US 5916309

I'll research this patent across USPTO records and litigation dockets.

Let me search specifically for litigation and Federal Circuit activity.

Let me do a few more targeted searches on litigation and PTAB/post-grant activity.

US Patent 5,916,309 — Analyst Summary

Scope note on sources: The authoritative record below is drawn from the full patent text supplied in your message (Google Patents, fetched 2026‑09‑24) plus the USPTO/EPO PDFs surfaced in my searches. I ran targeted searches for USPTO post‑grant proceedings and CAFC 2026 dockets on this exact number; results on that specific point are reported honestly at the end, including what I could not confirm.

---

1. Bibliographic Data (verbatim identifiers)

---

2. Abstract (as issued)

An improved communications buffer system dynamically allocates the size and number of buffers to optimally receive packets over a communications link while efficiently using available memory. It is readily adaptable to infrared communication ports used on printers. Once a session begins, the printer determines how much "free" memory exists in its "pool area" and divides that pool between transmit and receive buffers. Transmit buffers are sized and counted first, based on the pool size available at that moment. Receive buffers are configured next: a preliminary calculation ensures the receive buffer size does not exceed the amount of data receivable within the maximum turnaround time; the receive buffer size is then provisionally set to a power of 2, and the buffer count is set to the remaining pool size divided by that receive buffer size. After ensuring a minimum number of receive buffers, a "window size" is set to a certain maximum (for IR, capped per IrDA), after which the session proceeds to transfer actual print job data.

---

3. Plain‑Language Overview of the Independent Claims

There are six independent claims: three method claims (1, 6, 7) and three apparatus/system claims (12, 17, 18).

Claim 1 — Method (asynchronous)

A method, in a computer system with memory, a bi‑directional port used in asynchronous communications, and a processing circuit, of allocating the memory "pool area" into transmit and receive buffers by:

• (a) creating transmit buffers whose number and size depend on the total available pool size at the beginning of reception of an asynchronous message;
• (b) creating receive buffers whose size depends on communications parameters relating to the type of communications link providing the message, and on the total available pool size at that beginning time; and
• the number of receive buffers further depends on total available pool size and the receive‑buffer size, again at the beginning of reception.

Plain reading: the allocation is a one‑shot computation keyed to the start of receiving a message — not an ongoing, traffic‑adaptive adjustment — and the receive side looks at link parameters whereas the transmit side looks only at free pool size.

Claim 6 — Method + transmit‑size formula

Same allocation scheme as claim 1, but adds (c) a computed bound: the maximum transmit packet buffer size for the secondary equals the minimum of —

1. {(maximum turnaround time of the link − minimum turnaround time of the transmitting device's infrared communications port) × (transmission speed in bytes per millisecond) − number of BOF's of the transmitting device}, and
2. (maximum packet size of the transmitting device).

Note (literal reading): the preamble recites a generic "bi‑directional communications port," yet element (c) refers to an "infrared communications port" and to a "secondary" without antecedent definition in the claim — the "secondary" terminology is carried over from the specification's IrDA discussion.

Claim 7 — Method + receive‑size formula

Same allocation scheme, but adds (c) a bound on the maximum useful size of the receive buffers, equal to the minimum of —

1. {(maximum turnaround time of the link − minimum turnaround time of the infrared communications port of the printing apparatus) × (transmission speed in bytes per millisecond) − number of BOF's of the printing apparatus}, and
2. (maximum packet size of the printing apparatus).

Note (literal reading): the preamble is again a "computer system," while element (c) speaks of "said printing apparatus," which is not introduced in the claim — an antecedent‑basis inconsistency on its face.

Claim 12 — System (asynchronous)

The apparatus counterpart to claim 1: a memory circuit containing a pool area divided into transmit and receive buffers; a bi‑directional communications port in asynchronous communication with an external device and the memory; and a processing circuit configured to control data flow and to determine the number and size of the transmit/receive buffers, with the same dependencies as claim 1 (transmit count/size from pool size; receive size from link parameters plus pool size; receive count from pool size and receive‑buffer size, all at the beginning of reception).

Claim 17 — System + transmit‑size formula

Claim 12's system, with the claim‑6 transmit‑packet‑buffer‑size minimum‑of‑two formula added.

Claim 18 — System + receive‑size formula

Claim 12's system, with the claim‑7 maximum‑useful‑receive‑buffer‑size minimum‑of‑two formula added.

Dependent claim themes (2–5, 8–11, 13–16, 19–22)

• 2/13: one communication parameter is the maximum turnaround time specification.
• 3/14: that parameter is determined during a "discovery" procedure between the printer's port and the transmitting station.
• 4/15: the port is an infrared port.
• 5/16: the computer system is resident on a printing apparatus.
• 8/19: increase transmit buffer size when the connection exhibits a relatively slow data transmission speed (specification: byte‑stuffing at SIR speeds, doubling transmit buffers).
• 9/20: that connection is a serial infrared (SIR) link.
• 10/21: the apparatus enforces minimum counts of transmit and receive buffers even if buffer size must shrink.
• 11/22: the minimum number of receive buffers must exceed the window size of packets transmittable in a single turnaround time per IrDA.
• Literal‑reading observation: claims 8–11 are written as depending from claim 5, and claims 19–22 recite "printing apparatus" while depending from claim 16 — neither introduces "printing apparatus" in its own preamble, so these dependency/antecedent relationships read imperfectly on their face.

---

4. Implementation Detail Disclosed in the Specification (for context on the claims)

The "Reconfigure Buffers" flow (FIGS. 4A–4B) teaches concrete heuristics that inform the claimed computations: transmit buffer size typically 64 bytes; transmit buffer count 16 if pool > 20 KB, 8 if > 10 KB, else 4; transmit buffer size doubled for SIR; receive buffer size rounded up to the next power of two and capped at 2048 bytes (IrDA max data buffer); a loop halving receive buffer size until more than four receive buffers result; and window size = truncated[(number of buffers − 1) / 2], capped at 7 — the IrDA maximum packets per turnaround. The specification also explains the "discovery"/SNRM negotiation (default 9,600 bps NDM) that supplies the parameters claimed in claims 3/14, and stresses the advantage that received data is consumed directly from the packet buffer rather than copied into a ring.

---

5. USPTO Post‑Grant and CAFC 2026 Docket Search — Findings and Limits

I must be explicit about what the searches did and did not establish:

• No CAFC 2026 docket, appeal, or decision involving US 5,916,309 was found. My searches — including the Federal Circuit's own Scheduled Cases listing (July 2026 panel calendar, which shows unrelated matters such as Slingshot Printing LLC v. Canon U.S.A., SpaceTime3D v. Apple, FrenchPorte IP v. C.H.I. Overhead Doors) — returned no entry for patent 5,916,309. I cannot represent this as a verified negative; it reflects absence of hits in the sources I could retrieve, not confirmation that no such docket exists.
• No PTAB/IPR/PGR proceeding on this patent was surfaced. The only seemingly relevant hit was a third‑party USPTO PTACTS petition document containing a prior‑art search string "5916309".pn. alongside other patent numbers (6199106, 6396598, etc.) — that is a citation/search artifact in an unrelated petition, not a proceeding against this patent.
• Litigation history I could find centers on Lexmark's toner‑cartridge patents (e.g., ITC Inv. No. 337‑TA‑740 listing '032, '169, '233, '661, '432, '378, '291, '771, '015, '876, '692, '031, '510, '760, '204; and Lexmark Int'l v. Ink Technologies Printer Supplies, S.D. Ohio 1:10‑cv‑00564). US 5,916,309 does not appear in those charts. The high‑profile Lexmark appeals (Impression Products v. Lexmark, 137 S. Ct. 1523 (2017), and its Federal Circuit antecedent, 816 F.3d 721) concern toner‑cartridge exhaustion, not this patent.
• The strongest corroborating datum is that the patent is recorded as Expired – Lifetime with anticipated expiration May 12, 2017, which makes active 2026 appellate litigation on it unlikely.

Confidence statement: I am highly confident in the bibliographic data, the abstract, and the claim analysis (they come from the authoritative full text). I have low‑to‑moderate confidence in the completeness of my negative finding on CAFC/PTAB activity, because I could not query PACER, the CAFC CM/ECF docket system, or USPTO Patent Center/PTAB APIs directly — only indexed web sources. If you need a dispositive answer there, the appropriate next step is a direct PACER/CM‑ECF docket query for "5,916,309" and a USPTO Patent Center / PTAB API lookup on application 08/854,510 and patent 5,916,309.

I'll search multiple sources for litigation involving US patent 5916309.

Let me search more targeted litigation databases.

Let me check dedicated litigation databases for this specific patent.

Litigation Search Results for U.S. Patent No. 5,916,309

Bottom line: I found no evidence of any litigation involving U.S. Patent No. 5,916,309.

I must be candid that my searches did not surface any lawsuit, ITC investigation, PTAB proceeding, or other adversarial matter in which this specific patent number was asserted. Below I explain what I checked, what I found, and the important caveats.

---

What the patent is (to disambiguate)

• Patent: U.S. 5,916,309 — "System for dynamically determining the size and number of communication buffers based on communication parameters at the beginning of the reception of message"
• Application: US 08/854,510
• Filing date: 1997-05-12; Grant date: 1999-06-29
• Assignee (original & current): Lexmark International, Inc.
• Inventors: Rickey Carter Brown; Mark Lane Mayberry
• Status: Expired – Lifetime (anticipated expiration 2017-05-12)
• Subject matter: Dynamic allocation of transmit/receive buffers in a printer based on IrDA "discovery" communications parameters (max turnaround time, packet size, BOF count, etc.)

---

Sources checked and what they showed

1. Google Patents page for US5916309A (https://patents.google.com/patent/[US5916309](/patent/US5916309)/en) — The page lists "Cited By (54)" and "Families Citing this family (6)" references, and the patent family/applications (including the continuation-in-part US 6,046,817 and foreign equivalents EP0878757B1, JPH1195948A, KR19980086926A, CN1139034C, TW384428B, DE69818141T2). These are citations, not litigation. Google Patents' "Litigation" section did not appear for this patent, which is ordinarily populated when a patent is litigated. No litigation entries were returned.

2. General web searches for "5916309" patent lawsuit, US patent 5916309 litigation, and "5,916,309" — No case names, docket numbers, or complaints referencing this patent. The only hits with the literal string "5916309" were (a) an unrelated USPTO PTAB petition document where the number appears in a prior-art search string (i.e., the patent used as prior art by an examiner, not as an asserted patent), and (b) a Taiwan patent (TWI423032B / TW 1423032 B) that lists US 5916309 as a cited reference in its background art. Neither is litigation.

3. Lexmark patent-enforcement activity — Lexmark is an active litigant (e.g., the ITC toner-cartridge investigation; Lexmark Int'l v. Ink Technologies Printer Supplies, No. 1:10-cv-00564 (S.D. Ohio); Static Control Components v. Lexmark; Lexmark Int'l v. Laserland, No. 5:03-cv-00312 (E.D. Ky.); Impression Products v. Lexmark). However, none of these cases involved US 5,916,309. The patent lists asserted in the Lexmark ITC matter were toner-cartridge/imaging patents (e.g., 5,337,032; 5,634,169; 5,995,772; 6,397,015, etc.), not the infrared/buffer patent at issue here.

---

Important caveats — please read before relying on this

• I could not reach every litigation database directly. The available search tooling is limited, and dedicated resources such as the Unified Patents portal (portal.unifiedpatents.com), the Stanford NPE Litigation Database (npe.law.stanford.edu), PACER, and CourtListener's full docket search were only reachable indirectly and did not return records keyed to this patent number. I cannot rule out that a case exists that is not indexed in the sources I could reach.
• A granted, expired patent that was never asserted is entirely plausible here. US 5,916,309 expired in 2017 and issued to the manufacturer itself (Lexmark) rather than to an NPE. It has no NPE-style assignment history.
• A prior-art citation is not litigation. If you were pointed to a document mentioning "5916309," confirm whether it is a litigation complaint or merely a reference/prior-art listing — the sources I found were all of the latter type.

Recommendation

To be certain, the authoritative confirmations should be run directly against:

• Unified Patents litigation search: https://portal.unifiedpatents.com/litigation/caselist (filter patent number 5,916,309)
• Stanford NPE Litigation Database: https://npe.law.stanford.edu/ (patent lookup)
• CourtListener / RECAP and PACER docket searches (all districts) for "5,916,309"
• LexisNexis / Westlaw / Docket Navigator patent litigation searches

If a verified docket comes back from any of these, I'd be glad to summarize the plaintiff, defendant, jurisdiction, case number, filing date, and status for each case.

Summary: No litigation involving U.S. 5,916,309 was found; based on the sources available, the patent appears never to have been asserted in litigation.

Proceedings overview

Total AIA trial proceedings on file for US 5,916,309: 0. No IPRs, no PGRs, no CBM reviews. The USPTO Open Data Portal returns no AIA trial proceedings for this patent, and independent web searching surfaced none. Bottom line for a defendant: there is no PTAB record to leverage — the patent has never been tested at the Board, so you cannot point to canceled claims, and you also face no § 315(e)(2) estoppel shortcut from someone else's prior petition. The far more consequential fact is that this patent is expired (anticipated expiration 2017-05-12, per the Google Patents bibliographic record, consistent with a 20-year term from the 1997-05-12 filing date). Any current assertion is a past-damages play, not an injunction play.

One important caveat before the details: the search space is contaminated by a name collision. Lexmark owns a different patent, U.S. 9,245,591, which is routinely abbreviated "'591 patent" in PTAB papers. That patent — not US 5,916,309 — is the subject of active Lexmark enforcement (e.g., Lexmark Int'l, Inc. v. Universal Imaging Industries, LLC, M.D. Fla. 8:18-cv-01047-EAK-AEP, and IPR petitions styled against the '764/'591/'136 authentication family, including IPR2019-01378). Do not let an adversary's "the '591 patent" shorthand mislead you into thinking US 5,916,309 has a PTAB history. It does not.

---

No proceedings to report

There is no {PROCEEDING_NUMBER} to populate. Rather than fabricate one, here is what was checked and what was found:

⚠️ Near-miss (NOT this patent) — do not cite as US 5,916,309 activity

• IPR2019-01378 — involving U.S. 9,245,591 ("'591 Patent"), a Lexmark patent directed to memory-module intermediate-voltage signaling on a shared signal line. Petitioner relied on Blood (US 3,993,867), the I²C Spec, and Bruce (US 5,822,251) under § 103 against claims 1–16. The "'591" label in those papers refers to 9,245,591, and the "'764"/"'136" references are 9,400,764 / 9,837,136 — all from Lexmark's printer-cartridge authentication family, filed decades after US 5,916,309. Sources: PTAB petition documents and PTAB E2E.
• Lexmark's 2018–2019 enforcement campaign (M.D. Fla. 8:18-cv-01047) asserted 7,844,786; 8,966,193; 9,245,591; 9,400,764; 9,837,136 — US 5,916,309 is not among them. Lexmark's historical US litigation (ITC Inv. No. 337-TA-740, Lexmark v. Ink Technologies, Lexmark v. Static Control, Lexmark v. Impression Products) involved toner-cartridge patents, all unrelated to the infrared buffer patent here.

---

Strategic summary

Claim status: everything is UNTESTED. All 22 claims of US 5,916,309 remain as issued on 1999-06-29. Claim 1 (method of allocating transmit/receive buffers based on pool size and link parameters at the start of message reception), claims 6 and 7 (the max-transmit-buffer and max-receive-buffer formulas), and independent system claims 12, 17 and 18 are all presumptively valid — no PTAB panel has ever construed a term, ruled on a ground, or canceled anything. The practical significance is asymmetric: there is no in rem patentability finding to point to, but there is also no IPR-track record showing the claims are robust. The claims have simply never been stress-tested by anyone.

Estoppel landscape: a blank slate. Because no AIA petition has ever been filed against the '309 patent, no petitioner (or privy) is estopped under § 315(e)(1) or § 315(e)(2). Conversely, a defendant today is not helped by anyone else's estoppel, because there is none. That means the full field of § 102/§ 103 art — including any pre-1997 patents or printed publications on buffer pool partitioning for packet links — remains open at the district court on § 282 grounds, and even an IPR is theoretically available (subject to § 315(b) and the practical issues with expired patents, below). Note that IPR grounds are limited to § 102/§ 103 on patents and printed publications; § 112 and § 101 attacks must go through the district court (or, for pre-AIA patents, an expired CBM option that would fail the "financial product or service" eligibility test for a printer buffer patent).

Pattern signals: none. No serial petitioner, no defensive aggregator such as Unified Patents in the chain, no patent-owner appeal strategy (there being no FWD to appeal), and no reissue or reexam history surfaced. The patent's prosecution lineage is unusual in one respect that matters for claim scope: US 6,046,817 is a continuation-in-part of US 5,916,309's application, so the '309 claims are bounded by a 1997-05-12 priority date and carry no benefit of the later CIP disclosure. Separately, Google Patents lists 54 later documents citing the '309 patent as prior art (the usual "Cited By" list of subsequent patents) — those are citations, not challenges, and none of them is a PTAB proceeding.

Recommended next steps

1. No PTAB record exists to exploit — say so plainly, and don't let the "591" collision fool you. Any demand letter citing "591" should be checked against the patent number. If it says US 5,916,309, the patent expires/expired 2017-05-12 and the only live exposure is past damages.
2. Attack the damages window first, not validity. Under 35 U.S.C. § 286 there is a six-year lookback from the complaint; six years before 2026-10-01 is 2020-10-01, which falls after the patent's 2017-05-12 expiration. On those dates, a § 286 damages period for an expired patent asserted now is effectively empty unless a complaint was filed long ago. Verify whether any prior suit or tolling agreement fixes an earlier window.
3. Confirm expiration on the official record. Pull the fee/expiration status from Patent Center for US 5,916,309 and confirm the 2017-05-12 lapse (Google Patents shows "Expired – Lifetime" / "Anticipated expiration 2017-05-12"). If reinstatement via § 41(c) were ever attempted, that changes the posture — check for any petition for unintentional delay, none of which surfaced here.
4. If an IPR is contemplated anyway, note the practical obstacles: (a) an expired patent cannot be amended, so no substitute claims are possible; (b) the Board construes expired claims under the Phillips standard; and (c) § 315(b) would bar a petition filed more than one year after service of a complaint on a real party in interest or privy. But recognize the strategic oddity: filing an IPR against an expired patent mainly buys a district-court stay argument, not an answer to damages.
5. Do not cite IPR2019-01378 or the Universal Imaging/Miller petitions as authority against US 5,916,309. They construe U.S. 9,245,591, a different patent with a different specification, different claims, and a different priority date. Misstating this in a brief is an easy credibility hit.

Sources consulted: USPTO structured "PTAB proceedings on file" block for US 5,916,309; Google Patents US5916309A; PTAB E2E / Patent Trial and Appeal Board petition documents; USPTO AIA trial statistics. No Federal Circuit opinion or CourtListener docket tied to US 5,916,309 was located. Confidence: high that no AIA trial proceeding has been filed against US 5,916,309; moderate on the absence of reexamination/EPO opposition activity, since those are search-derived rather than from the structured ODP feed and should be confirmed in Patent Center / Global Dossier before being represented to a court.

I'll research the assignment record for US 5916309 using multiple sources.

Let me dig deeper into assignment records and the Lexmark portfolio transfer.

Let me search for the specific assignment records and check NPE/asserter databases.

Ownership reconstruction — US 5916309

Bottom line up front: There is no recorded post-issuance assignment in the materials I can reach. The only recordation in the chain is the original inventor→Lexmark assignment executed and recorded at filing (1997‑05‑12). US 5916309 stayed with Lexmark International, Inc. for its entire life and expired 2017‑05‑12 (20‑year term). I found zero NPE‑pattern signals. I could not retrieve the original reel/frame number from the sources available to me — I flag that as unverified rather than invent it.

---

Inventors

• Both are named as assignors in the 1997 "Assignment of Assignors' Interest" to Lexmark International, Inc. (Lexington, KY), which is the standard employer‑assignment signature block — this is the best available evidence they were Lexmark employees at filing.
• Rickey Carter Brown is credited with 4 US patents as a Lexmark inventor, the first granted 1999 and the most recent December 2008 (Patent Leaderboard) — i.e., he remained an active Lexmark inventor long after this filing.
• Unusual patterns: none observed. Neither inventor shows an early‑departure / pre‑fire‑sale signature (the classic tell would be all inventors leaving the assignee within 12 months; that is not documented here). I have no evidence about either inventor's later employment; treat that as unclear, not as a finding.

Note on scope: I also could not verify a second named inventor beyond the two listed in the assignment brief ("BROWN, RICKEY C., MAYBERRY, MARK L."); the authoritative patent text confirms exactly these two.

---

Original assignee

• Entity on the issued patent: Lexmark International, Inc., 740 West New Circle Road, Lexington, KY 40550. Assigned by the inventors on 1997‑05‑12 (filing date), recorded the same period.
• Product embodying the claims: Yes. The specification itself describes a laser printer (reference numeral 10) with an infrared IRDA port (18), and the claims recite a printer/computer system with a bidirectional infrared port and dynamically allocated transmit/receive buffers. Lexmark was shipping IrDA‑equipped laser printers (e.g., the Optra family) in this era, so the claims map to a real Lexmark product line.
• Primary line of business: laser/inkjet printers, multifunction peripherals and consumables (toner/ink). Lexmark was spun off from IBM in 1991 and was a publicly traded OEM (NYSE: LXK).
• Current status: Operating — now a private subsidiary. On 2016‑11‑29, a consortium led by Apex Technology Co., Ltd. (Zhuhai) — with PAG Asia Capital and Legend Capital — completed a ~$3.6B negotiated merger and took Lexmark private (Reuters/MarketScreener). In the 2017 litigation filings, Lexmark stated its 100% owner is Apex Swiss Holdings SARL, and that Ninestar Corporation (Shenzhen Stock Exchange) is the indirect owner of 51% of Apex Swiss Holdings (Commercial Copy Innovations v. Lexmark, 8:17‑cv‑01149, Dkt. 41). Lexmark is not in bankruptcy and has never been dissolved.
• Caution on the 2016 transaction: this was a merger in which Lexmark was the surviving entity, so the change of control happened at the corporate‑parent level and would not generate a per‑patent USPTO assignment record. That is consistent with the legal‑events history below showing nothing after 1999.

---

Assignment timeline

Recorded chain (as exposed by the authoritative Google Patents legal‑events record for US5916309):

• 1997‑05‑12 (executed) / recorded 1997 — Reel/Frame not determinable from the sources I can reach (the recordation is confirmed by the legal‑events entry "Assigned to LEXMARK INTERNATIONAL, INC. — ASSIGNMENT OF ASSIGNORS' INTEREST (SEE DOCUMENT FOR DETAILS)"; I could not retrieve the specific reel/frame, so I am not guessing one).
  • Conveyance: Assignment of Assignors' Interest (employer assignment)
  • Assignor: Rickey C. Brown; Mark L. Mayberry
  • Assignee: Lexmark International, Inc., Lexington, KY
  • Correspondent: not exposed in the materials available to me (flag: unverified; no recurrence to report on this chain)
  • Context: original inventor→employer assignment at filing.

Everything else in the record is corporate/family, not an ownership transfer:

• 1998‑01‑22 — Priority filing of US09/012,073 (continuation‑in‑part → US6046817, same Lexmark family). Internal prosecution, not an assignment.
• 1998‑05‑11 / 1998‑05‑12 — Foreign priority filings: KR 1019980016769, DE 69818141T, JP 16909698, CN 981083137, EP 98108583 (and TW 087107338, 1998‑07‑18). Family members, not assignments.
• 1999‑06‑29 — Patent granted.
• 2016‑11‑29 — Apex/Ninestar‑led merger takes Lexmark (the assignee) private. Corporate‑level change of control; no per‑patent recordation expected or found.
• 2017‑05‑12 — Anticipated expiration (20 years from filing). Status: Expired – Lifetime.

No post‑issuance assignment exists in the record I can reach. Per Google Patents legal events (authoritative here), the last ownership event is the 1997 assignment; there is no reassignment, security agreement, merger transfer, change of name, or license recordation afterward.

Additional corroboration that the patent remained in Lexmark's hands through the 2016 deal: Lexmark's US patent 08/854510 → US5916309 appears in the patent inventory tabled in Apex Technology's 重大资产购买报告书 (major‑asset‑purchase report) for the Lexmark transaction, listed with Lexmark's business unit tag "ISS" (Zhuhai Apex report, patent table). In other words, the asset being transferred to the Apex/Ninestar consortium was Lexmark's stock, not the patent title.

If the Assignment Center genuinely returns no record beyond the initial 1997 assignment on your query, the correct reading is: the original assignee still owned the patent at expiry. Verify at https://assignmentcenter.uspto.gov/ (or https://assignment.uspto.gov/patent/index.html) searching patent number 5916309.

---

Timeline diagram

timeline
    title Ownership of US 5916309
    1997 : Filed 12 May by Brown and Mayberry
         : Assigned to Lexmark International Inc
    1998 : CIP child US6046817 filed
    1999 : Patent issued 29 June
    2016 : Lexmark acquired by Apex led group
    2017 : Patent term expires 12 May

---

NPE / troll-pattern signals

1. Shell-entity transfer — not present. No assignment out of the operating OEM exists in the record. The assignee of record from 1997 to expiry is Lexmark International, Inc., an operating manufacturer — not an "IP / Holdings / Licensing / Ventures" LLC.
2. Known asserter in the chain — not present. No assignee matches Acacia, Marathon, IV, IPNav, Wi‑LAN, Mosaid/Conversant, Vringo, Pendrell, Innovatio, MPHJ, Lumen View, Round Rock, or any Spangenberg entity. The chain contains only Lexmark, an operating printer OEM.
3. Repeat correspondent across the chain — not present. There is exactly one recorded assignment, so there is no recurrence to measure. (Correspondent of record could not be retrieved; flagged unverified, not inferred.)
4. Cascading transfers — not present. Zero consecutive transfers through chained LLCs; no transfers at all post‑1997.
5. Pre-litigation transfer — not present. No transfer date exists to compare against any suit. US 5916309 is not the basis of any assertion I can attribute to it; the file history shows no reassignment en route to litigation.
6. Bankruptcy fire‑sale — not present. Lexmark's 2016 change of control was a negotiated ~$3.6B merger (Apex/Ninestar/PAG/Legend Capital), not a Chapter 7/11 sale; Lexmark survived as the entity. The patent expired the following year rather than being monetized.
7. Privateering — not present. No transfer from Lexmark to an assertion vehicle. (The inverse of privateering appears in the record: Wi‑LAN/Quarterhill's Commercial Copy Innovations, Inc. sued Lexmark in C.D. Cal. in 2017 — Lexmark as defendant, with Wi‑LAN entities identified as the plaintiff's interested parties (Dkt. 4). That case is about Wi‑LAN patents, not US 5916309, and does not put an NPE in this chain.)
8. Defensive aggregator — not present. The chain does not terminate at RPX, AST, LOT, Unified Patents, or OIN. It terminates at expiry (2017‑05‑12) while still held by Lexmark.

---

Verdict

Insufficient data — but with an affirmative, not an ambiguous, reading of the chain.

The only ownership event of record is the 1997‑05‑12 inventor→Lexmark assignment, followed by no reassignment, security interest, merger transfer, change of name, or license before the patent expired 2017‑05‑12 (Google Patents legal events; corroborated by US5916309's appearance in Lexmark's patent inventory in Apex's 2016 purchase report). Because there is only the original assignment and no post‑issuance records in the Assignment Center, the rubric classifies this as insufficient data rather than an NPE finding — and every one of the eight NPE signals is affirmatively not present. The patent never left an operating printer OEM, was never monetized, and lapsed at end of term; the 2016 Apex/Ninestar deal was a corporate‑parent merger, not a patent‑level transfer.

Verify: https://assignmentcenter.uspto.gov/ (search 5916309) · https://patents.google.com/patent/US5916309/en (legal events).

I'll research this patent and its cited prior art.

Let me gather details on the specific prior-art references cited in this patent.

Prior-Art Analysis — U.S. Patent No. 5,916,309

A. The patent at issue (verified)

Source: https://patents.google.com/patent/US5916309/en

Important caveat on sources. The authoritative full text I was given (the Google Patents page for US 5,916,309) reproduces the specification, abstract and claims of the patent, but the fetched HTML does not include the front-page "References Cited / U.S. Patent Documents" list. What the fetched text does contain is the Background of the Invention section, in which the patent expressly discusses nine U.S. patents as prior art. Those nine are the "patent citations" I can identify with confidence, and they are the subject of this analysis. I could not independently confirm from the fetched document whether the examiner's front-page citation list contained additional references. I flag this rather than invent a citation list.

Also note: many of the references below were cited in the patent against the 1989–1994 state of the art; several are therefore earlier than the 1997 priority date and are proper § 102/§ 103 prior art.

---

B. The cited prior-art references, one by one

1. U.S. Pat. No. 5,179,662 — Corrigan

• Full citation: U.S. Patent 5,179,662, "Optimized I/O buffers having the ability to increase or decrease in size to meet system requirements," Michael J. Corrigan et al., assigned to International Business Machines Corp.
• Dates: Filed 1989-08-31 (Ser. No. 07/401,335); issued 1993-01-12. (Corresponding EP 0 415 862 A2, publ. 1991-03-06.)
• Description: A double-buffering scheme for writing to/reading from auxiliary storage. Buffers are dynamically increased when the first buffer has not completed its asynchronous write before the second is filled, and decreased when the first completes before the second is filled past a threshold percentage (e.g., 50%). Size is adjusted iteratively until the system no longer waits synchronously.
• § 102 relevance: Potentially relevant to the concept of dynamically adapting buffer size (claim 1(a)/(b) in the abstract sense). It does not disclose the number of transmit buffers being set from total pool size, the receive-buffer size being set from communications-link parameters (turnaround time, packet size), or the "beginning of reception of a message" trigger. Therefore it does not anticipate claim 1; at most it is background/§ 103 art on "dynamic buffer resizing."

2. U.S. Pat. No. 5,289,470 — Chang et al.

• Full citation: U.S. Patent 5,289,470, "Flexible scheme for buffer space allocation in networking devices," Shu-Ping Chang & Ahmed N. Tantawy, assigned to International Business Machines Corp.
• Dates: Filed 1992-12-14 (Ser. No. 990,162); issued 1994-02-22.
• Description: A networking device keeps buffer sets of different sizes (e.g., 128/256/512/1024/2048 bytes). A predetermined ordered sequence of sets is established, and successive buffers for a packet are drawn from successive sets until the packet is stored. Buffer sizes/sequence are derived empirically from observed packet-size distributions.
• § 102 relevance: Relevant to "buffer size … dependent upon packet size" concepts and to memory-utilization optimization (claims 6–7 general background). It does not disclose dividing a pool into transmit vs. receive buffers based on remaining pool size at the beginning of a message reception, nor the max-transmit-buffer / max-receive-buffer formulas of claims 6/7/17/18. Not anticipatory of claim 1.

3. U.S. Pat. No. 5,303,347 — Gagne et al.

• Full citation: U.S. Patent 5,303,347, "Attribute based multiple data structures in host for network received traffic," David A. Gagne & Satish L. Rege, assigned to Digital Equipment Corp.
• Dates: Filed 1991-12-27 (Ser. No. 07/815,007); issued 1994-04-12.
• Description: Incoming packets are routed to host-memory buffers dedicated by packet attribute (packet size, user ID, size+user-ID, source ID, data type) using per-attribute receive rings plus a sequencing ring to preserve arrival order.
• § 102 relevance: Relevant to attribute-dependent placement of received data and use of receive rings (the Lexmark patent's "Ports Receive Buffer Queue"). But the routing is to attribute-dedicated buffers, not to a pool sized from total available memory and negotiated link parameters. Not anticipatory.

4. U.S. Pat. No. 5,566,315 — Milillo et al.

• Full citation: U.S. Patent 5,566,315 (Milillo et al.), "cache memory … controlled on a dynamic basis by adjusting its low and high thresholds."
• Dates: The patent describes it as issued Oct. 15, 1996. (I was not able to independently re-verify the inventor list, assignee and exact issue date from the search results; treat these as per the patent's own characterization.)
• Description: Cache free-space high/low thresholds are adjusted from an "allocation predictor" and a "blockage predictor"; cache space is released when free space falls to the low threshold and release stops at the high threshold.
• § 102 relevance: Relevant to dynamic memory reclamation based on demand — i.e., freeing memory and re-dividing it (Lexmark's "reclaim unused memory space"). It is a cache-admission-threshold scheme, not a transmit/receive packet-buffer allocator driven by link parameters. Not anticipatory.

5. U.S. Pat. No. 5,442,747 — Chan

• Full citation: U.S. Patent 5,442,747 (Chan), multimedia video processor chip with multi-port central cache memory.
• Dates: Issued ~1995. (Not independently re-verified in this session; description taken from the patent's characterization.)
• Description: A single central RAM cache queues all incoming/outgoing data; boundary areas are re-allocated when the mode of operation changes, each storage area holding data from one source only.
• § 102 relevance: Relevant to re-partitioning a shared memory among data streams upon a mode change — conceptually similar to re-partitioning the pool when connection parameters change. It does not teach the transmit/receive split or the pool-size-driven buffer count of claim 1.

6. U.S. Pat. No. 5,130,986 — Doshi et al.

• Full citation: U.S. Patent 5,130,986 (Doshi et al.), fiber-optic long-distance data communication with per-block acknowledgement.
• Dates: Issued 1992-07-14 (typical for this number/era). (Date not independently re-verified here.)
• Description: Two windows — a "network window" sized to the bandwidth-delay product and a "receiver flow-control window" ≥ twice the network window.
• § 102 relevance: Relevant to sizing a receive window/buffer from link characteristics (bandwidth × delay) — this is the closest conceptual cousin to Lexmark's "receive buffer not larger than the data receivable within the maximum turnaround time." Still, there is no disclose of dividing a printer/computer memory pool into transmit+receive packet buffers at the start of reception, nor of the receive-buffer-count computation. Not anticipatory of claim 1, but the strongest § 103 reference on the "size recv buffer to a turnaround-time/bandwidth limit" concept.

7. U.S. Pat. No. 5,440,692 — Janicek

• Full citation: U.S. Patent 5,440,692 (Janicek), DB2 buffer-pool expansion/contraction.
• Dates: Issued ~1995. (Not independently re-verified here.)
• Description: A DB2 buffer pool dynamically expands (rejected if <10 MB virtual storage remains; allocates 4K/32K pages) or contracts depending on the queue chain type ("in use" not released; "never used" released; "previously used" released when the application is done).
• § 102 relevance: Relevant to dynamic pool expansion with a minimum-free-space guard, analogous to Lexmark's minimum-buffer constraints (claims 10/21). Not anticipatory.

8. U.S. Pat. No. 5,046,039 — Ugajin

• Full citation: U.S. Patent 5,046,039 (Ugajin), buffer management with constant total size but variable transmit/receive buffer ratio.
• Dates: Issued 1991-08-27. (Not independently re-verified here.)
• Description: When receiver-buffer-busy events per interval exceed a preset value, the receiver buffer area (or number of receive buffers) is increased and the transmitter buffer area decreased by the same amount (and vice-versa).
• § 102 relevance: This is the most structurally on-point reference for Lexmark claim 1's "divide pool into transmit buffers and receive buffers," because it trades transmit vs. receive area within a fixed total. However, the ratio is adjusted reactively from measured busy statistics during operation, not prospectively at the beginning of message reception from negotiated link parameters and available pool size. Because claim 1 requires the size/number determinations to be keyed to "the beginning of the reception of an … asynchronous message" and to "communications parameters relating to the type of communications link," Ugajin does not disclose that limitation. It is a strong § 103 reference against the general transmit/receive-partitioning concept, but not a § 102 anticipation of claim 1.

9. U.S. Pat. No. 4,158,235 — Call et al.

• Full citation: U.S. Patent 4,158,235 (Call et al.), buffer storage system accessible by any of several I/O ports.
• Dates: Filed Dec. 19, 1977 (Ser. No. 861,710); issued 1979-06-19. (Filing date is my recollection; issue year confirmed by the reference's low number and the patent text's listing order — flagging the filing date as not re-verified.)
• Description: I/O buffers addressed by logical name; a memory cell is assigned/accessed when the requested name matches.
• § 102 relevance: Generic multi-port buffering of historical interest only. Not anticipatory of any claim.

---

C. § 102 anticipation assessment (claims of US 5,916,309)

Claim 1 (and its system counterpart claim 12) is the governing independent claim. Its elements are: dividing a pool area of memory into transmit and receive buffers, where

• (a) the number and size of transmit buffers depend on the total available pool size at the beginning of reception of an asynchronous message; and
• (b) the size of each receive buffer depends on (i) communications parameters relating to the type of communications link and (ii) the total available pool size, both at the beginning of reception; and
• the number of receive buffers further depends on the total pool size and the receive-buffer size.

Dependent-claim mapping that IS closely touched by prior art:

• Claims 2/13 & 6/7/17/18 (maximum turnaround time; max transmit/receive buffer formulas): Doshi (5,130,986) is the closest — it sizes a window from bandwidth×delay. But no reference discloses the specific formulas in claims 6/7/17/18 (subtracting minimum turnaround time and number of BOF's from the turnaround-time×rate product and taking the minimum with max packet size). These formulas derive from the IrDA protocol and appear novel.

Conclusion: None of the nine cited references, individually, discloses every element of claim 1, claim 12, or their independent counterparts under 35 U.S.C. § 102. They are properly characterized as § 103 (obviousness) background art, individually relevant to isolated sub-concepts (dynamic resizing, pool subdivision, free-memory guarding, link-rate-based window sizing), but not anticipatory. The novelty of US 5,916,309 resides in the combination: (i) splitting the printer's memory pool into transmit vs. receive packet buffers, (ii) determining transmit count/size from available pool size, (iii) determining receive size from negotiated IrDA link parameters, all (iv) computed at the beginning of message reception, and (v) setting a window size per the IrDA 7-packet turnaround limit.

---

D. Exchange/family citations (for completeness — not § 102 art against the US claims)

The same invention was prosecuted/issued abroad, and foreign search reports cite different art:

• EP 0 878 757 B1 (the family member) — the counterpart EP search report cites WO 92/02898 (Elite High Technology), US 4,680,626 (Deering), US 5,784,581 (Hannah), WO 96/32265 (Silverbrook/Eastman Kodak), US 5,438,438 (Lee), US 5,235,680 (Bijnagte), and US 5,666,495 (Yeh) against the corresponding claims.
• US 6,046,817 (the CIP child, filed 1998-01-22) is the continuation-in-part of this application and shares the 1997-05-12 priority date.

These are worth noting because US 5,784,581 (Hannah) and US 5,666,495 (Yeh) are not among the nine references in the US specification, and WO 92/02898 was treated in Europe as the closest art; the EP search report associates it with the broad claims (1, 4, 9, 11, 13–17).

---

E. Explicit uncertainties

1. Front-page citation list: I could not retrieve the exact "U.S. Patent Documents" list printed on the face of US 5,916,309 from the fetched document; the nine references above come verbatim from the patent's Background discussion. If a formal anticipation/invalidity opinion is needed, the face-of-patent citation list plus the file-wrapper IDS / PTO-892 should be pulled to confirm no additional references were cited.
2. Unverified bibliographic dates are flagged inline for US 5,566,315, 5,442,747, 5,130,986, 5,440,692, and 4,158,235. Dates that I verified during this session are: 5,179,662 (issued 1993-01-12, filed 1989-08-31), 5,289,470 (issued 1994-02-22, filed 1992-12-14), and 5,303,347 (issued 1994-04-12, filed 1991-12-27). I will not assert a date I have not seen.
3. "Anticipates" is used here in the strict § 102 sense (a single reference disclosing every claimed element). If the goal is an invalidity contention, the productive route is a § 103 combination, e.g., Ugajin (5,046,039) for the transmit/receive pool split + Doshi (5,130,986) for link-rate-based buffer sizing + Corrigan (5,179,662) for dynamic resizing — still facing the absence of any teaching to compute these at the beginning of message reception from IrDA-negotiated parameters.

Key URLs: patent text — https://patents.google.com/patent/US5916309/en; US 5,179,662 — https://patentimages.storage.googleapis.com/f6/6f/01/c506af2125b812/US5179662.pdf; US 5,289,470 — https://uspto.report/patent/grant/5289470; US 5,303,347 — https://worldwide.espacenet.com/publicationDetails/biblio?CC=US&NR=[5303347A](/patent/5303347A); EP family search report (EP 0910203 / related Lexmark filings) — https://patentimages.storage.googleapis.com/dc/0e/db/a8c1a19a4b9092/EP0910203A3.pdf.

I'll ground this analysis in the references actually cited on the patent's face/Background section. Let me verify key ones.

Let me verify the remaining cited references.

Obviousness Analysis — U.S. Patent No. 5,916,309 (Lexmark International)

"System for dynamically determining the size and number of communication buffers based on communication parameters at the beginning of the reception of message"
Inventors: Rickey Carter Brown; Mark Lane Mayberry — Appl. No. 08/854,510 — filed 1997‑05‑12 — granted 1999‑06‑29 — expired (anticipated expiration 2017‑05‑12)

Legal framework applied: 35 U.S.C. § 103; Graham v. John Deere Co., 383 U.S. 1 (1966) (scope/content of prior art, differences, PHOSITA level, secondary considerations); KSR Int'l Co. v. Teleflex Inc., 550 U.S. 398 (2007) (predictable combinations, design incentives, "a court must ask whether the improvement is more than the predictable use of prior art elements according to their established functions"). I analyze only the references appearing in the patent's own Background/Prior Art discussion and the references cited on the face of the document. I flag where I could not independently verify a reference through live search.

---

1. Scope and critical date

• Priority/filing date: 1997‑05‑12. Only references qualifying as prior art under § 102(a)/(b)/(e) before that date can be used, unless the claims' effective date is later.
• The patent describes itself as directed to a printer with an IrDA infrared port that, at the beginning of a message/reception, sets up a pool of transmit buffers and receive buffers. Independent claim 1 recites the timing and the parameter-driven sizing:

"…creating a plurality of transmit buffers in which the number of said transmit buffers and size of each of said transmit buffers is dependent upon the total available pool size at the beginning of the reception of an asynchronous message…; and (b) creating a plurality of receive buffers in which the size of each of said receive buffers is dependent upon communications parameters, at the beginning of the reception of said message, relating to the type of communications link… and dependent upon the total available pool size…; and in which the number of receive buffers… is further dependent upon the total available pool size and the size of said receive buffers…" (claim 1)

The alleged inventive hook is therefore a conjunction of four ideas, each individually old in the art:

---

2. The prior art on the face of the patent (as characterized by the patentee)

Note on verification: I confirmed the Corrigan, Chang, Gagne, Milillo, and Ugajin disclosures through live sources (URLs above). The Chan ('747), Doshi ('986), Janicek ('692) and Call ('235) characterizations below rest on the '309 Background text itself because my live verification calls were curtailed. A PHOSITA-level analysis should confirm those four against their full texts before relying on them as § 103 art.

---

3. Person Having Ordinary Skill in the Art (PHOSITA)

A practitioner with a bachelor's degree in electrical/computer engineering (or equivalent) and 2–3 years' experience in data‑communications buffer/driver design, familiar with packet‑switched and IR (IrDA) link protocols, including the concept of negotiating link parameters at session setup and sizing FIFO/packet buffers to link speed × turnaround time. This is the level at which the references would be read.

---

4. Primary § 103 combinations

Combination A (strongest) — Ugajin '039 + Corrigan '662, optionally + Milillo '315

What each teaches

• Ugajin '039 discloses splitting a RAM buffer into a transmit buffer area and a receive buffer area whose border moves — i.e., the ratio of TX:RX allocation is dynamically changed based on measured traffic (receiver-buffer-busy events) and an initial value setting table applied at reset. This is the closest art to claim 1's "dividing a pool area into a plurality of transmit buffers and receive buffers," and to sub‑element (ii) (TX allocation based on available pool).
• Corrigan '662 expressly discloses that "the size of the buffers is dynamically increased and decreased as needed to achieve an optimum size based on the unique and dynamically changing computer environment," and that the buffer size is set to the "Operation Size" at the start of a write/read — i.e., the size of the buffer is a function of the size of the thing being processed and the available system resources. This supplies claim 1's notion that buffer size is set from available memory + the workload's characteristics, and it is the art cited by the patentee for "the size of both buffers is increased/decreased."
• Milillo '315 supplies the idea of computing the threshold dynamically from measured/predicted workload rather than using fixed thresholds, reinforcing that a PHOSITA would treat buffer allocation as a computed, parameter‑driven quantity, not a fixed constant.

Why a PHOSITA would combine them. Both are buffer‑management systems in the same field (Ugajin in a communication adaptor's LLC layer; Corrigan in main‑storage I/O buffering); both address the identical problem the '309 patent states — memory is finite and the workload/link conditions change, so a fixed allocation wastes memory or drops data. Ugajin explicitly solves the TX/RX split; Corrigan explicitly solves the how‑big question. Combining them is the predictable use of two known techniques for their established functions (KSR). The only remaining gap is that both adjust during or after operation rather than at the beginning of message reception.

Gap: Claims 1/12 require the recomputation "at the beginning of the reception of… message." Neither Ugajin (adjusts at reset on a busy‑state timer) nor Corrigan (adjusts mid‑stream) performs it strictly at message‑reception onset.

---

Combination B — Doshi '986 + Ugajin '039 + Corrigan '662 (supplies the "at the beginning" + link‑parameter sizing)

This combination is aimed specifically at the limitation that distinguishes the '309 claims from A.

• Doshi '986 computes a window size as a function of the bandwidth‑delay product of the link — i.e., the amount of data that can be in flight. The '309 specification defines its own key variable in essentially the same terms: the "maximum turnaround time" is the time a station may hold the channel, and "it serves no purpose to have either transmit or receive buffers that are larger than the amount of data that can be transmitted (based on the data transfer rate) within this time period." Doshi thus teaches sizing a receive‑side storage window as a function of a link parameter (bandwidth × delay) computed at connection setup — squarely the "communications parameters relating to the type of communications link" of claim 1(b), determined "at the beginning."
• The '309 patent itself supplies the express bridge: its Background acknowledges the analogous reasoning in Doshi, and its claims define the buffer ceiling as (turnaround time − minimum turnaround) × transmission speed − BOFs. That is a bandwidth × time computation of the same genus as Doshi's bandwidth‑delay product.
• Ugajin + Corrigan then supply the pool split and the dynamic size adjustment.

Why combine. A PHOSITA designing an IrDA port would look to a system (Doshi) that already sizes receiver‑side flow‑control storage from a link's bandwidth‑delay product, and would find it obvious to compute that size once, at the beginning of the reception (when the parameters are known from discovery), rather than continuously — because recomputing after data has arrived risks losing already‑buffered data, and because the parameters are fixed for the session. That is a predictable design choice (KSR): move the computation to the point in time when its inputs become known and are stable.

---

Combination C — Chang '470 and/or Gagne '347 + Corrigan '662 + Ugajin '039 (supplies "size RX buffers from link/communications parameters")

• Chang '470 expressly ties buffer sizes and the allocation sequence to the protocol/environment: buffer sizes "are allocated on the basis of observed or expected traffic for the specific environment," with examples of FDDI/Ethernet/Token Ring and multi‑modal packet‑size distributions, and different sequences per network interface where "their characteristics are very different from one another." This maps to claim 1's "communications parameters relating to the type of communications link."
• Gagne '347 teaches selecting the buffer (or buffer set) by a packet attribute including packet size and "optimizing the amount of memory which must be dedicated to storing these packets" — the same objective as the '309 claims.
• Adding Corrigan (dynamic size) and Ugajin (TX/RX pool split and ratio) yields every element of claim 1(a)–(b) except the precise "beginning of reception" timing, which Doshi (Combination B) supplies.

Motivation. Chang, Gagne, Corrigan and Ugajin are all in the same field of buffer allocation for packet/communications I/O and address the same problem (wasted memory vs. dropped packets). A PHOSITA seeking to build a printer IR driver would naturally look to network‑adapter buffer‑allocation art (Chang, Gagne) and to dynamic pool art (Ugajin, Corrigan).

---

Combination D — Janicek '692 and/or Chan '747 + Corrigan '662 (pool expand/contract and mode‑based boundaries)

• Janicek '692 teaches a buffer pool that is dynamically expanded or contracted (4K/32K pages), with allocation conditions based on remaining storage — analogous to "the total available pool size" driving buffer creation (claim 1).
• Chan '747 teaches a central memory whose internal boundaries are changed when the mode changes, with buffer regions sized for worst‑case conditions of the data streams in the current mode. The '309's "discovery" step (selecting communications parameters) is functionally a "mode change," so Chan's mode‑boundary reallocation maps onto the '309's reconfiguration after discovery.

These two are best used as secondary references (to reinforce (i) pool‑size‑driven allocation and (ii) reallocation on a mode/timing change), rather than as the primary teaching of the link‑parameter sizing.

---

5. Claim‑by‑claim § 103 mapping

Legend: A = Ugajin '039; B = Corrigan '662; C = Chang '470; D = Gagne '347; E = Milillo '315; F = Doshi '986; G = Chan '747; H = Janicek '692; I = Call '235.

Independent claims 12 / 17 / 18 are apparatus counterparts of claims 1 / 6 / 7 and are rendered obvious by the same combinations (the structural elements — memory circuit with a pool area, bi‑directional port, processing circuit determining number and size — are conventional and taught by A/B/C/D/I).

---

6. Motivation‑to‑combine rationale (the heart of the § 103 case)

For each combination a PHOSITA would have had a concrete reason to combine, satisfying KSR:

1. Common field and common problem. All of A, B, C, D, E, F, G, H, I address buffer allocation in data I/O, where the shared problem is finite memory vs. variable workload/link conditions. Ugajin (communication adaptor LLC), Chang/Gagne (network adapters), Corrigan/Janicek (host‑main‑storage I/O), and Chan (video/multimedia cache) are all "buffering for high‑throughput I/O."
2. Complementary, non‑overlapping teachings. Ugajin solves the TX/RX split; Corrigan solves how big; Chang/Gagne solve sizing/selection from link protocol and packet size; Doshi solves how much in‑flight data the link can carry (bandwidth × delay); Milillo/Janicek solve dynamic thresholds and pool expand/contract. Combining them yields the predictable aggregate — no reference teaches away.
3. Design incentive / predictable result. The '309's own summary states the goal: "optimized memory utilization… the memory available for the port is always used in a manner that is most advantageous for the current connection parameters." That is precisely the motivation a PHOSITA would already have (reduce wasted buffers, avoid dropped packets) and is the explicit object of Chang, Gagne, and Corrigan. KSR: a "predictable use of prior art elements according to their established functions."
4. Recomputation at a known/stable point in time is an obvious design choice. Parameters are only known after the IrDA "discovery" negotiation; it is a trivial engineering decision to compute the pool split at the beginning of reception (Combination B), both because the parameters are constant for the session and because recomputing later risks losing buffered data. Doshi already computes its window at connection setup.
5. Application to IR/printers is obvious. The Background itself states the buffer scheme "works well for any packet oriented link protocol such as Ethernet, Token Ring or Local Talk" and that choice of port type is a design matter. Applying a known allocator to an IrDA printer port is an obvious application, not an inventive leap.

---

7. Where the § 103 case is weak / nonobviousness considerations

An examiner or court should weigh these against the prima facie case; they are material:

1. Claims 6 / 7 / 17 / 18 — the explicit min{} formulas, including the BOF term. I have found no reference that discloses sizing a buffer as min{(maxTAT − minTAT) × bytes/ms − BOFs; maxPacket}. The best art (Doshi) teaches the bandwidth‑delay‑product genus but not this closed form, and BOFs are IrDA‑protocol‑specific. This could be argued either way:
   • For obviousness: once the parameter set is known (IrDA defines maxTAT, minTAT, BOFs, maxPacket and speed), deriving a "maximum useful buffer = data transmittable in the available window" formula is a routine mathematical/engineering optimization — KSR permits obviousness where a result is the product of a known design incentive and a known technique; a formula that merely expresses a designer's recognized ceiling is, at minimum, prima facie obvious as an optimization.
   • Against: the specific subtraction of the BOF term and the min() against maxPacket may not have been recognized as the "useful" ceiling, and the BOF parameter is peculiar to the IR protocol — the Federal Circuit has found nonobviousness where the prior art did not disclose the specific parameter/relationship. This is the strongest nonobviousness foothold for the '309.
2. Claims 8/9 (SIR → double TX buffer size). The magnitude "double" is specific; art teaches increasing size for slow/overhead‑heavy links but may not show "doubling" for byte‑stuffing. Arguable, but doubling for byte‑stuffing (worst case 2 bytes/byte) is a predictable choice.
3. Claims 10/11/21/22 — predetermined minimum buffer counts and the 7‑packet IR window. The '309 Background states the 7‑packet window is an IrDA standard prior art fact; under § 103, standard/known protocol values used in the claimed manner are difficult to claim. Still, if the dependency of the minimum on the negotiated window is not in the art, claim 11/22 has some strength.
4. "Consumed directly out of the packet buffer" (specified as an advantage of the '309). Claims 1/12 do not recite it, so it cannot aid them under § 103 (claims measure the invention). But it is the sort of secondary‑consideration evidence (advantages/unexpected results) a patentee can advance for narrow claims. The instructions confirm claims are the measure.
5. Secondary considerations. No evidence of unexpected results, commercial success, copying, or long‑felt need specific to the buffer‑sizing feature is presented in the record I reviewed. Absent such evidence, the KSR rationales dominate.
6. Reference‑availability caveats. Four of the cited references (Chan '747, Doshi '986, Janicek '692, Call '235) were characterized from the patent's own text; actual disclosure scope must be confirmed against their full texts before a final rejection or validity opinion. Also confirm each reference's § 102(a)/(b)/(e) status relative to 1997‑05‑12 (the '309 priority date).

---

8. Bottom line

• Independent claims 1 and 12 would be prima facie obvious over U.S. 5,046,039 (Ugajin) in view of U.S. 5,179,662 (Corrigan), further in view of U.S. 5,130,986 (Doshi) (and/or U.S. 5,289,470 (Chang) and U.S. 5,303,347 (Gagne) for the "link‑parameter" sizing), with U.S. 5,566,315 (Milillo), U.S. 5,442,747 (Chan), and U.S. 5,440,692 (Janicek) as reinforcing secondary art. Motivation: same field, same finite‑memory problem, complementary teachings, and predictable results — plus the express design incentive stated in the '309 itself.
• Dependent claims 2–5, 13–16 fall with the independent claims; the "maximum turnaround time" and "discovery" elements are supplied by Doshi's bandwidth‑delay window and Chan's mode‑change reallocation, and choosing an IR port/printer is an obvious design choice.
• Claims 6, 7, 17, 18 (the min{} formulas) and, to a lesser degree, 8–11 and 19–22 present the strongest nonobviousness arguments, because no cited reference discloses the specific IrDA‑parameter formulas (maxTAT − minTAT) × speed − BOFs capped by maxPacket, nor the "double for SIR" and "minimum buffers > window" specifics. A validity attack on these claims should still be mounted on the "routine optimization" theory under KSR, but the outcome is less certain than for claims 1/12.

Sources

• US 5,916,309 (authoritative text) — https://patents.google.com/patent/US5916309/en
• US 5,179,662 (Corrigan) — https://patents.google.com/patent/US5179662
• US 5,289,470 (Chang) — https://patents.google.com/patent/US5289470A
• US 5,303,347 (Gagne) — https://www.freepatentsonline.com/5303347.html
• US 5,566,315 (Milillo) — https://patents.google.com/patent/US5566315A
• US 5,046,039 (Ugajin) — https://patents.google.com/patent/US5046039
• US 5,442,747 (Chan), US 5,130,986 (Doshi), US 5,440,692 (Janicek), US 4,158,235 (Call) — characterized from the '309 Background only; not independently re‑verified in this session.