US 5652916

I'll search for this specific patent number in the USPTO database and CAFC dockets.

Let me search for more details, particularly the complete claim set and any litigation/docket history.

Let me verify the complete claim set and check for any CAFC/appellate activity.

US Patent 5,652,916 (US5652916A) — Summary

Bibliographic data (verified against the USPTO patent image PDF and Google Patents)

Abstract (as issued)

"A prestaging method is adapted to a computer system comprising a central processing unit, a buffer and a secondary storage, where the prestaging method reads input data of a submitted job into the buffer from the secondary storage before executing the submitted job so as to carry out a prestaging. The prestaging method includes the steps of (a) obtaining a number of jobs for which prestagings are to be carried out simultaneously as the prestaging for the submitted job, (b) obtaining an effective data transfer speed based on the number of jobs and a data transfer speed of the secondary storage which stores the input data of the submitted job for which the prestaging is to be carried out, and (c) setting a prestaging start time from which the prestaging for the submitted job is to start by subtracting an effective data transfer time from a scheduled job start time from which the submitted job is to start, where the effective data transfer time is defined as a time which is obtained by dividing an amount of the input data by the effective data transfer speed."

Technology in plain language

The patent addresses the growing speed gap between CPUs and disk I/O. It "prestages" a job's input data from slow disk (14) into a faster buffer/disk cache (13) before the job starts, and solves the problem of when to start prestaging. Instead of using a fixed, static start time, the patent computes the start time dynamically from the disk's transfer speed, the amount of input data, the number of concurrently-prestaged jobs, and/or the current or predicted load on the data-transfer path — so prestaging finishes just-in-time (right before job execution), avoiding both premature cache occupation/takeover and incomplete prestaging. A second group of embodiments manages the buffer with three free lists ("used," "unused," "in-use") so sequentially-accessed blocks are recycled efficiently rather than using a plain LRU policy.

Independent claims — plain-language overview

• Claim 1 (prestaging start time from concurrent-job count): A method that (a) counts how many jobs will be prestaged simultaneously with the submitted job; (b) computes an "operative/effective" data-transfer speed as a function of the disk's transfer speed and that job count; and (c) sets the prestaging start time = scheduled job start time − (total input data ÷ effective transfer speed).

• Claim 10 (start time from current path load): At job submission, (a) detects the load on the data-transfer path between secondary storage and buffer; (b) computes effective transfer speed by subtracting that load from the disk's transfer speed; and (c) sets the prestaging start time by subtracting (total input data ÷ effective speed) from the scheduled start time.

• Claim 12 (start time from predicted path load at prestaging): (a) Calculates the predicted load on the data-transfer path at the time prestaging will begin; (b) calculates the data-transfer speed by subtracting that predicted load from the disk's transfer speed; and (c) sets the prestaging start time accordingly.

• Claim 14 (start time from reserved path): (a) Secures/reserves in advance the data-transfer path needed for prestaging; (b) obtains a "reserved" data-transfer speed based on the secured path; and (c) sets the prestaging start time by subtracting (input data ÷ reserved speed) from the scheduled start time (the reserved path guarantees that speed).

• Claim 16 (correction of prior jobs' start times): (a) Sets the prestaging start time for the newly submitted job; and (b) after doing so, corrects the prestaging start time of a previously submitted job to account for the change in data-transfer-path load caused by the new job.

The remaining claims (2–9, 11, 13, 15, and 17+) are dependent. Claims 2, 11, 13, and 15 limit the buffer to the computer's main storage. Claims 3–9 (depending from claim 2) add the buffer-management features: at least three free lists ("used," "unused," "in-use") with block release into a list chosen by buffer state (claim 3); the state comprising transfer kind and amount (claim 4); queue placement by block position within the file (claim 5); queue placement based on whether the whole block was transferred (claim 6); distinct queuing of invalidated blocks on file deletion (claim 7); allocating by searching used→unused→in-use lists in that order (claim 8); and using per-block map information/access counts to judge whole-block transfer completion (claim 9).

Uncertainty notes

• CAFC 2026 docket search: No results were found for patent 5652916 in any CAFC, PTAB, or other appellate docket search. This is consistent with the patent having expired (fee-related) around 2013, but I cannot affirmatively confirm the absence of any historical or current litigation; the docket search returned no hits either way.
• Complete claim count/text: I verified claims 1–10 verbatim from Justia's copy and claims 10–17 from EveryPatent's copy (both consistent with the Google Patents record). The exact text of any claims beyond 17 could not be fully retrieved in this search session, so I cannot state with certainty whether the patent contains, e.g., a separately independent file-system claim (the specification's Summary describes file-system embodiments) or a re-prestaging independent claim (sixth embodiment). The claim set includes at least 17 claims.
• The abstract quoted above is the official issued abstract as reproduced in the USPTO PDF and Google Patents record.

Sources: USPTO patent image PDF (patentimages.storage.googleapis.com/45/ab/88/e20fa8483136d6/US5652916.pdf); Google Patents record US5652916A; Justia patent record (patents.justia.com/patent/5652916); EveryPatent record (everypatent.com/comp/pat5652916.html).