Method and system for optimizing data caching in a disk-based computer system

• US 5,257,370 A
• Filed: 08/29/1989
• Issued: 10/26/1993
• Est. Priority Date: 08/29/1989
• Status: Expired due to Term

• Current Assignee: Microsoft Technology Licensing LLC (Microsoft Corporation)
• Original Assignee: Microsoft Corporation
• Inventors: Letwin, James G.
• Primary Examiner(s): Shaw, Gareth
• Assistant Examiner(s): Katbab, Abdollah
• Application Number:
  US07/401,315
• Time in Patent Office: 1,519 Days
• Field of Search: 364/200 MS File, 364/900 MS File, 395/650
• US Class Current: 711/113
• CPC Class Codes: G06F 12/0862 with prefetch
• CPC Class Codes: G06F 12/0866 for peripheral storage systems, e.g. disk cache
• CPC Class Codes: G06F 2212/311 In host system
• CPC Class Codes: G06F 2212/6024 History based prefetching
• CPC Class Codes: G06F 2212/6026 Prefetching based on access pattern detection, e.g. stride based prefetch

Abstract

A method for dynamic caching of data in a disk based operating system. Upon the opening of a file, an access log is created. The access log is updated while accessing the data from the file. Range of access and access time are determined from the access log, and based on the range of access and access time, it is determined whether access of the file was sequential and within a predetermined access time. Based on this determination, a predetermined amount of data from the file is stored in a cache memory when access was sequential and also within the predetermined access time.

Claims

1. A method, performed by a computer having a cache, for caching data from a file stored on disk, the method comprising the steps of;

monitoring, upon opening the file, file access behavior in reading data from the file;

maintaining, upon closing the file, access information corresponding to the file indicating the file access behavior monitored;

obtaining, upon reopening the file, the access information maintained; and

reading an amount of the data from the file into the cache, the amount based on the access information obtained.

2. The method of claim 1 wherein the step of monitoring file access behavior comprises monitoring a total number of reads in which substantially adjacent sectors are read while reading data from the file.

3. The method of claim 1 wherein the stpe of monitoring file access behavior comprises monitoring a total amount of time lapsed while reading data from the file.

4. The method of claim 1 wherein the step of reading an amount of the data from the file comprises

reading a first amount of data into the cache when a total number of reads in which substantially adjacent sectors are read, divided by a total number of reads while reading data from the file, exceeds a predetermined amount,

reading a second amount of data into the cache when the total number of reads in which substantially adjacent sectors are read, divided by a total number of reads while reading data from the file, exceeds a predetermined amount and a total number of sectors read while reading data from the file divided by a total amount of time lapsed while reading data from the file exceeds a predetermined amount, and

wherein the step of reading an amount of data comprises reading a second amount of data that is larger than the first amount of data when the total number of sectors read while reading data from the file divided by a total amount of time lapsed while reading data from the file exceeds a predetermined amount.

5. The method of claim 1 wherein the step of monitoring file access behavior comprises accumulating an amount of time lapsed and number of reads in which substantially adjacent sectors are read divided by a total number of reads during each access of the file.

6. The method of claim 5 wherein the step of maintaining access information comprises storing a flag in the file indicating whether a total of the accumulated number of reads in which substantially adjacent sectors are read divided by a total amount of reads during each access of the file, divided by a total amount of time lapsed while reading data from the file, exceeds a predetermined amount.

7. The method of claim 5 wherein the step of maintaining access information comprises storing a flag in the file indicating whether a total amount of time lapsed during each access of the file exceeds a predetermined amount.

8. A computer system for caching data, comprising:

a storage medium storing a file;

a computer connected to the storage medium and having a cache, the computer comprising

means for monitoring, upon opening the file, file access behavior in reading data from the file,

means for storing, upon closing the file, in the file access information indicating the file access behavior monitored, and

means for reading, upon reopening the file, an amount of the data from the file into the cache, the amount based on the access information stored in the file.

9. The computer system of claim 8 wherein the access information comprises a flag indicating whether a total number of reads in which substantially adjacent sectors are read, divided by a total number of reads while reading data from the file, exceeds a predetermined amount.

10. The computer system of claim 8 wherein the access information comprises a flag indicating whether the total number of sectors read from the file divided by a total amount of time lapsed while reading data from the file, exceeds a predetermined amount.

11. The computer system of claim 8 wherein

the access information comprises a total number of reads in which substantially adjacent secotrs are read while reading data from the file divided by a total number of reads, and a total amount of time lapsed while reading data from the file, and wherein

the means for reading an amount of the data from the file into the cache comprises

means for reading a first amount of data into the cache when the total number of reads in which substantially adjacent sectors are read while reading data from the file, divided by a total number of reads, exceeds a predetermined amount, and

means for reading a second amount of data into the cache when a total number of sectors read while reading data from the file, divided by a total amounnt of time lapsed while reading data from the file, exceeds a predetermined amount, the second amount larger than the first amount.