US 6792575

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

Let me check for the specific litigation/PTAB history and any 2026 appellate activity tied to this patent.

One more check on the PTAB proceedings and any 2026 appellate escalation.

US 6,792,575 B1 — Analyst Summary

Scope note on the search: I searched for patent number 6,792,575 specifically and did not surface any similar-but-different numbers as the subject of this report. I also found no Federal Circuit appeal docketed in 2026 that names US 6,792,575 as an asserted patent. If your workflow requires confirmation of "no CAFC activity," that is a negative finding, not a verified absence — see the uncertainty section below.

1. Bibliographic Data (authoritative, from the patent front page and USPTO sources)

Sources: Google Patents US6792575B1; issued-patent PDF (US006792575B1); USPTO Official Gazette entry.

2. Abstract

"A system using as input original media, an HTML document or browser language having proprietary tags, Web server traffic, and Web-client capabilities to generate an optimized Web media and HTML to refer to the generate[d] media, and to automatically deploy the HTML and media to the Web server is provided. A Web authoring process is provided for facilitating creation of the media, assignment of a unique name to the media, and modification of the HTML document or browser language to contain a proprietary tag. Viewing capability is provided by the Web server passing the HTML or browser language, client browser capabilities, and current server traffic to the system, which parses the HTML or browser language searching for the proprietary tags. If a proprietary tag is found, the tag is processed to generate the Web media. Information is stored in the system database in case identical proprietary tags are processed."

3. Independent Claims — Plain-Language Overview

(a) Claim 1 — Process for delivering an original media (a method claim).
Five practical steps:

1. Put the original media into a "network system."
2. Author an HTML (or other browser-language) document containing the system's proprietary media tags, and place it on a Web server.
3. On a customer's page request, have the Web server pass the requested page to an HTML parser on the network system.
4. The parser searches for the proprietary media tags; each tag is looked up in a media tags database on the network system.
5. For any tag not found: generate the Web media from that tag, put the generated media into a media cache, convert the proprietary tag into a standard HTML equivalent tag pointing at the cached media, and store the tag + HTML equivalent in the database. Then replace the proprietary tag(s) in the document with the standard HTML equivalent(s), return the modified document from the system to the Web server, and have the Web server deliver it to the customer's browser.

In short: server-side parse-and-substitute of custom tags into standard HTML plus cache-and-database-driven on-demand media generation.

(b) Claim 2 — Network system for delivering an original media (a "means-for" apparatus claim).
Mirrors claim 1 element-for-element, but recited as a system with "means for" performing each step: placement of original media, creating/placing HTML with proprietary tags, actuating the browser request to the parser, parsing, database lookup, the conditional generation/caching/conversion/storage block, generating the modified HTML, and the two delivery means. Under § 112(f), the disclosed structures map to the HTML parsing subsystem 610, media caching subsystem 620, media cache database 640, media creation subsystem 630, media processing engine 650, and media repository 660.

(c) Claim 3 — Media delivery system (structural/apparatus claim; Java servlet based).
Recites a system using a Java servlet engine, a Web server, and a customer browser, that generates HTML/browser language containing a proprietary image from an original document embedded with a proprietary tag. Components:

• an HTML parser subsystem that searches for the proprietary tag and replaces it with a standard HTML tag;
• a media caching subsystem that determines whether the proprietary image already exists or needs to be created, and forwards the standard HTML tag to the media creation subsystem;
• a media cache database storing data associated with the proprietary image;
• a media creation subsystem that deciphers the proprietary tag;
• a media processing engine that interprets the deciphered tag and generates the proprietary media; and
• a media repository storing the original media.

Claim 3 adds a specific database structure: a script table with columns for Media Processing Script, HTML Equivalent, Bandwidth, Generated File, and Dependency List; and a dependency table with File Name and Modification Date columns.

(d) Claim 9 — Process for creating a media (method claim; the cache-vs-regenerate decision logic).
This is the most granular independent claim and reads essentially as the Figure 10 flowchart:

• receive a request from the system for the HTML equivalent tag;
• combine the media tag with bandwidth information;
• determine whether a media tag entry exists in the media tag database;
• if it exists: check whether any original asset used to create the media has changed;
  • if not changed: retrieve and return the HTML equivalent tag;
  • if changed: delete the entry, re-create the media, cache it, generate HTML referring to the generated media, store the tag + HTML equivalent, and return the HTML equivalent;
• if it does not exist: create the media, cache it, generate the HTML/browser language referring to it, store the tag + HTML equivalent, and return the HTML equivalent.

In short: a dependency-aware cache-invalidation and media-regeneration loop keyed on the proprietary tag plus bandwidth.

Dependent claims 4–8 add: multiple tags/media/original media (4); stored data comprising the creation script + date, original media name + date, and generated HTML (5); means for detecting modification of script/original-media dates (6); parser means for sending tags, receiving standard tags, iterating, and detecting exhaustion (7); and means for determining whether the proprietary media has been modified (8).

4. Enforcement / Post-Grant Activity (relevant context, as of April 2026)

• District court: Equil IP Holdings LLC v. Akamai Technologies, Inc., No. 1:22-cv-00677 (D. Del.), filed May 23–24, 2022 before Judge Richard G. Andrews. Asserted patents: US 6,792,575, US 8,495,242, and US 9,158,745; accused product was Akamai's Image and Video Manager. The case was stayed on February 19, 2025 pending ex parte reexamination at the USPTO and then administratively closed — no merits ruling. (Unified Patents litigation portal; PatSnap case summary)
• PTAB: Google Patents flags IPR2023-00329 (status noted as "Settlement"). Related Akamai petitions in the same family include IPR2023-00330 (US 8,495,242) and IPR2023-00332 (US 9,158,745 — institution denied, Paper 15, July 21, 2023, on § 325(d) grounds). The '575 patent itself appears in the IPR2023-00330 record as Exhibit 1012, being cited as background/priority-chain evidence rather than as a challenged patent. (IPR2023-00332 institution denial; IPR2023-00330 exhibit list)
• Priority family: a large continuation/division family claims 1999-10-21 priority, including US 6,964,009 (continuation/CIP of 09/425,326), and later US 8,656,046, US 8,381,110, US 8,495,242, US 9,158,745, plus pending/abandoned applications. Note the '575 patent is a pre-AIA (1999) filing.
• Litigation flag: Google Patents also links a PTAB case IPR2023-00329 filed (Settlement) and a Darts-IP "first worldwide family litigation" entry for family 23686072.

5. Where I am NOT certain (explicit flags)

1. No 2026 CAFC docket found for '575. My searches did not identify a Federal Circuit appeal docketed in 2026 naming US 6,792,575. Given that the underlying Delaware case was stayed/administratively closed in 2025 and the '575 patent's term expired ~October 21, 2019, a 2026 CAFC appeal on the '575 patent is unlikely — but I could not conclusively verify a negative, so treat "no CAFC 2026 activity" as unconfirmed rather than established. (The 2026 Federal Circuit materials my searches surfaced — OpenSky/VLSI, Constellation Designs v. LG, Contour IP v. GoPro, etc. — are unrelated to this patent.)
2. Examiner discrepancy. The PTO issue file jacket (Form PTO-438A) shows Sanjiv Shah as Primary Examiner, while a third-party aggregator (SumoBrain) lists BASHORE, WILLIAM L. I treat the PTO file-jacket entry as more authoritative. (The PTO-438A scan also contains OCR artifacts — e.g., "506F" for "G06F" — which I have not auto-corrected in the quoted record.)
3. Assignee of record is layered. Google Patents lists two current assignees (Barger-Great South Ventures LLC and Equil IP Holdings LLC), and a 2022 security interest to Lit-US Chisum 22-A LLC. It is unclear from the public record which entity holds the operative enforcement rights today. Google Patents expressly disclaims having performed a legal analysis of assignee accuracy.
4. IPR2023-00329 subject matter. I could not retrieve the petition/challenged claims for IPR2023-00329 itself; the "Settlement" label comes from the Google Patents family page, and I did not independently verify the settlement or its terms.
5. Textual quality caveat. The '575 specification contains apparent typographical errors (e.g., "mhethodologies," "adiusting," "dicks," "The Web server 10") as they appear in the authoritative full text. I have quoted them literally rather than correcting them, per the instruction not to auto-correct identifiers/content.

Bottom line: US 6,792,575 is a 1999-filed, 2004-issued, now-expired Equilibrium Technologies patent on server-side parsing of proprietary HTML media tags into standard HTML, with on-demand media generation and a dependency-aware media cache. Its four independent claims (1, 2, 3, 9) respectively cover the method, the means-plus-function system, the component/system architecture with a specific cache-database schema, and the cache-vs-regenerate decision process. Its principal assertion was in Equil IP Holdings v. Akamai (D. Del. 1:22-cv-00677), which was stayed for ex parte reexamination and administratively closed; I found no 2026 CAFC appeal tied to this number.

I'll research litigation involving US patent 6792575 using multiple sources.

Let me dig deeper into specific cases and PTAB proceedings.

Bottom line

I found one district court lawsuit that asserted U.S. Patent No. 6,792,575, plus one PTAB proceeding directed at the ’575 patent. Both involved the same pair of parties: Equil IP Holdings LLC (plaintiff/patent owner) and Akamai Technologies, Inc. (defendant/petitioner).

I found no other litigation — no other defendants, no other districts, no CAFC appeal on the ’575 patent, and no 1999–2021 enforcement actions by the original owners (Equilibrium Technologies / Automated Media Processing Solutions).

1. Equil IP Holdings LLC v. Akamai Technologies, Inc.

Procedural history (context):

• The initial AO-120 form filed May 24, 2022 listed only the ’745 and ’242 patents against Akamai — the ’575 patent was added subsequently (the PatSnap litigation summary confirms three asserted patents; Equil's own PTAB mandatory notices list all three as at issue in 1:22-cv-00677).
• Claim construction opinion issued July 2, 2024 (D.I. 135), addressing the ’667 and ’745 patents.
• Aug. 28, 2024: the court granted Akamai's motion to strike Equil's infringement theory as to Akamai's Adaptive Image Compression ("AIC") product as untimely disclosed (D.I. 166).
• Feb. 19, 2025: the case was stayed pending ex parte reexamination at the USPTO (Akamai's ex parte reexam 90/019,461 against the '745 patent).
• Later in 2025 the cases were administratively closed — a procedural closure, not a merits ruling. The order states the cases "are ADMINISTRATIVELY CLOSED. The parties shall promptly notify the Court when the reexamination proceedings have been resolved so that this case may be reopened." Unified Patents lists the case status as "Closed."

Caution on the "win" narrative: Sterne Kessler's July 8, 2025 announcement that all claims were upheld concerns the ’745 patent (reexam 90/019,461), not the ’575 patent. I found no equivalent post-reexam validity ruling for the ’575 patent; the ’575 claims left the district court case via voluntary dismissal with prejudice.

Related second Delaware case (caveat): The administrative-closure order refers to "the above-captioned cases" and cites both C.A. 22-677-RGA and C.A. 22-1531-RGA. Both appear to be Equil v. Akamai matters (the claim-construction opinion notes the parties opposing Akamai "in these two cases" are the same). I could not confirm that C.A. 22-1531 asserted the ’575 patent — its claim-construction record references the ’667 and ’745 patents — so I am not listing it as ’575 litigation.

Sources:

• https://portal.unifiedpatents.com/litigation/caselist?patents=6792575&sort=-filed_date
• https://dockets.justia.com/docket/delaware/dedce/1:2022cv00677/79015
• https://storage.courtlistener.com/recap/gov.uscourts.ded.79015/gov.uscourts.ded.79015.75.0.pdf
• https://storage.courtlistener.com/recap/gov.uscourts.ded.79015/gov.uscourts.ded.79015.166.0.pdf
• https://www.patsnap.com/fr/resources/blog/litigation/equil-ip-holdings-v-akamai-technologies-cdn-edge-network-patents-patsnap/
• https://www.sternekessler.com/news-insights/news/sterne-kessler-scores-patent-reexamination-win-for-equilibrium-all-claims-upheld/

2. PTAB: Akamai Technologies, Inc. v. Equil IP Holdings LLC, IPR2023-00329

Note: IPR2023-00329 should not be confused with the companion Akamai petitions against the other patents in the same family — IPR2023-00330 (’242 patent) and IPR2023-00332 (’745 patent, institution denied on the merits in 2023). Those do not involve the ’575 patent.

Sources:

• https://patents.google.com/patent/US6792575B1/en (PTAB/litigation metadata block for the patent)
• https://www.docketalarm.com/cases/PTAB/IPR2023-00330/Akamai_Technologies_Inc._v._Equil_IP_Holdings_LLC/docs/01-31-2023-Patent_Owner/Notice__Mandatory_Notice-4-Notice__Mandatory_Notice.pdf
• https://www.sternekessler.com/news-insights/news/sterne-kessler-helps-secure-denial-ipr-against-patent-optimizing-media-content/

3. No other litigation found

• Unified Patents' "Litigation Case List" filtered to patents in case = 6792575 returns exactly one case: 1:22-cv-00677, Delaware, Equil IP Holdings LLC v. Akamai Technologies Inc., Infringement, status Closed.
• The Google Patents page for US6792575B1 shows only: (a) the PTAB case IPR2023-00329 (settlement), (b) the Delaware case 1:22-cv-00677, and (c) a Darts-IP "first worldwide family litigation" link for family ID 23686072 — consistent with a single enforcement campaign.
• I found no CAFC appeal docketed from 1:22-cv-00677, and no infringement suits against any defendant other than Akamai.

4. Look-alike numbers I deliberately excluded

To comply with your "no similar numbers" instruction, I filtered out these near-miss identifiers that surfaced in search results and do not relate to US 6,792,575:

• US 6,977,257 — the ABILIFY (aripiprazole) patent in DrugPatentWatch/ANDA litigation (e.g., Azurity Pharmaceuticals v. Annora Pharma, D. Del.) — a different patent, similar digits.
• US 9,836,257 — Flexiworld Technologies patent in Unified Patents v. Flexiworld, IPR2022-00775.
• US 6,792,575 ≠ any of the asserted patents in Ollnova v. ecobee (8,224,282 / 7,746,887 / 7,860,495 / 8,264,371), which appeared in results only because of proximity/indexing.

5. Confidence and caveats

• High confidence on the identity, court, case number, parties, and the ’575's dismissal with prejudice in 1:22-cv-00677 — grounded in the court's own D.I. 75 opinion and Unified Patents' docket data.
• Moderate confidence on the IPR2023-00329 outcome: I could verify the petition's existence and its reported "Settlement" disposition via Google Patents/Unified Patents metadata and Equil's mandatory notices, but I did not retrieve the Board's termination decision itself. Treat the "settlement" characterization as reported rather than independently confirmed.
• Some discrepancies exist in reported filing dates (May 23 vs. May 24, 2022) and in the apparent distinction between "Closed" (database status) and "administratively closed" (the court's actual order). The court's order is controlling: the case is stayed and administratively closed pending reexamination, not finally adjudicated.
• Databases such as Unified Patents and Darts-IP are generally reliable for post-2000 NPE campaigns but can be incomplete for older or non-electronic dockets. I found no indication of pre-2022 litigation, but I cannot affirmatively rule out an unindexed action in the 1999–2021 window.

I'll start by verifying the one proceeding flagged in the structured data and checking for any others.

Useful leads. Let me dig into the specific IPR numbers and the related litigation.

Let me pull the PTAB case pages and the Unified Patents litigation/PTAB records.

PTAB / AIA Trial Landscape — U.S. Patent No. 6,792,575 B1

"Automated processing and delivery of media to web servers" — App. No. 09/425,326, filed 1999-10-21, issued 2004-09-14, expired 2019-10-21 (anticipated). Owner of record: Equil IP Holdings LLC (assignment recorded 2021-12-23).

Sourcing note (read this first). The canonical U.S. Patent and Trademark Office Open Data Portal structured block supplied with this task returns zero AIA trial proceedings for US 6,792,575. Everything below in the Proceedings overview and the per-proceeding entries comes from web-sourced records that post-date or are not captured by that ODP ingest — specifically Google Patents' family/litigation metadata, Unified Patents' PTAB case portal, and Docket Alarm's PTAB docket mirror. I have flagged confidence levels at every field. I have not invented or inferred any proceeding number beyond the two that appear in those sources, and I have not found a Final Written Decision, an institution decision, or any Federal Circuit appeal for either case.

Proceedings overview

Total AIA trial proceedings identified against US 6,792,575: two (2) — IPR2023-00329 and IPR2023-00330.

Breakdown by status:

Bottom line defensive posture. This is not a "claims 1–5 have been canceled" situation, and it is not a "hardened patent that survived two IPRs" situation either. It is a third posture: no PTAB merits ruling exists at all. Both petitions were filed in early January 2023 and both were terminated by settlement before the Board ever decided institution — so claims 1–9 of the '575 patent are entirely UNTESTED at the PTAB and remain legally intact. For a defendant today: you get no free invalidity judgment to point at, but you also face no adverse PTAB precedent and no § 315(e) estoppel from these cases — the art and grounds that were in those petitions (and art that was never raised) are still fully available. The real validity pressure on this family has come from ex parte reexamination at the Central Reexamination Unit, which is what stalled the parallel district court case, not from the PTAB.

Proceedings

IPR2023-00329 — Petitioner not verified v. Equil IP Holdings LLC (patent owner of record)

• Type: Inter Partes Review (AIA trial; the structured Google Patents family record does not specify IPR/PGR/CBM, but the case-number prefix "IPR" does).
• Filed: Not verified. Petition numbering adjacent to IPR2023-00330 (whose petition exhibits are dated 2023-01-10) indicates a filing in the first half of January 2023. Confidence: medium.
• Status: "Settlement" — verbatim from the Google Patents structured litigation block: PTAB case IPR2023-00329 filed (Settlement). Plain-English gloss: the proceeding was terminated, most likely by a joint motion to terminate under 35 U.S.C. § 317(a) before an institution decision. Confidence in the "settled" label: high (it is the source's own status field); confidence in the mechanism (pre-institution § 317 termination): medium — the source does not give the termination paper.
• Judge panel: Not public / not found. No institution decision surfaced, so no panel has been publicly assigned in the records I could reach.
• Petition grounds: Not verified. I could not retrieve the petition, the challenged claims, or the art. Do not assume claims 1–9 were all challenged.
• Institution decision: None found. Consistent with a pre-institution settlement.
• Final Written Decision: None. No FWD exists. Consequently no claim of the '575 patent was canceled, and no claim was held patentable, by this proceeding.
• Settlement / termination: Occurred; terms are not public (settlement agreements filed under 35 U.S.C. § 317(b) / 37 C.F.R. § 42.74(c) are routinely treated as confidential business information, and I found no public copy). Confidence: medium-high.
• Appeal: None possible — an appeal to the Federal Circuit under 35 U.S.C. § 319 requires a final written decision, which does not exist here. No CAFC docket found.
• ⚠ Flagged ambiguity on the petitioner field. The Google Patents record renders a Petitioner: label immediately followed by the string "Unified Patents PTAB Data" by Unified Patents is licensed under a Creative Commons Attribution 4.0 International License. In my assessment that is Unified Patents' data-source attribution, not the name of the petitioner in this case. I cannot confirm whether Unified Patents LLC (the defensive aggregator) was the petitioner, a co-petitioner, or merely the metadata provider. Treat the petitioner identity in IPR2023-00329 as unverified. (Context, not proof: Unified Patents LLC does file its own Equilibrium/CDN-sector petitions — e.g., it is the named petitioner in IPR2023-00338, accorded a filing date of 2023-02-02, and it is a frequent § 315(b)-window filer against NPE-asserted patents — so the hypothesis is plausible but unsupported by anything I could verify.)
• Defensive value: Low as leverage, moderate as a signal. Nothing about this case invalidates or narrows a claim. But the fact that someone filed a third-party IPR against a 1999-priority patent within months of the Equil IP Holdings enforcement campaign, and that the patent owner then settled out of it, tells you the owner is willing to terminate PTAB exposure early rather than litigate validity to a judgment. For you, that means the prior art in the -00329 petition is still on the table for a new petition or a district court invalidity case.

IPR2023-00330 — Akamai Technologies Inc. v. Equil IP Holdings LLC

• Type: Inter Partes Review (AIA trial).
• Filed: Approximately 2023-01-10 (the petition's exhibit set is docketed 2023-01-10; the accorded filing date is not independently verified). Confidence: medium-high on the date; high on the petitioner.
• Status: "Settlement" (Google Patents family status field). Same gloss as above: terminated without an institution decision. Confidence: medium-high.
• Judge panel: Not public / not found.
• Petition grounds: Not verified at the claim/art level. What is verified is the paper trail: the petition's exhibit list (Docket Alarm, IPR2023-00330/Akamai_Technologies_Inc) includes Exhibit 1012 = US 6,792,575 B1 (the '575 patent itself) and Exhibit 1016 = the First Amended Complaint dated 2022-08-03 filed in D. Del. in the Equil–Akamai action. That exhibit set is strong evidence — though not conclusive — that IPR2023-00330 challenged the '575 patent. I could not retrieve the grounds, the challenged claim numbers, or the cited references, so I will not guess at § 102/§ 103 theories.
• Institution decision: None found.
• Final Written Decision: None. No claim of the '575 patent was canceled or confirmed by this proceeding.
• Settlement / termination: Occurred pre-institution; terms not public (standard § 317(b)/§ 42.74(c) confidentiality). Confidence: medium-high.
• Appeal: None. No FWD, so nothing appealable under § 319. No CAFC docket found.
• Defensive value: This is the more interesting of the two for you. Akamai was the defendant in the co-pending Delaware campaign, filed inside the § 315(b) one-year window (complaint served in 2022; petition ~2023-01-10), and then the whole thing resolved into settlement plus a reexamination-driven district court stay. The practical read: a well-resourced accused infringer built a full IPR record against the '575 patent and then settled rather than press for a judgment. That is not a loss for the patent owner, but it also means the '575 patent has never been tested by an Article III or PTAB tribunal on validity.

Related non-AIA USPTO activity (context, not a PTAB proceeding)

Not an AIA trial, but it is the single most important fact for assessing the '575 patent's exposure, and it is not in the structured PTAB block because it is an examination-side proceeding:

• Ex parte reexamination(s) pending at the USPTO drove the stay of the district court campaign. The Delaware court's stay order (quoted from the PACER docket for C.A. 1:22-cv-00677-RGA via the PatSnap litigation write-up) states: "WHEREAS, the above-captioned cases were stayed on February 19, 2025, due to Ex Parte Reexamination proceedings pending before the United States Patent and Trademark Office ('PTO') (D.I. 194 in C.A. 22-677-RGA; D.I. 80 in C.A. 22-1531-RGA); NOW THEREFORE, IT IS HEREBY ORDERED that, the above-captioned cases are ADMINISTRATIVELY CLOSED." The order is expressly a procedural closure, not a merits ruling, and preserves jurisdiction pending resolution of the reexaminations.
• Control numbers and outcomes: not verified. I could not retrieve the reexamination control numbers, their claim-by-claim determinations, or whether any certificate has issued. Do not assume any claim has been confirmed, narrowed, or canceled by reexamination.
• Litigation backdrop from the structured data: Equil IP Holdings LLC v. Akamai Technologies Inc, D. Del. 1:22-cv-00677, filed 2022-05-23, infringement, now closed/stayed; asserted patents included US 6,792,575, US 8,495,242, and US 9,158,745. A second caption (C.A. 22-1531-RGA) was stayed on the same order; the defendant in that second case is unverified.

Strategic summary

Claim-level status of US 6,792,575. No claim has been canceled, narrowed by certificate, or sustained by any PTAB tribunal on the record I could verify. All nine claims — independent claims 1, 2, 3, and 9, and dependent claims 4, 5, 6, 7, and 8 — are UNTESTED by any AIA trial. The claim set is, on paper, intact. The live claim-scope question is entirely a function of the ex parte reexamination files, which the district court deemed significant enough to stop the litigation in February 2025. If you are evaluating this patent for a licensing demand or a clearance opinion, your first document request is the reexamination file history, not the PTAB docket — the reexam, not the IPRs, is where the claims may actually be moving. Dates to police: any reexamination certificate issues → claims as amended become the new assertion target, and the Delaware case may be reopened (the order requires the parties to notify the court when the reexaminations resolve).

Estoppel landscape. This is arguably the most favorable part of the picture for a defendant. Section 315(e)(2) estoppel attaches only to "any ground that the petitioner raised or reasonably could have raised" after a final written decision. Neither IPR2023-00329 nor IPR2023-00330 produced an FWD — both terminated by settlement before institution. Accordingly:

• No § 315(e)(1) or § 315(e)(2) estoppel arises from either proceeding for the petitioners, their real parties in interest, or their privies.
• That cuts against a pure "stay and let the PTAB handle it" posture only because there is no dismissal-grade PTAB outcome to lean on; but it is good news for anyone who wants to run their own IPR, because the art assembled by the earlier petitioners is not procedurally spent, and there is no earlier FWD establishing (or foreclosing) claim constructions.
• Practical constraints on you: a fresh petition is still subject to § 315(b) (one year from service of a complaint on you) and the Board's discretionary-denial practice — which shifted sharply in 2025 (the Director rescinded the June 2022 Fintiv "compelling merits" interim guidance on 2025-02-28, and issued PTAB workload-management guidance on 2025-03-26), so a serial or late petition now carries more institution risk than it did in January 2023. And if the earlier petitioner was Akamai, a new Akamai petition (or one by its privies) would face the § 315(e) bar only if an FWD had issued — it did not.

Pattern signals.

• Two petitions, two different-looking filers, same ~two-week filing window, both settled pre-institution. That is a coordinated-looking pressure campaign on the owner rather than a single challenger grinding out a merits win.
• No serial petitioning by any one petitioner on this patent, and no repeat post-FWD filings, because nothing ever reached a merits decision.
• The patent owner has not pursued aggressive PTAB appeals of this patent — there was nothing to appeal. Its aggressive posture shows up in enforcement (Equil IP Holdings filed a multi-patent, multi-defendant Delaware campaign from 2022) and then in surrender-by-stay once reexamination started.
• Defensive aggregator in the chain: plausible but unconfirmed. Unified Patents appears in the structured data only as the data licensee/attribution, and it is not verified as a petitioner here. I flag this explicitly because it materially changes the "who's behind the challenge" narrative and I will not assert it without the petition cover page.
• Portfolio-wide litigation posture: this is a campaign patent, not a lone patent. The same owner asserted US 8,495,242 and US 9,158,745 alongside the '575 patent, and the structured data shows a nine-patent continuation family all claiming 1999-10-21 priority. Clearance or defense analysis on the '575 alone is incomplete — the family shares substantially the same specification.

Recommended next steps

1. Do not build a defense on a PTAB judgment that does not exist. There is no FWD to cite, so do not (and no one can) quote one. If a demand letter or complaint asserts claims 1–9 of the '575 patent, the correct response is not "the PTAB already killed this"; it is "the PTAB has never reached the merits, and here is our own art."
2. Pull the two petitions. Retrieve IPR2023-00329 and IPR2023-00330 from USPTO PTAB E2E (https://ptacts.uspto.gov/ptabweb/) and the Unified Patents portal entries at https://portal.unifiedpatents.com/ptab/case/IPR2023-00329 and the Docket Alarm mirror at https://www.docketalarm.com/cases/PTAB/IPR2023-00330/Akamai_Technologies_Inc/. You want four things: the identity of the real party in interest, the challenged claim numbers, the § 102/§ 103 grounds and exhibit lists, and the termination papers/§ 317 settlement filings. Those petition exhibit sets are free, ready-made prior-art packs — reusing them is the single highest-value, lowest-cost move available (with the caveat that any exhibit you rely on must be independently verified).
3. Pull the ex parte reexamination file(s) first in your ordering, though. Fetch the reexam control numbers via USPTO Patent Center / the reexamination docket for US 6,792,575 and check the current status of every claim. If a reexamination certificate issues narrowing or canceling claims, that — not the PTAB record — is the dispositive event. Also obtain the 2025-02-19 stay order and the 2025 administrative-closure order in 1:22-cv-00677 (PACER, D. Del.; the order is reproduced at https://www.patsnap.com/fr/resources/blog/litigation/equil-ip-holdings-v-akamai-technologies-cdn-edge-network-patents-patsnap/ and docket links are aggregated at https://portal.unifiedpatents.com/litigation/Delaware%20District%20Court/case/1%3A22-cv-00677).
4. Watch the two live clocks, not a PTAB 1-year clock. Neither IPR ever instituted, so no statutory 1-year trial deadline from § 316(a)(11) is running on this patent. The deadlines that matter are (a) the USPTO reexamination resolution, and (b) your own § 315(b) one-year window measured from service of any complaint on you. If a new complaint is served, calendar the § 315(b) date immediately — that is the only PTAB deadline you control.
5. Check whether the expired-patent timing helps you. The '575 patent's term expired 2019-10-21. Enforcement from 2022 onward can only reach pre-expiration past damages (subject to the § 286 six-year lookback). That does not bar an IPR (the Board does institute on expired patents), but it means the damages exposure is a closed, calculable historical period — a materially different negotiation posture than an in-force patent.
6. Verify, then re-verify, the two open questions. (i) Who was the petitioner in IPR2023-00329? (ii) Did IPR2023-00330 actually challenge the '575 patent, or a sibling? Both are answerable from the petition cover pages in PTAB E2E and neither is answerable from the sources I could reach. I have deliberately left both marked unverified rather than assert them.

I'll research the ownership chain using available sources. Let me start with parallel searches.

Good — I have litigation facts. Now let me dig for the recorded assignment details (reel/frame, correspondents) and entity backgrounds.

Note on source access: I could not open a live Assignment Center session from this environment, so the reel/frame and correspondent fields could not be pulled directly for most entries. The timeline below is built from the patent's own USPTO-derived legal-events record (reproduced in the authoritative document above), which does expose one reel/frame verbatim — Reel 015377 Frame 0392 — plus litigation and PTAB filings that restate the chain. Every point where a reel/frame or correspondent is unverified is flagged rather than filled in.

Inventors

Employer is determinable because all four appear as assignors to EQUILIBRIUM on the 1999-10-21 assignment of the application that issued as the '575 patent — the standard employee-obligation-to-assign pattern. The patent's front page lists the assignee as Equilibrium Technologies (Sausalito, CA).

Unusual pattern: not the "inventors depart" tell, but something adjacent. Within 67 days of filing (1999-12-27) the company pledged the patent as collateral to CMGI, Inc. — an investor/incubator taking a blanket security interest in the IP of a two-month-old portfolio. That is a capital-structure signal, not an inventor signal. I found no evidence, one way or the other, about how long any individual inventor stayed; I will not infer departures I cannot document.

Original assignee

Equilibrium Technologies (also rendered "Equilibrium Technology" and "Equilibrium" in the record), Sausalito, California — a small Web-media tooling company; the patent's own FIG. 11 shows an operator-facing administration/cache console, and the spec uses "freeride" as the internal product code name. Whether that shipped as a GA product embodying the claims is not verifiable from the sources I could reach — the only "product" evidence in the record is the patent's own GUI screenshot and the plaintiff's later litigation rhetoric, which is assertion language rather than product evidence.

Status: not an independent operating company today. The IP left the operating entity's hands: it was pledged to CMGI, Inc. (a publicly traded Internet holding company in the 1999–2001 period) on 1999-12-27, transferred out by CMGI to Barger-Great South Ventures, LLC on 2003-12-17, and placed into Automated Media Processing Solutions, Inc. dba Equilibrium ("AMPS, Inc.") in late 2004. That 2003→2004 sequence — a "Barger" entity receiving the asset from CMGI, followed by an assignment from BARGER, SEAN B. (later corrected to Equilibrium Technologies on Reel 015377/0392) into AMPS — reads as the original founder buying the asset back out of the investor's collateral position. Caveat: the founder identity link is an inference from the recorded surname plus the corrective assignment; I did not obtain a corporate record confirming it.

Assignment timeline

Data caveat, stated plainly: I could not retrieve USPTO Assignment Center reel/frame numbers or the "correspondent" field for these recordings. The dates below are the assignment events as surfaced in the patent's USPTO-derived legal-events feed; that feed does not distinguish execution date from recordation date, so treat each date as "assignment event date" unless a source says otherwise. The single verified reel/frame is 015377/0392, quoted in the corrective assignment itself. Correspondent entries are marked unverified where I have no recorded value.

• 1999-10-21 (executed) / recorded 1999-10-21 — Reel/Frame not retrieved

  • Conveyance: Assignment of Assignors' Interest
  • Assignor: Boyd, David R.; Offner, Nelson H. "Rocky"; Samaniego, Christopher; Thewlis, Adrian D.
  • Assignee: Equilibrium (Equilibrium Technologies), Sausalito, CA
  • Correspondent: unverified. For context only — the patent's prosecution attorney/agent of record is Glenn Patent Group, Seattle, WA (per SumoBrain's front-page extract); that is the patent-face representative, not necessarily the assignment recording correspondent. Do not conflate them.
  • Context: Founders' initial assignment to the operating company at filing.

• 1999-12-27 (executed) / recorded 1999-12-27 — Reel/Frame not retrieved

  • Conveyance: Collateral Assignment
  • Assignor: Equilibrium
  • Assignee: CMGI, Inc.
  • Correspondent: unverified
  • Context: Securitization — the investor CMGI takes a security interest in the patent ~2 months after filing.

• 2003-12-17 (executed) / recorded 2003-12-17 — Reel/Frame not retrieved

  • Conveyance: Assignment of Assignors' Interest
  • Assignor: CMGI, Inc.
  • Assignee: Barger-Great South Ventures, LLC
  • Correspondent: unverified
  • Context: Transfer out of CMGI's collateral position, consistent with a post-bubble workout/sale rather than an operating acquisition. (No Chapter 7/11 filing was located — see signal 6.)

• 2004-11-19 (executed) / recorded 2004-11-19 — Reel/Frame 015377/0392 ✅ (the one reel/frame confirmed from the document text)

  • Conveyance: Assignment of Assignors' Interest
  • Assignor: Barger, Sean B. (as originally recorded)
  • Assignee: Automated Media Processing Solutions, Inc. dba Equilibrium ("AMPS, Inc.")
  • Correspondent: unverified
  • Context: Asset moved from the Barger holding entity into the new operating vehicle AMPS.

• 2004-12-02 (executed) / recorded 2004-12-02 — Reel/Frame not retrieved

  • Conveyance: Corrective Assignment — "to correct the conveying party previously recorded on Reel 015377 Frame 0392"
  • Assignor: Equilibrium Technologies (substituted for Sean B. Barger)
  • Assignee: Automated Media Processing Solutions dba Equilibrium "AMPS Inc."
  • Correspondent: unverified
  • Context: Change-of-party correction only; no new consideration, no change in ultimate ownership.

• 2005-05-27 (executed) / recorded 2005-05-27 — Reel/Frame not retrieved

  • Conveyance: Security Agreement
  • Assignor: Automated Media Processing Solutions, Inc. dba Equilibrium, "AMPS, Inc."
  • Assignee: Remington Partners, Inc.
  • Correspondent: unverified
  • Context: Securitization — the operating entity pledges the patent to a lender.

• 2021-12-03 (executed) / recorded 2021-12-03 — Reel/Frame not retrieved

  • Conveyance: Release by Secured Party
  • Assignor: Remington Partners, Inc.
  • Assignee: Automated Media Processing Solutions, Inc.
  • Correspondent: unverified
  • Context: Extinguishment of the 2005 lien — the necessary clean-up step before sale of the asset.

• 2021-12-17 (executed, per Plaintiff's First Amended Complaint ¶ 20 dated Aug 3, 2022) / recorded 2021-12-23 (per legal-events feed) — Reel/Frame not retrieved

  • Conveyance: Assignment of Assignors' Interest
  • Assignor: Automated Media Processing Solutions, Inc.
  • Assignee: Equil IP Holdings LLC
  • Correspondent: unverified
  • Context: Transfer-to-asserter. The operating entity exits; a holding LLC with no products takes the whole family. Note the 20-day gap between the release (12-03) and the assignment event (12-17/12-23): lien cleared, then asset sold.

• 2022-04-07 (executed) / recorded 2022-04-07 — Reel/Frame not retrieved

  • Conveyance: Security Interest
  • Assignor: Equil IP Holdings, LLC
  • Assignee: LIT-US Chisum 22-A LLC
  • Correspondent: unverified
  • Context: Securitization on the assertion vehicle — a litigation-finance-style SPV takes a security interest 46 days before the Delaware complaint is filed. See signal 5.

No further post-2022 recordings appear in the legal-events feed I could reach. The patent's anticipated expiration is 2019-10-21 on the Google Patents page, which conflicts with the 2022 recordings and 2023–2025 litigation — that expiry field is unreliable and should not be relied on.

Timeline diagram

timeline
    title Ownership of US 6792575
    1999 : Inventors assign to Equilibrium
         : Collateral assignment to CMGI
    2003 : CMGI transfers to Barger Great South
    2004 : Assigned to AMPS dba Equilibrium
         : Corrective assignment on 015377 0392
    2005 : Security agreement to Remington
    2021 : Remington releases security
         : Assigned to Equil IP Holdings
    2022 : Security interest to LIT-US Chisum
         : Akamai suit filed in Delaware
    2025 : Case stayed and closed

NPE / troll-pattern signals

1. Shell-entity transfer — PRESENT.
Equil IP Holdings LLC took the family on 2021-12-17 (executed) / 2021-12-23 (recorded) from Automated Media Processing Solutions, Inc. The legal-events entry reads "ASSIGNMENT OF ASSIGNORS' INTEREST"; the First Amended Complaint confirms the transaction date. Supporting tells: "Holdings" suffix; the assignor operating entity exited the chain the same month the lien was released (2021-12-03); the LLC appears in court only as a patent plaintiff; and the asset was simultaneously pledged to a separate SPV (signal 5). Conflicting datum, disclosed for completeness: Unified Patents' litigation portal classifies the plaintiff entity type for 1:22-cv-00677 as "Operating Company," and the Ex Parte AI Lab party page labels it "Operating Company" as well. I treat those as third-party heuristics that conflict with the assignment record and with the absence of any product evidence; the assignment record controls. Do not rely on the "Operating Company" label without independent product evidence.

2. Known asserter in the chain — NOT PRESENT (against the enumerated lists).
No assignee anywhere in the chain matches Acacia Research, Marathon Patent Group, Intellectual Ventures, IPNav, Wi-LAN, Mosaid/Conversant, Vringo, Pendrell, Innovatio IP Ventures, MPHJ, Lumen View, Round Rock Research, Document Generation Corp, or any identified Erich Spangenberg entity. Equil IP Holdings LLC and LIT-US Chisum 22-A LLC are not on those lists. Equil is, however, a repeat patent plaintiff in D. Del. (multiple cases per the Ex Parte party page) — a "frequent plaintiff" designation, not membership in a named NPE fund.

3. Repeat correspondent across the chain — UNCLEAR (insufficient data).
I could not retrieve the recorded correspondent fields from Assignment Center, so I cannot show recurrence on the assignment records, which is where the signal lives. What I can document is recurrence of assertion counsel across two fora for the same owner: Charhon Callahan Robson & Garza, PLLC (USPTO Customer No. 172361) and Mauriel Kapouytian Woods LLP (USPTO Customer No. 51417) appear jointly on Equil IP Holdings LLC's Power of Attorney dated 2023-01-30 in IPR2023-00330, and the same two firms appear as plaintiff's counsel in the Delaware action (Dallas and San Francisco offices respectively); Potter Anderson & Corroon LLP (Wilmington) and Duane Morris LLP also appear for plaintiff in D. Del. Per the stated rule, a single appearance is not a finding — this is recurrence of litigation counsel, not evidence of a single attorney running a chain of shell LLCs. Marked unclear.

4. Cascading transfers — PRESENT (moderate).
Three recordings touching this asset in ~4 months: 2021-12-03 (Release by Secured Party, Remington Partners → AMPS), 2021-12-23 (Assignment of Assignors' Interest, AMPS → Equil IP Holdings), 2022-04-07 (Security Interest, Equil IP Holdings → LIT-US Chisum 22-A LLC). That is a classic "clean the lien, flip the asset, encumber it for litigation" sequence. It is not a chain of multiple successive assignee LLCs (only one new assignee appears), so this is a moderate, not strong, hit.

5. Pre-litigation transfer — PRESENT.
The assignment to Equil IP Holdings LLC executed 2021-12-17 (recorded 2021-12-23). The Delaware complaint Equil IP Holdings LLC v. Akamai Technologies, Inc., No. 1:22-cv-00677-RGA (D. Del.) was filed 2022-05-23 (court docket; the AO 120 form dated 5/24/2022 covers U.S. 9,158,745 and 8,495,242, with U.S. 6,792,575 added by the First Amended Complaint of 2022-08-03). Elapsed time: ~5 months — inside the 6-month window. Reinforcing this: the 2022-04-07 security interest to LIT-US Chisum 22-A LLC landed 6 weeks before the complaint. Both the clean-standing transfer and the financing encumbrance predate the suit.

6. Bankruptcy fire-sale — NOT PRESENT.
No Chapter 7 or Chapter 11 proceeding involving Equilibrium, CMGI, Barger-Great South Ventures, or AMPS was located in the sources reachable here. The 1999 collateral assignment to CMGI and the 2003-12-17 transfer out of CMGI are a workout / liquidation of an investor's collateral position in the post-bubble Internet downturn, which is a distinct fact pattern from a bankruptcy sale under §363. I am not upgrading this to "present" without a docket.

7. Privateering — NOT PRESENT.
No operating company has been identified as the beneficiary of Equil IP Holdings' assertions. The original operating entity (AMPS dba Equilibrium) assigned away the family and exited; there is no evidence it retains a license-back, a revenue share, or any role in the Akamai suit. Absent that, this is a straight sale to an assertion vehicle, not privateering.

8. Defensive aggregator — NOT PRESENT.
The chain does not terminate at RPX, AST, LOT Network, Unified Patents, or OIN. Note the inverse situation: Unified Patents — a defensive aggregator — petitioned for IPR2023-00329 against the '575 patent (Google Patents records the case as settled), and Akamai separately filed IPR2023-00330 ('242) and IPR2023-00332 ('745). Those are challenges to the patent, not acquisitions of it; the patent remains with the asserter.

Verdict

NPE — high confidence.

Three signals are documented, not inferred: (i) a shell-entity transfer on 2021-12-17 / recorded 2021-12-23, taking the family out of the operating entity AMPS into Equil IP Holdings LLC immediately after the 2021-12-03 release of the Remington lien; (ii) a pre-litigation transfer — that assignment predates the 2022-05-23 Akamai complaint by roughly five months, inside the six-month window; and (iii) cascading encumbrance — a 2022-04-07 security interest to LIT-US Chisum 22-A LLC, recorded six weeks before suit, which is the signature of litigation financing arranged in advance of filing. The chain's only verified reel/frame, 015377/0392 (with its 2004-12-02 corrective assignment), concerns an earlier, unrelated internal transfer and does not affect this conclusion. The assignment-correspondent analysis that would normally confirm the pattern (signal 3) could not be completed, so I flag that gap explicitly rather than letting a missing field read as a negative.

Verification links: USPTO Assignment Center — search patent number 6792575; legacy Assignment Search. Litigation/PTAB cross-references: Unified Patents litigation list for 6792575 and the Delaware docket at 1:22-cv-00677-RGA (case stayed 2025-02-19 pending ex parte reexamination and administratively closed, per the court's 2025-02-14 order — no merits ruling).

I'll research the prior-art references cited in US 6,792,575. Let me run targeted searches.

Now let me research the remaining cited references.

Prior-Art Analysis — US 6,792,575 B1

A note on sourcing and limits: I retrieved the authoritative full text and bibliographic data from the Google Patents page for US6792575 (source URL: https://patents.google.com/patent/US6792575/en), which mirrors the USPTO record. My tool budget was exhausted before I could separately query USPTO PatentCenter full-text or run independent searches on EP0843276 and AU-A-53031/98; for those two I rely on the dates and descriptions as recited in the '575 specification itself and flag them accordingly. I have not auto-corrected any patent number, and I interpret every identifier literally.

1. The patent at issue

Claim structure referenced below ("the claims"):

• Claim 1 – process: place original media in a network system; author an HTML/browser-language document carrying proprietary media tags; at a user request the Web server passes the page to an HTML parser; parser looks for proprietary tags; look up tags in a media tags database; on a cache miss, generate Web media from the tag, place it in a media cache, convert the tag to a standard HTML equivalent tag referring to the cached media, and store the tag + HTML equivalent in the database; replace tags in the document; return the modified document to the Web server and on to the customer browser.
• Claim 2 – means-plus-function system counterpart of claim 1.
• Claim 3 – system architecture: HTML parser subsystem, media caching subsystem, media cache database, media creation subsystem, media processing engine, media repository — plus a Script Table (Media Script / HTML Equivalent / Bandwidth / Generated File / Dependency List columns) and a Dependency Table (File Name / Modification Date columns).
• Claims 4–6, 8 – plural tags; stored script + original-media names/dates + generated HTML; means for detecting modification of script date / original-media date; means for detecting modification of the proprietary media.
• Claim 7 – parser subsystem means (send tag to cache subsystem, receive standard tag, find next tag, determine no more tags).
• Claim 9 – media-creation process: receive request for HTML equivalent; combine the media tag with bandwidth information; check the media tag database; on hit, check whether original assets changed; on change or miss, delete the entry, create the media, cache it, generate the referring HTML, store tag + HTML equivalent, and return the HTML equivalent.

Important framing: The specification's "Description of the Prior Art" section argues affirmatively that all of the disclosed prior art fails to deliver the claimed functionality. The references below are therefore mostly §103 combination art, not clean §102 anticipations — the applicant cited them to distinguish, not to admit anticipation. I note where each comes closest to a specific claim element.

2. Patent citations in / cited by US 6,792,575

2.1 US 5,355,472 — Lewis

• Full citation: Lewis, Jonathon R. T., System for substituting tags for non-editable data sets in hypertext documents and updating web files containing links between data sets corresponding to changes made to the tags, U.S. Pat. No. 5,355,472. Assignee: International Business Machines Corporation.
• Dates: Filed 11/19/1990; issued 10/11/1994; earliest priority 04/10/1990 (corresponding EP 0 451 384 A1, app. 90303870.1, filed 10.04.1990). (Confirmed via Google Patents / RPX / EPO publication server.)
• Description: Hypertext data-processing system that separates link information ("web file," an SQL table) from data sets; it injects tags into the data set at link locations to produce an editable file, permits conventional editing, then removes the tags and regenerates the output data set, updating the external link information. Tags may reference entries in a table to keep tags short.
• Potential §102 mapping: This is the closest cited reference to the tag-insertion / tag-substitution concept underlying claims 1 and 2 ("replacing said at least one media tag by said at least one standard HTML equivalent tag") and the table-referenced tag concept. However, it operates at authoring/edit time, not at Web-request time; it involves hypertext link tags, not proprietary media tags; and it discloses no media generation engine, no media cache, no bandwidth adaptation, and no Web server/parser transaction. It does not anticipate any of claims 1–9. Its realistic value is as §103 art on the tag-in/replace-tags-out element.
• Statutory note: Issued >1 year before 10/21/1999 → available as §102(b) art.

2.2 US 5,708,845 — Wistendahl

• Full citation: Wistendahl, Douglass A., et al., System for mapping hot spots in media content for interactive digital media program, U.S. Pat. No. 5,708,845.
• Dates: Application 08/536,107, filed 09/29/1995; issued 01/13/1998. (Confirmed.)
• Description: Media content ("Frame Data") is kept intact; interactivity is driven by object mapping data ("N Data") maintained separately from the frame data, specifying frame addresses and display coordinates of "hot spots." An IDM program links mapped objects to functions; in a network embodiment the Frame Data and N Data (and IDM program) are downloaded from a network server to a subscriber terminal on request.
• Potential §102 mapping: Relevant to the idea of storing original media separately in a repository and referencing it via separate metadata (cf. the media repository 660 / media cache database 640 concepts, claims 1, 3, 5). But it discloses no proprietary HTML media tags, no server-side media generation, no caching by tag, and no HTML tag substitution. Does not anticipate claims 1–9. Authors are the same Wistendahl family that later produced US 6,496,981 and US 7,577,978.
• Statutory note: Issued >1 year before 10/21/1999 → §102(b) art.

2.3 US 5,701,451 — Rogers et al.

• Full citation: Rogers, Richard M., et al., Method for fulfilling requests of a web browser, U.S. Pat. No. 5,701,451. Assignee: International Business Machines Corporation.
• Dates: Application filed 06/07/1995; issued 12/23/1997. (Confirmed.)
• Description: A Web browser makes requests to a Web server that fulfills them as an agent of the browser. A control program agent on the HTTPD invokes DIS capsule objects that retrieve data from multiple databases, perform calculations and formatting, and generate/report an HTML document back to the requesting browser. Claims recite displaying an HTML document, invoking the control program agent, receiving user-entered data as input parameters, and using an API to invoke command-file objects.
• Potential §102 mapping: Closest cited reference to the server-side, request-time dynamic generation of HTML delivered to a browser element of claim 1 ("delivering said modified HTML document ... to said Web server ... to said customer browser") and of claims 3/9's generation-then-return flow. It does not disclose proprietary media tags, a media tags database keyed by tag, a media cache, conversion of a proprietary tag to a standard HTML image tag, or dependency/date tracking. Does not anticipate claims 1–9; §103 art on the "dynamically build the response document from stored program objects at request time" limitation.
• Statutory note: Issued >1 year before 10/21/1999 → §102(b) art.

2.4 US 5,710,918 — Lagarde et al.

• Full citation: Lagarde, Konrad C., et al., Method for distributed task fulfillment of web browser requests, U.S. Pat. No. 5,710,918. Assignee: International Business Machines Corporation. (Companion to US 5,701,451 and US 5,752,246.)
• Dates: Application filed 06/07/1995; issued 01/20/1998. (Confirmed.)
• Description: DIS servers execute capsule objects that, on a single request arising from a Hypertext document, retrieve data from multiple geographies/types of databases via a database gateway, perform calculations and formatting, and report results to the browser (or to fax/printer/TV) in a selected format.
• Potential §102 mapping: Same family of teaching as Rogers; relevant only to the general request → server-side processing → formatted HTML/image result returned to browser architecture (claims 1, 3, 9). No proprietary tags, no cache, no dependency tracking. Does not anticipate claims 1–9.
• Statutory note: Issued >1 year before 10/21/1999 → §102(b) art.

2.5 US 5,844,084 — Cordell et al.

• Full citation: Cordell, John P., et al., Automatic data display formatting with a networking application, U.S. Pat. No. 5,844,084. Assignee: Microsoft Corporation.
• Dates: Application 08/634,407, filed 04/18/1996; issued 12/01/1998. (Confirmed.)
• Description: When a browser requests a document, the receive data transfer rate is monitored; if it is slow and the document embeds a graphic of unknown dimensions, a small placeholder image is displayed automatically in place of the real data, holding the layout until the real image arrives, whereupon the placeholder is removed and the display reformatted.
• Potential §102 mapping: This is the closest cited reference to the bandwidth/throttling thread running through the '575 claims. It is the only cited reference expressly responsive to transfer rate, which maps loosely to the "Bandwidth" column of the Script Table (claim 3) and the "combining said media tag with bandwidth information" step of claim 9. But it is a client-side display mechanism driven by measured rate, not a server-side, tag-driven generation of a bandwidth-appropriate media asset, and it stores no script/dependency records. Does not anticipate any claim; §103 art on bandwidth-responsiveness.
• Statutory note: Issued 12/01/1998, i.e., within one year of the 10/21/1999 filing — available under §102(a)/(e) (and as §102(e) art by virtue of its 04/18/1996 filing date), but not §102(b).

2.6 EP 0 843 276 — Tyan et al.

• Full citation: Tyan et al., HTML Generator, European Patent Application No. EP 0843276.
• Date: Published 05/20/1998 (date as recited in the '575 specification; I could not independently re-verify this publication date within my tool budget — treat as reported-by-the-patent, not independently confirmed).
• Description (per the '575 specification): "Generating an HTML file based on an input bitmap image ... automatic generation of an HTML file, based on a scanned-in document image, with the HTML file in turn being used to generate a Web page that accurately reproduces the layout of the original input bitmap image."
• Potential §102 mapping: Relevant only to the generic idea of automatically producing an HTML file that references/produces Web imagery (cf. step "generating an HTML document referring to said generated media," claim 9). It is a scanned-document/OCR-adjacent layout generator, with no proprietary tags, no parser intercepting a live page request, no cache and no dependency dates. Does not anticipate claims 1–9.
• Statutory note: If published 05/20/1998, that is >1 year before 10/21/1999 → potentially §102(b) art. (Verify publication date before relying on this.)

2.7 AU-A-53031/98 — Dudley

• Full citation: Dudley, John Mills, Network-Based Classified Information Systems, Australian patent application AU-A-53031/98.
• Date: Published 08/27/1998 (as recited in the '575 specification; not independently re-verified within my tool budget).
• Description (per the '575 specification): A system for "automatically creating databases containing industry, service, product and subject classification data, contact data, geographic location data (CCG-data) and links-to web pages from HTML, XML, or SGML encoded web pages posted on ... Internets or Intranets"; the databases are searchable for URLs via queries; "data duplication and coordination is reduced by including in the web page CCG-data display controls which are used by web browsers to format for display the same data that is used to automatically update the databases."
• Potential §102 mapping: Closest cited reference to parsing markup to populate a database at page-serving time (cf. claim 1's "parsing ... by looking for said proprietary media tags" and "looking up said proprietary media tags in a media tags database") and to embedding non-standard display controls in Web pages (loosely, the "proprietary tag" concept). It contains no image/media generation, no media cache, and no HTML-tag substitution. Does not anticipate claims 1–9.
• Statutory note: If published 08/27/1998, >1 year before 10/21/1999 → potentially §102(b) art. (Verify.)

2.8 Non-patent literature cited

• D. C. A. Bulterman, Models, Media, and Motion: Using the Web to Support Multimedia Documents, Proceedings of the 1997 International Conference on Multimedia Modeling, Singapore, 17–20 November 1997. Describes W3C Synchronized Multimedia working-group efforts toward a standard document format for rich multimedia presentations over the Internet. §102 relevance: none — it is a standards/format paper; no tag-substitution, caching, or generation system. Possible §103 "general knowledge" context only.
• "Text and Graphics on UMI's ProQuest Direct. The Best (yet) of both Worlds," Online, vol. 21, no. 2, pp. 73–77, March–April 1997. Describes an information system retrieving articles at varying detail levels with "Page Image" and "Text+Graphics" (graphics stored separately from text and manipulable as discrete items). §102 relevance: closest to the idea of storing media separately from text and serving variants on demand (claims 1, 5). Not anticipatory.
• TrueSpectra (products IrisAccelerate and IrisTransactive; described as "patent pending") — server-side image zoom/pan/transform, Flashpix-dependent, JPEG-only compression settings, a simple time-based cache with no dependency features for image propagation, and an XML-based compositing subsystem adding 5k–50k of XML per site. §102 relevance: this is the most substantively overlapping prior art discussed, because it (a) serves server-generated image derivatives, (b) caches them, and (c) composits layers. The specification expressly distinguishes it on the grounds that its cache lacks dependency/propagation features, requires Flashpix for performance, uses XML documents and hard-to-read compositing commands in HTML, and requires Javascript/ASP or a dedicated port. Not anticipatory of claims 1–9 as described; potentially strong §103 art on the caching + server-side generation + compositing aspects if the features were substantiated by dated documentation (the '575 text gives no publication date for the TrueSpectra materials — this must be established as prior art by corroborating evidence).

3. Bottom line on §102 anticipation

No single reference cited in US 6,792,575 discloses every element of any one of claims 1–9. Each of the claims 1, 2 and 9 requires a conjunctive set of limitations that no cited reference teaches together — specifically: (i) a proprietary media tag embedded in the authored HTML, (ii) a request-time Web-server→parser handoff, (iii) a tag-keyed media-tags database lookup, (iv) on a cache miss, generation of a media asset from the tag and its placement in a media cache, (v) conversion of the proprietary tag into a standard HTML tag referring to the cached media, and (vi) storage of the tag/HTML-equivalent pair in the database. Claims 3, 5–8 add the Script/Dependency table schema and modification-date dependency tracking, which no cited reference approaches.

Closest-element mapping (for §103, not §102):

Also noted: The 170 "Cited By" documents on the face of US 6,792,575 (e.g., US 2002/0116172, US 2002/0198961, US 2005/0097445) are later publications that cite the '575 patent; they are not prior art against the '575 claims and cannot anticipate them. Separately, US 2002/0078093 (Samaniego), appearing in the family/PTAB record (IPR2023-00330, Petitioner Akamai), is the publication of US 09/929,904 — a continuation of the '575 application itself — so it is a same-family member, not §102 prior art. The PTAB proceeding IPR2023-00329 was filed and terminated by settlement, and the Delaware case 1:22-cv-00677 is recorded; neither is itself anticipatory art.

Unverified items to confirm before any opinion is relied upon: the publication date of EP 0843276 (reported as 05/20/1998) and of AU-A-53031/98 (reported as 08/27/1998), and the date/authenticity of the TrueSpectra materials. My tool budget ran out before I could independently verify those three, and I will not assert them as confirmed.

I'll ground this analysis in the patent's own record and cross-check the litigation/PTAB history.

Obviousness Analysis — U.S. Patent No. 6,792,575 B1

"Automated processing and delivery of media to web servers"
Inventors: Samaniego, Offner, Thewlis, Boyd · Original assignee: Equilibrium Technology · Filed/Priority: 1999-10-21 · Granted: 2004-09-14 · Status: Expired – Lifetime (anticipated expiration 2019-10-21)
Source of record: https://patents.google.com/patent/[US6792575](/patent/US6792575)/en

0. Scope, method, and caveats

This analysis applies 35 U.S.C. § 103 as construed by Graham v. John Deere Co., 383 U.S. 1 (1966) and KSR Int'l Co. v. Teleflex Inc., 550 U.S. 398 (2007), and the framing of MPEP §§ 2141–2144. Because the patent is pre-AIA (effective filing date 1999-10-21), the pre-AIA §§ 102/103 framework governs.

Two procedural points that shape the analysis:

1. The Background section is an evidentiary gift to a challenger. Under Riverwood Int'l Corp. v. R.A. Jones & Co., 324 F.3d 1346 (Fed. Cir. 2003), and MPEP § 2129, an applicant's characterization of the prior art in the specification generally operates as an admission that the material exists and is prior art. Every reference analyzed below is described in the patent's own "Description of the Prior Art," which removes the "is this really prior art?" fight for most of them.
2. The public record on the IPR is thin. The Google Patents page records a PTAB case, IPR2023-00329 (Petitioner: Unified Patents), terminated by settlement, plus a Delaware case, Equil IP Holdings LLC v. Akamai Technologies, Inc., 1:22-cv-00677 (https://portal.unifiedpatents.com/litigation/Delaware%20District%20Court/case/1%3A22-cv-00677). My searches did not surface the petition's grounds. I therefore cannot represent that the grounds below mirror what Unified Patents actually argued — treat this as an independent § 103 reconstruction, not a report of the IPR record. (Note also that the D. Del. claim-construction order at D.I. 135 construed the '667 and '745 patents — continuations, not the '575 — so it does not construe the '575 claim terms.)

Assumed level of ordinary skill (POSITA): as of October 1999, a B.S. in computer science/EE or equivalent, with 2–4 years of experience building HTTP server-side applications (CGI/servlet/perl), HTML authoring, and image-processing/file-format handling. This POSITA understood HTTP request/response semantics, server-side includes and template preprocessors, database indexing, and basic image transforms.

1. The claim set, decomposed

2. Prior art available from the patent's own Background

Supplementing the Background (and confirming what a POSITA already knew in 1999): server-side template/preprocessor languages that detect non-standard proprietary tags in an HTML source and replace them with standard HTML before the page leaves the server were commercially ubiquitous by 1997–1999 — e.g., server-side includes (<!--#include-->, <!--#exec-->), Netscape/SSI LiveWire, ColdFusion CFML (1995), Microsoft Active Server Pages (1996), and PHP/FI (1995). This matters because "proprietary tag in, standard HTML out, at the server" is the single most-argued limitation of claims 1, 2 and 3, and it is squarely old.

3. Element-by-element § 103 chart for representative claim 1

Net: every element of claim 1 has an express antecedent in the Background references; none of the elements requires an image-processing operation, and the "proprietary tag" flavor is supplied by Dudley + Lewis + the then-ubiquitous server-side template-preprocessor art.

4. Specific proposed combinations and the motivation to combine

Ground 1 — Rogers '451 and/or Lagarde '918 + Tyan EP 0843276 (U.S. 5,893,127) + Dudley AU-A-53031/98 → claims 1, 2, 7 (and 4)

Where each fails alone / what the combination supplies: Rogers/Lagarde supply the architecture (browser request → intermediary server-side system → database access → processed output to the browser) but are silent on generating image derivatives and emitting the HTML that references them. Tyan supplies exactly that: an automated generator that creates an image file and generates an HTML file containing an HTML link to the generated image file (U.S. 5,893,127 claim 61). Dudley supplies the missing tag-driven server-side parsing → database → served page loop, and the idea of in-page controls that drive both display formatting and automatic database maintenance.

Motivation (KSR rationales):

• Known technique to improve a similar device: Rogers/Lagarde already obtain data "from one or more databases" to build a response; substituting Tyan's generate-and-link image pipeline and Dudley's page-derived database for static lookup is the use of a known technique for its known purpose — dynamic page composition — with predictable results. KSR, 550 U.S. at 416–17.
• Addressing a recognized deficiency: Rogers/Lagarde contemplate serving content to heterogeneous Web clients; the whole point of Tyan-style generation and Cordell-style adaptation is that a single stored source can be served at many sizes/qualities. A POSITA confronting "one source asset, many browser/bandwidth variants" would combine them.
• Design incentive: reducing per-page manual authoring labor is the stated "bottlenecked workflow" the patent itself identifies; both Rogers/Lagarde and Tyan are pointed at that exact problem.

Ground 2 — TrueSpectra IrisAccelerate/IrisTransactive + Lewis U.S. 5,355,472 + Cordell U.S. 5,845,084 → claims 1, 2, 3, 4, 7, 9

Where each fails alone / what the combination supplies: TrueSpectra supplies the server-side, on-demand generation of transformed/compressed image derivatives, with caching and HTML/XML parameterization — but it works through an XML document plus layer parameters, and the patent faults it for "a lot of overhead" (5–50 k XML per page) and for lacking "any dependency features for image propagation." Lewis supplies in-place tag insertion/removal with regeneration of a revised document — i.e., the tag→substitute→revised-document mechanic that lets the author keep the source document in a proprietary-tag form. Cordell supplies bandwidth-conditioned serving of images (monitor transfer rate → substitute a substitute image), which maps to claim 9's "combining said media tag with bandwidth information."

Motivation:

• Simplification of the prior technique while preserving function: replacing TrueSpectra's separate XML side-file with an inline proprietary tag that the server resolves is exactly the kind of "simplification" that is obvious where the elements work as expected — the patent's own stated advantage ("The need for the Web author to create different versions of a Web site is reduced," and eliminating XML overhead). KSR cautions that the availability of a design variation the inventor predicted and criticized does not itself create patentability. 550 U.S. at 418–19 ("if a technique has been used to improve one device, and a person of ordinary skill … would recognize that it would improve similar devices in the same way, using the technique is obvious").
• Obvious to try: once the POSITA sees TrueSpectra caching generated images keyed by parameters and no invalidation, the bounded set of design options for keeping a cache coherent — time-based expiry, checksum, or compare the modification timestamps of the inputs — is a small, finite, predictable set. See In re O'Farrell, 853 F.2d 894 (Fed. Cir. 1988) (predictability from a finite number of identified solutions).
• Cordell expressly motivates the bandwidth element: it teaches that response-time conditions should change which image variant is served. Applying Cordell's monitored-transfer-rate variable as a cache key component ("Bandwidth" column, claim 3; "combining said media tag with bandwidth information," claim 9) is the combination of two references directed to the same problem of web-image delivery latency.

Ground 3 — Wistendahl U.S. 5,708,845 + Dudley + Lewis → claims 1, 2, 3, 5, 6, 8

Wistendahl teaches author-defined mapping/parameter data stored separately from the media content, so the content "can be kept intact without embedded codes," with an authoring system to create it and network download of the mapping data plus program. That is structurally claims 3/5/6's "media repository + separately stored creation script + dependency metadata." Dudley adds parsing existing markup to populate a server database. Lewis adds the tag insertion/removal/revision mechanic. Motivation: Wistendahl's express purpose (preserving source assets, keeping mapping external) is identical to the patent's stated advantage of "preserv[ing] original image assets" and "automatically propagat[ing] changes of original assets throughout a Web site." Combining it with Dudley's server-side markup parsing to avoid the labor of hand-maintaining mapping data is a predictable advance.

Ground 4 — TrueSpectra's admitted caching + Wistendahl/Dudley + conventional database design → claims 3, 5, 6, 8, 9

This is where the patent's remaining differentiation lives, and it is the weakest ground for the patent owner:

• Claim 3's tables are: a script table with columns {Media Script, HTML Equivalent, Bandwidth, Generated File, Dependency List} and a dependency table with {File Name, Modification Date}. Once the caching of generated images is admitted (TrueSpectra), the only question is how to invalidate the cache. Selecting (i) a primary lookup table keyed by the request script, (ii) a column holding the resulting HTML, (iii) a column holding the generated filename, and (iv) a second table listing input files with their modification dates is a routine design choice / obvious arrangement of data with predictable results. The "Modification Date" value is nothing more than the standard filesystem mtime; reading it is not an inventive act. Cf. KSR, 550 U.S. at 417 ("a court must ask whether the improvement is more than the predictable use of prior art elements according to their established functions"); In re Kuhle, 526 F.2d 553, 555 (CCPA 1975) (obviousness may turn on the "arrangement of parts" where the modification is merely a matter of design).
• Claims 5, 6, 8, 9's invalidation logic (compare stored dates; if changed, delete entry and regenerate) is a functional restatement of the problem TrueSpectra's admitted cache left unsolved. Where the prior art identifies the deficiency (no dependency/propagation) and the solution is the most natural one available, the claim is obvious. See In re Kahn, 441 F.3d 977, 988 (Fed. Cir. 2006).
• Claim 9's "combine said media tag with bandwidth information" is met by Cordell (monitored transfer rate → different asset served) + TrueSpectra (per-image quality/compression settings). The claim adds only the sequencing of a memoization check around it.

Ground 5 — The "Java servlet engine" limitation (claim 3) and means-plus-function claims (2, 6, 7, 8)

• The patent itself states the Java servlet engine is "a third-party product that allows the system 100 to interface with the Web server 110 and execute Java servlet code." A claim element that recites a known, off-the-shelf third-party product as the integration substrate is not a source of patentable weight; the servlet API (JavaSoft, 1997) was published and widely deployed by 1999.
• Claims 2, 6, 7, 8 are § 112 ¶ 6 expressions of the same functions already recited in claims 1, 3 and 9. A means-plus-function limitation is only as patentable as the function it recites plus the disclosed structure; where the function is old (cache lookup, iterate to the next tag until none remain, compare timestamps, iterate over tags) the claim adds nothing to the § 103 analysis. Claim 7's "search for a next proprietary tag … [until] no more … exist" is the classic while(indexOf(tag) != -1) loop — a programming commonplace.

5. Why the combinations are proper (teaching-away / counter-argument rebuttal)

Secondary considerations. The record before me discloses (a) an expired patent (anticipated expiration 2019-10-21, so any commercial-success story in the record must span the pre-expiry window), (b) a 2022 Delaware infringement suit and (c) a PTAB petition that was settled rather than decided. Neither settlement nor license/assignment activity can carry nonobviousness without a demonstrated nexus and evidence of why the licensee acted — Iron Grip Barbell Co. v. USA Sports, Inc., 392 F.3d 1317, 1324 (Fed. Cir. 2004); In re GPAC Inc., 57 F.3d 1573, 1580 (Fed. Cir. 1995). The re-assignment chain visible on the record (Equilibrium → CMGI collateral → Barger-Great South Ventures → AMPS → Remington security interest → Equil IP Holdings, with Lit-US Chisum 22-A security interest, https://patents.google.com/patent/US6792575/en) reads as financial/servicing activity, not as industry praise or unexpected results. I found no evidence of unexpected results, long-felt need, or copying in the sources available to me.

6. Claim-by-claim bottom line

Overall: claims 1, 2, 4 and 7 look strongly vulnerable; claim 3 and its dependent database-schema/invalidation claims (5, 6, 8, 9) are the only portions with a meaningful defensive position, and even those turn on whether a two-table cache-key/dependency schema with mtime-based invalidation is credited as more than an obvious design choice over an admitted prior-art cache.

7. Explicit limitations on this analysis

• I did not retrieve the IPR2023-00329 petition, patent-owner response, or institution decision; my searches returned unrelated PTAB materials. Anything stated above about that proceeding is limited to the settlement status shown on Google Patents. Verify the IPR record before relying on it.
• The TrueSpectra material is described in the patent as product documentation and a "patent pending." Whether it qualifies as a printed publication or public use under pre-AIA § 102(a)/(b) requires documentary evidence (datasheets with dates, source code, or third-party corroboration) beyond the applicant's specification narrative. Its value here is chiefly as an admission of what the art contained.
• I have quoted AU-A-53031/98, U.S. 5,355,472, U.S. 5,708,845, U.S. 5,701,451, U.S. 5,710,918, U.S. 5,845,084 and EP 0843276 only as the '575 specification characterizes them, except for U.S. 5,893,127 (Tyan) and EP 0843276, for which I obtained the published text; the full texts of the remaining references should be obtained and cited by column/line before any paper is filed. No claim-by-claim invalidity contention should be finalized on this reconstruction alone.
• This is a technical/analytical assessment, not legal advice and not a legal conclusion of invalidity.