US 5652916

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

Let me check for litigation/docket activity and confirm the full claim set.

Let me confirm the remaining independent claims (the buffer management and file system claims).

US Patent 5,652,916 — Summary

Important scope note: My tools are web-search based, not direct portal access. I could not query USPTO PatentCenter/PAIR, PACER, or the CAFC docket directly. What follows is grounded in the authoritative full text provided plus corroborating public records (Google Patents, Justia, EveryPatent, and the USPTO-printed PDF). I found no evidence of any litigation, PTAB/IPR, or CAFC 2026 docket activity involving this patent — but I cannot certify a negative from web search alone, so treat that as "not found" rather than "confirmed none."

Bibliographic data (confirmed)

Abstract (as issued)

A prestaging method for a computer system having a CPU, buffer, and secondary storage, where input data of a submitted job is read into the buffer from secondary storage before job execution. The method: (a) obtains the number of jobs whose prestagings will be carried out simultaneously with the submitted job's prestaging; (b) obtains an effective data transfer speed based on that number and on the transfer speed of the secondary storage holding the job's input data; and (c) sets the prestaging start time by subtracting an effective data transfer time (total input data ÷ effective transfer speed) from the scheduled job start time.

Plain-language overview of the independent claims

The patent claims two families: prestaging timing and buffer block management.

Claim 1 — Prestaging start time from concurrent-job count (the flagship claim). Count how many jobs will be prestaged at the same time; derive a transfer speed from that count and the disk's transfer speed; set start time = job start time − (total input data ÷ that speed). (Note: the granted claim recites "operative data transfer speed," while the specification uses "effective data transfer speed.")

Claim 10 — Prestaging start time from measured current load. Detect the actual load on the storage→buffer transfer path when the job is submitted; effective speed = disk transfer speed − measured load; start time = job start − (input data ÷ effective speed).

Claim 12 — Prestaging start time from predicted/calculated load. Calculate the load on the transfer path at the time prestaging begins; compute speed by subtracting that load; set the start time accordingly. (This is the "predict the load" variant, e.g., by summing transfer speeds of overlapping prestagings.)

Claim 14 — Prestaging start time from a reserved transfer path. Reserve/secure the transfer path needed for prestaging in advance; derive a guaranteed "reserved" transfer speed from the secured path; set start time = job start − (input data ÷ reserved speed). This turns an estimate into a guaranteed bandwidth.

Claim 16 — Retroactive correction of prior jobs. After setting a prestaging start time for a newly submitted job, correct the prestaging start time of a previously submitted job to reflect the new job's impact on transfer-path load — i.e., reschedule earlier jobs rather than statically fixing their start times.

Claim 18 — Re-prestaging after early completion / data eviction. Perform prestaging; check whether it finished on or before a predetermined reference time; if so, prestage again to re-read the data into the buffer. This guards against the classic failure mode where prestaged data is evicted by another job before the job runs.

Claim 20 (numbering not fully verified) — Buffer management method. Manage buffer regions in blocks using at least three free lists holding block information for blocks subject to takeover: a "used list" (low likelihood of reuse), an "unused list" (unused blocks), and an "in-use list" (high likelihood of reuse). When a block's data transfer ends, release it by queuing its block information into the free list selected according to buffer state (e.g., type and amount of data transfer).

Claim 21 (numbering not fully verified) — File system apparatus. A file system implementing the same three-free-list buffer management described in Claim 20, expressed as structural means rather than method steps.

Dependent claims 2–9, 11, 13, 15, 17, 19 add: using main storage as the buffer; file-position-based queuing into the unused list (leading blocks queued toward the tail so they survive longer); completion-based queuing (used list vs. in-use list); distinct handling of invalidated blocks on file deletion; searching free lists in the order used → unused → in-use; and map-information tracking of whether an entire block's I/O has completed.

Caveats and uncertainties

1. Claim-set numbering for the buffer/file-system claims is not certain. I directly verified independent claims 1, 10, 12, 14, 16, and 18. The buffer management and file-system claims appear in the claim set as the final independents (likely 20 and 21), but my sources truncated the list at claim 19, so I flag their exact numbers as unverified.
2. Apparent transcription inconsistency in Claim 3. The retrieved claim text describes the "unused list" as including block information "having a high possibility of being used again" — which conflicts with the specification (unused = not yet used; in-use = high reuse likelihood). This is almost certainly an OCR/transcription artifact in the source, not the operative claim language, but I note it rather than silently correcting it.
3. Legal status. "Expired – Fee Related" with anticipated expiration in 2013 is Google Patents' stated status and is an assumption, not a legal conclusion. Given expiration, no live enforcement activity would be expected.
4. Priority date. Google Patents lists 1992-06-29 as priority; the U.S. filing chain is the 1993-06-16 parent application (08/077,056) and the 1995-12-21 continuation (08/576,402). The 1992 date presumably reflects a Japanese priority filing, which I did not independently verify.