US 6658526

I'll research this patent. Let me search the available sources.

Let me check specifically for any litigation or CAFC docket activity tied to this patent number.

US Patent 6,658,526 — "Network Attached Virtual Data Storage Subsystem"

Note on scope of search: I searched for the exact number 6658526 (US patent). I found no CAFC 2026 docket, appeal, or pending litigation tied to this specific patent number. Based on the record I can retrieve, the patent's enforceable life ended years ago. Details and caveats below.

Bibliographic data

Priority chain: This patent is a continuation-in-part of application Ser. No. 09/343,512 filed June 30, 1999 (which issued as US 6,529,996 B1, "Network attached virtual tape data storage subsystem"), itself a continuation of Ser. No. 08/815,739 filed March 12, 1997 (abandoned). The March 12, 1997 date is the priority date shown on the face of the record.

Assignment chain: Assigned to Storage Technology Corporation by the three inventors (assignment paperwork executed Oct. 19–20, 1999); later recorded transfers to Sun Microsystems, Inc. (merger) and then Oracle America, Inc. (merger/name change). Google Patents lists Oracle America Inc. as current assignee. Caveat: I am reporting the recorded chain from the patent record, not independently verifying current title.

Abstract

The patent describes a data storage subsystem that creates multiple virtual storage devices on demand, available to any and all of the system's host data processors. It uses a plurality of tape devices and automated tape libraries, with the tape devices connected to the data processors via a high-speed network. A Network Storage Manager (NSM) — with Resource Allocation, Resource Configuration, Resource Management, and Security Management software modules — controls system operations. Depending on requested performance, one or more tape devices are assigned to a virtual device, and data striping across multiple tape devices lets the storage bandwidth of a virtual device be changed on demand. Capacity and performance scale by adding/removing storage devices. The NSM lets devices be managed as a pool while still attached directly to the network as individual resources.

Plain-language overview of the independent claims

This patent has 20 claims, of which 10 are independent (claims 1, 2, 3, 6, 10, 14, 17, 18, 19, 20). All are drafted in broad "means for" (functional) style, which matters for how they would have been construed. Summaries:

Claim 1 — Core system claim. A network; multiple data processors on it; multiple storage devices on it; and a mechanism that dynamically builds a virtual device out of at least two of those storage devices in response to a storage request from a processor — where the virtual device is built to deliver a requested data storage characteristic selected from capacity, data rate, fault tolerance, and security. In plain terms: "put a pool of drives on a network and let a controller stitch several together on demand to hit the capacity/speed/redundancy/security the requester asked for."

Claim 2 — Tape-specific variant. Same structure, but the storage devices are tape drives and tape media managed by at least one automated tape cartridge library system, with the library system coupled to the virtual-device-creating mechanism.

Claim 3 — Dual-network variant. Claim 1's structure plus a control network connecting the data processors to the virtual-device-creating mechanism, where the data network is a high-speed network and the control network runs at a lower speed. (Dependent claims 4 and 5 narrow this to an Ethernet control network and a Fibre Channel high-speed data network, respectively.)

Claim 6 — Control-block-at-host variant. The virtual-device-creating mechanism is made up of (a) a network storage manager that receives the storage request and (b) multiple control blocks, each sitting between a respective data processor and the network, that redirect that processor's data to the storage devices as previously configured by the manager. Dependent claims 7–9 place those control blocks inside the data processor, inside its network interface card, or inside the network itself.

Claim 10 — Control-block-at-storage variant. Same as claim 6 but the control blocks sit between the respective storage devices and the network and redirect data coming off the network to the storage devices. Dependent claims 11–13 again allow the blocks to be integral with the storage device, its NIC, or the network.

Claim 14 — Control block inside the network. Structure of claim 1 where the redirecting control block is integral with the network itself. (Claim 15 repeats the "integral with the network" limitation — it appears redundant with claim 14; claim 16 narrows to a network fabric switch.)

Claim 17 — Pooling/management claim. A network, storage devices on it, and a management mechanism that manages the storage devices as a pooled resource, comprising a network system management function plus control blocks that translate data received from data processors to the appropriate storage devices as previously configured by that management function.

Claim 18 — Dynamic reconfiguration of virtual devices. A mechanism for dynamically managing a virtual device composed of at least two storage devices, specifically including the ability to dynamically change either the configuration or the performance of that virtual device as specified by a storage request.

Claim 19 — Method claim (creation + redirect). In a system of network + data processors + storage devices: receive a storage request; in response, dynamically create a virtual device by configuring multiple control blocks on the network to redirect data to one or more storage devices; then redirect data from a processor to the storage devices via those configured control blocks.

Claim 20 — Method claim (pooled resource, reconfiguration). In a system of a network and physical storage devices: configure at least two devices as a virtual device with a first storage characteristic; receive a storage request from a processor; in response, dynamically reconfigure those devices as a different virtual device with a second, different storage characteristic; and translate data received from the processor to the appropriate storage devices. This is the claim most directly aimed at the "change performance on demand" idea. (Drafting note: the claim text on the record contains a stray colon — "receiving a storage request from a processor:" — which appears to be a typographical artifact, and I have not treated it as substantive.)

Points of uncertainty / things I could not confirm

1. No CAFC 2026 docket found. I found no Federal Circuit appeal, oral argument, or 2026 docket entry involving US 6,658,526. Given the patent's March 12, 2017 expiration date and "Expired – Lifetime" status, it should not be collecting new infringement damages. If you need a definitive negative, this must be confirmed against PACER/CAFC docket systems directly rather than a general web search.
2. Litigation context is adjacent, not identical. StorageTek/Sun/Oracle did litigate storage patents in this era (e.g., StorageTek v. Cisco; Sun/StorageTek v. NetApp declaratory judgment over US 5,403,639, 5,410,667, and 6,581,185). I found no source tying US 6,658,526 specifically to an asserted patent in those cases, and I will not assert a connection.
3. Priority-date discrepancy in third-party databases. One aggregator (Unified Patents) lists the priority date as 1997-03-11 and assignee as Oracle America Inc. The authoritative patent record and USPTO assignment data show 1997-03-12. Per your instruction not to auto-correct identifiers, I am flagging the discrepancy rather than silently harmonizing it.
4. Legal status is not a legal conclusion. The "Expired – Lifetime" and expiration-date entries come from the patent record's own disclaimer that legal status is an assumption and not a legal conclusion.
5. I did not perform a formal claim-construction or validity analysis, and the claim summaries above are paraphrases for orientation, not constructions.

Sources: Google Patents record for US6658526B2 (patents.google.com/patent/US6658526/en); USPTO Patent Assignment 10331/286 (as indexed by PlainSite); related family member US 6,529,996 B1.

I'll research litigation involving US patent 6658526. Let me run targeted searches.

The initial searches returned other unrelated "'526" patents (e.g., 5,682,526, 6,594,526). Let me target the specific number and litigation databases.

Let me try the dedicated litigation databases directly.

Litigation Search — US Patent No. 6,658,526

Bottom line

No litigation involving US Patent No. 6,658,526 was identified in the sources I was able to search. I am reporting this as a "no hits found" result, not as an affirmative guarantee that none exists — see the confidence/limits section below.

What I verified about the patent itself (for disambiguation)

Because several different patents share the "'526" shorthand and appear throughout the litigation literature, here is the identifier set I searched on, confirmed against the authoritative patent text:

Cases found: none

I could not identify any district court action, ITC investigation, or Federal Circuit appeal in which US 6,658,526 (or its application 09/421,916) was asserted as a patent-in-suit, challenged in a declaratory-judgment counterclaim, or reviewed in a post-grant proceeding.

Important disambiguation — the "'526" hits you will see are different patents

My searches for "526" in patent litigation returned numerous matches that are not US 6,658,526. A patent number must be read literally; these are different documents:

• U.S. 5,682,526 — SpaceLabs Medical, "Method and System for Flexibly Organizing, Recording, and Displaying Medical Patient Care Information Using Fields in a Flowsheet" (named in i4i/Uniloc-era EHR litigation materials).
• U.S. 7,765,107 and the Fresenius/Par Pharmaceutical "'526 patent" — an Hatch-Waxman pharmaceutical patent (Fresenius counterclaim, D. Del.).
• Japanese Patent No. 6658526 (P6658526) — Mitsubishi Chemical, a biosubstance adsorbent/desorbent sheet. A completely unrelated grant in a different jurisdiction.
• Chinese CN application 201821226581.0 (utility model) and various commercial listing IDs.

None of these is US 6,658,526.

Confidence and limits (please read)

• Confidence that no major, reported litigation exists: moderately high. The patent expired for failure to pay maintenance fees / reaching end of term on 2017-03-12, it was a corporate portfolio asset of StorageTek → Sun → Oracle (holders that generally did not run serial assertion campaigns on this asset), and it surfaces in the record only as prior art.
• Confidence that no litigation exists anywhere: not high. My tool access was cut off before I could query PACER directly or run targeted queries against Docket Navigator, Lex Machina, RPX/Unified Patents' full litigation database, or the ITC EDIS. Keyword web search is an unreliable instrument for finding quiet litigation (e.g., a suit settled or dismissed within months, or a case where the patent is pled under the application number 09/421,916 rather than the patent number).
• The one litigation-adjacent hit I saw was the Unified Patents litigation portal and the UPC document server, but neither returned a record tied to US 6,658,526.

Recommended next steps to close the gap

1. PACER / CourtListener RECAP: search docket text for "6,658,526", "6658526", "09/421,916", and "09/343,512".
2. Unified Patents litigation portal (portal.unifiedpatents.com/litigation/caselist) and Docket Alarm — filter by patent number. Note that Docket Alarm's coverage of pre-2010 cases is thinner.
3. USPTO Patent Center / Global Dossier — check the "Litigation" and "Assignment" tabs, since a change of ownership (Oracle) can follow a settlement or acquisition.
4. Lex Machina / Docket Navigator — the only databases with reliable historical depth for 2003–2012 storage-sector suits.

If you need a litigable-certainty answer (e.g., for a freedom-to-operate or acquisition opinion), I should not be the sole source — commission a PACER/Docket Navigator search.

Practical note for your purposes

Even if a case exists, the patent's Expired – Lifetime status (term ran 2003-12-02 to 2017-03-12) means it can no longer be asserted for prospective infringement, and any past damages window is now well outside the 35 U.S.C. § 286 six-year lookback. Any litigation of consequence would necessarily predate March 2017.

I'll start by verifying the structured-data baseline against what web sources surface for this patent.

Let me run more targeted searches to test whether any proceeding exists under alternate citations.

Proceedings overview

Zero AIA trial proceedings on file. The structured "PTAB proceedings on file" block — sourced from the USPTO Open Data Portal — returns no IPR, PGR, or CBM for US 6,658,526, and my independent web searches (Google Patents family/citation data, PTAB E2E document hits, CourtListener, and general docket queries) surfaced no petition, institution decision, Final Written Decision, or Federal Circuit appeal naming this patent. Breakdown: 0 active, 0 claims invalidated, 0 claims sustained, 0 settled, 0 institution denied. The defensive posture this gives a defendant is not "the patent is hardened" — it is "the patent was never challenged at the Board, and, more importantly, all twenty claims remain untested and un-narrowed, but the patent expired on (or about) 2017-03-12, so the practical question is not validity but whether any enforceable damages window remains."

Search methodology and its limits (stated plainly): I could not query the PTAB API directly in this session. I relied on (i) the canonical structured block in the prompt, (ii) Google Patents' Cited By table for US 6,658,526 (351 citing entries, all patents/publications — no PTAB paper, no FWD, no CAFC citation), (iii) PTAB E2E / ptacts.uspto.gov search hits for this patent number and title, and (iv) general docket searches. None produced a proceeding. I did not find a proceeding that the ODP has simply failed to index, and I am not reporting one.

No proceeding sections are provided because no proceeding exists.

I will not invent proceeding numbers, panels, or dispositions to fill the requested per-proceeding template. For completeness, here is what the record does show, which is the closest thing to "procedural history" available:

• Patent identity: US 6,658,526 B2, "Network attached virtual data storage subsystem," inventors Thai Nguyen, Robert M. Raymond, Michael L. Leonhardt; Original Assignee Storage Technology Corp.; Current Assignee Oracle America, Inc. — reached via re-recorded assignments for (i) Storage Technology Corp. → Sun Microsystems, Inc. (merger, recorded 2016-02-04) and (ii) Sun / Oracle USA → Oracle America, Inc. (merger and change of name, recorded 2016-02-04).
• Priority / filing: Priority date 1997-03-12; application US 09/421,916 filed 1999-10-20; granted 2003-12-02. It is a continuation-in-part of US 09/343,512 (filed 1999-06-30, now US 6,529,996 B1, "Network attached virtual tape data storage subsystem"), itself a continuation of US 08/815,739 (filed 1997-03-12, abandoned).
• Legal status: "Expired - Lifetime"; front page shows Anticipated expiration 2017-03-12.
• No litigation-driven PTAB activity appears in the record. The citation graph shows 6,658,526 cited by later patents (e.g., Hitachi US 7,636,820; Commvault's family; SanDisk Enterprise IP LLC patents) — i.e., it is a heavily-cited piece of foundational prior art, not a heavily-challenged assertion vehicle.
• No Federal Circuit appeal exists, correspondingly, because there was no FWD or Board decision to appeal. (Nor did I find a CAFC or district-court decision construing these claims that would function as an estoppel anchor.)

A note on the sibling patent: the parent/CIP US 6,529,996 is a separate patent in the same priority chain and is not covered by this prompt's structured proceedings block. If Oracle (or any successor asserting tape-library/networked-virtual-tape claims) sends a demand, verify whether the number cited is 6,658,526 or 6,529,996 — they are different patents with different claims, and I have no PTAB data on the '996 in this record. Do not assume activity on one implies activity on the other.

Strategic summary

Claim status: all twenty claims are UNTESTED. No claim of 6,658,526 has been canceled, confirmed, or even construed by the Board. Independent claims 1, 2, 3, 6, 10, 14, 17, and 18 and dependent claims 4, 5, 7, 8, 9, 11, 12, 13, 15, 16 are all live on their face; method claims 19 and 20 as well. This cuts both ways: there is no IPR-narrowing to exploit (no prosecuted disclaimer, no canceled claim, no Board construction binding a district court), but also no defensive estoppel cushion and no tested validity holding to lean on. A defendant cannot say "claims 1–5 are dead." It can say "nothing about these claims has ever been adjudicated."

The dominant fact is expiration, not invalidity. Under 35 U.S.C. § 154(a)(2), because the application specifically references earlier applications under § 120 (a CIP of 09/343,512, itself a continuation of 08/815,739), the term runs 20 years from the earliest such filing — 1997-03-12 — i.e., 2017-03-12, subject to any patent term adjustment (PTA). The front page labels this an "anticipated expiration" and separately lists "Expired - Lifetime," which is consistent but not the same as a PTA-recomputed date. Verify the actual expiration date in USPTO Patent Center before relying on this. If the patent did expire on or about 2017-03-12, then a complaint filed in 2026 reaches back only six years under § 286 — i.e., to 2020 — and that entire lookback window post-dates expiration. On those facts there is no recoverable damages period at all, and no prospective injunction is available against an expired patent. That is a dispositive, non-validity defense that costs a fraction of an IPR.

Estoppel landscape: no estoppel attaches — everything is available. Because there have been no petitioners, § 315(e)(2) estops no one. Every § 102/§ 103 ground, every combination, and every prior-art reference (including the ~351 citing documents and the dense StorageTek/IBM/Commvault tape-virtualization art) remains fully available to a current defendant in district court or in a de novo IPR. There is also no prior PTAB or district-court claim construction binding anyone, so a defendant is free to advance a § 112(f) / § 112(b) theory: claims 1, 2, 3, 6, 10, 14, 17 and 18 are drafted in means-plus-function form ("means ... for dynamically creating," "network storage manager means," "control block means"), which invites a § 112(f) construction fight — and, if the specification lacks corresponding structure for a given "means," an indefiniteness attack. That is a litigation posture, not a holding; no tribunal has ruled on it.

Pattern signals: none of the usual markers. No repeat petitioner (there are no petitioners at all), no patent-owner PTAB appeal practice, no defensive aggregator (no Unified Patents or similar filing) in the chain. Ownership sits with a large operating company (Oracle America, Inc., via the Sun merger), not a shell or monetization vehicle — which is consistent with the observed pattern: expired in 2017 and never became an AIA-era assertion target. The absence of IPRs here is therefore a weak signal about the patent's strength and a strong signal about its commercial relevance window having closed. Note also that CBM review was never a fit (this is a storage-architecture patent, not a "financial product or service"), and the CBM program sunset on 2020-09-16 in any event.

Recommended next steps

If you are a defendant and the demand letter cites US 6,658,526:

1. Attack the term date first, before anything else. Pull the Patent Center file wrapper and confirm the actual expiration date, accounting for PTA. The Google Patents front page (https://patents.google.com/patent/US6658526/en) states an anticipated expiration of 2017-03-12 and lists the patent as "Expired - Lifetime." If that date holds, a 2026 complaint has a six-year § 286 lookback beginning in 2020 — entirely post-expiration — leaving zero recoverable damages. This is checkable in an afternoon and may end the dispute.
2. Confirm which patent is actually being asserted. Distinguish US 6,658,526 from its CIP parent US 6,529,996. The claim sets are different; a demand letter citing the wrong number is itself a signal about the plaintiff's diligence.
3. There is no FWD to cite, no IPR estoppel, and no canceled claim. Say so in your response rather than implying otherwise. The honest framing is: "no claim of this patent has ever been adjudicated; the patent expired in 2017; and there is no damages window."
4. Preserve the full prior-art case. With no § 315(e)(2) estoppel and no Board constructions on the record, every § 102/§ 103 ground remains open in district court: the citing-patent universe (e.g., Hitachi US 7,636,820; the Commvault and SanDisk Enterprise families) plus 1990s StorageTek/IBM/ADIC networked tape-library and RAIT art. Run a fresh § 102/§ 103 search on the asserted claims; do not assume a challenger's work already exists.
5. Develop the § 112(f) / § 112(b) theory. The "means ... for" limitations in claims 1, 2, 3, 6, 10, 14, 17 and 18 will need corresponding structure identified in the specification; where the specification is silent as to a recited function, indefiniteness is on the table. Useful both defensively and as leverage.

If Oracle or any successor appears poised to assert a patent in this family:

1. Check US 6,529,996 separately for PTAB activity and litigation history — it is a distinct patent and is not addressed by the structured proceedings block in this prompt.

On IPR as a tool: an IPR on an expired patent is legally permissible (§ 311 places no live-infringement requirement on the patent, though a petitioner must still satisfy the § 315(a)/(b) real-party-in-interest and one-year-bar rules), and the Board applies Phillips to expired claims. But if the expiration date forecloses the damages window, an IPR buys you very little: it costs several hundred thousand dollars to invalidate claims in a patent from which no damages can be recovered. Litigation-cost arbitrage — not Board invalidation — is the right play on these facts.

No active proceedings are pending, so there are no trial-stage milestones to calendar (no institution deadline, no oral hearing date, no statutory one-year FWD date). If the structured data ever changes — a petition is filed — the § 315(b) one-year clock begins on service of a complaint alleging infringement, and the Board's FWD is due within one year of institution (§ 316(a)(11)).

Sources: Google Patents, https://patents.google.com/patent/US6658526/en (claims, family, assignments, legal status, "Anticipated expiration 2017-03-12" / "Expired - Lifetime"); USPTO PTAB E2E, https://ptacts.uspto.gov/ (no proceedings retrieved); USPTO Patent Center, https://patentcenter.uspto.gov/ (file wrapper — verify actual expiration/PTA); CourtListener, https://www.courtlistener.com/ (no docket or opinion for this patent); uspto.report, https://uspto.report/patent/grant/6,658,526 (bibliographic record).

I'll start by trying to pull the actual assignment records and any assertion history for this patent.

Good — I found a real reel/frame for the original assignment. Let me chase the Sun and Oracle merger recordings and any assertion history.

I have the original assignment's reel/frame. Let me try to pin down the merger recordings and check for any assertion of this patent specifically.

US 6,658,526 B2 — Assignment & Ownership Analysis

Patent: US 6,658,526 B2 — "Network attached virtual data storage subsystem"
Application: 09/421,916, filed 1999-10-20; issued 2003-12-02; priority 1997-03-12
Status: Expired – Lifetime (20-yr term from 1997-03-12; anticipated expiration 2017-03-12)
Family: Continuation-in-part of 09/343,512 (filed 1999-06-30, issued as US 6,529,996) which is a continuation of 08/815,739 (filed 1997-03-12, abandoned).

Inventors

Employer attribution is grounded, not assumed: the recorded assignment (USPTO Patent Assignment 10331/286) names all three as assignors to Storage Technology Corporation, 1 StorageTek Drive, Louisville, CO 80028, and the correspondent of record is at that same corporate address (see timeline). The specification's named system components — ACS/ACSLS, Network Storage Manager, RA/RC/RM/SM software modules — are StorageTek's own product vocabulary, so these are in-house engineers, not outside contributors.

Unusual patterns: None detected. I found no evidence of inventors departing StorageTek within 12 months of the 1999-10-20 filing; the record contains no inventor-side re-assignments, no employment-agreement disputes, and no later individual-assignor entries. Note the unusual family structure rather than an unusual inventor pattern: this patent is a CIP that was filed the same day the inventors signed, and its entire family (08/815,739 → 09/343,512 → 09/421,916) shares one 1997 priority date and stayed inside the same corporate chain for 18 years. That is the opposite of a pre-fire-sale signature.

Original assignee

Storage Technology Corporation (StorageTek) — named on the issued patent.

• Primary line of business: Enterprise data storage hardware and software; founded 1969 in Boulder/Louisville, CO by four ex-IBM engineers. Dominant in automated tape libraries (the SA- and SL-series robotic silos), tape drives (9840, 9940, T10000), tape management software, and — directly on point here — the StorageTek Virtual Storage Manager (VSM) and Virtual Library Extension (VLE) virtual-tape products, plus the ACS/ACSLS automated cartridge system and library software that the specification describes.
• Did they ship a product embodying the claims? Substantially yes. The claimed subject matter — pooled physical tape drives, dynamically configured into virtual devices with selectable capacity/data rate/fault tolerance, controlled by a Network Storage Manager over a high-speed network — corresponds to StorageTek's ACS + VSM virtualization architecture and, after acquisition, to the products still sold as Oracle StorageTek VSM 5/5e and the SL3000/SL8500 libraries with VLE. There is no evidence this patent was a paper-only asset at StorageTek.
• Current status: Operating, but no longer independent. StorageTek filed Chapter 11 in 1984 and emerged in 1987 (a full decade before this family's 1997 priority date — it is not a fire-sale antecedent for this patent). It was acquired by Sun Microsystems for US$4.1B cash, deal announced 2005-06-02, closed 2005-08-31; Sun was in turn acquired by Oracle Corporation on 2010-01-27 for US$7.4B. The StorageTek brand and line survive as Oracle StorageTek under Oracle America, Inc. No bankruptcy, dissolution, or assignment-for-the-benefit-of-creditors event appears at any point after 1987.

Assignment timeline

Coverage note / limitation on this section. The USPTO Assignment Center record for this patent could not be fully retrieved by the tools available to me. Only one post-filing recording (10331/286) exposes a reel/frame in any indexed copy I could reach. The two merger recordings below are reported from Google Patents legal events, which lists the conveyance type, parties, and recordation date but not the reel/frame. I have not fabricated reel/frame numbers for them; they must be pulled directly from Assignment Center (search patent 6,658,526) to complete the record. This is a data gap, not a "no records" finding — there are recorded assignments.

1999-10-19 / 1999-10-20 (executed) — recorded 1999-10-20 (per Google Patents legal events) — Reel 10331/286

• Conveyance: Assignment of Assignors' Interest (first assignment, inventors → company)
• Assignor: Thai Nguyen (paperwork executed 1999-10-19); Robert M. Raymond (executed 1999-10-20); Michael L. Leonhardt (execution date not separately exposed in the indexed copy)
• Assignee: Storage Technology Corporation, 1 StorageTek Drive, Louisville, CO 80028
• Correspondent: TIMOTHY R. SCHULTE, ESQ., Mail Station 4309, One StorageTek Drive, Louisville, CO 80028-4309. Flag: this is StorageTek's own headquarters address with an internal mail stop — an in-house patent attorney filing the company's standard employee assignment, not an outside recording agent. Schulte appears once in this chain with no recurrence and no appearance in any NPE assertion listing I could find. No repeat-correspondent signal.
• Context: Ordinary employee invention assignment, filed contemporaneously with the application — not a fire-sale, securitization, or transfer-to-asserter.

2005-08-31 (transaction closed) — recorded 2016-02-04 (per Google Patents legal events) — Reel/frame not exposed in indexed sources

• Conveyance: Merger (Storage Technology Corporation merged into/acquired by Sun Microsystems, Inc.)
• Assignor: Storage Technology Corporation
• Assignee: Sun Microsystems, Inc.
• Correspondent: Not exposed in the indexed record; obtain from Assignment Center.
• Context: Internal corporate reorganization — a US$4.1B all-cash strategic acquisition of an operating tape-storage vendor by an operating systems/servers vendor. Anomaly worth verifying: Google Patents shows the recordation date as 2016-02-04, roughly a decade after the 2005 closing. That pattern (late bulk chain-of-title clean-up, likely swept in alongside the Oracle name-change filings) is common when a successor records legacy merger evidence in one batch; it is not itself evidence of NPE behavior, but it does mean the recorded date does not track the operative transaction date.

2010-01-27 (transaction closed) — recorded 2016-02-04 (per Google Patents legal events) — Reel/frame not exposed in indexed sources

• Conveyance: Merger and Change of Name (Sun Microsystems, Inc. → Oracle America, Inc., via Oracle USA, Inc.)
• Assignor: Sun Microsystems, Inc. (record also lists ORACLE USA, INC. and ORACLE AMERICA, INC. among the assignor/assignee parties)
• Assignee: Oracle America, Inc.
• Correspondent: Not exposed in the indexed record; obtain from Assignment Center.
• Context: Internal corporate reorganization following Oracle's US$7.4B acquisition of Sun. Same 2016-02-04 batch recordation as above.

2017-03-12

• Conveyance: None — legal-status event: anticipated expiration / patent expired for failure to reach end of term. No assignment associated. Current assignee of record remains Oracle America, Inc.

Post-issuance enforcement note (not an assignment): StorageTek sued Quantum Corp. in August 2003 (Denver), asserting patents relating to reading/writing data on the backside of data-storage tape — a media-technology suit, not the virtualization subject matter of US 6,658,526. Sun (as StorageTek's acquirer) settled in late February 2006 for $25M with a limited patent cross-license. Separately, Sun asserted StorageTek-acquired patents against Network Appliance (the '639/'667/'857/'095 patents) in its 2007–2008 counterclaims. I could not confirm that US 6,658,526 itself was ever asserted in any litigation.

Timeline diagram

timeline
    title Ownership of US 6658526
    1997 : Priority application filed by StorageTek
    1999 : CIP filed and assigned to StorageTek
         : Reel 10331 0286 in-house counsel Schulte
    2003 : US 6658526 issues
    2005 : StorageTek acquired by Sun Microsystems
    2010 : Sun acquired by Oracle Corporation
    2016 : Sun and Oracle merger records entered
    2017 : Patent reaches end of term and expires

NPE / troll-pattern signals

1. Shell-entity transfer — not present. The chain contains no "IP / Patents / Licensing / Holdings / Ventures" entity, no single-purpose LLC, and no registered-agent service address. Every assignee in the chain (Storage Technology Corporation, Sun Microsystems, Inc., Oracle America, Inc.) is a named operating company with products, employees, and SEC-reporting parents. Reel 10331/286 (inventors → StorageTek) and the two merger entries all run to operating entities.

2. Known asserter in the chain — not present. No match to Acacia, Marathon, Intellectual Ventures, IPNav, Wi-LAN, Mosaid/Conversant, Vringo, Pendrell, Innovatio, MPHJ, Lumen View, Round Rock, Document Generation Corp, or any Spangenberg entity. The terminal assignee is Oracle America, Inc. — a product company. Note the shape contrast: StorageTek and Sun did assert patents (Quantum, Network Appliance), but as competitor-vs-competitor product disputes, which is operating-company assertion, not NPE conduct, and US 6,658,526 is not confirmed as one of the asserted patents.

3. Repeat correspondent across the chain — not present. Exactly one correspondent is exposed: Timothy R. Schulte, Esq., Mail Station 4309, One StorageTek Drive, Louisville, CO 80028-4309, on Reel 10331/286 only. He appears once. The address is the assignee's own corporate campus with an internal mail stop, indicating first-party in-house prosecution/assignment practice. There is no recurrence within this chain and no indication he appears as correspondent on an NPE assertion list. A single appearance is expressly not a finding under the stated criteria.

4. Cascading transfers — not present. Two successive transfers exist, but they are 5 years apart (2005 and 2010), both public-company mergers with regulatory filings (EU merger case COMP/M.3866 — Sun/StorageTek), not a sub-24-month chain of LLCs. There is no shared correspondent address or common-principal pattern. The 2016-02-04 batch recordation of both mergers is a recording-household artifact, not a substantive cascade.

5. Pre-litigation transfer — not present. No assignment in this chain sits within 6 months of any infringement suit naming this patent; the two transfers are separated from any identified litigation by years, and the patent expired 2017-03-12, extinguishing assertion value.

6. Bankruptcy fire-sale — not present (for this patent). StorageTek's Chapter 11 was in 1984, thirteen years before this family's 1997 priority date and fifteen years before the 1999 filing; it emerged in 1987. The 2005 exit was a US$4.1B cash merger, not a distressed sale, and the 2010 Sun exit was a US$7.4B acquisition in Sun's own decline, not a bankruptcy auction. No reel/frame in this chain records a trustee, receiver, or §363 sale.

7. Privateering — not present. No operating company transferred this patent to an outside assertion vehicle. The chain is strictly vertical (target absorbed by acquirer), with no license-back, no revenue-share to a litigating NPE, and no third-party asserter appearing as assignee.

8. Defensive aggregator (anti-NPE) — not present. The chain does not terminate at RPX, AST, LOT Network, Unified Patents, or OIN. The patent was never neutralized by defensive aggregation; it simply ran to term inside Oracle and expired.

Verdict

Operating-company assertion — with one express caveat.

Justification: the entire recorded chain is operating companies — inventors to Storage Technology Corporation (Reel 10331/286, executed 1999-10-19/20), then Sun Microsystems, Inc. (merger, closed 2005-08-31), then Oracle America, Inc. (merger/name change, closed 2010-01-27) — and there is not one shell-entity transfer, NPE-asserter assignee, recurring NPE correspondent, sub-24-month LLC cascade, bankruptcy sale, privateering arrangement, or defensive-aggregator terminus anywhere in the record. The original assignee genuinely built and sold the claimed technology (ACS/ACSLS, StorageTek VSM/VLE, SL-series libraries, still shipped today as Oracle StorageTek products), and StorageTek did litigate its tape portfolio against a real competitor (Quantum, settled February 2006 for $25M) while this patent was in force.

Caveat, stated plainly: I could not confirm that US 6,658,526 itself was ever asserted — the Quantum and NetApp matters involved different patents, and the patent expired 2017-03-12. So the accurate reading is "operating-company-held and operating-company-enforced portfolio, specific patent assertion unverified," not a documented suit on this number. Had the patent been sold into a licensing vehicle, the tell would have been a 2010–2017 entry naming an IP-holdings LLC and an outside correspondent; no such entry exists.

Verification links:

• USPTO Assignment Center — search patent number 6,658,526: https://assignmentcenter.uspto.gov/
• Legacy Assignments on the Web search: https://assignment.uspto.gov/patent/index.html
• The one assignable reel/frame record I could confirm (Assignment 10331/286): https://www.plainsite.org/patents/assignment.html?id=[1775711](/patent/1775711)
• Google Patents legal events / reassignment entries for this patent: https://patents.google.com/patent/US6658526/en

Open items for a follow-up pull from Assignment Center: (1) the reel/frame for the Sun Microsystems merger recording; (2) the reel/frame for the Oracle America merger-and-change-of-name recording; (3) the correspondents of record on both merger entries (expected to be Oracle/Sun in-house or outside M&A counsel, which would further confirm the absence of a repeat NPE filer).

I'll start by pulling the authoritative record for this patent number and its cited references (the prior art listed on the face of the patent).

This gives me the key front-page reference list. Let me pull the complete "References Cited" list for 6658526 and confirm the PCT search report (which cites the actual prior art of record).

Prior‑Art Analysis — US 6,658,526 (number interpreted literally as "6658526")

0. Record identification (as returned)

I did not substitute or "correct" any number. Note that the only patent in this family with a similar number is US 6,529,996 (the parent/sibling) — it is a related application, not prior art against 6658526, and is treated separately below.

1. Threshold correction about what "cited by" means here

Your premise — "look at each patent citation for 6658526" — needs one clarification, because the two citation sets point in opposite directions:

• The "Cited By (351)" list on Google Patents (and the truncated US2005… lists returned in search) are forward citations — patents that issued after 2003 and cite 6658526. They are not prior art to 6658526.
• The "Prior art keywords" on the Google Patents page (network, data, storage, storage devices, tape) are machine-generated indexing terms — also not prior art.
• The actual prior art of record is the "(56) References Cited" block on the face of the patent, which the search cleared up. That block is reproduced below in full.

A second correction on law: § 102 requires a single reference to disclose every element of a claim as arranged. Since 6658526 issued over all of these references, none of them actually anticipated it — the examiner used them singly for dependent features and in combination for the independent claims. So everything below is a candidate mapping ("potentially anticipates"), not a finding.

Because the § 102 availability of each reference depends on its own effective date versus the 1997‑03‑12/1999‑10‑20 critical dates, I flag the dating problem explicitly in § 5.

2. Prior art of record — the US patent documents (all 58 on the face)

Grouped by relevance. "Anticipates?" = which claim(s) it is potentially most relevant to under § 102 (based on title/subject matter — see caveat § 5).

Group A — Closest art: virtualization of storage / virtual tape & virtual libraries

Group B — Closest art: the network‑attached / distributed mass‑storage architecture

Group C — RAIT/RAID, striping and parity background (claims 1, 6, 10, 14, 20)

(Titles for the 1964–1985 block were only partially returned; see § 5 for what I could not verify.)

Group D — Robotic library / media management (claims 2, 19)

4,945,429 Munro; 5,181,909‑class art; 5,613,154 Burke et al., "System and method for management of transient data storage mediums in an automated data storage library" (1997‑03‑18); 5,689,481 Tamura et al., "Method of accessing multifile storage medium and library system" (1997‑11‑18); 5,303,214 Kulakowski et al. (1994‑04‑12); 5,410,000‑series.

Group E — Snapshot / configuration / miscellaneous (claims 14–18)

5,649,152 Ohran et al., "Method and system for providing a static snapshot of data stored on a mass storage system" (1997‑07‑15) — static snapshot (contrast with dynamic reconfiguration in claim 18); 6,041,363 Schaffer (VxD calls in a DLL); plus the pre‑1990 hierarchy/cache patents (3,139,387, 3,699,533, 3,806,888, 3,909,799, 3,949,377, 3,976,977, 4,021,782, 4,040,026, 4,054,951, 4,080,651/652, 4,084,234, 4,086,629, 4,110,823, 4,123,795, 4,126,893/894, 4,158,235, 4,215,400, 4,228,501, 4,241,420, 4,246,637, 4,276,595, 4,298,932, 4,310,883, 4,318,184, 4,327,408, 4,412,285, 4,414,644, 4,533,995, 5,129,088, 5,164,909, 5,386,516, 5,388,260, 5,412,791, 5,454,103).

Foreign patent documents of record

CA 892798 (Feb 1972); CA 907211 (Aug 1972); GB 1167762 (Oct 1969); GB 1353770 (May 1974); GB 1359662 (Jul 1974); GB 1496779 (Jan 1978); GB 1496780 (Jan 1978); GB 1547381 (Jun 1979); GB 2063532 (Jun 1981); JP 51‑18409 (Feb 1976); JP 52‑106641 (Sep 1977); JP 53‑22331 (Mar 1978); JP 53‑84632 (Jul 1978); JP 53‑98741 (Aug 1978); JP 53‑108747 (Sep 1978); JP 55‑153058 (Nov 1980); JP 55‑164958 (Dec 1980); JP 4‑48250 (Aug 1992). Relevance: background in virtual/ hierarchical storage; JP 4‑48250 (1992) is the only one close in time to the invention and is potentially relevant to claims 1/17.

Non‑patent literature of record (all listed on the face)

• Sam Coleman & Steve Miller, Mass Storage System Reference Model: Version 4, IEEE TC on Mass Storage Systems (38 pp.) — claims 1, 17.
• J.L. Sloan, B.T. O'Lear, D.L. Kitts, B.L. Irwin, "MaSSIVE™: The Mass Storage System IV Enterprise," Proc. IEEE, vol. 81, no. 4, pp. 621‑630 (1993) — claims 1, 3, 17, 18; the companion publication to Irwin '331 and the architecture the '996 spec distinguishes.
• Garth Gibson, "Network‑Attached Secure Disks (NASD)" — claims 1, 3, 6, 10.
• Nagle, Gibson, Ganger & Courtright, "Networking for Network‑Attached Storage" — claims 1, 3.
• Masstor Systems Corp., "Shared Virtual Storage System" System Description (1979) and System Overview (1980), Sunnyvale CA — claims 1, 17, 18.
• Fourth IEEE Symposium on Mass Storage Systems, Boulder CO (Apr 1980); FPEXEC File Processor Executive Program Logic Manual (Masstor, 1982); IBM 3850 MSS Introduction (GA32‑0038‑1) and Principles of Operation: Theory (GA32‑0035‑1); IBM 3830/3330 Reference Manual (GA26‑1592‑5); IBM System/370 Principles of Operation; Liptay, "Structural Aspects of the System/360 Model 85," IBM Systems Journal vol. 7 no. 1 (1968); Meade, Electronics (Jan 17, 1972); Smith, "Directions for Memory Hierarchies…" (1978); Christman et al., "A Network File Storage System" (LA‑8887‑MS); Chorn et al., "The Standard File Transport Protocol" (LA‑7388‑MS, 1978); Collins, "Mass Storage at LASL"; Waddell & Way, "Channel Scheduling by a Control Unit," IBM TDB vol. 17 no. 9 (Feb 1975); Boggs, "Virtual Input/Output Channels for a Digital Computer," IBM TDB vol. 20 no. 1 (Jun 1977); Nishimukai et al., "Universal I/O Device Controller Using Programmable Logic Arrays," IEEE (1978); plus Ampex TBM/NCAR/Shell‑Masstor purchase‑order materials. Relevance: these are the classic virtual‑storage, channel‑virtualization and staging references underpinning the "virtual device" concept in claims 1/17/18.

3. The strongest "anticipation‑type" candidates — the PCT examiner's picks

The international counterpart of this family, WO1998040810A2/A3 ("Network attached virtual tape data storage subsystem," Storage Technology Corp.), published 1998‑09‑17, carried a search report signed off by the ISA. That ISR named five references as category Y (relevant in combination) — these are, in effect, the examiner's closest art for the same disclosure:

Bottom line on the single best § 102 candidate: US 5,566,331 (Irwin) — it is the reference the examiner treated as most nearly disclosing the network‑attached, manager‑controlled, pooled mass‑storage architecture, and it is the architecture the patent's own specification works around. Under a strict § 102 test it lacks the tape‑device + connection‑block translation + on‑demand characteristic‑setting features that distinguish claims 1/6/19; it is therefore more naturally a § 103 (obviousness) reference than a pure § 102 reference.

4. Claim‑by‑claim "closest prior art" summary

5. Caveats you must carry into any opinion (stated explicitly, per the operating rules)

1. This is a relevance map, not a § 102 finding. I have titles, dates and claim‑numbers‑of‑record but not the full disclosure text of each reference. A real anticipation opinion requires reading each reference against each claim element. I am not asserting any reference discloses every limitation of any claim — and the fact that 6658526 issued in 2003 over this exact art is strong evidence none of them did.
2. Dating problem for the post‑priority references. The following references issued after the 1997‑03‑12 priority date: 5,689,481 (11/1997); 5,790,852 (8/1998); 5,805,864 (9/1998); 5,878,232 (3/1999); 5,881,311 (3/1999); 5,954,796 (9/1999); 5,996,024 (11/1999); 6,041,363 (3/2000); 6,044,442 (3/2000); 6,141,729 (10/2000); 6,154,854 (11/2000); 6,216,202 (4/2001). These can only be prior art under pre‑AIA § 102(e) (or § 102(a)/(b) if their publication predates the critical date), which turns on their effective U.S. filing dates — dates I did not retrieve and will not guess. Verify each one's filing date before relying on it.
3. Means‑plus‑function. Claims 1, 2, 3, 6, 10, 14, 17 and 18 are drafted in "means for…" form and are governed by 35 U.S.C. § 112 ¶ 6. Anticipation therefore requires the reference to disclose the corresponding structure (NSM, Resource Allocation/Configuration/Management software, connection blocks CB0–CB9, striping/parity engine), not merely the recited function.
4. Family member ≠ prior art. US 6,529,996 B1 and WO1998040810A2/A3 share the 1997‑03‑12 priority and the same inventors/assignee. They are the parent/companion disclosures, not § 102 art by "another," and should not be tabulated as anticipating references.
5. Unverified items. I could not verify (a) the titles of the pre‑1990 U.S. patents (3,139,387 through 5,454,103) or of 5,214,768 / 4,945,429 / 5,454,103; (b) the exact filing dates of any reference; (c) the substance of the foreign patent documents. Where the tables above show a title, it came from the patent's own face (FreePatentsOnline/uspto.report); where a description is generic, it is flagged as background.

Sources

• US 6,658,526 face + claims: https://patents.google.com/patent/US6658526/en ; https://uspto.report/patent/grant/6,658,526 ; https://FreePatentsOnline.com/6658526.html
• Family/priority: https://patents.google.com/patent/US6529996B1/en ; https://patents.justia.com/patent/[6529996](/patent/6529996)
• PCT search report (WO1998040810A3): https://patentimages.storage.googleapis.com/43/01/7a/6fa8a1aa5d5394/WO1998040810A3.pdf

I'll research the prior art landscape for US6658526 before building the obviousness analysis.

Let me get the details of the specific prior-art references cited during prosecution.

Obviousness Analysis — US 6,658,526 B2

0. Identification of the patent and the record used

Important sourcing note (honesty caveat). The "Prior Art" material visible in the supplied page is: (a) the auto-generated Prior art keywords field ("network, data, storage, storage devices, tape"), (b) the Prior art date (1997-03-12), and (c) a Cited By (351) table. The Cited By table is a list of forward citations. Scanning it, no entry has a priority date earlier than 1997-03-12 (the earliest are US 2008/0037777 A1, priority 1998-03-11; US 7,962,642 B2 and US 8,019,963 B2, priority 1997-10-30; US 8,234,477 B2, priority 1998-07-31). Forward citations are therefore not §102/§103 prior art against this patent. The substantive prior-art basis for a §103 analysis must come from the prosecution record, and the family's record is available: the international search report on the parent's PCT publication WO 1998/040810 A3 (published 1998-09-17), which cites against the parent's claims 1–14:

• US 5,566,331 A (Irwin, Jr. et al.) — 1996-10-15 — "Y" (obviousness-type), claims 1–14
• US 4,945,429 A (Munro et al.) — 1990-07-31 — "Y", claims 1–14
• US 5,630,067 A (Kindell et al.) — 1997-05-13 — "Y,P", claims 1–14
• US 5,214,768 A (Martin et al.) — 1993-05-25 — "Y", claims 1–14
• US 5,506,986 A (Healy) — 1996-04-09 — "Y", claims 1–5 and 10–14
  (Source: https://patentimages.storage.googleapis.com/43/01/7a/6fa8a1aa5d5394/WO1998040810A3.pdf)

I could not retrieve the disclosure content of US 5,214,768 and US 5,506,986 in the available material. I do not know their contents with high confidence, so I use them only as "record" references (with the passages the ISA relied on noted) and do not build any ground on unverified disclosure. Everything below that I attribute to Irwin, Munro and Kindell is grounded in text I retrieved.

A useful threshold point: every reference in the ISA list is dated 1993–1997-05-13, i.e. more than one year before the 1999-06-30 CIP/continuation filing and before the 1999-10-20 filing of the '526. They are consequently §102(b) statutory-bar art under either the 1997-03-12 priority date or the later 1999 dates — they cannot be sworn behind or antedated. That removes the priority-date fight from most of the analysis.

1. Claim 1 construed, and what the prior art must supply

Claim 1 (broadest claim) requires:

1. a network;
2. a plurality of data processors coupled to it;
3. a plurality of storage devices coupled to it;
4. means … for dynamically creating at least one virtual device from at least two of the plurality of storage devices;
5. the creating means being responsive to a storage request from a data processor; and
6. the virtual device being established with at least one data storage characteristic of (i) capacity, (ii) data rate, (iii) fault tolerance and (iv) security as specified by the storage request.

Two construction notes that drive the whole analysis:

• Means-plus-function. Every "means" element is a §112 ¶6 limitation. The corresponding structure disclosed in the '526 specification is the NSM (Resource Allocation / Resource Configuration / Resource Management software resident in the ACSC controller) plus the connection blocks CB0–CB9, and the '526 §112 ¶6 algorithm is the sequence described at col. "Referring to FIG. 5" (RM receives request → RM consults ACSLS via RA → RM determines drive count/configuration → RA reserves devices → RC configures the connection blocks → RM notifies the host). Prior art must therefore disclose a controller programmed to allocate, configure and manage a logically unified device out of plural physical devices, not merely the physical plumbing.
• "at least one … characteristic of (i) … (ii) … (iii) … and (iv)" is best read under the broadest reasonable interpretation as "at least one of (i)–(iv)." This makes claim 1 broad: satisfying any one of capacity, data rate, fault tolerance, or security in response to the request suffices. (If instead read to require all four simultaneously, the claim would be narrower — and then the Munro/Kindell/RAID combination becomes more important, not less.)

2. The primary references

2.1 US 5,566,331 A — Irwin, Jr. et al. ("Mass storage system for file-systems") — https://patents.google.com/patent/[US5566331A](/patent/US5566331A)/en

Discloses almost the entire preamble of claim 1 and much of its substance:

• A mass storage system consisting of "a plurality of data storage devices that are interconnected via a high speed network or a data switching-channel to a plurality of client data processors" (Problems section).
• A mapping/data-flow controller (the "bitfile server" / "regulating means") that "maintains mapping data independent of said plurality of processors" (Irwin claim 18) — i.e., a centralized logical-to-physical map, which is the essence of virtualization.
• Logical units (file-systems stored as "bitfiles") that are staged onto shared direct-access storage and channel-attached to a requesting processor, which "allows the requesting client data processor to treat the retrieved file-system as if the requested file-system resided on a direct access storage device that was locally attached." This is the "host does not need to know how the virtual device is realized" concept of the '526.
• Explicit support for plural, differing backend configurations that a POSITA would recognize as RAID: "for illustrative purposes … a RAID-3 disk array data storage subsystem … a RAID-5 disk array direct access storage subsystem … and a tape drive sequential access storage device" (Irwin Fig. 1 discussion), plus claim 17 ("said second data storage means comprises: tape media data storage subsystem") and claim 19 ("said exchanging means comprises a nonblocking network").
• "means in said exchanging means for concurrently transferring data … between said first data storage means and said requesting processor with other data being transferred from said first and second data storage means to other said processors" (Irwin claim 14) — i.e., simultaneous, multiplexed use of the pooled devices.

What Irwin teaches against claim 1: network + processors + storage devices + dynamic, request-driven mapping of a logical device image (bitfile) onto selected ones of plural physical devices, with mapping held independently of the hosts. What it does not squarely teach: treating "at least two" devices as a single aggregated device to satisfy a performance characteristic (data rate), and a demand-driven selection of the number/configuration of devices based on a requested data rate. That gap is filled by Kindell.

2.2 US 5,630,067 A — Kindell et al. (IBM, "System for the management of multiple time-critical data streams") — https://patents.google.com/patent/US5630067 and https://worldwide.espacenet.com/publicationDetails/biblio?CC=US&NR=[5630067A](/patent/5630067A)&KC=A&FT=D

Discloses the requested-performance-driven, pooled multi-device allocation missing from Irwin:

• A networked system that "distributes that data among multiple storage devices" in which "each viewing station has access to any file on any of the storage devices and can retrieve and display any file on demand" (abstract) — a pooled, shared storage substrate, not dedicated per-host devices.
• Each computer "capable of acting as both a client and a server" and retrieving a stream "either from its own local storage or from the storage associated with another computer" — logical device decoupled from physical location.
• Resource-management facilities that "allocate … resources", with each allocating mechanism maintaining "an internal table of the utilization level of each of its critical local system components (disk drive, computer processor, CPU/BUS, LAN adapter, etc.)" and, "when a request is made … the allocation mechanism allocates a portion of this maximum throughput value necessary to service that request"; if the record shows over-commitment, the request is rejected.
• A single computer that "supervises and controls access to the LAN" and rejects requests exceeding network capability — i.e., a network storage manager allocating pooled resources in response to a request with a throughput/performance dimension, and re-allocating as requests arrive and depart.

What Kindell teaches against claim 1: dynamically creating, in response to a request specifying a data-rate/throughput requirement, an allocation of a selected number of pooled storage devices ("at least two") to serve that request, under a central manager. This is the "translate the desired bandwidth … into the number of tape devices" step of the '526.

2.3 US 4,945,429 A — Munro et al. (Storage Technology Corp., "Data security arrangement for an automated magnetic tape cartridge library system") — https://patents.google.com/patent/[US4945429A](/patent/US4945429A)/en

Same assignee as the '526. Discloses the tape-library and security half of the invention:

• An automated magnetic tape cartridge library system of interconnectable library modules with a robot and tape drive units — the ACS/ACSC of the '526.
• "The control and data signaling paths are completely independent of each other" and the library modules "are not electrically connected to the tape drive units or the data retrieval bus" — the separate control network (CN) and high-speed data network (FN) of '526 claims 3–5, and the "network security can be implemented" passage.
• Library software in the host translates data-file identity into tape-cartridge identity, "the library module is unaware of the identity of the data stored … and the host computer is unaware of the identity of the magnetic tape cartridge" — the '526's central mapping/migration function ("cartridge movement and data migration").
• The whole reference is a data-security arrangement: preventing unauthorized access by separating control from data paths and by making the media inaccessible to the data path.

2.4 Record references US 5,214,768 (Martin et al.) and US 5,506,986 (Healy)

Categorised "Y" by the ISA against claims 1–14 and 1–5/10–14 respectively, with the ISA pointing to Martin cols. 5–6 and Healy col. 3 line 49 – col. 5 line 38. I did not retrieve these documents, so I do not assert their contents. They are best treated as cumulative, and as evidence that the examiner/ISA considered the striping/parallel-transfer and media-management features (relevant to claims 6, 7, 10, 14, 19, 20) to be within the ordinary skill level.

3. The §103 framework applied

Under 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 rationales catalogued at MPEP §§2143, 2144, the combination is proper if the references are analogous art, each element is disclosed or suggested, and there is an articulated reason (explicit or from the knowledge and common sense of a POSITA) to combine with a reasonable expectation of success. The available rationales here are:

• (A) Combination of known elements according to known methods yielding predictable results — Irwin (networked pooled storage with a central logical→physical map) + Kindell (performance-driven allocation of a selected number of pooled devices under a central manager) + Munro (tape library with split control/data paths and library-level mapping/security) are all components of a data-storage system; their aggregation is the routine design of a networked storage subsystem.
• (B) Known technique improving a similar device in the same way and (C) application of a known technique to a known device ready for improvement — Irwin expressly contemplates the tape-media backend (Irwin claim 17) as one of its "data storage subsystems," so the improvement of substituting/adding a robotic tape library per Munro requires no change in the principle of operation.
• (D) "Obvious to try" over a finite number of identified, predictable solutions (KSR) — the '526 itself lists the candidate configurations as RAIT 0, 1, 3, 4, 5, and Irwin already names RAID-3 and RAID-5. Choosing a stripe/parity organization to hit a requested data rate or fault-tolerance level is a finite, predictable design choice.
• (E) Express admission in the specification — the '526 states RAIT is arranged "in a manner analogous to the Redundant Array of Independent Disks (RAID) which is a well known architecture in the disk device technology," and that Fibre Channel is "a bandwidth scalable system" with switched point-to-point fabric. These are applicant admissions of known art (MPEP §2129) usable against the claims.
• (F) Simultaneous advantages (In re Merck): separate control and data paths (Munro) simultaneously improve security and offload the data fabric (Kindell/Irwin), so there is no teaching away.

4. Proposed grounds of rejection

Ground 1 — Irwin in view of Kindell → claims 1, 3, 6–9, 10–13, 17, 18, 19, 20 (at least)

Motivation: both references address the same problem (serving multiple processors from shared storage over a network while meeting a per-request performance level); Irwin's mapping/server layer is precisely the "means for creating" and Kindell supplies only the allocation arithmetic and central control that Irwin's system lacks. No reference teaches away; the result (a host sees one logical device whose delivered capacity/rate equals the sum of the allocated devices) is the predictable consequence of the two disclosed functions.

Ground 2 — Irwin + Kindell + Munro → claims 1, 2, 3–5, 14–17, 18, 20

Adds the tape library, split control/data network, and security limitations:

• Claim 2 (plurality of tape drives and tape media manipulated by ≥1 automated tape cartridge library system operatively coupled to the creating means): Munro (automated cartridge library, robot, tape drives, library software translating file identity to cartridge identity) + Irwin claim 17 (tape media data storage subsystem) + Kindell's manager.
• Claims 3–5 (low-speed control network separate from high-speed data network; ethernet control; fibre channel data): Munro — "the control and data signaling paths are completely independent of each other" and the library "not electrically connected to the tape drive units or the data retrieval bus." Munro's control path is a low-rate library-command path; the data path is the channel-attached high-bandwidth path. Combining with Irwin's "high speed network or switched-channel" (Irwin cl. 19: "nonblocking network") and the '526's own admission that Fibre Channel is a switched, bandwidth-scalable fabric supplies claims 4–5.
• Claim 1's "security" characteristic and claims 14–17: Munro is entirely about data security by architectural separation and by keeping the library ignorant of data content; the '526's Security Management module and "authentication between the data processors and the NSM" is the same function; and "managing the plurality of storage devices as a pooled resource" (claim 17) is Munro's central library controller + Irwin's bitfile server.
• Motivation: Munro is the same assignee's tape-library art in the identical field; the '526's stated purpose is to cure the "fixed, static configuration" of SCSI-attached tape libraries, which presupposes the library of Munro as the starting point. A POSITA seeking to scale a Munro library would naturally apply Irwin/Kindell's networked pooled mapping — that is the very improvement the '526 describes as needed.

Ground 3 — Irwin + Kindell + Munro + a RAID/RAIT reference (e.g. Martin, US 5,214,768, per the ISA's "Y" citation; or the applicant's own admitted RAID art) → claims 1, 6, 10, 14, 18, 19, 20

• The fault-tolerance characteristic (claim 1(iii)), the striping/reconfiguration features (claims 18, 20 "second storage characteristic different from the first"), and the connection-block redirection of striped data (claims 6, 10, 14, 19) are supplied by RAID/RAIT striping with parity, which Irwin already names (RAID-3, RAID-5) and which the '526 admits is "well known."
• Motivation: KSR "finite number of identified, predictable solutions." Once the logical device is composed of ≥2 physical devices (Irwin/Kindell), varying parity/striping across a requestable set {single drive, RAIT 0, 1, 3, 4, 5} is the routine selectable design space; each yields the predictable fault-tolerance/rate trade-off.

Ground 4 (contingent) — add US 6,094,605 A ("Virtual automated cartridge system," Storage Technology Corp.) if the claims are not entitled to the 1997-03-12 date

Because the '526 is a CIP, claims that find §112 support only in the 1999-06-30 / 1999-10-20 disclosures (in particular the broad, tape-agnostic claims 1, 3–5 and 18) may take a 1999 effective filing date. On that assumption, US 6,094,605 A (filed 1998-07-06, granted 2000-07-25; listed in the family's citation record) would be available as §102(e) art for its "virtual cartridge/automated cartridge system" disclosure. Caveat: I have not retrieved its disclosure and do not assert what it teaches; this ground is flagged for verification.

5. Why a POSITA would have combined the references (the articulated rationale)

1. Same field, same problem. Irwin, Kindell and Munro are all in the field of data storage subsystems serving plural processors from a shared, centrally managed store; each is directed to the identical deficiency — the mismatch between fixed/dedicated storage configurations and widely varying per-request performance, capacity and access-security needs. KSR, 550 U.S. at 417 (predictable use of prior-art elements).
2. Complementary, non-overlapping teachings with no change in principle of operation. Irwin supplies network pooling + logical→physical mapping + plural backend types; Kindell supplies request-driven allocation of a number of pooled devices to a requested throughput; Munro supplies the tape library, its split control/data paths, its secure architecture and library-level media mapping. Each is "reasonably pertinent to the particular problem" (In re Bigio; In re Clay).
3. Design incentive expressly recognized in the references themselves. Kindell's rationale for pooling (eliminating the expensive dedicated server, allowing any station to reach any file) is the same rationale the '526 gives for pooling tape devices ("the use of dedicated tape devices is an expensive proposition … especially if the data access loads … are erratic").
4. The '526 itself concedes the components are known. RAID is "well known"; RAIT is "analogous"; Fibre Channel is a "bandwidth scalable system"; the virtualization layer "can be housed in separate physical boxes or located inside of DP1–DP4 (on a network interface-like card) or integrated into the network FN's hardware/software" — the claim-7/8/9/11/12/13/15/16 placements are explicit alternative locations, i.e. admitted design choices.
5. Reasonable expectation of success. There is nothing unpredictable in aggregating N physical devices behind a mapping controller to obtain N× bandwidth, or in adding parity, or in placing the translation logic in a host, a storage device, or the fabric. The outputs (capacity sum, rate sum, parity-based fault tolerance, access control) are the arithmetic and known-protocol consequences of the disclosed structures.
6. No teaching away. Nothing in Irwin (server-mediated) contradicts placing the translator in the connection block; Munro's independence of control and data paths is reinforced, not undermined, by Irwin's separate high-speed fabric.

6. Claim-by-claim summary of the strongest grounds

7. Counterarguments a patent owner would raise, and the response

1. "No single reference teaches a single aggregated virtual device made from two or more physical devices with a requested data rate." Response: Kindell expressly allocates "a portion of this maximum throughput value" of plural resources to satisfy a request, and Irwin's bitfile staging occupies plural backend devices under one logical image; the combination is the ordinary aggregation of the two. Also, the '526's own specification says the virtual device's rate is obtained by striping "across devices TD1 + TD2 + … +TD5 as needed to achieve the required data rate," which is mechanical load-splitting.
2. "The references are all in disk/remote-file or video-streaming fields, not tape." Response: analogous art — Irwin expressly claims a "tape media data storage subsystem," and Munro is tape; Kindell is drawn to the same problem (a network-served pooled store with per-request performance guarantees), and the '526 treats the storage medium as interchangeable ("the same system architecture could be applied to other removable storage device types … and … fixed media storage devices").
3. "Means-plus-function: the disclosed NSM software architecture (RA/RC/RM/ACSLS interaction) is not shown." Response: the claim's function — creating a virtual device of ≥2 physical devices, responsive to a request, with a requested characteristic — is performed by Irwin's bitfile server + Kindell's allocator; the particular division into RA/RC/RM modules is unclaimed (claims 6, 10, 14, 17 recite only "network storage manager means" generically).
4. "Dynamic reconfiguration on demand" (claims 18, 20) is the point of novelty. Response: this is the weakest redoubt, and it is still met by Kindell's allocate/reallocate/de-allocate on successive requests and by Irwin's mount/dismount-per-request staging; the '526 adds only the label "virtual device" to a mechanism whose constituents the references disclose. Post-KSR, a known mechanism applied to a known environment for its known advantage does not become non-obvious by renaming.
5. "The claims require the request itself to specify the characteristic." Response: met by Kindell (throughput requirement drives allocation) and by Irwin's named-file request driving backend selection; even if one insists on an explicit performance parameter, the '526's request form ("data transfer rate, data access time, data compression, and the like") is exactly the parameter Kindell's allocator consumes.

Where the patent is strongest: the combination of (a) tape libraries (not disk), (b) a separate low-speed control network with (c) an explicit "security" characteristic specified by the request and (d) placeability of the translator in the host, the device, or the fabric switch as pleaded alternatives. These are best attacked by Munro (control/data path independence + security) for (b)/(c) and by the specification's own admission for (d). I would expect a PTAB/§103 posture to succeed on claims 1, 3–7, 9–20 and to require the Munro combination for claim 2 and the fault-tolerance reading of claim 1.

8. Confidence and open items

• High confidence: the identity, dates and content of Irwin (US 5,566,331), Munro (US 4,945,429) and Kindell (US 5,630,067); the fact that the ISA examined the parent family and cited these as Y (obviousness) art; the effective-date analysis and the §102(b) status of that art.
• Not verified — do not rely without checking: the disclosures of US 5,214,768 (Martin) and US 5,506,986 (Healy) (cited by the ISA to claims 1–14 and 1–5/10–14 respectively; I only have the citations and column references), and US 6,094,605 (used contingently in Ground 4). I have deliberately not attributed any specific teaching to them.
• Structural point worth flagging to a decision-maker: because US 6,658,526 is expired (2017-03-12), §103 analysis serves reexamination/invalidity-libel or defensive purposes; the appropriate standard is a preponderance (or clear-and-convincing in litigation), not the examiner's "substantial new question of patentability."
• The "Prior Art" section of the Google Patents page as supplied does not contain a pre-priority art list; anyone proceeding to an actual paper should pull the file wrapper / IDS of US 09/421,916 and US 09/343,512 from USPTO Patent Center to confirm the exact art actually before the examiner, and pull the WO 1998/040810 A3 search report (linked above) for the ISA's categories.