion by the database's auto-summarization. This must be resolved against the '187's full text before any § 102 position is taken. (US20030189919A1, Google Patents; US7289456, FreePatentsOnline)

(B) The '187's inventive entity is not identical to the '664's. The previously generated bibliographic section treats the '187 as "the incorporated '187 application" without noting inventorship. Secondary sources attribute US 7,289,456 to Gupta, Raheja, and Ferrer, while the '664 names Gupta alone. That difference matters: a multi-inventor reference is "by another" for pre-AIA § 102(e)/(a)(2) purposes even though the '664's sole inventor is one of them. Conversely, both were commonly owned by Telcordia Technologies, which triggers the pre-AIA § 103(c) question discussed in § 9 below. (Golden wiki, US 7,289,456)

---

2. Framework, POSITA, and construction of the dispositive terms

Legal framework. Graham v. John Deere Co., 383 U.S. 1 (1966) (scope/content, differences, PHOSITA level, secondary considerations); KSR Int'l Co. v. Teleflex Inc., 550 U.S. 398 (2007) (motivation may be supplied by "design incentives and other market forces," by the "interrelated teachings of multiple patents," by "any need or problem known in the field," and by "common sense"; predictable variation of a known technique is obvious); MPEP §§ 2141–2144 (enumerated rationales). Reasonable expectation of success is judged from the perspective of a POSITA, In re O'Farrell, 853 F.2d 894 (Fed. Cir. 1988).

POSITA. A network engineer/architect with a B.S. in EE or CS and roughly 3–5 years of experience designing or operating transport/connection-oriented networks, familiar with the OSI layering model, SONET/SDH and DCS/optical cross-connects, NMS/EMS management hierarchies, and graph-based shortest-path routing (the last of which the patent itself invokes by modeling a "graph" of nodes and links). Notably, the '664's own specification treats layer-1 transport management as routine engineering: "Other uses of DCS are well known and will be obvious to one skilled in the art." That sentence is an admission of the ordinary skill level in exactly the field of the claimed subject matter.

Claim construction of the dispositive terms:

---

3. The prior-art corpus on the face of the '664

Patent citations (17 shown). US4284852A (Northern Telecom, alternate routing); US4669113A (AT&T, dynamic nonhierarchical routing); US4788421A (Mitsubishi — "Apparatus for controlling relative movement of an optical head to an optical disk with velocity detection"); US5297137A (IBM, routing data packets around a multi-node network); US5377262A (AT&T, adaptive routing switching nodes); US5526414A (Northern Telecom, dynamically controlled routing using virtual nodes); US6091720A (Nortel, dynamic management of intra-link traffic to virtual destination nodes); US5764740A (Ericsson, logical network capacity dimensioning); US20020029298A1 (Wilson, management communication); US20030071840A1 (Huang, GUI for building/monitoring a telecom network); US20040107277A1 (Fujitsu, EMS with tree-structured representations); US7173912B2 (Fujitsu, modeling/advertising asymmetric topology of a node in a transport network); US6981065B1 (Nortel, provisioning of connection through a SONET/SDH network); US20030189919A1 (Gupta, determining and provisioning paths within a network of communication elements — the " '187 application"); US20060015617A1 (Castro, service-subscription provisioning); US20050097108A1 (Oracle, network data model for an RDBMS); US20050169179A1 (Ericsson, cluster-based network provisioning).

I note literally that US4788421A is, on its face, optical-disk-head velocity control and has no apparent relationship to network path modeling. I do not auto-correct it; I simply observe that it is the one facially irrelevant citation in the list and adds nothing to either side.

Non-patent citations. EPO search report for EP 05857725.5 (2009-08-25); ISR for PCT/US2005/034418 (mailed 2006-12-27); JP Notice of Rejection for JP 2007-534687 (2009-07-31, with English translation). These are procedural, but the JP rejection is a lead: it means at least one foreign examiner raised art against a family member with the same disclosure, and that art (and the applicant's distinguishing arguments) should be pulled from the JP file.

---

4. What the key references actually teach (with the verbatim findings that matter most)

4.1 US 2003/0189919 A1 — Gupta ('187 application; granted as US 7,289,456 B2)

This is the single most damaging reference, and the confirmation search produced two findings that are close to dispositive.

(i) The claim-1 architecture, near-verbatim. The '187's claim 1 recites: "an inventory subsystem for modeling the network as a graph of nodes and links that interconnect the nodes; and a routing engine that uses the graph for determining the path between points in the network; wherein the inventory subsystem models part of an element or a set of one or more elements as the nodes… and wherein the inventory subsystem further models the network links as the links of the graph." That maps, element for element, onto '664 claim 1's "routing manager computer comprising an inventory subsystem and a routing engine," "model each of said at least a first network element … as one or more routing nodes in a graph," and "routing engine … configured to use said graph for determining a path." (RPX Insight, US20030189919A1; US20030189919A1)

(ii) The "cross connection status database" limitation, near-verbatim. The '187's Summary states: "In accordance with a further aspect of our invention, the status of each cross-connection comprising a provisioned virtual trunk/circuit is maintained, which status indicates whether a cross-connection has been successfully provisioned. In the event a virtual circuit/trunk is not successfully provisioned because of one or more failed cross-connections, the circuit/trunk can be re-provisioned by noting the failed cross-connections." Compare '664 claim 1: "a cross connection status database configured to store a status of each interconnection of said plurality of routing nodes, wherein the status indicates whether a cross-connection … was successfully provisioned." The correspondence is not thematic; it is textual. (US 7,289,456, FreePatentsOnline)

The '187 also discloses the data structures the '664 claims: the graph "is maintained, illustratively, in three database tables: routing link table 214, routing node table 216, and NMS/EMS table 218," and FIG. 7 depicts "an illustrative database … for implementing the model." That squarely meets claim 10 ("storing said routing links in a routing link table") and supports the "database" recitations of claims 1, 4, 7, 9 and 14.

What the '187 does not appear to disclose (the gap): modeling the DCS device as a link instead of a node. It models network links and provisioned virtual trunks as routing links, and models network elements as routing nodes. It does not, in the portions I retrieved, say "model the cross-connect as a link."

4.2 US 7,173,912 B2 — Jaber et al. (Fujitsu)

Teaches modeling a node's internal connectivity as links and using those links to determine routing paths through the network: "determining asymmetric connections between traffic bearing components in a network node[;] [a]n intranode connectivity is determined between the traffic bearing components based on the asymmetric connections[;] [a] model of the node indicative of the intranode connectivity is distributed to a disparate node… [and] used at a disparate node in determining a routing path through the network." It further teaches weighting connections by speed and running OSPF over the weighted connections. This is the closest thing in the cited art to an express teaching that the internal cross-connection fabric of a transport node is represented as links in a routing model — i.e., it supplies the missing "DCS-as-link" step. (US7173912B2, Google Patents; US 7,173,912 PDF)

4.3 US 6,981,065 B1 — Nortel

"Provisioning of connection through a SONET/SDH network formed from multiple sub-networks." It expressly frames the problem in OSI layer terms (the same layers the '664 uses to distinguish its layer-1 contribution) and addresses provisioning a connection across sub-network boundaries through intermediate nodes. It is a natural secondary reference for the "first network / second network interconnected via a cross-connect" preamble of claims 1, 2, 5, 7, 9 and 14, and for the provisioning/status aspect. (US6981065, Google Patents)

(Caveat: my confirmation search on US 6,981,065 returned the specification's framing sections rather than its claims. I have not verified whether it describes a cross-connection status database in terms. Its use below is as a layering/sub-network-crossing teaching, not as the primary status reference.)

4.4 US 5,526,414 A and US 6,091,720 A — Northern Telecom / Nortel ("virtual nodes")

These teach the graph-abstraction principle directly: the graph need not mirror physical topology. US 5,526,414 claims a network where a "listing of destination nodes … comprises a virtual destination node representing a group of two or more components, each component being a distinct physical network element," and where "each link from a particular origin node to the virtual destination node is a set of circuit groups." In other words, a physical network element can be collapsed or re-expressed in the routing graph as a node or as a link depending on routing needs. US 6,091,720 is the same family with dynamic intra-link management. (US5526414, SumoBrain; US6091720, FreePatentsOnline)

4.5 US 5,297,137 A (IBM), US 5,377,262 A (AT&T), US 4,669,113 A (AT&T), US 4,284,852 A (Northern Telecom)

Collectively they establish that modeling a communications network as nodes and links and computing paths across that graph was routine by the early 1990s — including hierarchical/collapsed representations (US 4,284,852, cited within US 5,526,414 as an early dynamically-controlled-routing system) and adaptive, periodically-recomputed routing tables (US 5,377,262). They are § 102(b) art of record and provide the "known technique" half of several KSR rationales.

4.6 US 2005/0097108 A1 — Oracle ("Network data model for relational database management system")

Teaches persisting a network model of nodes and links in relational database structures. This is a direct structural teaching for the "routing link table" of claim 10 and for the claimed "database" components of claims 1 and 4. (I flag that my verification search on this reference returned no usable text — treat this as a title/specification-abstract-level characterization only.)

4.7 Secondary references (title-level, lower weight)

US 2004/0107277 A1 (Fujitsu, EMS with tree-structured representations) — management-system representation of network elements and their state, relevant to the inventory/status-DB elements. US 2005/0169179 A1 (Ericsson, cluster-based network provisioning) — clustering network elements into logical provisioning units, further evidence of the abstraction principle. US 2002/0029298 A1 (Wilson); US 2003/0071840 A1 (Huang); US 5,764,740 A (Ericsson); US 2006/0015617 A1 (Castro) — background on management-system provisioning and topology presentation.

Not cited art, but relevant context (from the "Similar Documents" list, not the citation list): Katz et al., SDH management network: Architecture, routing and addressing (1993); Da Costa et al., Planif 2000, a tool for planning SDH networks (2001); Jajszczyk, The ASON approach to the control plane for optical networks (2004). These show that modeling SDH transport networks — including their cross-connects — as routed graphs was an established 1990s–2000s art. I list them as corroboration of the state of the art only; they are not part of the record's citation list and are not relied on below.

---

5. Element-by-element mapping — independent claims

Claim 1 (system)

Claim 4 (means-plus-function routing manager)

Same mapping; the three "means" correspond to the '187's inventory subsystem (means for creating the graph; means for modeling the DCS as a link), its routing engine, and a status store (means for storing the status). The gap is identical: DCS-as-link.

Claim 7 (method — determine DCS interconnections, represent as links, store status)

• "determining … the interconnections created by said digital cross connect": US 6,981,065 (determining connections across sub-networks and intermediate nodes) and '187 (determining "a set of cross-connections required to provision" the trunk/circuit).
• "representing each of said interconnections as a link": US 7,173,912 (node internal connectivity as links); '187 (provisioned connections as routing links).
• "storing a status … whether a cross-connection … was successfully provisioned": '187 — verbatim.

Claim 9 (method — nodes, routing links, DCS as routing link, store status)

Identical to claim 1 minus the database-naming; the '187 supplies four of the five steps verbatim and US 7,173,912 supplies the DCS-as-link step. Dependent claims 10–12 add the routing link table ('187's routing link table 214), determining the path using the routing link ('187 claim 17), and provisioning the path ('187 service activation system 210 / element adapter 212).

Claim 14 (method — generic "common network device")

The broadest and weakest claim: the intermediate device is not limited to a DCS at all. Any teaching of representing an intermediate network device as a link (US 7,173,912) or of graph abstraction decoupled from physical topology (US 5,526,414 / US 6,091,720) reaches it. Claim 14 is the easiest of the independent claims to invalidate.

---

6. Proposed grounds of rejection

Ground A — Primary: Gupta '187 in view of Fujitsu US 7,173,912 (and optionally Nortel US 6,981,065)

The '187 teaches every limitation of claims 1, 4, 7, 9 and 14 except "modeling the DCS/common device as a link." US 7,173,912 teaches exactly that step in the same field — representing a transport node's internal connectivity as links that are used at another node "in determining a routing path through the network." Motivation: both references attack the same problem — reducing the size of the routing graph and the number of hops traversed — and the '187 itself already models physical links and provisioned virtual trunks as routing links, so extending the link abstraction to the statically-provisioned cross-connection fabric of a DCS is the application of a known technique ('187) to a known device (DCS) with a predictable result (fewer nodes, less routing processing), squarely within KSR. Expectation of success is high because both references are graph-modeling references operating on the same transport-network data.

Ground B — '187 in view of Northern Telecom US 5,526,414 (and/or US 6,091,720)

The virtual-node references establish that graph representation is an engineering choice independent of physical topology — a "group of two or more components, each … a distinct physical network element" can be represented as a single node, and a "link" can be a defined set of circuit groups. Motivation: if a POSITA will collapse multiple physical elements into one node to reduce graph complexity, the symmetric step — collapsing a statically-provisioned intermediate device into a link — is an obvious design alternative for the same purpose. This ground is useful because it does not depend on the Fujitsu reference at all, which matters if US 7,173,912 is distinguished on the ground that its "node" is a datagram transport router rather than a layer-1 cross-connect.

Ground C — '187 in view of Oracle US 2005/0097108 (for the database limitations) and Fujitsu US 2004/0107277

Directed at the structural database recitations ("cross connection status database," "routing link table"). The '187 already discloses storing the graph "in three database tables," and Oracle teaches the express node/link relational data model. Motivation: persisting an already-defined graph model in a relational store is a routine implementation choice (KSR; MPEP 2144.04 — rearrangement of parts / obvious use of a known data structure for its established function).

Ground D — '187 in view of Nortel US 6,981,065

Directed at the multi-network preamble (claims 2, 5) and the provisioning steps (claims 8, 12): SONET/SDH provisioning across multiple sub-networks through intermediate nodes, in the same OSI-layer-1 context the '664 invokes.

Ground E — Admitted prior art, standing alone or as a supplement

The '664 specification itself describes the '187 approach as the known prior attempt and states that DCSs "are very well known in the art and … obvious to one skilled in the art." Under In re Constant and the "admitted prior art" line, those statements are usable as the teachings of the '187 regardless of its prior-art qualification (see § 9). This is the fallback ground and it is the one most likely to survive a § 103(c) common-ownership challenge.

---

7. Why a POSITA would have been motivated to combine — the affirmative case

1. Same problem, same field, same stated goal. The '187's stated purpose is to overcome "scalability issues" and "performance issues" caused by modeling ports as nodes and "traversing each physical hop." The '664 recites the identical problem and the identical benefit ("fewer nodes are represented in a network graph … route processing is reduced"). A POSITA optimizing the '187 model has a direct, reference-supplied incentive to keep removing hops. KSR: "any need or problem known in the field."
2. The '187 has already done the same thing twice. It models physical links as routing links, and it models provisioned virtual trunks as routing links. A provisioned cross-connection inside a DCS is, functionally, a semi-permanent connection between two ports — the '664's own specification says a DCS "may be treated as links, such as would be formed by a physical cable." Since the '187 already treats a cable as a link, treating a DCS cross-connection as a link is a substitution of one known element for another to obtain the same predictable benefit (MPEP 2144.04).
3. The '664 supplies the operative fact as an admission. The '664 admits DCS paths "are provisioned in a relatively static manner" and that a DCS is "typically not used to dynamically alter switching over a short time period." A device that makes no routing decision per traffic flow contributes a hop cost with no routing value. Using that admitted fact as the reason to remove the hop is precisely the "predictable variation" rationale.
4. The art expressly frames graph abstraction as a modeling choice, not a physical constraint (US 5,526,414; US 6,091,720; US 2005/0169179). The '664's own FIG. 4→FIG. 5 transition is nothing more than re-labeling an existing FIG. 4 element as a link.
5. Express link-modeling of node internals exists in the same art (US 7,173,912), so the DCS-as-link step is not merely "obvious to try" but is disclosed in a routing-path-determination context.
6. No contrary teaching of record. Nothing in the cited art warns against representing a cross-connect as a link. The closest thing to a "teach away" is US 7,173,912's detailed modeling of node internals — but teaching a more granular model is not a teaching that a coarser model will fail; the '187 and the '664 both embrace the coarser model for the same stated reason. Under Winner Int'l Royalty Corp. v. Wang, 202 F.3d 1340 (Fed. Cir. 2000), a reference does not teach away absent a clear discouragement of the claimed approach.

---

8. Counterarguments for the patent owner, and how they fare

---

9. Prior-art qualification — the real risk to these grounds

This is where the analysis can fail, and it should be stress-tested first.

1. Pre-AIA § 103(c) (common ownership). The '664 issued from an application filed 2005-04-07 and is governed by pre-AIA law. Both the '187 and the '664 were assigned to Telcordia Technologies. Pre-AIA § 103(c) provides that subject matter "developed by another person, which qualifies as prior art only under one or more of subsections (e), (f), and (g) of section 102, shall not preclude patentability under this section" where the subject matter and the claimed invention were commonly owned at the time the invention was made. If the '187 qualifies only under § 102(e), the '187 is barred as § 103 art. This is very likely the reason the examiner did not reject over the '187.
2. The escape routes. (a) If the '664's claims are not entitled to the 2004-09-30 provisional's benefit for this subject matter, the § 102(b) critical date becomes 2004-04-07 and the '187's 2003-10-09 publication becomes § 102(b) art — § 103(c) does not reach § 102(b) art. (b) The '187's content is also in the '664's own specification as the described prior approach, making it admitted prior art usable notwithstanding § 103(c). (c) Grounds B, C, D and the non-commonly-owned portions of A (Fujitsu, Nortel, Oracle, IBM, AT&T) can be re-run with references that raise no § 103(c) issue at all.
3. Different inventive entity. The '187's multi-inventor entity (Gupta/Raheja/Ferrer, per secondary sourcing) makes it "by another" for § 102(e), which is necessary but not sufficient given point 1. This should be confirmed from the face of US 7,289,456.
4. Dating. The '187 was filed 2002-04-08 and published 2003-10-09; both precede the '664's 2004-09-30 priority and 2005-04-07 filing. Dating is not in dispute.

Practical consequence: the examiner's allowance is explained less by the '187's absence of teaching than by its prior-art qualification. A district-court invalidity case should therefore lead with the non-commonly-owned references (Fujitsu US 7,173,912, Nortel US 6,981,065, Northern Telecom US 5,526,414/6,091,720, Oracle US 2005/0097108, IBM US 5,297,137, AT&T US 5,377,262/4,669,113) and use the '187 primarily as admitted prior art and as a roadmap.

---

10. Secondary considerations

• Long-felt need / failure of others: there is no evidence of either; the same problem was already being solved by the '187 and by the virtual-node art, and the '664's own specification concedes that DCS provisioning practices were well known.
• Industry praise / copying / licensing: the assertion campaign (CommWorks-driven, per the litigation section) is licensing by an NPE, not market acceptance, and has no nexus to the claimed path-modeling technique. Under In re GPAC, commercial success without a nexus to the claimed feature gets no weight.
• Reexamination confirmation: addressed in § 8 — weak, no adversarial testing, no § 103(c) adjudication.
• Skepticism/adoption: none of record.

---

11. Bottom line

Claims 1, 4, 7, 9 and 14 are vulnerable to a § 103 challenge, and the vulnerability is concentrated in one limitation. On the face of the record, the '187 application (US 2003/0189919 A1 / US 7,289,456 B2) discloses the "inventory subsystem + routing engine + graph of routing nodes and routing links" architecture and the "cross-connection status … successfully provisioned" limitation in near-verbatim terms. The only ground the '664 genuinely occupies over the '187 is "modeling the digital cross connect as a link instead of a node," and that step is supplied by US 7,173,912 (node-internal connectivity modeled as links for path determination) combined with US 5,526,414/US 6,091,720 (graph representation is a modeling choice independent of physical topology), and is independently motivated by the '664's own admission that DCS cross-connections are static. Claim 14 is the weakest of all, because it recites "a common network device" rather than a DCS.

Two things must be verified before a § 103 position is committed to:

1. Whether the '187 itself describes modeling a DCS/common network device as a link. The record's Definitions block says it does, which would convert this from an obviousness case into an anticipation case under § 102. My retrieval of the '187 publication's own text does not support that reading, but is not a complete review. Pull the '187 full text (US 2003/0189919 A1 and US 7,289,456 B2) and read it end to end.
2. The pre-AIA § 103(c) common-ownership status of the '187. If the '187 qualifies only under § 102(e) and was commonly owned with the '664 by Telcordia when the '664 invention was made, the '187 cannot support a § 103 rejection — which is likely why it was not used during examination. Build the primary grounds on the non-commonly-owned references and use the '187 as admitted prior art.

Evidentiary next steps: pull (a) the '187/US 7,289,456 full text and its filing/priority chain; (b) the JP file (Notice of Rejection, JP 2007-534687) for the art the Japanese examiner applied to the family member; (c) the EPO search report for EP 05857725.5 (2009-08-25) for the EPO's art of record; (d) the ex parte reexamination file 90/015,057 to see exactly which references the examiner considered and whether the '187 was addressed; and (e) the provisional 60/614,609 to test whether the claims are entitled to the 2004-09-30 date — because if they are not, the '187's 2003-10-09 publication becomes § 102(b) art and the § 103(c) problem disappears.