Electrical computers and digital processing systems: memory – Storage accessing and control – Specific memory composition
Reexamination Certificate
1998-08-04
2001-04-17
Chan, Eddie P. (Department: 2751)
Electrical computers and digital processing systems: memory
Storage accessing and control
Specific memory composition
C711S111000
Reexamination Certificate
active
06219752
ABSTRACT:
BACKGROUND OF THE INVENTION
The present invention relates to disk array storage devices such as RAID (Redundant Array of Inexpensive Disk), etc. using a plurality of disk memories and more particularly to storage data reading methods from disk array memories and control devices. The RAID system storage disclosed in U.S. Pat. No. 5,124,987, Japanese Patent No. 214720/1994 and 266510/1994, proposes a high speed writing method to memories. A method to write updated-data collectively in previously arranged separate empty areas in disk memory units (hereinafter simply referred to as disk memory units) instead of rewriting old data areas and a rewriting method of old data areas in idle times when made available thereafter were proposed.
The conventional methods described above will be briefly described using FIG.
18
. In
FIG. 18
, an example involving updating data blocks stored in logic block addresses (hereinafter simply referred to as logic addresses) L
6
, L
4
, L
2
, L
132
, L
7
and L
11
is considered. In these logic block addresses L
6
, L
4
, L
2
, L
12
, L
7
and L
11
, old data is stored in physical block addresses (hereinafter simply referred to as physical addresses) P
6
, P
4
, P
2
, P
12
, P
7
and P
11
in 3 disk units
181
,
182
and
193
. First, the data blocks
16
Data, L
4
Data, L
2
Data, L
12
Data, L
7
Data and L
11
Data that are newly to be updated are stored temporarily in a buffer memory
184
which is normally composed of a non-volatile memory. These data blocks are collectively written into physical addresses P
51
, P
52
, P
53
, P
54
, P
55
and P
56
which are previously arranged separate empty areas instead of directly replacing data that are the contents of the physical block addresses P
6
, P
5
, P
2
and P
12
in which old data to be updated are stored, with old data left therein. As these data are written into continuous physical addresses P
51
-P
52
, P
53
-P
54
and P
55
-P
56
in 3 disk units
181
,
182
and
183
, 6 times of writing operation required for the direct rewriting are physically reduced to 3 times of the writing operation and thus, the writing performance is largely improved.
On the other hand, in this type of conventional disk array storage systems, there is provided a conversion map
185
, that is a table showing the correspondence between logical addresses and physical addresses in which data blocks are stored. When updating data, as latest data in the logical addresses L
6
, L
5
, L
2
, L
12
, L
7
and L
11
are actually existing in the physical addresses P
51
, P
52
, P
53
, P
54
, P
55
and P
56
in the disk units as described above, the contents of the conversion map are rewritten to indicate proper disk locations. In other words, for instance, data blocks in the logical address L
6
must originally be stored in the physical address P
6
in the disk unit
181
but as actually they are stored in the physical address P
51
, the physical address P
6
corresponding to the logical address L
6
in the conversion map
185
is rewritten to P
51
. Similarly, the physical addresses corresponding to the logical addresses L
5
, L
2
, L
12
, L
7
and L
11
in the conversion map
185
are rewritten to P
52
, P
53
, P
54
, P
55
and P
56
, respectively.
Further, when reading data stored in the disk array storage, data is read out by obtaining physical addresses wherein latest data blocks corresponding to designated logical addresses are stored are read out and therefore, there is no possibility to read out old data.
Further, in order to make the explanation simple, although only two physical blocks were written in the example shown in
FIG. 18
as data blocks that are stored in one disk unit, several ten blocks are actually written in one disk unit.
SUMMARY OF THE INVENTION
In the conventional technique described above, there was a problem of data maintenance, in that all data stored in a disk memory unit was lost if latest data was erased by failure or erroneous operation of an conversion map as the positional information of latest data was managed by the conversion map. Further, there was also a problem that a conversion map becomes highly expensive because it was needed to provide an conversion map for all logic blocks and to retain the conversion map in a large capacity for power source failure.
The present invention solves the problems described above and it is an object of the present invention to provide an inexpensive and high speed data updating method for a control system for a disk storage system.
The disk storage data updating method of the present invention is characterized in that, the disk memory is equipped with N units of disk unit, a control unit to write data to the N units of disk unit or read data therefrom according to instructions from a host equipment, a volatile memory connected to this control unit, including a time stamp memory and a conversion map memory and a non-volatile memory connected to the control unit, including a write buffer memory having a memory capacity equivalent to N×K (an integer) logical block data and a buffer management table memory. The data updating method accumulates logical block data to be updated in the write buffer memory until the number of logical blocks reaches N×K−1, generates logical addresses for these logical blocks and logical address tag blocks including time stamps stored in a time stamp memory and adds them to the N×K−1 logical blocks to a total N×K logical blocks and sequentially writes into empty address areas separate from logical address areas that are retaining data to be updated on the N units of disk storage successively.
In the disk storage data updating method of the present invention, it is further characterized in that data is written in a stripe area extending over a plurality of disk units.
In the disk storage data updating method of the present invention, it is further characterized in that the time stamp memory is incremented whenever N×K logical blocks accumulated in the write buffer memory are written into N units of disk unit.
In the disk storage data updating method of the present invention, it is further characterized in that physical storage locations on the disk units corresponding to logical addresses are detected by reading and inspecting logical address tag blocks in the stripe areas recorded in the disk units and data is written in or read out of the detected storage locations.
In the disk storage data updating method of the present invention, it is further characterized in that when inspecting the logical address tag blocks, if there are a plurality of stripe areas containing the same logical address, when the time stamp in the logical address tag block is latest, the logical address blocks in that stripe area are judged to be valid blocks and blocks having the same logical address as those in other stripe areas are judged to be invalid blocks.
In the disk storage data updating method of the present invention, it is further characterized in that when inspecting the logical address tag blocks, by detecting the maximum time stamp value, a time stamp to be added in the next writing is reproduced.
In the disk storage data updating method of the present invention, it is further characterized in that when inspecting the logical address tag blocks, the time stamp value that becomes the writing sequence standard is obtained by detecting the minimum time stamp value.
In the disk storage data updating method of the present invention, it is further characterized in that by reading out the logical block data in a plurality of stripe areas stored in the disk storage and inspecting the logical address tag blocks, only valid logical blocks in the stripe areas are moved in the write buffer memory, new logical address tag blocks corresponding to these valid logical blocks are generated, and by sequentially writing logical blocks for one stripe comprising the valid data moved in the write buffer and newly generated logical address tags in empty areas separate from a plurality of read out stripe areas, an empty area is prod
Chan Eddie P.
Finnegan Henderson Farabow Garrett & Dunner L.L.P.
Kabushiki Kaisha Toshiba
Williams, II Jan S.
LandOfFree
Disk storage data updating method and disk storage controller does not yet have a rating. At this time, there are no reviews or comments for this patent.
If you have personal experience with Disk storage data updating method and disk storage controller, we encourage you to share that experience with our LandOfFree.com community. Your opinion is very important and Disk storage data updating method and disk storage controller will most certainly appreciate the feedback.
Profile ID: LFUS-PAI-O-2544527