US 7173912

I'll search for current information on this patent number before summarizing.

Search results / docket check — up front

I ran targeted searches for the literal string "7173912" and for "7173912" combined with CAFC/2026 docket terms. Findings:

• No CAFC 2026 docket activity for US 7,173,912 was returned. The CAFC 2026 material my searches surfaced concerns unrelated patents (e.g., the AudienceScience v. Google § 101 appeal involving 7,747,676 / 7,882,175 / 8,082,298; ASSA ABLOY v. CPC; Slingshot v. Canon). None of these mention 7,173,912.
• No district-court litigation or USPTO post-grant proceeding for 7,173,912 was returned either. The only directly relevant hit was the Google Patents page for US7173912B2 itself, which is consistent with the authoritative full text supplied in the prompt.
• Caution: my search results were noisy (largely unrelated foreign patent-office publications). I therefore cannot confirm from live sources the absence of any 2026 filing; I can only report that nothing for this patent number appeared. Given the recorded status — Expired – Fee Related, adjusted expiration 2021-07-10 — active 2026 litigation or CAFC activity would be unusual, since a lapsed patent has no injunctive reach going forward (past damages for the pre-expiration window are theoretically possible but not indicated by anything I found).

Everything below is grounded in the authoritative patent text provided in the prompt.

US 7,173,912 B2 — identification

Sole listed reference: the patent carries no "References Cited" entries in the provided text; the Google Patents page lists 42–46 cited US/EP/WO documents (e.g., US 5,818,840 "Asymmetric ATM switch"; US 6,529,300 "WDM optical network with passive pass-through at each node"; US 6,516,966 "System and method for application object transport" by the same Fujitsu family) and 13 citing documents, none of which I independently verified.

Abstract (verbatim)

"A method and system for providing an internal topology of a node within a network includes determining asymmetric connections between traffic bearing components in a network node. An 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 in a network with a node. The model is used at a disparate node in determining a routing path through the network."

Plain-language overview of the independent claims

Claim 1 — Method (the only method claim)

A four-step method carried out at a network node (the specification's "Internet Protocol Transport node," e.g., an add/drop multiplexer built from RTPs):

1. Discover the asymmetric internal links. Determine the connections between receiver-transmitter pairs (RTPs) inside the node. The claim specifies that each RTP itself contains higher-speed intra-RTP connections joining its internal components, and that those intra-RTP connections run faster than the inter-RTP "asymmetric connections." It further requires that the internal RTP components include an optical receiver and an optical transmitter for interfacing with a WDM system (in the spec: 10 Gb/s receiver/transmitter without SONET knowledge, optionally amplified and directly interfacing a WDM system, with 155 Mb/s mux/demux ports).
2. Derive intranode connectivity. From those asymmetric links, work out which RTPs can reach which other RTPs inside the node (the spec's approach: weight the links and run an internal OSPF instance).
3. Advertise a model of the node. Distribute a model reflecting that internal connectivity to another ("disparate") node in the same network — per the spec, encoded/flooded in opaque LSAs.
4. Use the model for routing. The receiving node uses the model when computing a routing path through the network.

Claim 8 — System, "means for" form

Substantively the same four functions as claim 1, but recited as a system using means-plus-function language: means for determining the asymmetric inter-RTP connections (same intra-vs-inter speed relationship and same optical receiver/transmitter-with-WDM requirement), means for determining intranode connectivity, means for distributing the node model, and means for using the model at the disparate node to determine a routing path.

Claim 15 — System, "logic encoded in media" form

Also substantively the same four functions as claim 1, but as a system comprising logic encoded in media (i.e., software/firmware-implemented logic) operable to determine the asymmetric RTP connections, determine intranode connectivity, distribute the node model, and use it at the disparate node for routing. Same intra/inter speed asymmetry and same optical-receiver/optical-transmitter-with-WDM limitation as claims 1 and 8.

Effect of the dependent claims (2–7, 9–14, 16–21, paralleling one another across the three independent claims): they add that the RTPs include lower-speed interfaces to external nodes (2/9/16); that all possible connectivity between RTPs is determined (3/10/17); that the model is distributed via opaque LSAs (4/11/18); that the network is a private network (5/12/19); that connectivity is determined by weighting links by speed (6/13/20); and that a lower weight is assigned to higher-speed links and a higher weight to lower-speed links, with OSPF then run on the weighted connections at the receiving node to pick the path (7/14/21).

Observations and uncertainty flags

• The core inventive distinction is the asymmetry. The specification notes the node's intra-RTP links are high-speed (~10 Gb/s) while inter-RTP connections and external (POS) interfaces are lower-speed (155 Mb/s / 2.5 Gb/s and below). The patent's contribution is treating that non-uniform internal fabric as a routable topology, weighting it, and flooding it to peers, rather than hiding it behind a flat node abstraction. The specification's framing: "the transport network provides a flexible node topology… Asymmetric intranode connectivity may be modeled and advertised within the transport network and used for internal routing."
• Terminology inconsistency worth noting. The claims recite "intranode connectivity" in claims 1, 8, and 15, but dependent claims 3, 6, 10, 13, 17, and 20 refer to "internode connectivity between the RTPs" — which reads as internal-connectivity discovery despite the "inter" prefix. The specification itself contains the same looseness ("Internode topology is identified by the RTP/interface connections"). I have not verified how this was resolved during prosecution or in any later construction, so I flag it as a drafting ambiguity rather than a legal conclusion.
• RTP expansion varies. The claims say "receiver transmitter pairs"; the specification's Detailed Description says "receiver-transceiver pairs." I am not auto-correcting either form — both appear in the authoritative text.
• Unverified items. I could not independently verify the assignment chain, the cited art list, or the "expired – fee related" status from a live USPTO source in this session; those details come from the supplied Google Patents text. Treat them as the authoritative record provided, but confirm via USPTO Patent Center / Global Dossier if you need a certified current status.

I'll search for litigation involving US patent 7173912 across patent litigation sources.

Let me try more targeted litigation searches.

Let me check patent litigation databases directly.

Search Results: Litigation Involving U.S. Patent No. 7,173,912

Bottom line: I found no evidence of any litigation involving U.S. Patent No. 7,173,912 in the sources I searched. I'm reporting this as a negative finding rather than guessing.

What I searched

• General web searches for the patent number with litigation/lawsuit terms (e.g., US patent 7173912 litigation, "7173912" patent lawsuit, "7,173,912" patent infringement case)
• Searches combining the assignee (Fujitsu) with the patent number and litigation terms
• Searches targeting the patent's title ("Modeling and advertising asymmetric topology")
• Search results returning court documents (CourtListener/RECAP, PACER-derived dockets, District Court opinions)

Why the negative result is plausible

Several data points from the authoritative patent record (Google Patents) are consistent with the patent never having been asserted:

1. The patent is expired and was never part of a litigation family. The record shows legal status "Expired – Fee Related," with an adjusted expiration of 2021-07-10. Sole original assignee/current assignee is Fujitsu Limited (Fujitsu Network Communications, Inc. before a 2005 reassignment).
2. The "Cited By" list (13 items) contains no litigation-driven follow-on filings, and none of the third-party citations are patent-infringement complaints naming this patent.
3. The priority chain (Provisional 60/202,190, filed 2000-05-05; PCT/US2001/014626; AU2001257550) shows no related enforcement actions in the record.
4. The prior-art/keyword content ("network," "node," "connections," "rtps," "transport") is infrastructure/telecom routing technology; the co-owned applications cited in the specification (filed Jun. 6, 2000) are the ones more commonly associated with Fujitsu's transport-network work.

Important caveats — what I could NOT verify

• Unified Patents litigation portal (portal.unifiedpatents.com/litigation/caselist) and Docket Navigator were not successfully queried in my available searches. These are the most reliable sources for exhaustive U.S. patent litigation histories, and I could not confirm a null result from them directly.
• I could not run a full PACER/CAFC docket search for this patent number.
• Search engines returned several false positives for the string "7173912" (a Lithuanian business registry notice, a Delhi University exam roll number, a New York LLC DOS ID, a CodeCanyon asset ID, and Brazilian notary records). None of these relate to the patent and should be disregarded — I'm listing them only to flag that raw-number hits are noise.

Recommended next steps for a definitive answer

To confirm the negative result with authority, run these directly:

1. Unified Patents litigation caselist — filter by patent number 7173912.
2. PACER / CourtListener RECAP — search "7,173,912" and "7173912" across all district courts.
3. CAFC docket search for any appeal naming the patent.
4. USPTO Patent Center / Assignment records — check for any recorded security interests or assignments that might signal licensing/enforcement activity (none appeared in the record above).
5. Docket Navigator / Lex Machina if accessible.

Conclusion: Based on the searches performed, there is no known litigation involving U.S. Patent No. 7,173,912. However, because I could not directly query the Unified Patents portal, PACER, or CAFC dockets — the definitive sources — I would characterize this as "no litigation found" rather than a fully verified "no litigation exists." I did not identify any plaintiff, defendant, jurisdiction, case number, filing date, or outcome for this patent.

Proceedings overview

Total AIA trial proceedings on file: 0. The USPTO Open Data Portal (ODP) API returns no IPR, PGR, or CBM proceeding for US 7,173,912, and independent web searching surfaced no PTAB proceeding number naming this patent — so the breakdown is 0 active / 0 claims invalidated / 0 claims sustained / 0 settled / 0 institution denials. The bottom line for a defendant is not "hardened patent" but something stronger: the patent expired on 2021-07-10, and it has never been tested at the PTAB — so there is no IPR record to lean on, but also no live claim to enjoin, and only a ~9.5-month sliver of pre-expiration damages remains numerically recoverable if a complaint were filed today.

Data provenance note: the structured block in this prompt is the canonical proceedings list and it is empty. My searches (Google Patents family/PTAB data for US7173912; queries combining the patent number with "IPR," "PTAB," "inter partes review," "reexamination," and the inventor names Jaber/Harbin) returned no AIA trial proceeding. I am reporting "no PTAB activity on file." If your docket or a demand letter says otherwise, treat that as new information to be verified on PTAB E2E before relying on this memo.

Proceedings detail

No proceedings on file

There is nothing to report per proceeding. For completeness, the non-events that matter:

• No IPR / PGR / CBM ever filed. Tabulating from the record: US 7,173,912 issued 2007-02-06, and IPR/PGR/CBM practice opened 2012-09-16. That is roughly a nine-year window (2012-09-16 → 2021-07-10) in which any competitor could have petitioned, and none did.
• No ex parte or inter partes reexamination surfaced in the prosecution-file metadata (the 42-references examination-art list on Google Patents shows only original examination citations, not reexam certificates).
• No Federal Circuit appeal exists, because there was no Board decision to appeal.

Status of the patent itself (verbatim from the structured data, with gloss):

• Status: "Expired - Fee Related" / "Adjusted expiration 2021-07-10" — gloss: the 20-year term from the 2001-05-04 filing date, plus patent term adjustment, ran out on 2021-07-10. As of today (2026-09-27) the patent is expired.
• Filing: 2001-05-04 (App. No. 09/848,871) · Priority: 2000-05-05 (Prov. 60/202,190) · Granted: 2007-02-06
• Assignee: Fujitsu Limited (originally Fujitsu Network Communications, Inc.) — the patent sits in an operating company's portfolio, not a troll's.
• Foreign family: WO2001086867A2 (PCT/US2001/014626, "Ceased"), AU2001257550A1 ("Abandoned").
• Claims: 21 total — independent claims 1, 8, 15; dependent claims 2–7, 9–14, and 16–21.

Wrong-patent flag — read this before anything else. Search traffic for "the '912 patent" is heavily polluted by US 7,619,912 (Netlist, Inc.), a memory-module patent whose claim 16 was held unpatentable in IPR2022-00615 on 2024-04-17, and which is the subject of the Netlist v. Samsung appeal-stay fight, e.g. CourtListener docket 64861664. That is a different patent with a similar number and no relationship to this Fujitsu transport-network patent. If a demand letter or an internal summary cites "the '912 patent" as already-invalidated, verify the full number before relying on it.

Strategic summary

Claim status: nothing canceled, nothing sustained, everything UNTESTED. Because no AIA trial was ever instituted, all 21 claims — independents 1, 8, and 15 and their dependents — stand in their original, never-adjudicated form. There is no IPR estoppel, no Board claim construction, and no FWD to cite. That cuts both ways: you cannot tell a court "the PTAB already killed claim 1," but the patent owner also has zero PTAB validation to wave at a jury. Any validity attack you mount would be the first-ever merits adjudication of these claims, litigated de novo in district court under clear-and-convincing evidence.

Estoppel landscape: clean, but largely academic. § 315(e)(2) estoppel is petitioner- and privy-specific, and with zero petitioners there is zero estoppel — no ground is foreclosed to you. Every § 102/§ 103 combination, plus §§ 101 and 112, remains available. The real constraint is not estoppel but economics: because the patent expired 2021-07-10, prospective relief is off the table, and the § 286 six-year damages lookback from a complaint filed today (2026-09-27) reaches back only to 2020-09-27 — leaving roughly a nine-month accrual window (2020-09-27 → 2021-07-10) before the term ran out. (That window shrinks every day and vanishes entirely for complaints filed after 2027-07-10. This is my derivation from § 286, not a holding I can cite.) Practically, that means the cheapest resolution is usually a short pre-suit damages math problem, not a merits defense.

Pattern signals. No repeated petitioner, no defensive aggregator (no Unified Patents, RPX, or similar entity) in the chain, and no patent-owner PTAB appeals — because there was never a proceeding to appeal. The patent is a 2000-vintage, pre-AIA Fujitsu Network Communications filing; its "Cited By" list is dominated by large operating companies building on the disclosure (Cisco's US 20080080494 and US 20080198854 on OSPF transit-network hiding and distributed-router local packet transport; Cisco US 2015/0010301 on advertising Layer 0 topology into Layer 3; ADVA's US 9628367 on traffic engineering over asymmetrical network switches; STMicroelectronics US 20060123154 on modular data transfer architecture). That citing pattern fits a foundational, widely-read disclosure that nobody bothered to challenge — consistent with a patent that was never a litigation cash cow.

Recommended next steps

1. If you are a defendant today, lead with expiration and the damages window, not invalidity. Confirm the term on USPTO PatentCenter (application 09/848,871) and the maintenance-fee/expiration entries, then run the § 286 math for your proposed filing date. On the face of the structured data, the recoverable period is at most 2020-09-27 → 2021-07-10.
2. Do not cite an IPR to the court. There is no FWD. Any statement that claims 1–21 (or any subset) were invalidated at the PTAB would be a misstatement — and if the source of that belief is "the '912 patent," you are almost certainly looking at Netlist's US 7,619,912, not this patent.
3. If you nonetheless need an invalidity position, note that a § 102/§ 103 attack is fully open — no estoppel, no prior Board construction, and a 2000 priority date with a deep field of 1990s ATM/MPLS/WDM routing art (the examiner's own 42 references, including US 5,818,840 "Asymmetric ATM switch," are a starting point, not a ceiling). Weigh that against the fact that claim 1 was drafted with the WDM optical receiver/transmitter interface language and the intra-RTP-higher-speed-than-inter-RTP limitation, which is where a § 103 case will be won or lost.
4. If you have a demand letter citing US 7,173,912, treat it as a stale-claim assertion: the claims cannot support ongoing royalties or an injunction. Ask for the accused-instrumentality period and check whether any of it postdates 2021-07-10.
5. Keep watching for AIA activity only as a formality. With the patent expired, IPR/PGR filings are commercially implausible; a fresh search of PTAB E2E and CourtListener is a five-minute check if the situation changes.

I'll research the assignment record for US 7173912 across the sources requested.

Let me dig further into the assignment records and related litigation history.

Let me verify the inventors and the corporate status of the original assignee.

Ownership & Assignment Analysis — US 7,173,912 B2

Data-sourcing caveat (read first). I was unable to retrieve the raw USPTO Assignment Center records (assignmentcenter.uspto.gov / assignment.uspto.gov) directly — repeated queries returned the Google Patents mirror and third-party indexes but not the Assignment Center's reel/frame detail records, and I hit the tool-call ceiling before completing the cross-reference. Everything below is grounded in the Google Patents legal-events mirror bundled with the authoritative patent text, plus corroborating public records. Where the mirror does not expose a data element (notably correspondent of record, execution dates, and the 2005 reel/frame), I say so explicitly rather than guess. Those gaps should be closed by opening the two records at the Assignment Center before relying on this as a litigation-grade ownership opinion.

Inventors

Employer at time of filing: Fujitsu Network Communications, Inc. (FNC), Richardson, Texas — all five appear as assignors to FNC on the same 2001-09-04 record, which is the signature pattern of a standard employee invention-assignment program rather than an inventor-held portfolio.

Unusual patterns: None detected. There is no evidence of inventors departing within 12 months of filing — the opposite, in fact: co-inventor Harbin resurfaces on a contemporaneous, co-owned FNC application in the same "IP transport" family (the specification also cross-references four co-owned FNC applications filed 2000-06-06, including one that issued as US 6,515,966 B1 to FNC). A team that keeps filing together is not a pre-fire-sale signature. I cannot verify post-2001 employment histories from the sources retrieved, so I am flagging the inventor-departure question as not observable, not as no departures occurred.

Original assignee

Entity on the issued patent (2007-02-06): Fujitsu Limited, Kawasaki, Japan (1-1, Kamikodanaka 4-chome, Nakahara-ku, Kawasaki-shi, Kanagawa 211-8588) — per the Google Patents "Original Assignee / Current Assignee" fields and the recorded 2005 assignment.

Entity on the application at filing (2001-05-04): Fujitsu Network Communications, Inc. (FNC), Richardson, TX — the inventors assigned to FNC on 2001-09-04, and FNC then conveyed to Fujitsu Limited on 2005-12-06, before issuance. So "original assignee" depends on the reference date: FNC at filing, Fujitsu Limited at grant.

• Primary line of business: FNC was the U.S. optical/IP transport arm of the Fujitsu group — carrier-class transport gear and management software (FLASHWAVE-family optical transport, NETSMART management). Fujitsu Limited is the global parent (computing, networking, telecom infrastructure).
• Did they ship a product embodying the claims? The patent claims an intranode topology model for a transport node built from 10 Gb/s receiver-transmitter pairs (RTPs) with internal optical receiver/transmitter interfaces to a WDM system, feeding OSPF-based path selection (claims 1, 7, 21). This is the architecture of the Fujitsu "IP transport" node described in the same specification and in the co-owned applications referenced above — i.e., a real, shipped product line, not a paper asset. I state this as a strong inference from the specification's internal consistency, not as a product-name citation I could independently verify within the search budget.
• Current status: Operating. Fujitsu Limited remains an active global public company; the patent itself is recorded by Google Patents as "Expired – Fee Related," adjusted expiration 2021-07-10. FNC operated as a Richardson, TX subsidiary and, per ARIN records, still held its own network/ASN registration (AS16831, NET-168-127-0-0-1, registered 1994-04-21) — concrete evidence FNC was an operating carrier-equipment business with its own network presence, not a shell. FNC's U.S. operations were subsequently consolidated into the Fujitsu group's North American business; I could not verify the exact restructuring date(s) and will not invent them.

Assignment timeline

Two recorded events, both pre-issuance and both intra-group:

• Executed date not disclosed / recorded 2001-09-04 — Reel 012151/0242

  • Conveyance: Assignment (Assignment of Assignor's Interest). Recorded as a re-record "to correct the number of microfilm pages from 6 to 7 at reel 012151, frame 0242." The source does not disambiguate whether 012151/0242 is the original frame or the corrected frame.
  • Assignor: Abed Mohd Jaber; Edward Alton Harbin; Nimer Ibrahim Yaseen; Li Mo (NMI); George E. Bucklin
  • Assignee: Fujitsu Network Communications, Inc.
  • Correspondent: Not exposed in the sources retrieved — unknown/unverified. I will not name an attorney I cannot source. (The mirror supplies no correspondent field; this is precisely the field the Assignment Center record would supply.)
  • Context: Standard employee invention assignment to the employer at filing — not an acquisition, not a fire-sale.

• Executed date not disclosed / recorded 2005-12-06 — Reel NNNNNN/NNNN: not disclosed in the sources retrieved

  • Conveyance: Assignment (Assignment of Assignor's Interest)
  • Assignor: Fujitsu Network Communications, Inc.
  • Assignee: Fujitsu Limited
  • Correspondent: Not exposed in the sources retrieved — unknown/unverified.
  • Context: Internal corporate reorganization / consolidation of U.S. patent holdings under the parent, effected ~14 months before issuance (2007-02-06). No third party entered the chain.

No post-issuance assignment is recorded. No security agreement, license, merger, name change, or release appears anywhere in the record. The patent is recorded as Expired – Fee Related, adjusted expiration 2021-07-10, with Fujitsu Limited still shown as current assignee — i.e., the chain never left the Fujitsu corporate family and the asset lapsed rather than being monetized. The PCT/AU siblings in the same priority family (WO 2001/086867 A2; AU 2001257550 A1) are recorded as Ceased / Abandoned with no assignments of their own.

Timeline diagram

timeline
    title Ownership of US 7173912
    2000 : Provisional filed by Fujitsu Network Communications
    2001 : Non-provisional application filed
         : Inventors assign rights to Fujitsu Network Communications
    2005 : Assigned to parent Fujitsu Limited
    2007 : Patent issued to Fujitsu Limited
    2021 : Patent recorded as expired

NPE / troll-pattern signals

1. Shell-entity transfer — NOT PRESENT. The only two assignees are Fujitsu Network Communications, Inc. (an operating carrier-equipment subsidiary at Richardson, TX, holding its own ARIN network registration AS16831) and Fujitsu Limited (Kawasaki, JP). No suffix "IP / Patents / Licensing / Holdings / Ventures" appears; no registered-agent address; no single-member LLC. Basis: reel 012151/0242 (2001) and the 2005-12-06 record.
2. Known asserter in the chain — NOT PRESENT. No assignee in the record matches Acacia, Marathon, Intellectual Ventures, IPNav, Wi-LAN, Mosaid/Conversant, Vringo, Pendrell, Innovatio, MPHJ, Lumen View, Round Rock, Document Generation Corp, or any Spangenberg-linked entity. Chain is Fujitsu-only.
3. Repeat correspondent across the chain — UNKNOWN / NOT ASSESSABLE. The correspondents of record are simply not exposed in the sources I could retrieve. With two intra-group records and no correspondent data, recurrence cannot be tested. Flagged as a data gap, not as a negative finding — this is the single field most worth pulling from the Assignment Center.
4. Cascading transfers — NOT PRESENT. Two assignments total, ~4.3 years apart (2001-09-04 → 2005-12-06), both to related Fujitsu entities. No chained LLCs, no common-agent addresses, no sub-24-month sequence.
5. Pre-litigation transfer — NOT PRESENT. No infringement suit naming US 7,173,912 was located. The one Fujitsu-network litigation surfaced in search (Tellabs Operations, Inc. v. Fujitsu Limited and Fujitsu Network Communications, Inc., N.D. Ill. Civil Action No. 08 CV 3379) has Fujitsu as the accused defendant on Tellabs' patents, not Fujitsu asserting the '912. A 2005 assignment cannot be "pre-litigation" for a suit that does not exist.
6. Bankruptcy fire-sale — NOT PRESENT. No Chapter 7/11 proceeding, no patent sale in bankruptcy, no Kodak/Nortel/Polaroid-style transaction. Assignor FNC and assignee Fujitsu Limited were both solvent operating entities throughout.
7. Privateering — NOT PRESENT. The transferee is the parent operating company, not an NPE, and there is no evidence of assertion by an NPE on Fujitsu's behalf. The chain's direction (subsidiary → parent) is the opposite of the privateering pattern (operating company → shell asserting against competitors).
8. Defensive aggregator (anti-NPE) — NOT PRESENT. The chain does not terminate at RPX, AST, LOT Network, Unified Patents, or OIN. It terminates at Fujitsu Limited and the patent lapsed for fee.

Verdict

Insufficient data — for NPE-screening purposes, and I want to be precise about why this is the honest pick rather than a hedge.

The record contains no post-issuance assignment at all; the only two entries are pre-issuance, intra-group predecessor records (reel 012151/0242 dated 2001-09-04, and the 2005-12-06 transfer to Fujitsu Limited) between two operating companies in the same corporate family. None of the four substantive verdict categories fits the facts: there is no second assignee, no LLC, no known asserter, no defensive aggregator, and — critically — no infringement suit naming this patent, so "operating-company assertion" (which requires the current assignee to be suing actual competitors) is not supported either. The affirmative finding worth carrying forward is negative and clean: no NPE pattern is present, the patent never left Fujitsu, and it is recorded as expired for failure to pay maintenance fees as of 2021-07-10.

Verify at the Assignment Center search page (search by patent number 7173912): https://assignmentcenter.uspto.gov/ (legacy interface: https://assignment.uspto.gov/patent/index.html). When you pull the two records, capture the correspondent of record for each — that field is the one gap that could change signal #3 from unknown to a finding.

I'll start by verifying the patent record and gathering prior-art details on the cited references.

I have the patent record. Let me pull details on the most technically relevant citations.

1. Patent record confirmed (identifiers read literally)

Source: https://patents.google.com/patent/US7173912/en and the granted PDF (US7173912.pdf). This is the correct record — no similar-number substitution was made.

Critical date for prior-art purposes: because the application claims benefit of provisional 60/202,190, the effective date for §102 analysis of the claims is May 5, 2000 (subject to whether each claim is actually supported by the provisional — see caveats in §5).

2. Scope / methodology note (important)

Two citation lists appear in the record and they do not fully agree, which I am not auto-correcting:

• "Citations (42)" and "Patent Citations (46)" — the retrieved text was truncated partway through the 46-item list (it cuts off at US6205158B1), so I could not verify every one of the 46 entries byte-for-byte. I have reproduced every reference actually visible in the record below.
• The "Cited By (13)" and "Families Citing this family (29)" lists are forward citations (later documents citing US 7,173,912). Those are not prior art to US 7,173,912 and are excluded from the §102 analysis.

I could not independently re-verify every bibliographic field against USPTO PatentCenter within this session; the dates/assignees below are taken from the citation tables in the patent record. Where a field is the priority date rather than the publication date, I say so.

3. Prior art assessment — Tier 1 (references most relevant to the claimed subject matter)

4. Prior art assessment — Tier 2 (internal-topology, distributed-routing and modeling art)

5. Prior art assessment — Tier 3 (remaining cited references; background/general art)

These are the balance of the cited list. Each is a real §102(b)-eligible reference (all published before 2000-05-05 unless noted), but none appears to touch the specific asymmetric-intranode-topology modeling/advertising combination:

6. Bottom line on anticipation (§102) vs. obviousness (§103)

No single cited reference appears to disclose every limitation of independent claim 1 (or its counterparts, claims 8 and 15), so a clean §102 anticipation is not evident from the cited art. Claim 1 requires, in combination:

1. asymmetric connections between receiver-transmitter pairs (RTPs) in a network node;
2. intra-RTP connections between internal RTP components that are higher speed than the inter-RTP asymmetric connections;
3. internal RTP components comprising an optical receiver and an optical transmitter for interfacing with a WDM system;
4. determining intranode connectivity between the RTPs from those asymmetric connections;
5. distributing a model of the node to a disparate node; and
6. using the model at the disparate node to determine a routing path.

• The "intranode connectivity → model → routing" core (elements 4–6) is most closely met by US 2002/0080775 A1 / US 6,996,116 B2 (IBM) — but that reference is dated 2000-11-22, i.e., after the 2000-05-05 provisional date, so it is §102(e)/§103 material only for any claim not entitled to the provisional's benefit.
• The WDM optical receiver/transmitter element (3) is most closely met by US 6,529,300 B1 (Nortel) — relevant to that sub-element, not to the whole claim.
• The "asymmetric connections" element (1) is most closely met by US 5,818,840 A (Time Warner) — but in an ATM-switch context, with no intranode-topology advertisement.

The realistic risk profile is therefore §103 obviousness (e.g., IBM's intra-node topology routing in view of Nortel's WDM add/drop optical interfaces and Time Warner's asymmetric switch), not §102 anticipation. Elements 2 (RTP-specific speed asymmetry) and 5 (opaque-LSA distribution of a node model) are the strongest differentiators for patentability and are not clearly shown by any cited reference.

Dependent-claim notes: the weighting limitations (claims 6, 7, 13, 14, 20, 21 — first weight for higher-speed connections, higher weight for lower-speed connections, then OSPF over the weighted graph) map onto standard link-metric/OSPF practice, and US 5,781,534 A (path characteristics) plus US 6,016,306 A (bandwidth-reserved routing) are the most plausible §102(b)/§103 supporting art. The opaque-LSA limitation (claims 4, 11, 18) has no close single-reference match among the cited art.

7. Explicit caveats / things I could not confirm

• Truncation: the retrieved record's "Patent Citations (46)" table was cut off after US 6,205,158 B1; I confirmed roughly 44 unique cited references across the two lists but cannot certify the exact 46th/45th entries. I did not invent or backfill them.
• List discrepancy: "Citations (42)" and "Patent Citations (46)" differ (e.g., US 5,483,522 A, US 6,262,976 B1, US 2001/0025310 A1, US 6,331,905 B1, US 2002/0080775 A1, US 6,947,415 B1, US 6,359,857 B1 appear in only one of the two). I have preserved both rather than reconciling them.
• Co-owned same-priority documents (US 6,515,966 B1; US 2001/0052029 A1) are not §102 prior art to this patent despite appearing in the citation lists.
• Forward citations ("Cited By (13)", "Families Citing this family (29)") are excluded — they are not prior art.
• I could not complete a live USPTO PatentCenter fetch of the examiner-cited "References Cited" list within this session; the analysis above is grounded in the patent record text provided and the corroborating searches cited (patents.google.com/patent/US7173912/en; patents.google.com/patent/US6529300; patents.google.com/patent/US5818840; patents.google.com/patent/US20020080775; patents.justia.com/patent/6996116).

If you want, I can next produce a formal element-by-element (limitation-by-limitation) claim chart mapping each claim-1 limitation against IBM US 6,996,116 and Nortel US 6,529,300 for §103 purposes.

I'll ground the analysis by verifying what the key cited references actually disclose.

Now let me verify several more of the cited references that map to specific claim limitations.

Obviousness Analysis — U.S. Patent No. 7,173,912 B2

Claimed subject matter: modeling and advertising asymmetric intranode topology (RTP-to-RTP connectivity) so remote nodes can compute end-to-end paths through the node's internal fabric.

Key bibliographic facts (per the page): App. No. 09/848,871, filed 2001-05-04; priority 2000-05-05 (prov. 60/202,190); granted 2007-02-06; assignee Fujitsu Ltd (originally Fujitsu Network Communications, Inc.); adjusted expiration 2021-07-10, status "Expired – Fee Related." 21 claims: independent method claim 1, means-plus-function system claim 8, and logic-in-media system claim 15, with parallel dependent sets.

Analysis framework: Graham v. John Deere factors as applied under KSR Int'l v. Teleflex (2007) and MPEP § 2143 — scope/content of the prior art, differences from the claims, PHOSITA level, and objective indicia. Where I could not independently verify a reference's disclosure during this session, I say so explicitly.

1. Preliminary caveats on the "Prior Art section"

Three important filters before the mapping:

1. "Cited By (13)" and "Families Citing this family (29)" are forward citations. They post-date 2000-05-05 and are not prior art against the '912 patent. This matters because one of them — US 2015/0010301 A1 (Cisco, "Advertising Layer 0 Network Topology Information to a Layer 3 Network") — is superficially the closest-sounding art of all, but is 15 years too late.
2. Two items on the page appear to be sibling/commonly-owned applications, not prior art. US 2001/0052029 A1 ("Unique address space and method for a transport network," Harbin — a named '912 inventor, priority 2000-05-05) and US 6,515,966 B1 (Fujitsu Network Communications, "System and method for application object transport," priority 2000-05-05) are not "by another" and share the same priority date; they should not be treated as § 102/§ 103 art. Under pre-AIA § 103(c)(1), the commonly-owned Fujitsu references of record (e.g., US 6,317,426 B1; US 6,353,593 B1) may also be disqualified for § 103 combination purposes if commonly owned at the time the invention was made.
3. The specification contains significant admitted prior art — OSPF/IS-IS/RIP peering, NAT, SONET framing, class-of-service mapping, and the co-owned applications filed Jun. 6, 2000. This is usable as "applicant's own admission" prior art even though those applications are not in the "Patent Citations (46)" list.

2. Claim 1 element mapping

Bottom line: no single reference of record discloses all of claim 1; a three-way combination does.

3. Recommended combinations and the motivation to combine

Combination A (primary — strongest)

US 5,818,840 A + US 6,529,300 B1 + US 5,825,772 A, optionally + US 6,016,306 A or US 5,951,634 A.

• Motorola-style "predictable combination of known elements" (KSR) — each reference is drawn from the same art (H04L/H04Q packet/cell switching and optical transport) and each element performs precisely its known function: the asymmetric fabric supplies asymmetric connectivity; the WDM interface supplies optical TX/RX to the WDM plant; the link-state protocol supplies discovery + distribution + metric-based path computation. No element's function is altered.
• Articulable motivation for the asymmetry element: add/drop traffic in a transport node is inherently asymmetric (heavy express/through traffic vs. lighter local add/drop), so an asymmetric fabric reduces port count and cost — this is US 5,818,840's express rationale ("size and complexity of ATM switches for such networks can be reduced by reducing the number of input connectors on one side … and output connectors on the other").
• Articulable motivation for the WDM element: both references target the same problem — removing SONET framing/termination penalty and obtaining protocol/bit-rate-independent high-speed bit pipes. US 6,529,300 B1 frames this explicitly ("each wavelength essentially provides a protocol independent high speed bit pipe"; SONET "makes upgrading the capacity of the ring a slow and costly process"). The '912 specification states the same goal (direct optical transmission "without SONET framing and its associated overhead").
• Articulable motivation for the advertising element: once the node's internal fabric is asymmetric and non-uniform in speed, a topology model that treats the node as a black box produces incorrect path computations — a path that appears to traverse the node may be internally impossible or slower than an alternative. A PHOSITA seeking to make end-to-end shortest-path computation accurate is therefore led directly to advertise the internal connectivity. Cabletron supplies the mechanism and, critically, already models the network at port granularity with metric/cost/bandwidth link states, which is the same abstraction the '912 uses ("the router is a point without latency between points with bandwidth restrictions").
• Reasonable expectation of success: all three technologies were mature and interoperable by the 2000 priority date (ATM/POS over SONET and WDM; OSPF/IS-IS-area link-state flooding); there are no disclosed incompatibilities.

Combination B (alternative, all "node-internal resource management" lineage)

US 5,818,840 A + US 6,529,300 B1 + US 5,483,522 A / US 5,425,021 A (IBM, "Packet switching resource management within nodes") + EP 0 512 495 A2 (Fujitsu, "Switching node in a network with label multiplexed information").

• EP 0 512 495 A2 verified: a switching node in a label-multiplexing network coupling m input and n output logical communication paths, with separate call-control (connection-oriented) and routing (connectionless) means — i.e., a node whose internal path structure is explicitly modeled and differently handled per traffic class. This is a same-field, same-assignee reference and is useful for the "internal topology is itself a routable structure" proposition.
• Motivation: the IBM references supply the node-internal resource/bandwidth bookkeeping that makes identifying possible intra-node connectivity meaningful (claim 3/10/17), and a node whose internal paths are separately controlled per class (EP '495) naturally wants that structure advertised for end-to-end path selection.

Combination C (for the weighting/shortest-path dependent claims)

US 5,825,772 A + US 5,951,634 A (IBM, multi-priority line-bandwidth optimization) + US 6,016,306 A (IBM, bandwidth-reserved routing), on top of A or B.

• Cabletron's verified disclosure that link-state "changes include … metrics, 'cost' or bandwidth" and that path selection returns the "lowest aggregated metric" path, including equal-cost paths, is a direct read on claims 7/14/21 ("first weight for higher speed connections and a second higher weight for lower speed connections … utilizing open shortest path first on the weighted connections"). Weighting a faster link with a lower OSPF cost and a slower link with a higher cost is the definition of an OSPF link metric — a design choice within the ordinary skill, and the '912 specification treats it as such.

4. Dependent claims

Note a drafting defect worth recording: claims 1 and 8 recite "intranode" connectivity, while dependent claims 3, 6, 10, 13, 17 and 20 recite "internode" connectivity "between the RTPs." Read literally, the dependents do not further limit the independents' intranode determination. That is a § 112 written-description/enablement and indefiniteness concern, not a § 103 concern — but it changes how much weight the dependents carry.

5. Counter-arguments and residual non-obviousness risk

1. Granularity gap. No reference of record clearly models the node at RTP granularity, let alone the specific "intra-RTP faster than inter-RTP" relationship. If the patentee argued that the invention resides in the unit of abstraction rather than in the asymmetry per se, an examiner might have been persuaded — but KSR rejects the notion that a change in the level of abstraction of an otherwise known mechanism is enough to confer non-obviousness.
2. Possible "teaching away" from Nortel. US 6,529,300 B1 emphasizes passive forwarding at intermediate nodes ("communication can be established directly between a pair of nodes … without the active intervention of any intervening node"). A skilled advocate could argue this teaches away from a node that internally switches and must advertise granular intranode connectivity. Rebuttal: the '912 node is an add/drop multiplexer that does terminate/add/drop traffic; the Nortel teaching is about bands not destined for the node, so it is not inconsistent and does not criticize, discredit or otherwise discourage the '912 approach.
3. No objective indicia identified. There is no evidence on the face of the record of commercial success, long-felt need, licensing, or industry praise having a nexus to the claims. The status "Expired – Fee Related" (adjusted expiration 2021-07-10) is consistent with low commercial significance, though this is not a legal test and should not be over-read.
4. No prosecution-history or examiner-reason-for-allowance review. I have not examined the file wrapper, the claimed priority to prov. 60/202,190, or whether any of the 46 references were actually applied. Any real § 103 opinion would require that, plus a full independent prior-art search (which would almost certainly find RFC 2370, RFC 2328, and the OSPF-metric art outside the page's list).

6. Conclusion

Claims 1, 8 and 15 are likely obvious over US 5,818,840 A in view of US 6,529,300 B1 and US 5,825,772 A (optionally with US 6,016,306 A or US 5,951,634 A). The combination supplies (i) an asymmetric node-internal connection topology, (ii) optical receive/transmit interfaces to a WDM system on the node's internal components, and (iii) distributed, link-state-based advertisement of per-port topology followed by metric-based path computation at a remote node — with a concrete, articulable motivation (accurate end-to-end path computation across a non-uniform internal fabric) and a predictable result. Dependent claims 2–7 and 9–14 and 16–21 add only conventional refinements (lower-speed external interfaces, full-connectivity determination, link metrics, OSPF) that are each separately taught or admitted. The principal soft spots are claim 4/11/18 (opaque LSAs, which I could not tie to any reference on the page's citation list) and claim 5/12/19 (private network), which rest on weaker art of record.

References cited in this analysis

• US 7,173,912 B2 — https://patents.google.com/patent/US7173912/en (authoritative text used throughout)
• US 5,818,840 A (Time Warner Entertainment Co. L.P., Asymmetric ATM switch) — https://patents.google.com/patent/US5818840
• US 6,529,300 B1 (Nortel Networks Ltd., WDM optical network with passive pass-through at each node) — https://patents.google.com/patent/US6529300 ; family counterpart EP 0 905 936 A3 — https://patentimages.storage.googleapis.com/b3/d3/aa/04215acdd6a300/EP0905936A3.pdf
• US 5,825,772 A (Cabletron Systems, Inc., Distributed connection-oriented services for switched communications networks) — https://patents.google.com/patent/US5825772 ; related continuation disclosure (metric-based path selection) — https://patentimages.storage.googleapis.com/ba/72/b4/b5da37e05f361e/[US7720076](/patent/US7720076).pdf
• EP 0 512 495 A2 / B1 (Fujitsu Limited, Switching node in a network with label multiplexed information) — https://worldwide.espacenet.com/patent/search/family/026442417/publication/EP0512495A2 ; http://data.epo.org/gpi/EP0512495B1-Switching-node-in-a-network-with-label-multiplexed-information.html
• US 5,461,624 A; US 5,483,522 A; US 5,425,021 A; US 5,852,606 A; US 5,951,634 A; US 6,016,306 A; US 5,781,534 A; US 5,432,783 A; US 5,432,783 A; WO 1998/000954 A2 — listed on the US 7,173,912 page's "Patent Citations (46)"; content characterized from title/field, not independently verified in this session.