US 5652908

I'll search for current information on this patent, including any litigation or docket activity.

US Patent 5,652,908 — Analyst Summary

Scope note: I searched for this specific number. The authoritative patent full text was supplied and is used below; I also searched for USPTO/CAFC activity. I found no CAFC 2026 docket entry, no IPR/PGR, and no district-court litigation naming US 5,652,908. Any statement below about 2026 activity is therefore "not found," not "does not exist." I also want to flag that searching "5652908" as a bare string returns false positives in other identifier spaces — e.g., OSTI ID 5652908 (which is actually US 4,835,670, a Honeywell fuel-burner control) and JP5652908B2 (a Japanese publication number). Those are not this patent.

Bibliographic Data (from the patent record)

Abstract (verbatim from the patent)

"A system for providing distributed control of a resource with centralized configuration management. A network of distributed workstations is provided for controlling a resource such as a large mainframe computer. Server workstations for applying control commands to the resource are attached to each resource, preferably through redundant connections. Remote client consoles are defined which may be connected to the server for control of that resource. Server and resource location and primary and fallback connection paths are maintained by a centralized control server. Upon client console request, the central control server causes the server workstation associated with a particular resource to establish a control session between the server and the client console. A record of each dynamic connection is maintained by the central control server. Upon failure of a client console, the server accesses central server fallback information and establishes a fallback connection with an operational client console. Failure of a server workstation causes the resource to establish a link with a backup server which is then able to recover all sessions with client consoles by accessing the dynamic configuration database of the central controller. Configuration data may be updated dynamically taking effect when the next request for establishment of a session is made. In an alternate embodiment, interprocess security is optionally implemented in control servers and hardware resource servers."

Plain-Language Overview of the Independent Claims

Claim 1 — Client-side failure recovery (method).
Directed to recovering when the console (remote client) drops off. The control server keeps "fallback configuration data" that names a substitute fallback client. A central control facility holds the administration programs and has direct access to the configuration data. The resource server itself detects that its session with the remote client has failed; it then asks the control server over the network for fallback configuration data; the control server looks it up and sends it back; the resource server uses it to establish a new session with the fallback client.

Claim 4 — Server-side failure recovery (method).
Directed to recovering when the server workstation fails. Fallback configuration data identifies a fallback resource server for the resource; the resource can be controlled by that fallback server if the primary fails. The resource itself detects the primary server's failure and notifies the fallback server. The fallback server then requests fallback configuration data from the control server, receives it, and initiates a new session with the client. (Note: the claim's preamble/body text also contains cut-and-paste "remote client and said fallback client" phrasing carried over from claim 1 — a drafting artifact worth reading literally rather than harmonizing.)

Claim 8 — Apparatus version of Claim 1.
Same client-failure scenario, expressed as a system: a control server with means for maintaining fallback configuration data; a central control facility with direct access to all configuration data; means at the resource server for detecting session failure; means for requesting fallback data; means at the control server for accessing/transmitting it; and means at the resource server for establishing the new session with the fallback client.

Claim 9 — Apparatus version of Claim 4.
Same server-failure scenario, expressed as a system: means at the resource for detecting primary-server failure; means for notifying the fallback server; means at the fallback server for requesting fallback configuration data; means at the control server for supplying it; and means at the fallback server for initiating a new session with the client.

Dependent claims (for completeness): 2 and 5 add an acknowledgment from the newly connected client back to the control server. 3 and 6 add updating the configuration data with the new session's information. Claim 7 adds that where the primary server held sessions with multiple clients, the fallback server re-establishes sessions with each of them.

Notable Technical Context

• Preferred embodiment: IBM 3090 mainframe as the controlled resource; IBM PS/2 workstations as servers (108/110/112) and client consoles (130/132); dual IBM Token Ring LANs for redundancy; OS/2 Remote Data Services for database updates; configuration database 162 holding "static" (resource locations, primary/fallback access paths) and "dynamic" (current session state) data.
• Cited prior art of record includes US 5,005,122 (Digital Equipment Corp., cited and distinguished in the background), US 5,157,663 (Novell), US 5,341,477 (DEC broker), and US 5,127,787/5,023,873/5,027,269 (IBM).
• Interprocess security (FIG. 10) — the frontend/backend key built from a shared-memory-segment address + data checksum + random number — is described in this patent's specification but is not claimed here; it is claimed in the sibling family member US 5,537,642. This is the main reason to read the two specs together.

Uncertainty / Limitations of This Analysis

• I could not access a paywalled litigation database; my litigation/docket conclusion rests on open web searching only. I found nothing for CAFC 2026 or any other forum.
• The RPX Corporation assignment (2009) is as listed on the patent record; I did not independently verify current chain of title.
• Because the patent's anticipated expiration was 2014-07-29 and its status is "Expired – Lifetime," new 2026 appellate activity would be unusual; appeals from long-pending litigation are possible in theory but I found no evidence of any.
• Claim scope observations above are plain-language paraphrases of the literal claim text, not a legal claim construction.

Sources: US5652908A on Google Patents; Justia patent listing showing "5652908 | July 29, 1997 | Douglas et al."; Patent Leaderboard co-inventor listing.

I'll research litigation involving US patent 5652908.

Let me search litigation databases specifically.

Let me search more targeted litigation databases.

Summary of Findings

Based on my searches of publicly available patent litigation sources, I was unable to identify any litigation involving US Patent No. 5,652,908. No district court actions, ITC investigations, Federal Circuit appeals, or PTAB proceedings asserting or otherwise involving this specific patent were found.

What I searched and what I found

Important caveats and context

1. No case identified means no plaintiff, defendant, jurisdiction, case number, filing date, or outcome can be reported. I will not fabricate a case or infer one from a similar-looking patent number. I specifically avoided results for the commonly confused numbers that appear in search noise, e.g.:

   • 9,566,290 — Anacor v. Mylan/Lupin (D. Del., 2018) pharmaceutical ANDA cases
   • 5652908 (OSTI ID) — a library record for Honeywell's US 4,835,670 fuel-burner patent (not a patent number)
   • JP5652908B2 — a Japanese switching power supply patent
   • 6,538,908 / 6,963,908 / 5,999,908 — unrelated "908" patents

2. Ownership is consistent with non-assertion. Per the face of the patent record, US 5,652,908 was assigned from International Business Machines Corporation to RPX Corporation on 2009-07-14. RPX is a defensive patent aggregator that acquires patents to take them off the assertion market; the absence of litigation is consistent with that defensive "catch-and-hold" model.

3. Family/related records show no litigation link. The patent family includes:

   • US 5,537,642 ("Method for authenticating messages passed between tasks")
   • EP 0537903 A2
   • JP 2566711 B2

   None of these surfaced as patents-in-suit in my searches.

4. Why this patent appears in search results. US 5,652,908 shows up almost entirely as prior art — it is cited against other patents/applications (e.g., Nortel Networks' WO 00/26806 and WO 00/26807 "Server Manager" PCT applications cite it as an "X" reference), and it is cited in the "Cited By" tables of later IBM, Micron, and Nomadix patents. Being cited as prior art is not litigation.

Confidence level

• High confidence that no litigation involving US 5,652,908 appears in the major free/public sources I could reach.
• Caveat: Free sources are incomplete. A definitive negative would require a paid subscription search of Docket Navigator, LexisNexis IP Litigation (Lex Machina), or PACER for the patent number, plus the USPTO Patent Center "Litigation" view and PTAB/ITC records. If you need an exhaustive negative or a freedom-to-operate/clearance opinion, I'd recommend running those databases directly. I cannot guarantee the absence of a case filed after my knowledge cutoff or one that never generated a public, indexed docket.

If you have a specific case number or a party name you believe is associated with this patent, provide it and I can search for that specifically.

I'll verify the structured "no proceedings" default with targeted web searches before finalizing.

Let me run additional targeted checks against PTAB dockets, aggregators, and the sibling patent before finalizing.

PTAB Proceedings Analysis — US 5,652,908 A

Patent: Method and apparatus for establishing communications sessions in a remote resource control environment
Inventors: Douglas et al. | Original assignee: International Business Machines Corp. | Current assignee: RPX Corporation (assignment recorded 2009‑07‑14)
Priority: 1991‑10‑02 | Granted: 1997‑07‑29 | Status: Expired – Lifetime (anticipated expiration 2014‑07‑29)
Claims: 9 (independent claims 1, 4, and 8; apparatus claims 8–9)
Source of record: US5652908A on Google Patents; USPTO Open Data Portal structured "PTAB proceedings on file" block.

Proceedings overview

Total AIA trial proceedings on file: zero (0). The USPTO ODP structured data returns no IPR, PGR, or CBM proceedings against US 5,652,908; I found no proceeding in web searches of PTAB dockets, aggregator portals (Unified Patents, RPX), or third‑party petition repositories, and no Federal Circuit appeal or PTAB Final Written Decision naming this patent. Accordingly: claims invalidated — none; claims sustained — none (never tested at the PTAB); settled — none; institution denied — none; active — none.

Bottom line for a defendant: the patent is not "hardened" by surviving PTAB attacks — it has simply never been attacked, and the reason is structural, not tactical. Two facts dominate any assertion analysis: (1) the patent expired on 2014‑07‑29, so there is no prospective infringement and any recovery is limited to pre‑expiration acts still inside 35 U.S.C. § 286's six‑year lookback; and (2) it is now held by RPX Corporation, a defensive patent aggregator that acquired it from IBM in 2009 — consistent with the absence of assertion activity and the absence of IPRs. The practical posture is therefore stale patent, defensive holder, not battle‑hardened patent.

No proceedings to report

Because the structured source lists none, there are no proceeding entries (IPR/PGR/CBM numbers) to detail. I am deliberately not supplying proceeding numbers, panels, grounds, or dispositions — none exist in the record, and inventing them would be fabrication. Blog and docket search hits referencing a "'908 patent" refer to different patents (e.g., Neo Wireless' communications patent, Google's keyfact‑retrieval patent) and are not US 5,652,908.

For corroboration, the canonical places to confirm the null result directly:

• USPTO Patent Trial and Appeal Board End‑to‑End (PTAB E2E) system: https://ptab.uspto.gov
• USPTO PTAB decisions repository: https://www.uspto.gov/patents/ptab/decisions
• USPTO Open Data Portal – PTAB API (the source for the structured block): https://developer.uspto.gov
• Patent file wrapper via Patent Center: https://patentcenter.uspto.gov

Strategic summary

Claim status: all of claims 1–9 are UNTESTED at the PTAB, none CANCELED, none adjudicated SUSTAINED. There is no narrowing from any AIA trial, and no estoppel‑generating Final Written Decision. Claim 1 (client‑session fallback recovery via a control server), claim 4 (fallback resource‑server recovery), and claims 8–9 (apparatus counterparts) stand exactly as issued in 1997 — for whatever that is worth given the 2014 expiration. The only prior art of record that has ever been cited against this family is prosecution/foreign‑search art, not IPR art: US 5,005,122 (Digital Equipment) and the IBM automated‑operations application Ser. No. 07/577,967 were addressed in the specification's Background, and the JPO/EPO cited US 5,652,908 as prior art against others (e.g., WO 00/26806, WO 2005/051580), not vice versa. Notably, the sibling in this family — US 5,537,642 ("Method for authenticating messages passed between tasks," the FIG. 10 security‑key invention) — likewise shows no IPR/PGR in the searches I ran.

Estoppel landscape: there is no § 315(e)(2) estoppel, and there is no prior petitioner. Because no IPR/PGR ever reached a Final Written Decision, no party is barred from raising § 102/§ 103 grounds — in the PTAB or in district court. If you were (counterfactually) facing an assertion, the entire prior‑art universe remains open, and this is a 1991‑priority patent whose claims recite generic client/server/fallback‑session management using ordinary workstations (IBM PS/2) over a token‑ring LAN — a subject matter where the 1991 art is dense and the claims would face a serious § 103 fight. The FIG. 10 authentication subject matter is separately claimed in US 5,537,642, so watch for family‑level assertion rather than a single‑patent campaign.

Pattern signals: none of the usual ones. No repeat petitioner (there is no petitioner at all); no patent‑owner appeal activity; no Unified Patents or other defensive aggregator IPR in the chain. The one aggregator signal is ownership: RPX Corporation took assignment from IBM on 2009‑07‑14, five years before the patent's 2014 expiration — a classic defensive‑aggregation move to warehouse the patent rather than assert it. Combined with the expired legal status, this is a patent with essentially no live enforcement footprint.

Recommended next steps

1. Confirm the null result yourself, in the primary sources. Pull the patent's trial‑history tab on PTAB E2E (https://ptab.uspto.gov) and the PTAB decisions search (https://www.uspto.gov/patents/ptab/decisions). A clean result there, matched to the ODP structured block, is your defensible record.

2. Do not treat a demand letter citing this patent as an IPR candidate. There is nothing to invalidate that isn't already dead: the patent expired 2014‑07‑29. Any damages theory is capped by 35 U.S.C. § 286 (six years pre‑complaint) and, more fundamentally, by the expiration date — post‑2014 conduct cannot infringe. An IPR on an expired patent is generally a poor investment (no prospective injunctive relief to defeat, and claims cannot be amended to survive).

3. If (hypothetically) an assertion surfaces, attack standing/ownership first. Ask for the chain of title from IBM to RPX (recorded 2009‑07‑14) and confirm the plaintiff actually holds enforceable rights. Also check whether the same campaign bundles US 5,537,642 and the family members EP 0 537 903 A2 (withdrawn) and JP 2566711 B2 (expired/fee‑related) — those statuses independently limit foreign exposure.

4. Because no proceeding exists, there is no FWD to quote and no appeal docket to pull. Anyone who tells you this patent "survived two IPRs" or was "narrowed by the Board" is describing a different patent. The correct, defensible statement is: no PTAB activity on file; patent expired 2014; currently held by a defensive aggregator — an absence that, for this particular patent, is best explained by expiration and defensive ownership rather than by the strength of the claims.

Caveat on completeness: ODP ingest can lag, and older or recently filed papers may not be indexed. My searches surfaced no such late‑filed proceeding, but if you need a litigation‑grade record, run a native PTAB E2E party‑name and patent‑number search plus a docket sweep on the day you file your response.

I'll research the assignment record for US 5,652,908. Let me search multiple sources.

Ownership & Assignment Reconstruction — US 5,652,908

Sourcing limitation up front (read before relying on the chain): In this session I could not complete a direct query of the USPTO Patent Assignment Center (assignmentcenter.uspto.gov / assignment.uspto.gov) — the tool calls ran out before the record detail returned. My free-source searches did not surface a reel/frame, execution date, or correspondent for any link in this chain. Accordingly, I report no reel/frame numbers, because I do not have them, and I will not invent them. The one post-issuance transfer below is documented on the patent record itself (Google Patents legal events) and is corroborated by the "Current Assignee" field, but its reel/frame and correspondent are unretrieved. Treat the timeline as provenance-documented, reel-verified: no.

Inventors

Seven named inventors on the face of the patent:

• Priority/filing era: Inventive work predates the 1991-10-02 priority date. All seven appear in IBM-inventor aggregations, consistent with a single-employer (IBM) team on an IBM product (the IBM 3090 / PS/2 / OS/2 resource-control environment described in the spec).
• Unusual-pattern check — departures within 12 months: No evidence found. I could not retrieve employment-departure data, and no inventor was found to have exited IBM in the 1991–1992 window in the sources I reached. This is a "not found," not a negative finding. Notably, the inventors' post-filing careers are not what moved this patent — the portfolio transfer happened 18 years later at the corporate level (see below), so this patent does not fit the "inventor departure precedes fire-sale" precursor pattern.

Original assignee

International Business Machines Corporation (Armonk, NY) — named on the issued patent (issued 1997-07-29).

• Primary line of business: mainframe computer systems, software, and services. IBM is an operating company and a top-tier filer.
• Did IBM ship a product embodying the claims? The preferred embodiment is squarely a shipped IBM product family: the IBM 3090 mainframe as the controlled resource, IBM PS/2 workstations as servers and consoles, dual IBM Token Ring LANs, and OS/2 Remote Data Services for configuration-database updates. The spec describes a real IBM "central control facility (CAA)" resource-control system. On that basis the claims read on IBM's own operations-automation product line — i.e., yes, an operating-company embodiment is described. (Caveat: I did not locate a specific IBM product announcement naming this feature, so this rests on the specification's own product identifications.)
• Current status: IBM remains an operating public company (ticker IBM). No bankruptcy, merger, or dissolution affecting this patent.
• Note on the original inventor→IBM assignment: IBM employee-inventor assignments are standard practice, so an inventor-to-IBM assignment for application 07/771,064 almost certainly exists in the Assignment Center, but I did not retrieve it, and I will not state a reel/frame for it.

Assignment timeline

Chronological, every documented post-issuance event I could reach:

• Executed: not retrieved / recorded 2009-07-14 — Reel ___/____ (reel/frame not obtained; do not cite a number)
  • Conveyance: Assignment of assignors' interest ("ASSIGNMENT OF ASSIGNORS INTEREST (SEE DOCUMENT FOR DETAILS)").
  • Assignor: International Business Machines Corporation.
  • Assignee: RPX Corporation.
  • Correspondent: not retrieved. (This is the field you most wanted — I could not capture it, so I make no recurrence claim.)
  • Context: Defensive aggregation — the patent was moved from an operating company (IBM) into a defensive patent aggregator's catch-and-hold portfolio; RPX's model is to take patents off the assertion market for subscribers, so this is the inverse of a transfer-to-asserter.

Prior links (pre-issuance, unretrieved):

• Inventor(s) → International Business Machines Corporation, recorded at/near filing of 07/771,064 or at issue of 08/474,528 — reel/frame not retrieved.
• Continuity chain (not assignments, but relevant to title): 07/771,064 (filed 1991-10-02, abandoned) → 08/222,496 (filed 1994-04-04, abandoned) → 08/474,528 (filed 1995-06-07, issued as 5,652,908). Continuations do not create assignment records but do affect which record carries title.

Bottom line for this section: the Assignment Center does have at least one post-issuance record for this patent (the 2009 RPX record is reflected on the patent face and in Google Patents legal events), so I do not stop here — but I could not pull reel/frame or correspondent, so the chain is not reel-verified in this report.

Timeline diagram

timeline
    title Ownership of US 5652908
    1991 : Priority filing by IBM
    1994 : First continuation filed
    1995 : Second continuation filed
    1997 : Patent issued to IBM
    2009 : Assigned to RPX Corporation
         : Chain ends at defensive aggregator

NPE / troll-pattern signals

1. Shell-entity transfer — not present. The 2009-07-14 transfer ran from an operating company (IBM) to a defensive aggregator (RPX), not to a licensing-only shell. No "IP / Holdings / Ventures" LLC appears, no registered-agent address is documented, and RPX is a known corporate entity with an operating insurance/services business — not a single-purpose Delaware/Texas LLC. Insufficient record detail (no reel/frame) to go further.

2. Known asserter in the chain — not present. Neither link matches a public NPE list: IBM is an operating company; RPX appears on the anti-assertion side (RPX built a defensive subscription business and a 2014 NPE-insurance service, and acquired PatentFreedom, an NPE-intelligence service). No Acacia / Marathon / IV / Wi-LAN / Mosaid / Vringo / Pendrell / Round Rock / Spangenberg entity appears. (Prior-family note: sibling US 5,537,642 shows the same IBM origin; its post-issuance chain was not retrieved, so this call covers 5,652,908 only.)

3. Repeat correspondent across the chain — unclear. I retrieved no correspondent data at all this session. I therefore make no recurrence claim, and none should be inferred. This is the single biggest evidentiary gap, because correspondent recurrence is the field you flagged as most diagnostic.

4. Cascading transfers — not present. Only one recorded post-issuance transfer is documented (2009-07-14 IBM→RPX). No chain of LLC-to-LLC hops, no <24-month cascade, no common-address cluster. (Also note the 18-year gap between issue in 1997 and transfer in 2009 — the opposite of a rapid cascade.)

5. Pre-litigation transfer — not present. The 2009-07-14 transfer is not dated within 6 months of any identified suit, and per the prior litigation section no infringement suit naming this patent was found. With no first suit to anchor against, the 6-month test cannot be satisfied on the available record.

6. Bankruptcy fire-sale — not present. IBM was solvent and operating in 2009; no Chapter 7/11 of the assignor is associated with this patent. Not a Kodak/Nortel/Polaroid-style estate sale.

7. Privateering — not present on the evidence. A privateering pattern requires the operating company to hand patents to an NPE that asserts on its behalf. Here the transferee is a defensive aggregator that neutralizes rather than asserts; no SEC-filing or press record of IBM-directed assertion against competitors was found. (Unclear only in the sense that RPX's later disposition of any individual asset is not publicly itemized.)

8. Defensive aggregator (anti-NPE) — present. The chain terminates at RPX Corporation per the 2009-07-14 assignment on the patent record. RPX is a defensive patent aggregator whose stated purpose is to reduce NPE litigation risk for members. This is the classic "patent has been neutralized" endpoint.

Verdict

Defensive / non-asserting.

Justification: the only documented post-issuance link is the 2009-07-14 assignment from IBM to RPX Corporation "ASSIGNMENT OF ASSIGNORS INTEREST (SEE DOCUMENT FOR DETAILS)," i.e., the chain terminates at a recognized defensive aggregator whose business model is to take patents off the assertion market (signal 8 present); every assertion-facing signal is absent — no shell-entity transfer, no known NPE, no cascading LLC hops, no bankruptcy sale, and no first suit to anchor a pre-litigation-transfer call. The main weakness in this verdict is evidentiary: I could not retrieve reel/frame or correspondent from the Assignment Center, so the RPX link is document-sourced but not reel-verified, and the inventor→IBM link was not retrieved at all.

Verify at:

• USPTO Assignment Center — search by patent number 5652908: https://assignmentcenter.uspto.gov/
• Legacy Assignment Search (same underlying records): https://assignment.uspto.gov/patent/index.html
• Google Patents legal events for the 2009-07-14 RPX entry: https://patents.google.com/patent/[US5652908A](/patent/US5652908A)/en

To close the gaps I could not close here, pull from the Assignment Center: (a) the reel/frame and correspondent for the 2009-07-14 IBM→RPX record; (b) any inventor→IBM record with its execution date; and (c) the equivalent record for sibling US 5,537,642, since a shared correspondent/address across the two family members would be the one coupling most worth testing.

I'll research the USPTO record and the prior-art citations for US 5,652,908. Let me run several targeted searches.

Let me gather details on the remaining cited references.

Prior-Art Analysis — US Patent 5,652,908

Scope / method note: The authoritative full text of US 5,652,908 (supplied in the record above) lists 19 examiner-cited U.S. patent references. I treated those as the prior art to analyze, since they are the references "of record" against this patent. My live searches confirmed the identification and content of several (US 5,005,122, US 5,341,477, US 4,872,106, US 5,157,663). I hit my tool-call limit partway through and could not independently pull the full text of US 5,129,080, US 5,121,486, US 5,043,881, US 5,027,269, US 5,343,477, and US 5,023,873. For those six I rely on the title/assignee/date in the patent record plus the description in the specification's Background; I flag each such instance below. This is a technical anticipation screen, not a legal opinion or claim construction.

1. The prior-art date rule applied here (and a correction worth flagging)

Because US 5,652,908 is a continuation (08/474,528 → 08/222,496 → 07/771,064), the critical date is the earliest parent that supports the claims: 1991-10-02. For pre-AIA § 102:

• § 102(b) (statutory bar): reference must have issued/published before ~1990-10-02 (one year before the priority date).
• § 102(a) / § 102(e): a U.S. patent qualifies as of its issue date (if before the invention) or, more usefully here, as of its filing date (§ 102(e)), provided it was filed before 1991-10-02.

Every one of the 19 references was filed before 1991-10-02, so all 19 are available as § 102(e) art; those issued before 1990-10-02 are also § 102(b) art. Two references issued after the priority date (US 5,058,108 on 1991-10-15; US 5,341,477 on 1994-08-23) rely on § 102(e) via their earlier filing dates.

Correction to a prior section: The earlier "Patent summary" and "Litigation summary" flagged OSTI ID 5652908 / JP5652908B2 as false positives, which is correct and should be preserved. I found no new litigation and continue to treat the 2026 docket/litigation result as "not found," not "nonexistent."

2. The 19 cited references — full citations and § 102 screen

Additional references cited inside the specification (not in the numeric "References Cited" table):

• Co-pending IBM application Ser. No. 07/577,967, filed 1990-09-04 (automated operations controller with remote workstation access) — cited in the Background and a candidate § 102(e) reference; commonly owned.
• Foreign family citations (relevant mainly to the security sibling US 5,537,642): SE 426128 B, US 4,423,287, US 6,649,233/… (IBM composite session keys), JP S63-56738 A, NL 8702012 A, and others — these appear in the "Family Cites Families" table and bear on authentication/session-key concepts, not on this patent's claims.

3. Bottom line on anticipation

No single one of the 19 references appears to anticipate any of claims 1–9 in full. Each independent claim requires, in combination:

• (Claims 1 & 8 — client failure) a control server that maintains fallback configuration data naming a fallback client; the resource server itself detects the session failure and requests that fallback data over the network; the control server returns it; and the resource server establishes a new session with the fallback client.
• (Claims 4 & 9 — server failure) the resource detects primary-server failure and notifies a fallback server; the fallback server requests fallback configuration data from the control server; and then initiates a new session with the client.
• Dependent 2/3/5/6 add the fallback-client acknowledgment to the control server and the dynamic-data update; claim 7 extends re-establishment to multiple clients.

The references that come closest each miss at least one essential element:

• US 5,341,477 and US 5,005,122 (both DEC) disclose a central mediator that picks/directs servers — but not a resource server that detects its own session failure and not distribution of fallback-client configuration data.
• US 4,872,106, US 5,157,663, US 5,129,080 disclose primary/backup failover — but with no intervening central control server and no client-session re-establishment.
• US 5,121,486 and US 5,043,881 disclose dynamic session/connection control — but not the fallback-configuration-driven recovery of claims 1/8.

That is consistent with the examiner having cited these references, then allowing claims 1, 4, 8, and 9 — i.e., the references were treated as background/§ 103-context rather than § 102 anticipatory art. If forced to rank them as the most relevant prior art for a § 102/§ 103 challenge, my order would be:

1. US 5,341,477 (broker/central selection + failover suggestion)
2. US 5,005,122 (central management server directing service servers)
3. US 5,121,486 (dynamic logical-connection switching)
4. US 5,157,663 (server fault tolerance / failover)
5. US 4,872,106 (resource-side backup takes over from failed primary)
6. US 5,023,873 / US 5,343,477 / US 5,027,269 (link/transmission-failure recovery and availability)

4. Caveats and where this analysis is weak

• Six references (US 5,129,080, US 5,121,486, US 5,043,881, US 5,027,269, US 5,343,477, US 5,023,873) were analyzed from title/date metadata only because I reached the tool-call limit before retrieving their full text. My claim-mapping for those six is provisional and could over- or under-state relevance — especially US 5,121,486 and US 5,343,477, which by title are the most likely to contain session/failure-recovery detail.
• § 102 vs. § 103: Because every claim is a combination, the realistic attack vector for these references is § 103 obviousness (e.g., DEC's broker/management-server teachings + Novell's server failover), not § 102 anticipation. I flag that because the question asked specifically about § 102, and the honest § 102 answer is largely negative.
• Who cited whom: The record's "Patent Citations" table marks entries with an asterisk ("cited by examiner"). I could not verify from the live USPTO page which of the 19 were examiner-cited vs. applicant-listed; the distinction matters for prosecution-history weight but not for whether the art is "prior art."
• Verification of the number: The supplied full text and Google Patents record confirm US 5,652,908 with application 08/474,528, priority 1991-10-02, issued 1997-07-29, IBM → RPX (2009), status Expired–Lifetime (anticipated expiration 2014-07-29). No similar-number result (e.g., OSTI ID 5652908, JP5652908B2) was substituted.

Sources used: the supplied authoritative full text of US5652908A; Google Patents US5005122A; Google Patents US5341477A; Google Patents US4872106A; Google Patents US5157663A; unifiedpatents.com US-5157663-A; everypatent.com listing for US 5,005,122.

I'll research the prior-art references of record to ground the obviousness analysis.

Let me verify the remaining key references.

Obviousness Analysis — US 5,652,908 under 35 U.S.C. § 103

Note on dates: The task header states "Current Date: April 26, 2026," while my operating environment says 2026-09-24. I flag the discrepancy rather than silently harmonize it. It does not affect a § 103 analysis of a patent that expired 2014-07-29.

Note on scope: Builds on the prior summary and litigation summary. The two pivotal facts from those sections carry forward: (i) the patent's effective priority is 1991-10-02, and (ii) the FIG. 10 interprocess-security subject matter is not claimed here — it is claimed in sibling US 5,537,642. Accordingly, § 103 is assessed only against the four independent claims (1, 4, 8, 9) and their dependents, and only against the "Prior Art section" of record.

1. Governing framework

Date-qualification caveat (important)

Four cited references issued after the 1991-10-02 critical date and therefore qualify only under pre-AIA § 102(e) if their underlying U.S. filing dates precede it:

• US 5,121,486 (Hitachi) — issued 1992-06-09; continuation chain back to a 1987-11-20 priority (confirmed in search results). Qualifies.
• US 5,341,477 (DEC) — issued 1994-08-23; continuation of Ser. No. 07/314,853, filed 1989-02-24, priority 1989-02-24 (confirmed, incl. EP 0384339 priority 1989-02-24). Qualifies.
• US 5,129,080 (IBM) — issued 1992-07-07; per the record, filed 1990-10-17. Qualifies (title/date only; I did not retrieve full text — searches were cut off).
• US 5,157,663 (Novell) — issued 1992-10-20; filing date not independently verified in my searches. Presumed to qualify on the record's own citation, but flag as unverified.

All remaining cited references issued 1991 or earlier and are plainly prior art.

2. Element-by-element mapping of the independent claims

Claim 1 / Claim 8 — client-console failure recovery (method / apparatus)

Claim 4 / Claim 9 — resource-server failure recovery (method / apparatus)

Dependents

• Claims 2, 5 (acknowledgment to the control server): US 5,005,122 — the network service node "generate[s] said status message identifying the status of an operation" and transmits it over the link, and the management server "maintain[s] the status of said network service in response thereto." Direct correspondence.
• Claims 3, 6 (update configuration data with the new session): US 5,005,122 (status maintained centrally; task schedule records "the last time each said task was performed"); US 5,023,873 (dynamic configuration database constituting a live network map); US 5,341,477 (broker datastructures/preview window updated from node status messages).
• Claim 7 (multiple clients; fallback restores each): US 5,121,486 — the CCP identifies each terminal logically connected to the failed processing unit and re-establishes connections for the devices, holding "histories of command data transactions corresponding to the respective terminal devices" (plural); US 5,005,122 (plural client nodes); US 5,027,269 (IBM) — title "providing continuous availability of applications in a computer network" (content not independently verified).

3. The combinations, and why PHOSITA would have made them

Ground A — Claims 1–3, 8 obvious over US 5,023,873 + US 5,005,122 + US 5,341,477

Rationale (KSR-valid): applying a known centralized-directory/failure-recovery technique to a known distributed session architecture to yield a predictable result.

1. US 5,023,873 supplies the failure-detection + central-configuration-database core: a communication link manager holds a dynamic physical configuration data base including "spare resources that are available for automatic recovery," and on a failure indication it "accesses these data files for an identified target node" and applies "the appropriate recovery procedure" without terminating sessions. That is the heart of claim 1.
2. US 5,005,122 supplies the central control facility that the claim requires: a management server node with centrally maintained status, task schedules, and master/slave redundancy, which directs other nodes to perform service and receives status back. It supplies claims 2 and 3's acknowledgment-and-record steps almost verbatim.
3. US 5,341,477 supplies the fallback-identity feature: a central broker that holds a data structure of server entries and deliberately provides a second, fallback identity so a connection can be re-routed on failure "without the need to recontact the broker."

Motivation. All three are network-management/fault-tolerance references; all three state the same objective the patent's own Background identifies as the unmet need — "recovery from the failure of any component," "control system component failure must not cause the failure of control of the larger system." A PHOSITA reading 5,023,873 (recovery at the layer of detection, using a central config DB) would naturally look to 5,005,122's management-server model for who holds the directory, and to 5,341,477 for how a substitute endpoint is designated. There is no new cooperation of elements; it is arrangement of known building blocks with predictable results. The examiner already cited all three, confirming they are in the field of search.

Ground B — Claims 4–7, 9 obvious over US 4,872,106 + US 5,121,486 + US 5,341,477 (or + US 5,023,873)

Rationale: substitution of one known failover architecture for another in a resource-control setting, plus use of a known central directory to supply the replacement identity.

1. US 4,872,106 discloses the server-failure half of the invention with striking specificity: the controlled station itself "detect[s]" primary-processor failure, generates a failover signal, and switches the back-up processor into active control. That is exactly claim 4's "detecting from said resource a failure in said primary resource server" and "transmitting a notification … from said resource to said fallback resource server." Note also that the patent's own specification expressly extends the resource to "a production assembly line, chemical processing facility," i.e., the industrial-process-control subject matter of 4,872,106 — which strengthens the field-of-endeavor link.
2. US 5,121,486 discloses the session-recovery half: a central facility (communication control processor) storing (a) per-terminal command histories and (b) replacement processing-unit identities, then "automatically establishing a logical connection between a replacing processing unit … and the identified terminal device" and delivering the histories to the replacement. That covers the claim's "fallback resource server … establishes new connections with each client console previously connected" and claim 7's multi-client restoration.
3. US 5,341,477 / US 5,023,873 supply the "request fallback data from a central server" leg — the broker/manager is queried for the identity of the entity to connect to.

Motivation. Both 4,872,106 and 5,121,486 are automated-takeover systems in which a backup controller is substituted for a failed controller without operator intervention; both predate the critical date; both are cited of record. A PHOSITA asked to give a mainframe console server a fallback path would combine 4,872,106's "controlled entity detects and switches" with 5,121,486's "central table of replacement units + preserved session history," and would add a broker/manager (5,341,477/5,023,873) to centralize the lookup. Predictable result, no new function.

Ground C — Corroborating single-reference pressure on claims 4/9

US 5,121,486 alone discloses nearly every element of claim 4 except the inversion of the request direction (a fallback server querying a central control server for fallback data). Its CCP-centric design makes it more naturally a § 103 reference than a § 102 reference — but it is the closest single-art mapping in the record, and it is the strongest candidate for a "one reference plus the knowledge of PHOSITA" attack under KSR. The remaining gap (the resource-server-initiated request to a separate control server) is filled by 5,023,873 and 5,341,477, which are demonstrably directed to exactly that query-and-redirect step.

Corroborating references for redundancy

• US 4,780,821 (IBM) — server computer + plural remote computers with multiple-program management: independent support for claim 1's "plurality of clients" / console architecture.
• US 5,043,881 / 5,048,081 (Hitachi) — "session control method for rewriting information in a session control device in an interactive processing system": independent support for centrally rewriting session state on failure.
• US 5,027,269 (IBM, "continuous availability of applications in a computer network") and US 5,129,080 (IBM, "increasing the operational availability of a system of computer programs operating in a distributed system") — titles and dates indicate squarely relevant session-recovery disclosures. I could not retrieve their full texts (search budget exhausted); they are candidates that a full analysis should chart, not confirmed § 103 anchors.

4. Common motivation narrative (applies to all grounds)

Under KSR, these are the recognized rationales that hold here:

• Known technique, same problem. Every reference addresses failure recovery in a networked multi-node system. The patent's Background concedes the three unresolved needs: (1) locate the resource and link it to a console; (2) recover from failure of any control-system component; (3) allow configuration changes without disruption. The references collectively answer all three.
• Predictable arrangement. Placing a directory/configuration database in a central management node and having endpoints query it on failure was a routine architecture by 1991 (see 5,005,122; 5,023,873; 5,341,477).
• Design incentive — "not terminating sessions." 5,023,873 articulates the very design goal the patent asserts, i.e., that recovery should preserve a user's session rather than force a new login. That is a motivation statement supplied by the art itself.
• The patent's own admissions. The Background disparages remote-console approaches only for lack of redundancy and reconfigurability, not for the session-establishment mechanics — which is an admission that the architecture itself was known.

5. Where the claims could survive (rebuttal points the patentee would press)

I want to be balanced rather than one-sided:

1. The request-direction inversion (claims 1, 4, 8, 9). In 5,121,486 and 5,341,477 the client or the communication control processor initiates; in the patent the resource server (or the resource) detects and initiates the query to the control server. A patentee will argue this "server-originated, network-routed fallback request" is the point of novelty. Counter: 5,023,873 expressly performs recovery "at the layer where the failure is detected," which points to the detector initiating — cutting against this distinction.
2. Fallback client (console) rather than fallback server (claims 1, 8). The cited failover art is overwhelmingly server/processor-oriented; a fallback operator console is a more distinctive choice. A patentee could argue the art does not suggest substituting a different console as the fallback endpoint. Counter: 5,005,122 centrally enables service "with respect to a client node" among plural client nodes, which is the mechanism.
3. No secondary considerations of record. From the prior sections: no litigation, no licensing evidence, no asserted commercial-success or long-felt-need nexus I could locate. The 2009 RPX acquisition is a defensive-aggregation fact and is not probative of non-obviousness either way. Absent a nexus, objective indicia do not rescue the claims.
4. The examiner's allowance cuts both ways. All of these references were of record and the patent still issued — which the patentee will cite as evidence of non-obviousness, but which is weak given that the examiner's own citation list shows the references were considered.

6. Explicit uncertainties and contradictions

• Claim 4/9 drafting artifact (carried over from the prior summary, and it matters here). Claim 4's body recites "said remote client and said fallback client" although the preamble names only "a client" — a copy-through from claim 1. Under a literal reading, claim 4 may require both a remote client and a fallback client to be network-connected to the "central control facility." That makes the § 103 mapping ambiguous and is separately a § 112(b) indefiniteness problem. My mapping above reads the limitation as "the client(s) already in the system," consistent with the specification, but a court construing it literally would narrow the claim and strengthen the patentee's position.
• "Central control facility" vs. "control server." Both appear in every independent claim, apparently as the same entity (CAA 160). If construed as two distinct structures, the § 103 mapping needs an additional reference; if construed as one, the mapping above holds.
• Verification status. Directly confirmed by search: US 5,341,477, US 5,121,486, US 4,872,106, US 5,005,122, US 5,023,873. Relied on title/date from the patent record only: US 5,157,663, US 5,027,269, US 5,129,080, US 4,780,821, US 5,043,881/5,048,081. My searches for US 5,027,269, US 5,129,080, US 5,157,663, and US 5,023,873 content were truncated by the tool step limit, so I state their relevance with moderate confidence, not high.
• Commonly owned art is off the § 103 table. The Background's Ser. No. 07/577,967 (IBM, filed 1990-09-04) and the sibling US 5,537,642 (same 1991-10-02 priority, same family) cannot be combined for obviousness against these claims under pre-AIA § 103(c) (common ownership/obligation), even though 5,537,642 discloses the FIG. 10 security subject matter. Do not use them as § 103 references.
• This is an analyst opinion, not a legal conclusion. No claim-construction ruling exists; the patent expired 2014-07-29, so any live § 103 dispute would have to arise collaterally (e.g., validity as prior art against a later patent).

7. Bottom line

On the record as cited, claims 1–3 and 8 are obvious over US 5,023,873 in view of US 5,005,122 and US 5,341,477, and claims 4–7 and 9 are obvious over US 4,872,106 in view of US 5,121,486 and US 5,341,477 (alternatively US 5,023,873). In each case the combination merely arranges known network-management and failover building blocks — centralized configuration/status management, a replacement-endpoint directory, and automatic session re-establishment without user logout — to achieve the predictable result of surviving a component failure, which is precisely the stated aim of the references themselves. The only features with genuine non-obviousness potential are (a) the server-originated fallback query routed over the network to a control server, and (b) the use of a fallback console rather than a fallback server; both are partially, though not completely, met by US 5,023,873's "recover at the layer of detection" teaching and US 5,005,122's client-node-directed service model.

Sources: US5652908A — Google Patents; US 5,341,477 — Google Patents; US 5,341,477 PDF; US5121486 — FreePatentsOnline; US 4,872,106 — Google Patents; US 4,872,106 PDF; US5005122 PDF; US 5,005,122 — FreePatentsOnline; US5023873 PDF.