US 7593458

I'll search for authoritative information on this specific patent number, including USPTO records and any CAFC 2026 docket activity.

Let me search for the PTAB case and any related Federal Circuit appeal, plus the independent claims specifically.

Let me get the exact independent claim language and check the Federal Circuit docket directly.

US Patent 7,593,458 — Analyst Summary

Strict-ID note: All identifiers below are reproduced literally as found; no auto-correction applied.

Bibliographic Data

Source: https://patents.google.com/patent/US7593458/en ; https://patentimages.storage.googleapis.com/db/f5/44/07189cdd955779/US7593458.pdf

Abstract (verbatim, from the authoritative patent text)

"Operational data is utilized to determine the FEXT interference induced by one line into the other DSL line. FEXT interference can be calculated using the NEXT interference measured between the two lines at the upstream ends of the loops and the downstream channel transfer function of one of the loops. Because the NEXT and transfer function constitute a linear time-invariant system, as does the FEXT interference between the lines, the NEXT interference and line transfer function can be multiplied (if in linear format) or added (if in logarithmic format) to approximate the FEXT interference between the lines. The collection of data, calculations and other functions performed in these techniques may be performed by a system controller, such as a DSL optimizer. An Xlog(u,n) quantity is a decibel-magnitude representation of the insertion-loss equivalent of FEXT transfer functions and is defined as the ratio of (1) a line u's source power into a matched load of 100 Ohms when no binder is present to (2) the power at the output of the subject line when line u is excited with the same source and the binder is present. Xlin(u,n) is the linear equivalent of Xlog(u,n). The Xlog(u,n) and Xlin(u,n) quantities may be represented in specific formats that assist in their use in DSL and other systems. When defined as a line's insertion loss, Xlin (or equivalently Xlog) does not include the effect of any transmit filter."

Core inventive mechanism: the patent exploits the fact that downstream FEXT, upstream-end NEXT, and the line channel transfer function are all linear time-invariant systems, so FEXT can be approximated by combining measured NEXT with the victim line's downstream channel transfer function — avoiding direct FEXT measurement and enabling single-ended (upstream-side) determination.

Plain-Language Overview of the Independent Claims

⚠️ Uncertainty flag: The authoritative full claim set (44 claims) was not fully reproduced in the fetched text. The overviews below are reconstructed from the patent's own description, the Google/JUSTIA claim text, and the IPR2025-00492 petition record. Verbatim claim-1 language could not be independently confirmed, so treat the precise wording as provisional while the substance is well-supported.

Claim 1 — Method (evaluating a multi-line, vectored DSL system).
A method for characterizing crosstalk in a vectored DSL system: the crosstalking lines in a channel are excited with an input signal; output data is acquired from a "primary" (victim) line that reflects the crosstalk coupling; the crosstalking lines are ordered/ranked by index u (u = 1 being the largest/worst crosstalker); and a DSM-data quantity representative of the crosstalk coupling — Xlog(u,n) (decibel form) or Xlin(u,n) (linear form) — is generated per tone n. In the related medium/controller versions the generated quantity is then used to send operating instructions to the modems.

Claim 16 — Controller.
A controller (e.g., a DSL optimizer / dynamic spectrum manager) comprising collecting means (to collect operational data from the DSL system) and calculating means that determine a DSM-data crosstalk quantity (Xlog(u,n) or Xlin(u,n)) for a primary line relative to each crosstalking line, with the ordered-index convention (u = 1 worst) and defined data formats (e.g., Xlog(u,n) = 6 − m(u,n)/10 for a 10-bit unsigned integer m(u,n); Xlin(u,n) via a scale(u) factor and 16-bit complex components).

Claim 25 — Computer program product.
A machine-readable medium carrying program instructions that perform a method of evaluating a multi-line, vectored DSL system: acquiring data representative of crosstalk coupling between a primary line and U crosstalking lines, ordering those lines by index u (u = 1 largest crosstalker), and generating a DSM-data quantity (Xlog/Xlin) representative of that coupling.

Claim 27 — Method of determining downstream FEXT interference.
A method of determining the downstream FEXT generated by a first DSL line into a second DSL line, comprising: obtaining a downstream value of the second line's channel transfer function (Hlog/Hlin); obtaining an upstream-end NEXT interference value generated by the first line into the second line; and combining the two to approximate the downstream FEXT.
• Dependent claims add the domain rule: add the log-domain values (claim 28) or multiply the linear-domain values (claim 29).

Claim 29 — Computer-readable medium.
A computer-readable medium with instructions for evaluating a multi-line, vectored DSL system that: excite the crosstalking lines, acquire output data from the primary line, order the crosstalking lines by index, generate a data quantity representative of crosstalk coupling, and send instructions to one or more modems based on the generated quantity (e.g., operational-condition values, maximum data-rate limit, upstream transmission band limits, transmit power levels, coding/latency requirements). Dependent claims (30–44) add the Xlog/Xlin data formats, XPSD reporting, log/linear combination, and single-sided/common-clock vectored-system limitations.

Litigation / PTAB / CAFC Status

• PTAB: IPR2025-00492, Nokia of America Corp. v. Adaptive Spectrum and Signal Alignment, Inc., filed January 22, 2025 against U.S. Patent 7,593,458; challenged claims 1–2, 4–5, 7, 9–12, 14–15; status Terminated–Settled (termination date June 24, 2025). Grounds (as summarized in petition-analysis dockets) were obviousness over Rezvani (US 7,356,049) in view of Cendrillon I ("Partial Crosstalk Cancellation Exploiting Line and Tone Selection in VDSL," 2003), and further in view of G.992.3 (ADSL2) for the Xlin(u,n) data-format claim.
  • https://ipverse.greyb.com/ptab-web/cases/case-details/IPR2025-00492
  • https://portal.unifiedpatents.com/ptab/case/IPR2025-00492
• District court: Adaptive Spectrum and Signal Alignment, Inc. v. AT&T Inc. et al, 2:24-cv-00029 (E.D. Tex.), filed January 19, 2024; '458 is among the asserted patents (with 7,428,669; 7,809,996; 7,991,122; 10,848,398; 11,050,654; 11,477,108; 11,770,313; 9,954,631).
  • https://npe.law.stanford.edu/patent/7593458
• CAFC 2026 dockets: No 2026 Federal Circuit appeal or docket entry specifically involving U.S. Patent 7,593,458 was found in the searches performed. Searches of 2026 CAFC materials returned only unrelated cases (e.g., AeroVironment, Global Tubing, Teva, Enviro Tech, TJTM). I cannot confirm the existence (or non-existence) of any non-public or newly docketed 2026 CAFC proceeding from the sources available. Note the IPR settled in 2025, which typically forecloses a subsequent CAFC appeal of that IPR.

Confidence & Caveats

• High confidence: title, inventor, assignee chain, application number, filing/priority/issue dates, abstract, PTAB/litigation identifiers.
• Moderate confidence: claim count (44), the set and gist of independent claims (1, 16, 25, 27, 29), and the dependent-claim domain rules.
• Lower confidence: the verbatim text of claim 1 and the exact independent-claim boundaries — the full 44-claim text was not captured in the fetched material. For any legal or validity analysis, verify claim language against the USPTO PatentCenter record (application 11/122,365) or the issued patent PDF before relying on it.
• Per operating rules, conflicting live-search results were preferred over prior knowledge; the identifiers above were transcribed literally and not auto-corrected.

I'll search for litigation involving US patent 7593458 across patent litigation sources.

Let me dig deeper into the specific cases found and check additional litigation sources.

Let me check the Charter Communications case and any other cases to confirm whether '458 was asserted beyond the AT&T action.

Litigation Involving U.S. Patent 7,593,458

I searched patent-litigation sources (Stanford NPE Litigation Database, Unified Patents/PTAB records, Docket Alarm / PACER dockets, Patexia, and the Google Patents litigation metadata for US7593458). Based on those results, US 7,593,458 ("FEXT determination system") appears in exactly one district-court action and one related PTAB proceeding. Details below.

⚠️ Caveat on the record: I was able to retrieve the case-identifying data and status but not every underlying docket entry. Items I could not independently confirm from a second source are flagged. All identifiers are reported exactly as found.

1. District Court Litigation

Adaptive Spectrum and Signal Alignment, Inc. v. AT&T Inc. et al

Notes on this case:

• US 7,593,458 was one of nine ASSIA patents asserted in this single complaint (per the Stanford NPE database and Patexia's patent list for the case).
• The case did not reach trial. A jury trial had been scheduled for September 22, 2025, and a Markman hearing was held March 19, 2025, but the parties settled before trial.
• Separately, the '669 patent was dismissed from the case by joint motion in February 2025 (D.I. 119, order granting D.I. 112), before the overall settlement.
• ASSIA is classified as an "8 Product company" (practicing-entity status) in the Stanford NPE database, not a classic NPE, and no declaratory-judgment action was filed.

Sources: Stanford NPE Litigation Database – patent page https://npe.law.stanford.edu/patent/7593458 and case page https://npe.law.stanford.edu/case/[285351](/patent/285351) ; Patexia case summary https://services.patexia.com/lawsuits/Adaptive-Spectrum-and-Signal-Alignment-Inc-v-AT%26T-Inc-et-al-id-[196104](/patent/196104) ; Docket Alarm docket https://www.docketalarm.com/cases/Texas_Eastern_District_Court/2--24-cv-00029/

2. Related PTAB Proceeding (not a district-court suit, but directly on this patent)

Nokia of America Corp. v. Adaptive Spectrum and Signal Alignment, Inc.

Sources: GreyB IP Verse https://ipverse.greyb.com/ptab-web/cases/case-details/IPR2025-00492 ; Ex Parte / ai-lab case page https://ai-lab.exparte.com/case/ptab/IPR2025-00492/nokia-of-america-corp-v-adaptive-spectrum-signal-alignment-inc ; Google Patents litigation metadata for US7593458 (noting "PTAB case IPR2025-00492 filed (Settlement)").

3. Items I could NOT confirm as involving US 7,593,458

• ASSIA v. Charter Communications (E.D. Tex., 2:24-cv-00124). This companion ASSIA campaign exists (a complaint was cited as an exhibit in a related PTAB petition regarding ASSIA's '996 patent), and Unified Patents ran prior-art contests covering "ASSIA … DSL" patents "asserted against AT&T and Charter Communications." However, I could not confirm that US 7,593,458 specifically was asserted against Charter. The Stanford NPE database lists only the single AT&T case for this patent, so I am treating a Charter assertion of '458 as unverified.
• ASSIA v. BT (England & Wales). There was ASSIA/BT litigation in the UK High Court (e.g., [2022] EWHC 1707 (Ch) and related judgments), but that involved ASSIA's EP patents and DSL management technology. US 7,593,458 is a U.S. patent, so the UK action is not litigation "involving US 7,593,458."
• Other jurisdictions / CAFC appeals. I found no Federal Circuit appeal and no other district-court case listing this patent. The Google Patents page shows a "First worldwide family litigation" link (Darts-IP family 34968490), which is consistent with the AT&T action being the family's principal litigation.

Summary

If the standard is US district-court infringement litigation, the answer is: one known case — ASSIA v. AT&T, 2:24-cv-00029 (E.D. Tex.), filed January 19, 2024, ended by settlement (dismissed) around April 2025. If PTAB validity challenges count, add IPR2025-00492 (Nokia v. ASSIA), filed January 22, 2025, terminated-settled June 24, 2025.

I found no other court cases asserting this specific patent. Note that my searches relied on third-party litigation databases rather than a direct PACER/CAFC full-text query, so if you need a fully exhaustive record (e.g., for freedom-to-operate or damages diligence), I'd recommend a direct PACER docket search on "7,593,458" and a Lex Machina / Darts-IP report, which I could not run here.

PTAB Activity — U.S. Patent No. 7,593,458 ("FEXT determination system")

Data-source note (read first). The structured "PTAB proceedings on file" block states that the USPTO ODP API returns no AIA trial proceedings for this patent. That is incomplete. A web search surfaces IPR2025-00492, which is recorded on Google Patents itself (the patent page carries the tag "PTAB case IPR2025-00492 filed (Settlement)"). I therefore flag one proceeding below rather than asserting a clean "no PTAB activity" result. Separately, the "Petitioner: Unified Patents PTAB Data" line on the Google Patents page is a data-attribution/license credit for the litigation dataset, not a petitioner of record.

Proceedings overview

There is one (1) AIA trial proceeding on US 7,593,458: IPR2025-00492, filed by Nokia of America Corp., which was terminated on settlement before institution — meaning zero claims were canceled, zero claims were sustained, no institution decision ever issued, no Final Written Decision exists, and no estoppel attached. The bottom-line defensive posture is therefore the least favorable of the possible outcomes for an accused infringer: the patent is completely untested at the PTAB, all fifteen-plus claims remain live and presumptively valid, and a defendant gets no free invalidity ruling — only the knowledge that a well-resourced petitioner (Nokia, likely in a DSL-equipment indemnity posture for AT&T) thought the claims were worth attacking with Rezvani + Cendrillon art, and that ASSIA chose to buy peace rather than defend them on the merits.

IPR2025-00492 — Nokia of America Corp. v. Adaptive Spectrum and Signal Alignment, Inc.

• Type: Inter Partes Review (AIA), 35 U.S.C. §§ 311–319
• Filed: 2025-01-22
• Status: "Terminated-Settled" (third-party PTAB databases; Google Patents records the litigation tag as "IPR2025-00492 filed (Settlement)"). The Board's own disposition paper is captioned "Decision — Settlement Prior to Institution of Trial, 37 C.F.R. § 42.74." Termination date: 2025-06-24.
• Judge panel: Jameson Lee, Jason J. Chung, and Brian P. Murphy, Administrative Patent Judges. APJ Chung authored the termination decision (Paper 9).
• Petition grounds (sourced from a third-party petition-analysis summary — see reliability caveat below):
  • Ground 1 — § 103 obviousness of claims 1–2, 4–5, 7, 9–10, 12, 14–15 over Rezvani (U.S. Patent No. 7,356,049) in view of Cendrillon I, Partial Crosstalk Cancellation Exploiting Line and Tone Selection in VDSL (2003).
  • Ground 2 — § 103 obviousness of claim 11 over Rezvani + Cendrillon I + ITU-T G.992.3 (ADSL2), the last reference supplying the 16-bit a(u,n)+j*b(u,n) / scale(u) linear data format recited in claim 11.
  • Challenged claim set, as reported: 1–2, 4–5, 7, 9–12, 14–15 — i.e., claims 3, 6, 8, and 13 were not challenged.
  • No § 112 grounds were asserted; every ground was § 103.
  • ⚠️ Reliability caveat — do not cite this as gospel. The summary source misstates the patent's title (calling it "Method for Evaluating Operational Characteristics of a Multi-Line DSL System" when the granted title is "FEXT determination system"), so I treat its claim/ground mapping as indicative but unverified. It is internally consistent (Ground 1 claims + claim 11 = the reported challenged set), which is reason for moderate confidence, but a defendant should pull Paper 1 (the Petition) from PTAB E2E before relying on the art combination for anything in a brief.
  • Preliminary Response: the docket shows Patent Owner activity and a document mailed 2025-08-04 bearing Sterne Kessler's IPR2025-00492 service block; because no institution decision was ever reached, I cannot confirm what substantive POPR was filed. No POPR-based merits reasoning is publicly attributed to the Board.
• Institution decision: None. The Board expressly stated: "We have not yet instituted trial in this proceeding." The statutory 6-month institution window under § 314(b) was never triggered to a decision — the parties pulled the plug first.
• Final Written Decision: None issued. Accordingly there is no claim-level verdict — no independent claim canceled, no dependent claim canceled, no claim held patentable. Anyone who tells a defendant that "claims 1–5 of the '458 patent are invalid" is wrong.
• Settlement / termination: On 2025-06-18 the parties filed a Joint Motion to Terminate (Paper 8) together with a Settlement Agreement (Ex. 1026) and a joint request to keep it confidential. On 2025-06-24 the Board (Paper 9) granted termination as to both Petitioner and Patent Owner, granted the 42.74(c) request, ordered the agreement kept separate from the file of Patent 7,593,458, and made it available only to Federal Government agencies on written request or to any person on a showing of good cause. The parties represented that Ex. 1026 is all agreements made in connection with termination. The terms — license, royalty, covenant, scope — are not public. The joint motion represents that the agreement resolves "their disputes regarding the '458 patent," which suggests the deal is broader than this one patent.
  • Termination decision: https://www.docketalarm.com/cases/PTAB/IPR2025-00492/Nokia_of_America_Corp._v._ADAPTIVE_SPECTRUM_AND_SIGNAL_ALIGNMENT_INC/docs/06-24-2025-Board/Termination_Decision__Pre_DI_settlement-9-Termination_Decision__Settlement_Prior_to_Institution_of_Trial_37_CFR_sec_4274,_Granting_Request_to_File_Settlement_Agreement_as_Business_Confidential_Information_37_CFR_sec_4274c.pdf
  • PTAB E2E case lookup: https://ptacts.uspto.gov/ptab/cases
• Appeal: No Federal Circuit appeal. There was no appealable Final Written Decision under 35 U.S.C. § 319. No CAFC docket number exists for this proceeding.
• Counsel: Petitioner — M. Scott Stevens, Christopher T.L. Douglas, Nicholas Marais, Daniel O'Connor (Alston & Bird LLP); also Matthew Turk (Foley). Patent Owner — Michael D. Specht, Jason A. Fitzsimmons, Richard M. Bemben (Sterne, Kessler, Goldstein & Fox PLLC). Same Alston & Bird / Sterne Kessler lineup appears in the parallel ASSERTIA IPRs on sibling patents — this was a coordinated, high-budget campaign, not a pro se shot in the dark.
• Defensive value: Near-zero as precedent, meaningful as intelligence. Because the case died before institution, the Art Unit never weighed in on Rezvani or Cendrillon-vis-à-vis these claims, and § 315(e)(2) estoppel never attached (estoppel requires a final written decision — § 315(e)(1)–(2)). A defendant today can re-run the same Rezvani + Cendrillon I + G.992.3 grounds in district court or in a fresh IPR without any estoppel obstacle, and can also argue the settlement is evidence that the patent owner was not confident in the claims. But it must prove unpatentability from scratch, on a clear-and-convincing (litigation) or preponderance (new IPR) burden.

Strategic summary

Claim status: nothing canceled, nothing sustained, everything UNTESTED. US 7,593,458 has never had a single claim adjudicated by the PTAB. The reported challenge in IPR2025-00492 swept claims 1–2, 4–5, 7, 9–12, and 14–15, leaving claims 3, 6, 8, and 13 outside the attack. The patent's claim set appears to run to at least claim 15, but I could not verify the total issued claim count from the sources retrieved — do not assume there is no claim 16+ without pulling the printed patent. For a defendant being asserted against today, the practical position is: the entire asserted claim set remains presumptively valid and fully pleadable by the patent owner. The only narrowing that has occurred is prosecution narrowing, if any — not PTAB narrowing.

Estoppel landscape — a genuinely open field. This is the single most useful takeaway. Because IPR2025-00492 terminated pre-institution, Nokia is not estopped under 35 U.S.C. § 315(e)(2) from raising the Rezvani/Cendrillon/G.992.3 grounds in the E.D. Tex. action or any other forum; estoppel attaches only "after the petitioner… receives a final written decision." Equally, no estoppel runs against any other defendant — a new party can file a fresh IPR on the same art, or on entirely different art (including the Cendrillon material already floating around the district court docket). The countervailing risks are § 325(d) (art already presented/substantively considered by the Office — weak here, since no FWD ever resolved it) and General Plastic / Fintiv-style discretionary denial, which the Board has been applying aggressively in 2025–2026, particularly against follow-on petitions challenging the same claims.

Pattern signals. This was a multi-patent, multi-petitioner coordinated campaign by AT&T and Nokia against the ASSIA DSL portfolio, all triggered by Adaptive Spectrum and Signal Alignment, Inc. v. AT&T Inc. et al., No. 2:24-cv-00029 (E.D. Tex., filed 2024-01-19, Judge Rodney Gilstrap), which asserted the '458 patent alongside U.S. Patent Nos. 7,428,669; 7,809,996; 7,991,122; 9,954,631 and several later patents:

• IPR2025-00395 — AT&T Services Inc. and Nokia of America Corp. v. ASSIA, on U.S. 7,428,669 (filed 2025-01-03);
• IPR2025-00451 — AT&T Services Inc. v. ASSIA, on U.S. 9,954,631 (filed 2025-01-17);
• IPR2025-00428 — Nokia of America Corp. v. ASSIA, on U.S. 7,991,122, challenging claims 1–20 over Tsatsanis + Savoor and Tsatsanis + Rezvani (filed 2025-01-22) — also reported "Terminated-Settled";
• IPR2025-00492 — Nokia v. ASSIA, on the '458 patent (filed 2025-01-22) — terminated 2025-06-24.
  The uniformity of outcome (the '122 IPR settled too) points to a global commercial resolution in 2025 — consistent with a third-party docket aggregator showing 2:24-cv-00029 as closed — but the terms are confidential and I could not confirm a district-court dismissal order or judgment. No defensive aggregator (Unified Patents, RPX, AST) is in the chain. Nokia is a large operating company acting in a vendor/indemnity posture for AT&T's DSL and VDSL deployments. Separately, the E.D. Tex. docket shows AT&T's invalidity-contentions reply attaching Cendrillon articles on MIMO-DSL power allocation and zero-forcing crosstalk cancellers — the same academic-art family Nokia used at the PTAB, which tells you the defendants' engineers considered Rezvani + Cendrillon the core attack.
• Litigation context: https://npe.law.stanford.edu/patent/7593458
• Patent record (with the PTAB/litigation tags): https://patents.google.com/patent/US7593458/en

The absence of a merits ruling is itself a signal — in both directions. Well-asserted patents do attract IPRs, and this one did, from a deep-pocketed defendant with competent counsel. But the case never got past the POPR stage, which means either (a) the parties settled for commercial reasons before the Board could rule, or (b) patent owner bought out the challenge because its preliminary response was weak. We cannot tell which from the public record, and any advocate claiming to know is overreading a redacted docket.

Recommended next steps

1. Pull the primary documents — do not rely on aggregator summaries. Get Paper 1 (Petition), Paper 5 (POPR, if filed), Paper 8 (Joint Motion to Terminate), and Paper 9 (Termination Decision, 2025-06-24) from PTAB E2E at https://ptacts.uspto.gov/ptab/cases. Confirm with your own eyes the exact challenged-claim list and the Rezvani/Cendrillon I/G.992.3 mapping, and confirm whether a POPR exists (it materially affects whether the settlement reflects weakness or commercial pragmatism).
2. If you have been newly accused: there is no FWD to quote, so you cannot cite a cancellation. Your best PTAB play is a fresh IPR pressing the Rezvani + Cendrillon I + G.992.3 combination (no estoppel blocks you), or better, different art to sidestep § 325(d) and General Plastic discretionary denial — e.g., the additional Cendrillon papers already cited in the E.D. Tex. contentions ("Simplified Power Allocation and TX/RX Structure for MIMO-DSL"; "The Linear Zero-Forcing Crosstalk Canceler is Near-Optimal in DSL Channels"). Note § 311(b) limits you to patents and printed publications.
3. Pre-date the § 315(b) clock. File within one year of service of any complaint alleging infringement of the '458 patent, or your IPR right is extinguished.
4. Check the assignment/licensing chain for the settlement's reach. The public reassignment record is unusually active (ASSIA SPE LLC → MGG California → Muzinich BDC → Valuegate Astro SPV1 → ASSIA Incorporated between 2016 and 2024). If the 2025 settlement included a license to AT&T and/or Nokia, confirm whether it runs to you as a supplier or customer — that is frequently the fastest exit and it is not discoverable from the redacted Ex. 1026.
5. Track the sibling proceedings. IPR2025-00428 ('122), -00395 ('669) and -00451 ('631) all run to the same patent owner and, in the case of the '122, the same Rezvani reference. Any institution decision, POPR, or FWD in those cases is admissible context for how the Board reads this shared Rezvani art — and a Board finding on Rezvani in a sibling case is worth far more to you than anything in the settled '458 case.
6. Do not overlook reexamination. With IPR2025-00492 gone without a merits ruling, an ex parte reexamination request (or a well-supported § 282 invalidity case) may be the lower-risk path where you want art that isn't patents/printed publications.

Bottom line for a defendant: the troll is not holding a dead patent — but it also does not hold a PTAB-hardened one. US 7,593,458's claims are wholly untested, and the one petitioner who took a serious swing walked away before the Board could rule. That is an opening.

I'll research the assignment record and litigation history for US 7593458 before assembling the report.

Let me dig for reel/frame specifics and the ASSIA litigation campaign.

Assignment & Ownership Analysis — US 7,593,458 B2 ("FEXT determination system")

Sourcing note / limitation. The USPTO Assignment Center (https://assignmentcenter.uspto.gov/, mirrored at https://assignment.uspto.gov/patent/index.html) exposes reel/frame numbers and a "correspondent of record" field per recording. In this session I could not pull the Assignment Center's field-level output directly — the indexed legal-event records I could retrieve (Google Patents "Legal Events," which mirror the USPTO assignment filings) show conveyance type, assignor, assignee, and event date, but not reel/frame numbers or correspondent attorney names. I therefore present the recorded assignment chain with those two fields flagged as not verifiable from the sources accessed. I have not invented reel/frame numbers or correspondent names. Where a name is not directly reported, I say so.

Inventors

• Single named inventor. The face of the patent (75) lists only Cioffi. Priority to US provisional 60/572,051 (2004-05-18); application filed 2005-05-05.
• No departure pattern. With one inventor who is also the founder of the assignee, the "all inventors left the assignee within 12 months" fire-sale tell does not apply. Cioffi remained the public face of ASSIA well past grant (described as founder/CTO-level figure and as Patent Owner's principal in ASSIA's later enforcement campaigns).

Original assignee

Adaptive Spectrum and Signal Alignment, Inc. ("ASSIA"), Redwood City, CA — named as assignee on the issued patent (73).

• Line of business: DSL/broadband network optimization software — Dynamic Line Management (DLM), dynamic spectrum management (DSM), and later CloudCheck/Wi-Fi diagnostics. The patent itself is core DSM/vectoring technology.
• Founded: 2003, described in Cioffi's own account as "formed in partnership with the USA's largest DSL service provider AT&T," which became the first DSM adopter. The company grew to a mid-size vendor with a 500+ patent family and worldwide operator deployments.
• Did it ship a product embodying the claims? Probable but not confirmed at claim level. ASSIA's DLM/DSM software operates on the same DSL line-management and crosstalk data the '458 claims address, but I could not tie the specific claims of the '458 to a shipped executable in the sources accessed. Treat embodiment as likely, unverified.
• Current status: Operating. ASSIA continues to prosecute new applications (last filings surfaced into 2024–2025) and is an active plaintiff/patent owner. No bankruptcy found. The portfolio has been cycled through secured-financing vehicles (see below) and the patent was returned to the operating parent in 2024.

Assignment timeline

Dates and parties below are those surfaced by the indexed legal-event records. Reel/frame and correspondent fields could not be retrieved; they are marked ⚠︎ where the Assignment Center would normally supply them.

• 2005-10-13 (recorded) ⚠︎ Reel not retrievable

  • Conveyance: Assignment of Assignors' Interest
  • Assignor: John M. Cioffi
  • Assignee: Adaptive Spectrum and Signal Alignment, Inc.
  • Correspondent: ⚠︎ not retrievable (no recording attorney visible in the retrieved record)
  • Context: Initial prosecution assignment — inventor assigns to his own startup.

• 2015-01-13 (recorded) ⚠︎ Reel not retrievable

  • Conveyance: Security Interest (secured financing)
  • Assignor: Adaptive Spectrum and Signal Alignment, Incorporated
  • Assignee: PARTNERS FOR GROWTH IV, L.P.
  • Correspondent: ⚠︎ not retrievable
  • Context: Securitization — venture-debt lender takes a security interest in the portfolio.

• 2016-10-03 (recorded) ⚠︎ Reel not retrievable

  • Conveyance: Assignment of Assignors' Interest
  • Assignor: PARTNERS FOR GROWTH
  • Assignee: Adaptive Spectrum and Signal Alignment, Inc.
  • Correspondent: ⚠︎ not retrievable
  • Context: Internal reorg / financing unwind — prior lender interest returned/cleared.

• 2016-11-16 (recorded) ⚠︎ Reel not retrievable

  • Conveyance: Assignment
  • Assignor: Adaptive Spectrum and Signal Alignment, Inc.
  • Assignee: ASSIA SPE, LLC (1209 Orange Street, Corporation Trust Center, Wilmington, DE 19801)
  • Correspondent: ⚠︎ not retrievable
  • Context: Ring-fencing into a special-purpose entity — IP moved into a Delaware SPE ahead of a new secured financing ("SPE" = special purpose entity; address is CT Corporation's registered-agent service).

• 2016-11-30 (recorded) ⚠︎ Reel not retrievable

  • Conveyance: Release by Secured Party
  • Assignor: PARTNERS FOR GROWTH IV, L.P.
  • Assignee: Adaptive Spectrum and Signal Alignment, Incorporated
  • Correspondent: ⚠︎ not retrievable
  • Context: Securitization cleanup — prior lien formally released.

• 2016-12-05 (recorded) ⚠︎ Reel not retrievable

  • Conveyance: Grant of a Security Interest — Patents
  • Assignor: ASSIA SPE, LLC
  • Assignee: MGG CALIFORNIA LLC, as Collateral Agent
  • Correspondent: ⚠︎ not retrievable
  • Context: Securitization — SPE pledges the portfolio as collateral for new debt.

• 2020-12-04 (recorded) ⚠︎ Reel not retrievable

  • Conveyance: Security Interest
  • Assignor: Adaptive Spectrum and Signal Alignment, Incorporated
  • Assignee: MUZINICH BDC, INC.
  • Correspondent: ⚠︎ not retrievable
  • Context: Securitization / refinancing — new lender.

• 2020-12-07 (recorded) ⚠︎ Reel not retrievable

  • Conveyance: Release by Secured Party
  • Assignor: MGG CALIFORNIA, LLC
  • Assignee: ASSIA SPE, LLC
  • Correspondent: ⚠︎ not retrievable
  • Context: Release of prior lender on refinancing.

• 2022-07-21 (recorded) ⚠︎ Reel not retrievable

  • Conveyance: Release by Secured Party
  • Assignor: MUZINICH BDC, INC.
  • Assignee: ASSIA SPE, LLC
  • Correspondent: ⚠︎ not retrievable
  • Context: Release of prior lender.

• 2022-10-29 (recorded) ⚠︎ Reel not retrievable

  • Conveyance: Security Interest
  • Assignor: Adaptive Spectrum and Signal Alignment, Incorporated and ASSIA SPE, LLC
  • Assignee: VALUEGATE ASTRO SPV1
  • Correspondent: ⚠︎ not retrievable
  • Context: Securitization / refinancing — new lender over both entities' rights.

• 2023-08-17 (recorded) ⚠︎ Reel not retrievable

  • Conveyance: Release by Secured Party
  • Assignor: VALUEGATE ASTRO SPV1
  • Assignee: ASSIA SPE, LLC / Adaptive Spectrum and Signal Alignment, Incorporated
  • Correspondent: ⚠︎ not retrievable
  • Context: Release of prior lender.

• 2024-01-18 (recorded) ⚠︎ Reel not retrievable

  • Conveyance: Assignment of Assignors' Interest
  • Assignor: ASSIA SPE, LLC
  • Assignee: Adaptive Spectrum and Signal Alignment, Incorporated
  • Correspondent: ⚠︎ not retrievable
  • Context: Transfer back to the operating parent out of the financing SPE — recorded one day before ASSIA's infringement complaint against Charter was filed (2024-01-19), consistent with establishing a clean ownership/standing record pre-suit.

Net ownership today: Adaptive Spectrum and Signal Alignment, Inc. (consistent with ASSIA being listed as Patent Owner in IPR2025-00492).

Timeline diagram

timeline
    title Ownership of US 7593458
    2005 : Inventor assigns to ASSIA Inc
    2009 : Patent issued
    2015 : Security interest to Partners for Growth
    2016 : IP moved into ASSIA SPE LLC
         : Security interest to MGG California
    2020 : Security interest to Muzinich BDC
    2022 : Security interest to Valuegate Astro
    2024 : IP returned to ASSIA Inc
         : ASSIA sues Charter in EDTX
    2025 : Nokias IPR filed then settled

NPE / troll-pattern signals

1. Shell-entity transfer — UNCLEAR (form present, purpose = financing).
   The patent moved from the operating company to ASSIA SPE, LLC on 2016-11-16, a Delaware entity at 1209 Orange Street, Corporation Trust Center, Wilmington DE 19801 (CT Corporation's registered-agent address — a classic anonymous-shell tell). However, the SPE was the company's own financing vehicle (it immediately granted a security interest to MGG California LLC as collateral agent on 2016-12-05), it held the whole portfolio rather than cherry-picked patents, and the patent was returned to ASSIA Inc on 2024-01-18. This is a securitization structure, not a transfer-to-asserter.

2. Known asserter in the chain — PRESENT (with caveat).
   ASSIA does not appear on the legacy NPE lists (Acacia, Marathon, Intellectual Ventures, Wi-LAN, Conversant, Vringo, Pendrell, Round Rock, Spangenberg, etc.). But Unified Patents expressly designates ASSIA "an NPE" in its public PATROLL program, stating its patents "have all been asserted against AT&T and Charter Communications." That is a high-frequency-plaintiff designation by a defensive aggregator. Countervailing: ASSIA is a genuine product/R&D company, not a shell. Mixed.

3. Repeat correspondent across the chain — INSUFFICIENT DATA.
   Correspondent names are not exposed in the records I could retrieve, so recurrence cannot be tested. (For completeness, litigation counsel — not assignment correspondents — do recur: Sterne, Kessler, Goldstein & Fox represents ASSIA as Patent Owner in PTAB, and Alston & Bird appears for Nokia as Petitioner. Those are prosecution/litigation counsel and are not evidence of a repeated assignment-recording attorney.)

4. Cascading transfers — PRESENT (but securitization, not an NPE cascade).
   Multiple recorded transfers cluster tightly: 2016-11-16 → 2016-11-30 → 2016-12-05 (three in ~3 weeks), 2020-12-04 → 2020-12-07 (two in 3 days), and 2022-07-21 → 2022-10-29. The pattern is a lender-release / new-lender cycle across four successive lenders (Partners for Growth → MGG California → Muzinich BDC → Valuegate Astro), not a chain of unrelated LLCs under common principals.

5. Pre-litigation transfer — PRESENT (timing coincidence).
   The 2024-01-18 assignment returning the IP from ASSIA SPE, LLC to ASSIA Inc is contemporaneous with ASSIA's infringement complaint against Charter Communications filed 2024-01-19 (No. 2:24-cv-00124, E.D. Tex.). The '458 itself was the subject of a separate challenge (see below), but the one-day gap indicates ownership was consolidated into the asserting entity immediately before litigation.

6. Bankruptcy fire-sale — NOT PRESENT.
   No Chapter 7/11 proceeding for ASSIA appears in the record; the assignee continued operating and continued filing new patents.

7. Privateering — NOT PRESENT.
   ASSIA asserts in its own name; there is no operating-company→NPE transfer to assert against the operating company's competitors. (ASSIA's UK licensing relationship with BT — where a "patent laundering"/re-licensing clause was litigated — is a licensing dispute, not privateering.)

8. Defensive aggregator — NOT PRESENT.
   The chain does not terminate at RPX, AST, LOT, Unified Patents, or OIN. To the contrary, Unified Patents is asserting prior art against ASSIA patents, i.e., ASSIA is the target side of the defensive ecosystem.

Related enforcement context (not assignment records):

• IPR2025-00492, Nokia of America Corp. v. Adaptive Spectrum and Signal Alignment, Inc., filed 2025-01-22, terminated-settled 2025-06-24 — the only IPR naming the '458. The full-text banner's "Petitioner: 'Unified Patents PTAB Data'" is the data-source attribution (Unified's CC-licensed dataset), not the petitioner; the actual petitioner is Nokia of America Corp. per PTAB records.
• ASSIA v. Charter Communications, No. 2:24-cv-00124 (E.D. Tex.), filed 2024-01-19, with Charter-filed IPRs IPR2024-01379 / IPR2025-00012 / IPR2025-00013 / IPR2025-00088 all denied institution (2025).
• Parallel European enforcement of the same family (EP 1 756 963 B1 is the EP member of this patent): ASSIA v. BT and ASSIA v. Orange/Alcatel-Lucent (Nokia) in the UK/France.

Verdict

Operating-company assertion.

The chain begins and ends with Adaptive Spectrum and Signal Alignment, Inc. — the original assignee, founded by the named inventor in 2003 as a DSL-optimization software vendor, still filing new applications into 2024–2025, and asserting its own patents in its own name (Charter, 2:24-cv-00124, filed 2024-01-19; Nokia IPR2025-00492). The intervening assignments (2016-11-16 into ASSIA SPE, LLC; security interests to Partners for Growth, MGG California, Muzinich BDC, and Valuegate Astro) are a secured-financing / securitization cycle, not a transfer to a licensing shell, and the IP was returned to the operating parent on 2024-01-18. The one genuinely NPE-adjacent fact is Unified Patents' designation of ASSIA as an NPE; combined with the SPE/CT-Corporation structure this makes ASSIA a hybrid — an operating company running an aggressive licensing program — but the operating-company evidence (real products, ongoing R&D, self-assertion, inventor-founder retained) predominates.

Verify at: USPTO Patent Assignment Search — https://assignment.uspto.gov/patent/index.html#/patent/search (search patent number 7593458) and https://assignmentcenter.uspto.gov/; corroborating legal events at https://patents.google.com/patent/US7593458/en. Reel/frame and correspondent fields should be pulled from the Assignment Center directly, since they could not be verified here.

I'll research this systematically. Let me start by searching for the patent's citation records.

Let me gather the full citation list and claim set for this patent.

Prior-Art Analysis — U.S. Patent No. 7,593,458 B2

1. Patent identification (verified, not auto-corrected)

• Patent number: US 7,593,458 B2 — "FEXT determination system"
• Inventor: John M. Cioffi (Atherton, CA)
• Assignee (as issued): Adaptive Spectrum and Signal Alignment, Inc. (Redwood City, CA)
• Application: US 11/122,365, filed 2005-05-05; published as US 2005/0259725 A1 on 2005-11-24
• Provisional priority: U.S. 60/572,051, 2004-05-18
• Date of Patent (grant): 2009-09-22; 44 claims, 9 drawing sheets
• PCT counterpart: PCT/IB2005/001544 → WO 2005/114861 A1; EP 1756963 B1; JP 5057971 B2; CN 1965498 B; CA 2566499 C
• Post-grant proceedings of record: PTAB IPR2025-00492 (Nokia of America Corp. v. ASSIA), filed 2025-01-22, terminated/settled 2025-06-24, targeting application 11/122,365 / patent 7,593,458 (per GreyB/IPVerse). District-court suit: Adaptive Spectrum and Signal Alignment, Inc. v. AT&T Inc. et al., 2:24-cv-00029 (E.D. Tex., filed 2024-01-19).

Note on the ID: The string "7593458" also appears in unrelated contexts (a Japanese patent number 7593458 for a semiconductor package, a CommScope tool installation-bulletin number, a bushing part). These are not the patent in question and are excluded, per your instruction. The US patent number 7,593,458 is the relevant document.

2. What is on the face of the patent (source of the citation list)

US 7,593,458 lists, under "(56) References Cited," U.S. patent documents (both granted patents and published applications) plus one non-patent item. This list is reproduced on Justia and in the patent PDF. The U.S. list is:

Granted U.S. patents:

Published U.S. applications (pre-grant publications):

Other publication cited: International Search Report, International Application No. PCT/IB2005/001544 (3 pages).

3. Critical framing before any §102 mapping (please read)

Three points govern the honest answer to "which claim(s) each reference anticipates under 35 U.S.C. § 102":

1. A "References Cited" list is not a list of anticipations. References on the face of a patent are citations considered by the examiner and/or supplied by the applicant. Because US 7,593,458 issued with 44 claims, none of these references was found to anticipate the claims as allowed — otherwise the claims would have been rejected and would not appear in the granted patent. So the correct statement is: on the record available, no listed reference anticipates any granted claim. Labeling a reference "anticipating" here would be an overstatement not supported by the prosecution record.

2. Several listed items are probably not prior art at all relative to the 2004-05-18 priority date (pre-AIA §102 analysis applies; this case was filed 2005-05-05):

   • US 2005/0259725 A1 is the patent's own pre-grant publication (same application 11/122,365). It is the applicant's own disclosure of the same invention and cannot anticipate the patent.
   • US 2005/0123027 A1 and US 2005/0123028 A1 (Cioffi et al.) published 2005-06-09 — after the priority date; if they share inventorship/obligation with the patent they are also not "by another," and in any event too late for §102(b).
   • Items 28–33 (and the applications published in 2004–2006: Kerpez, Tzannes, Fishman, Kimble, Savoor, Rhee) published after 2004-05-18. They can only be prior art under pre-AIA §102(e) if their filing dates precede the priority date, and several are applicant-affiliated. I do not have reliable filing dates for each and therefore will not assert §102(e) status for them.

3. I could not retrieve the full text or the examiner's reasons for citation for each reference. The front-page list does not indicate which claims (if any) each reference was applied against. My claim mappings below are therefore relevance assessments, not established anticipations, and several descriptions are flagged as low-confidence.

4. Per-reference analysis (relevance + §102 assessment)

A. Closest art — crosstalk identification / cancellation / channel estimation

US 5,887,032 — Cioffi — granted 1999-03-23. Crosstalk/echo cancellation in DSL (methods for estimating and removing crosstalk coupling between subscriber lines). Relevance: addresses crosstalk estimation/removal, i.e., the general field. §102: Does not disclose deriving downstream FEXT from an upstream-end NEXT measurement multiplied/added to a line transfer function. No anticipation of independent claims (the "obtain H, obtain upstream-end N, combine → X" family). (Inventor-common with the patent, further weakening any §102(a) argument.)

US 6,885,746 — Hausman et al. — granted 2005-04-26 (family member US 2003/0112967 A1, 2003-06-19). Crosstalk identification for spectrum management in broadband/DSL systems. Relevance: high — directed to identifying crosstalk (incl. FEXT) characterization among lines; the closest subject-matter overlap. §102: Could be argued against narrower dependent claims concerning identifying/ordering crosstalkers and reporting crosstalk quantities, but it does not appear to teach the specific NEXT(Hlog)-approximates-FEXT derivation or the Xlog(u,n)/Xlin(u,n) ratio definitions of the independent claims. Assessment: no full-claim anticipation; at most a §103/§102(b) reference against specific sub-features. Filing date not verified.

US 7,016,822 — Bosley et al. — granted 2006-03-21. Crosstalk/channel modeling for communication channels. Relevance: moderate; modeling of crosstalk coupling. §102: does not disclose the claimed NEXT-plus-Hlog combination. No anticipation. (Filing date not verified; issue date is post-priority.)

US 6,990,196 — Zeng et al. — granted 2006-01-24 (family US 2002/0136397 A1, 2002-09-26). Crosstalk cancellation / channel estimation in multi-line DSL. Relevance: moderate. §102: no disclosure of the claimed upstream-NEXT + downstream-H derivation. No anticipation.

US 7,274,734 — Tsatsanis — granted 2007-09-25. Crosstalk cancellation in multi-line systems. Candidate for the "crosstalk cancellation methods" to which the specification says Xlin is useful. §102: does not disclose the claimed derivation. No anticipation. (Post-priority issue date.)

US 7,158,563 — Ginis et al. — granted 2007-01-02 (family US 2003/0086514 A1, 2003-05-08). Dynamic spectrum management / crosstalk mitigation across DSL lines. Relevance: high for DSM context and crosstalk coupling handling. §102: no teaching of the specific NEXT→FEXT approximation or Xlog/Xlin reporting format. No anticipation.

US 7,302,379 — Cioffi et al. — granted 2007-11-27. "DSL system estimation" (DSM estimation of line/channel data). Relevance: high, but same inventor/assignee and post-priority; not "by another" for §102(a) purposes and likely disqualified as prior art. No anticipation.

B. DSL channel/multicarrier background art

US 5,479,447 — Chow et al. — granted 1995-12-26; filed 1993-05-03. "Method and apparatus for adaptive, variable bandwidth, high-speed data transmission of a multicarrier signal over digital subscriber lines" — DMT, adaptive bit/power allocation, channel SNR characterization. Relevance: background for Hlin/Hlog channel-transfer concepts; co-invented by Cioffi. §102: discloses line channel characterization but nothing about FEXT-vs-NEXT derivation; no anticipation.

US 6,259,746 — Levin et al. — 2001-07-10; US 6,377,640 — Trans — 2002-04-23 (and 2002/0181633 A1); US 6,748,016 — Tzannes — 2004-06-08 (and 2004/0136463 A1); US 7,035,400 — Gaikwad et al. — 2006-04-25; US 7,050,489 — Sjöberg et al. — 2006-05-23; US 7,133,441 — Barlev et al. — 2006-11-07; US 7,190,665 — Warke et al. — 2007-03-13; US 6,870,901 — Gudmundsson et al. — 2005-03-22; US 6,999,583 — Valenti et al. — 2006-02-14. These relate to DSL transmission, resource allocation, loop diagnostics, and (some) crosstalk. Relevance: general background; none discloses the claimed core (upstream-end NEXT + line H → FEXT). §102: no anticipation of any claim.

• Confidence caveat: I was able to verify exact titles for only some of these (e.g., US 5,479,447, US 6,885,746). I cannot confirm the exact titles/filings for US 6,259,746, 6,377,640, 6,748,016, 6,870,901, 6,999,583, 7,035,400, 7,050,489, 7,133,441, 7,190,665 from the sources retrieved; treat their one-line descriptions as indicative, not verbatim.

C. Published applications

US 2004/0095921 A1 — Kerpez — published 2004-05-20. Crosstalk estimation in DSL. Relevance: high — timing is notable: it published two days after the 2004-05-18 priority date, so it is prior art only if its filing date predates the priority date (§102(e)), which I could not verify. §102: bears on crosstalk estimation generally, not the NEXT→FEXT combination. No anticipation asserted.

US 2003/0086362 A1 (Hasegawa et al., 2003-05-08); US 2003/0202459 A1 (Warke et al., 2003-10-30); US 2004/0157566 A1 (Fishman, 2004-08-12); US 2005/0129218 A1 (Kimble et al., 2005-06-16); US 2006/0072722 A1 (Savoor et al., 2006-04-06); US 2006/0098725 A1 (Rhee et al., 2006-05-11); US 2002/0041565 A1 (Valenti et al., 2002-04-11). Background art on DSL management, line data collection, and crosstalk/monitoring. The 2005–2006 publications post-date the priority date and are not usable against the claims absent verified earlier filing dates. §102: no anticipation.

US 2005/0259725 A1 — Cioffi — 2005-11-24: the patent's own publication — expressly not prior art.

5. Bottom line

• No reference cited on the face of US 7,593,458 anticipates any of its 44 claims. The defining independent-claim combination — (i) obtaining an upstream-end NEXT value between two DSL lines, (ii) obtaining one line's downstream channel transfer function (Hlog/Hlin), and (iii) combining them (add in log domain / multiply in linear domain) to approximate the downstream FEXT, plus the related Xlog(u,n)/Xlin(u,n) definitions and reporting formats (Eqs. 1–3) — is not disclosed by any single cited reference.
• The most relevant cited references for subject-matter overlap are the crosstalk-identification/cancellation and DSM items: US 6,885,746 (Hausman et al.), US 7,158,563 (Ginis et al.), US 5,887,032 (Cioffi), US 6,990,196 (Zeng et al.), US 7,016,822 (Bosley et al.), US 7,274,734 (Tsatsanis), and US 2004/0095921 (Kerpez). If anything here were ever pressed under §102, it would be against narrow dependent claims (e.g., identifying/ordering crosstalkers; using a measured coupling value in cancellation), not the independent claims.
• Claim families most likely to be the subject of any §102 argument (by analogy to the family publication US 2005/0259725 A1's claim set) are the method/controller/computer-program claims reciting "obtain H, obtain upstream N, combine to get X" and the insertion-loss + NEXT → FEXT claim — but again, the cited art does not appear to teach those combinations.
• Timing disqualifications to keep in view: US 2005/0259725 A1 is the applicant's own publication; US 2005/0123027 A1, US 2005/0123028 A1, US 7,302,379 and most 2005–2006 items post-date the priority date and/or are applicant-affiliated; several 2005–2007 patents are usable, if at all, only as §102(e) art on the basis of their (unverified) filing dates.

Confidence/limitations: I verified the patent's bibliographic data, claim count (44), citation list, and family/litigation record. I could not retrieve the full text of every cited reference or the examiner's rejection rationales, so exact titles for the references flagged above are uncertain, and the §102 assessments are relevance judgments, not adjudicated anticipations. This is technical analysis, not legal advice.

I'll first try to pin down the specific prior-art references for this patent, since the task asks me to build on "the Prior Art section." The fetched patent text I was given does not contain a prior-art / cited-references section, so let me research what references are actually cited against US 7,593,458.

I have the IPR grounds and the patent's own cited-references list now. Let me pull the actual claim language of US 7,593,458 and details on the key prior-art references (Rezvani, Cendrillon I, G.992.3).

Let me get the exact text of independent claim 1 (and the key dependent claims) so the § 103 analysis maps element-by-element.

Obviousness Analysis — U.S. Patent 7,593,458 B2 ("FEXT determination system")

Analyst note / scope caveat up front. The task instructs me to "use the results from the Prior Art section of this page." The page content supplied for this patent contains a Litigation summary section but no Prior Art section — so there is nothing to build on for prior art. Rather than invent one, I reconstructed the prior-art record from three verifiable sources: (1) the patent's own "References Cited" (56) list reproduced on the face of US 7,593,458; (2) the grounds actually pleaded in the co-pending IPR2025-00492 petition; and (3) the standards and publications cited inside the '458 specification. Where I am relying on a party's allegation rather than an adjudicated finding, I say so.

Two contradictions to flag explicitly:

1. The PTAB case-summary page I retrieved describes the '458 title as "Method for Evaluating Operational Characteristics of a Multi-Line DSL System." That is not the title of US 7,593,458, which is "FEXT determination system." The "operational characteristics" language is the preamble of the granted EP 1756963 B1 claim 1 (same family). This looks like a database-labeling error; I treat "FEXT determination system" as authoritative per the granted face of the patent.
2. IPR2025-00492 was terminated by settlement on June 24, 2025 before institution (no Final Written Decision). Likewise ASSIA v. AT&T settled before trial. No tribunal has ever adjudicated the validity of these claims. Everything below is therefore a predictive, counsel-style § 103 analysis, not a restatement of a holding.

1. Legal framework applied

• 35 U.S.C. § 103 asks whether the claimed subject matter as a whole would have been obvious to a person having ordinary skill in the art (POSITA) at the effective filing date, in view of the prior art and the ordinary creativity of the skilled artisan — Graham v. John Deere, 383 U.S. 1 (1966); KSR Int'l Co. v. Teleflex Inc., 550 U.S. 398 (2007).
• A combination is obvious where a POSITA had a rationale to combine (MPEP § 2143, rationales A–F: predictable results from combining known elements; simple substitution; known technique improving a similar device; applying a known technique to a known device ready for improvement; "obvious to try"; and explicit teaching/suggestion/motivation in the references).
• Effective priority date: the '458 claims priority to provisional 60/572,051, filed May 18, 2004, and was filed May 5, 2005. Earlier art (e.g., ANSI T1.417-2003) is clearly prior art; post-May-18-2004 art is prior art only against claim subject matter that is not entitled to the provisional's date.

2. The claims at issue

US 7,593,458 has 44 claims in three families:

Independent claim 1 (reconstructed from the granted-claim text and the IPR mapping) recites, in substance:

1. exciting the plurality of crosstalking lines in the channel with a known sequence of input symbols applied to the crosstalking lines;
2. acquiring output data from a primary line, the output data comprising the known sequence of input symbols after having been affected by crosstalk coupling;
3. ordering the plurality of crosstalking lines based on a comparison between the known input sequence and the output data;
4. generating a data quantity representative of the crosstalk coupling between the primary line and each crosstalking line based on the ordering; and
5. sending instructions to one or more modems based on the generated data quantity.

Representative dependent limitations (relevant to the combination analysis):

• Zeroing all other inputs / 100-Ω differential termination during the measurement (claims 19, 31).
• Threshold-based selection of U crosstalkers, and ordering u = 1 … U by crosstalk magnitude (FR4/Fig. 3 at 370–380).
• Xlog(u,n) = decibel-magnitude insertion-loss equivalent of FEXT/NEXT (claims 22–23, 33, 37).
• Xlog(u,n) = 6 − m(u,n)/10, m a 10-bit unsigned integer (claims 24–25, 34, 36).
• Xlin(u,n) represented by a scale(u) factor and a normalized complex number a(u,n)+j·b(u,n) coded as 16-bit 2's-complement integers (claim 11; claims 39, 44).
• u = 0 reserved for the line's own insertion loss including transmit/receive filter effects (claims 26, 44).

3. The prior-art landscape

3(a) References cited on the face of the '458 patent (56)

3(b) Non-patent literature cited in the specification (all pre-date the priority date)

• ANSI T1.417-2003, "Spectrum Management for Loop Transmission Systems" (Sept. 3, 2003) — the 1 % worst-case FEXT coupling model.
• W. Yu & Cioffi, "Distributed Multiuser Power Control for DSL," IEEE JSAC (June 2002) and "An Adaptive Multiuser Power Control Algorithm for VDSL," Globecom '01 (Nov. 2001) — DSM iterative spectrum balancing using channel/crosstalk data.
• Ginis & Cioffi, "Vectored Transmission for DSL Systems," IEEE JSAC (June 2002) and "Vectored-DMT: A FEXT Canceling Modulation Scheme for Coordinating Users," ICC 2001 — vectored DSL, per-tone crosstalk characterization.
• Zeng, "Crosstalk Identification and Cancellation in DSL Systems" (July 2001) — identifying which lines are crosstalking and ranking them.
• Cioffi et al., "Example Improvements of Dynamic Spectrum Management," T1E1.4/2001-089 (Feb. 2001).
• ITU-T Recommendations G.992.3 (ADSL2) & G.992.5 (ADSL2+) — define and standardize the Hlog[n] / Hlin[n] per-tone channel-characteristic data objects; G.997.1 defines the MIB management model.

3(c) References actually pleaded in IPR2025-00492 (Nokia v. ASSIA)

As pleaded, the Petitioner's mapping is: Rezvani injects known training sequences into subscriber lines and measures the resulting interference on other lines to build an interference matrix, and ranks the interferer channels "in order of importance" (reads on the "exciting," "acquiring," and ordering-adjacent limitations); Cendrillon I cancels only the largest crosstalkers on each tone, "sorted in order of crosstalk strength" (reads squarely on claim 1's "ordering … based on a comparison" and the threshold/U-selection dependents). Source: https://ai-lab-cl-prod.azurewebsites.net/case/ptab/IPR2025-00492/doc/summary/1 ; termination decision at https://www.docketalarm.com/cases/PTAB/IPR2025-00492/.

3(d) Additional art a § 103 ground could add (not pleaded, but available)

• EP 0 985 933 A2 / Koeman, "Modeling and Measuring ELFEXT Link Performance," NOC '98 — expressly treats the identity ELFEXT ≈ FEXT ÷ insertion loss (equivalently FEXT = ELFEXT + insertion loss in dB). This is the exact algebraic-relationship reasoning the '458 asserts as its core.
• IEEE 802.3 / TIA cabling practice — defines ELFEXT = FEXT normalized by insertion loss (cited in US 2007/0258514 as eq. (4): "ELFEXT = FEXT · Insertion_loss"). Same relationship, different physics domain (Ethernet cabling).

4. Mapping claim 1 to the primary combination (Rezvani + Cendrillon I)

Why a POSITA would combine them. Both references are in the same field (mitigating FEXT in bundled, co-terminated DSL), address the same problem, and the combination is a predictable use of a known technique (partial, largest-first crosstalk selection) applied to a known device (Rezvani's training/interference-matrix engine) that was ready for improvement — MPEP § 2143 rationales (A), (C), and (D). Rezvani's own ranking "in order of importance" plus Cendrillon's express instruction to sort by crosstalk strength makes the ordering limitation a design choice, not an inventive contribution.

5. Mapping the FEXT-from-NEXT core to cited art (the strongest § 103 theory)

The '458's central technical assertion is that, because the NEXT coupling, the loop's channel transfer function, and the FEXT coupling are each linear time-invariant (LTI) systems, the downstream FEXT can be obtained by the LTI cascade of the upstream-end NEXT and the victim line's downstream insertion loss — add in dB / multiply in linear:

Xlog(u,n) ≈ Nlog(u,n) + Hlog(0,n) (or Xlin = Nlin · Xlin(0,n)/T(n)).

This is uniquely vulnerable to § 103 for two independent reasons grounded in the patent's own text:

1. The specification itself treats the enabling principle as old. It states the LTI basis "is well known in the art" and that "different paths between two points will yield the same attenuation values." Under In re Boesch/MPEP § 2129, an admission of what was known in the specification operates as prior art against the claims.
2. The FEXT/insertion-loss/NEXT algebraic relationship was conventional in the cabling arts well before 2004. Koeman (NOC '98) and the TIA/802.3 ELFEXT = FEXT ÷ insertion loss definition establish that expressing one crosstalk quantity as the insertion loss combined with a normalized coupling quantity was known practice. ANSI T1.417-2003 (cited in the specification) models 1 % FEXT coupling as a function of the line's own channel response and a NEXT-derived coupling constant.

Combination: Rezvani (or the cited Zeng 2001 thesis) as to identifying and ranking the crosstalking lines and measuring the coupling × G.992.3/G.997.1 as to obtaining the victim line's per-tone insertion loss Hlog/Hlin and reporting it as standard MIB data × T1.417-2003 / Koeman as to the LTI/insertion-loss relationship that yields FEXT. Motivation: the '458 itself concedes the data are "readily available at the upstream end" and that using them avoids "disrupt[ing] normal operation" — i.e., the combination achieves a known, predictable benefit (single-ended estimation instead of disruptive direct FEXT measurement) with a reasonable expectation of success.

Note a date hazard worth flagging: the Draft DSM Report T1E1.4/2003-018RA is dated May 24, 2004 — six days after the May 18, 2004 provisional. A challenger should therefore rely on the earlier contribution revisions (T1E1.4/2003-018 and its R-series) and on T1.417-2003 and G.992.3 (2004)* as of their own publication dates, rather than the RA revision, to avoid a priority-date fight.

6. Mapping the format claims (claim 11, 24–25, 39, 44) — all obvious over standards art

These claims recite, in substance, that the crosstalk quantity is (a) Xlog = 6 − m/10 with m a 10-bit unsigned integer, or (b) Xlin = (scale(u)/2¹⁵)·(a + j·b)/2¹⁵ with a, b 16-bit 2's-complement integers, and (c) that the index u = 0 carries the line's own insertion loss including filter effects.

The '458 specification is its own worst enemy here:

• It says Xlog(u,n) is "analogous to the Hlog[n] of Section 6.3.1.1 in the [DSM] Report" and to G.992.3 § 8.12.3.1 / G.992.5.
• It says Xlin(u,n) "can be specified … in a format identical to Hlin[n]" and "is analogous to the Hlin[n] … in ADSL2/2+."
• It says Xlin(0,n) is "the insertion loss of the line without any removal of transmit or receiver filters."

Combination: The Petitioner's Ground 2 (Rezvani + Cendrillon I + G.992.3) is well-taken in principle: G.992.3 standardizes the Hlin(f) complex-per-tone data object in exactly the scale/normalized-integer form, and it was a routine and obvious design choice to apply an existing standardized reporting format to a newly generated data quantity of the same kind — MPEP § 2143 rationale (C)/(F), and KSR's "familiar elements according to known methods." Motivation exists because interoperability and MIB-compatibility favor standardized formats, and the patent itself identifies the analogous format as the industry norm.

Caveat / potential weak spot: claim 11's exact tie between the scale factor chosen so max(|a|,|b|) = 2¹⁵−1 and the NEXT/Nlin-based derivation could support a narrow non-obviousness argument if the prior art's Hlin formatting is applied only to direct channels; a Patent Owner would argue no reference teaches applying that format to a crosstalk channel. That is a genuine, if narrow, distinction the challenger must rebut with a "design choice / obvious application" theory.

7. Motivation-to-combine and expectation-of-success summary

8. Counter-considerations (arguments the Patent Owner would press)

1. No merits adjudication. Both the IPR (terminated pre-institution, June 24, 2025) and the district-court case (settled ~April 2025) ended without any validity ruling. The PTAB's settlement decision expressly states "We have not yet instituted trial." A challenger gets no estoppel benefit and must prove the case fresh.
2. The single-ended insight may be framed as non-obvious. The patent frames the advance as estimating FEXT from one end of the loop, which the Background characterizes as a long-felt need ("systems … that permit measurement of FEXT interference … from one side of the DSL loops … would represent a significant advancement"). If a POSITA was previously thought to require coordinated (both-ended) training, that can be a secondary-consideration argument.
3. ASSIA is a practicing entity ("8 Product company," per the Stanford NPE database from the litigation summary), which can support commercial-success secondary considerations — though the evidence here shows settlement rather than long-felt-need testimony, and settlement is generally weak evidence of validity.
4. Certain claims were never challenged. IPR2025-00492 attacked only claims 1–2, 4–5, 7, 9–12, 14–15. Claims 3, 6, 8, 13, and 16 were not in the challenge and may contain the narrower FEXT = NEXT + Hlog cascade / ΔX (phantom-mode) limitations that are harder to invalidate. Any invalidity position should be built on those narrower claims too, not just claim 1.

9. Bottom line

• Claim 1 (and dependents 2, 4–5, 7, 9–10, 12, 14–15): Likely obvious under § 103 over Rezvani (US 7,356,049) in view of Cendrillon I — the combination supplies every limitation of claim 1, and the ordering limitation is expressly taught by Cendrillon I and corroborated by Rezvani's own "order of importance" ranking. This is the most probative ground; it was pleaded, is coherent, and matches the claim language.
• Claim 11 (and the format claims): Likely obvious over Rezvani + Cendrillon I + G.992.3, subject to the narrow "crosstalk-channel-vs.-direct-channel application of Hlin formatting" counter-argument.
• The FEXT-from-NEXT LTI core: Probably obvious over the 47 cited references / LTI-admission in the specification / T1.417-2003 / Koeman-ELFEXT line of reasoning — but this ground is analytically unadjudicated and would need careful priority-date work (see the May 24, 2004 DSM-Report hazard in § 5).
• The narrowest claims (3, 6, 8, 13, 16, u = 0 / ΔX phantom mode): Not established as obvious on the record I could verify — these were never charted in the IPR.

Residual uncertainty I cannot eliminate from here: (i) Rezvani's actual § 102(e) prior-art date and specification disclosure must be verified against the reference itself (I could not pull the full text); (ii) Cendrillon I's exact sorting language should be quoted from the paper; (iii) the "Prior Art section" the task referenced was not present in the supplied page, so if a different prior-art list exists, this analysis should be reconciled against it.

Recommended next steps for a rigorous FTO/validity opinion: obtain the IPR2025-00492 petition and exhibits (Exs. 1001–1025) from PTAB/PACER, confirm Rezvani's filing date, and run the claim charts against claims 3, 6, 8, 13, and 16 — the unadjudicated claims are where validity, if any, now resides for this patent.