Dense allocation for data storage
View Patent ↗A method for increasing locality of disk references for data storage in a single file system is provided. This locality results in higher system throughput for delivery and access to the data. The data is separated into meta data and content. Both meta data and the content are densely packed in quasi-reserved sections of the disk space. The content is allocated using time based locality.
1. A method of allocating data in an application, the data including immutable content and meta data associated with the immutable content, the method comprising:
designating a first quasi-reserved section for the meta data;
designating a second quasi-reserved section for the immutable content; and
allocating the meta data and the immutable content to the first and second quasi-reserved sections, respectively,
wherein, subject to space unavailability, at least one of the first quasi-reserved section stores immutable content and the second quasi-reserved section stores meta data, and
wherein the immutable content is allocated based on a time of delivery, thereby providing fast write and read processes in a multiple user access application.
2. A method of allocating data in an application, the data including immutable content and meta data associated with the immutable content, the method comprising:
designating a first quasi-reserved section for the meta data;
designating a second quasi-reserved section for the immutable content; and
allocating the meta data and the immutable content to the first and second quasi-reserved sections, respectively,
wherein, subject to space unavailability, at least one of the first quasi-reserved section stores immutable content and the second quasi-reserved section stores meta data,
wherein the immutable content is allocated based on a time of delivery, thereby providing fast write and read processes in a multiple user access application, and
wherein determining space unavailability includes determining whether free space in a quasi-reserved section is larger than an average amount of free space across that quasi-reserved section.
3. A method of allocating data in an application, the data including immutable content and meta data associated with the immutable content, the method comprising:
designating a first quasi-reserved section for the meta data;
designating a second quasi-reserved section for the immutable content; and
allocating the meta data and the immutable content to the first and second quasi-reserved sections, respectively,
wherein, subject to space unavailability, at least one of the first quasi-reserved section stores immutable content and the second quasi-reserved section stores meta data,
wherein the immutable content is allocated based on a time of delivery, thereby providing fast write and read processes in a multiple user access application, and
wherein determining space unavailability includes determining whether a free space in a quasi-reserved section subtracted from an average free space in that quasi-reserved section is smaller than a tunable constant.
4. The method of claim 1 , wherein allocating the immutable content to the second quasi-reserved section includes attempting allocation from a first region pointed to by an index.
5. The method of claim 4 , wherein the index moves to a second region in the second quasi-reserved section after allocation, wherein the second region is at a first free space after the first region.
6. The method of claim 4 , further including deleting immutable content, thereby freeing space from its region.
7. The method of claim 6 , wherein if the free space in the region does not exceed a computed size, then leaving the index at the first region.
8. The method of claim 6 , wherein if the free space in the region exceeds a computed size, then determining whether the freed region precedes the first region and is within the second quasi-reserved section.
9. The method of claim 8 , wherein if the freed region precedes the first region, then setting the index to the freed region.
10. The method of claim 9 , wherein if the freed region does not precede the first region, then leaving the index at the first region.
11. A software program stored on a computer-readable storage medium and executable on a computer comprising a processor, the software program for allocating data in an application, the data including immutable content and meta data associated with the immutable content, the software program comprising:
a source code segment for designating a first quasi-reserved section for the meta data;
a source code segment for designating a second quasi-reserved section for the immutable content;
a source code segment for allocating the meta data and the immutable content to the first and second quasi-reserved sections, respectively,
wherein the software program creates the first and second quasi-reserved sections on a disk, wherein the first quasi-reserved section preferentially stores meta data and the second quasi-reserved section preferentially stores immutable content,
wherein, subject space unavailability, at least one of the first quasi-reserved section stores immutable content and the second quasi-reserved section stores meta data,
wherein the source code segment for allocating the immutable content the second quasi-reserved section allocates immutable content based on a time of delivery, thereby providing fast write and read processes in a multiple user access application.
12. A software program stored on a computer-readable storage medium and executable on a computer comprising a processor, the software program for allocating data in an application, the data including immutable content and meta data associated with the immutable content, the software program comprising:
a source code segment for designating a first quasi-reserved section for the meta data;
a source code segment for designating a second quasi-reserved section for the immutable content;
a source code segment for allocating the meta data and the immutable content to the first and second quasi-reserved sections, respectively; and
a source code segment for determining whether free space in a quasi-reserved section is larger than an average amount of free space across that quasi-reserved section, thereby determining space unavailability,
wherein the software program when executed by a computer creates the first and second quasi-reserved sections on a disk, wherein the first quasi-reserved section preferentially stores meta data and the second quasi-reserved section preferentially stores immutable content,
wherein, subject space unavailability, at least one of the first quasi-reserved section stores immutable content and the second quasi-reserved section stores meta data, and
wherein the source code segment for allocating the immutable content the second quasi-reserved section allocates immutable content based on a time of delivery, thereby providing fast write and read processes in a multiple user access application.
13. A software program stored on a computer-readable storage medium and executable on a computer comprising a processor, the software program for allocating data in an application, the data including immutable content and meta data associated with the immutable content, the software program comprising:
a source code segment for designating a first quasi-reserved section for the meta data;
a source code segment for designating a second quasi-reserved section for the immutable content;
a source code segment for allocating the meta data and the immutable content to the first and second quasi-reserved sections, respectively; and
a source code segment for determining whether a free space in a quasi-reserved section subtracted from an average free space in that quasi-reserved section is smaller that a tunable constant, thereby determining space unavailability,
wherein the software program when executed by a computer creates the first and second quasi-reserved sections on a disk, wherein the first quasi-reserved section preferentially stores meta data and the second quasi-reserved section preferentially stores immutable content,
wherein, subject space unavailability, at least one of the first quasi-reserved section stores immutable content and the second quasi-reserved section stores meta data, and
wherein the source code segment for allocating the immutable content the second quasi-reserved section allocates immutable content based on a time of delivery, thereby providing fast write and read processes in a multiple user access application.
14. The software program of claim 11 , wherein the source code segment for allocating the immutable content to the second quasi-reserved section includes a source code segment for attempting allocation from a first region pointed to by an index.
15. The software program of claim 14 , further including a source code segment for the moving the index to a second region in the second quasi-reserved section after allocation, wherein the second region is at a first free space after the first region.
16. The software program of claim 14 , further including a source code segment for deleting immutable content, thereby freeing space from its region.
17. The software program of claim 16 , further including a source code segment for leaving the index at the first region if the free space in the region does not exceed a computed size.
18. The software program of claim 16 , further including a source code segment for determining whether the free space in the region precedes the first region if the free space in the region exceeds a computed size.
19. The software program of claim 18 , further including a source code segment for setting the index to the free space in the region if the free space in the region precedes the first region and is within the second quasi-reserved section.
20. The software program of claim 19 , further including a source code segment for leaving the index at the first region if the freed region does not precede the first region.
21. A method of storing data in at least one disk subsystem, the data including content and meta data associated with the content, the method comprising:
preferentially allocating the meta data to a first section; and
preferentially allocating the content to a second section,
wherein if a space is unavailable for the meta data in the first section, then finding the space in the second section,
wherein if a space is unavailable for the content in the second section, then finding space in the first section, and
wherein the immutable content is allocated based on a time of delivery, thereby providing fast write processes in a multiple user access application.
22. The method of claim 21 , wherein preferentially allocating the content to the second section includes attempting allocation from a first region pointed to by an index.
23. The method of claim 21 , wherein if multiple disk subsystems are provided, then performing the allocations according to an allocation algorithm.
24. A disk space for storing data, the data including content and meta data associated with the content, the disk space comprising:
a first section preferentially dedicated to meta data; and
a second section preferentially dedicated to content,
wherein the meta data and the immutable content are allocated to the first and second sections, respectively,
wherein, subject to space unavailability, meta data is storable in the second section and content is storable in the first section, and wherein the immutable content is allocated based on a time of delivery, thereby providing fast write and read processes in a multiple user access application.
25. The disk space of claim 24 , wherein content in the second section is placed in chronological order.
26. The disk space of claim 24 , wherein the meta data in the first section is densely packed together using a first space available allocation.