Electrical computers and digital processing systems: memory – Storage accessing and control – Specific memory composition
Reexamination Certificate
2002-01-10
2004-06-08
Moazzami, Nasser (Department: 2187)
Electrical computers and digital processing systems: memory
Storage accessing and control
Specific memory composition
C711S112000, C714S006130
Reexamination Certificate
active
06748489
ABSTRACT:
BACKGROUND OF THE INVENTION
In general, the present invention relates to a volume management method for defining a volume on a physical storage device of a storage apparatus. More particularly, the present invention relates to a processing technique for creating and moving a volume by consideration of effects on performance of other volumes.
There exists a technology called RAID (Redundant Array of Inexpensive Disks) for improving reliability and performance by grouping two or more physical disks to provide redundancy. In using a storage apparatus by adoption of the RAID technology, first of all, two or more physical disks (physical storage devices) of a storage apparatus are collected to form a group called a parity group. Then, logical storage areas are defined in the parity group. The logical storage areas are each called a volume to be used by a computer making accesses to the storage apparatus. In many cases, at least two volumes can be defined in a parity group. FIG.
1
(
a
) shows typical definition. As shown in FIG.
1
(
a
), four physical disks
1001
,
1002
,
1003
and
1004
constitute a parity group
1005
in which three volumes
1006
,
1007
and
1008
are defined. The volumes
1006
,
1007
and
1008
defined in the parity group
1005
all use the physical disks
1001
,
1002
,
1003
and
1004
. A state of letting different volumes share the same physical disk prevails not only in a case wherein a storage apparatus adopting the RAID technology is used, but also a case in which two or more volumes are defined on a physical disk.
There are several ways to define the method of composing a parity group of a plurality of physical disks by using such a RAID technology. A standard used in the definition of the method is referred to as a RAID level. At RAID level 0, which is also called a stripe, for example, data is written sequentially into a plurality of physical disks in predetermined-size units. At RAID level 1, which is called a mirror, on the other hand, data is read out from and written into 2 physical disks. At RAID levels 4 and 5 for storing data redundantly, data is stored by being distributed among a plurality of disks except one specific disk so that data can be recovered onto the specific disk in case another physical disk fails and becomes unusable. RAID levels 4 and 5 differ from each other in that they employ different ways of holding redundant data.
There also exists a storage apparatus in which a parity group is composed by grouping partial areas of physical disks instead of whole areas of physical disks. In this case, a physical disk may belong to a plurality of parity groups. An example is shown in FIG.
1
(
b
). In this example, a parity group
1015
is composed of partial areas of physical disks
1009
,
1010
,
1011
and
1012
. A volume
1016
is defined in the parity group
1015
. A parity group
1013
is composed of partial areas of physical disks
1009
and
1010
and no volume is defined in the parity group
1013
. A parity group
1014
is composed of partial areas of physical disks
1011
and
1012
and a volume
1017
is defined in the parity group
1014
. In this storage apparatus, the physical disks
1009
,
1010
,
1011
and
1012
each belong to 2 parity groups.
In some storage apparatuses adopting the RAID technology, a volume is not directly defined in a parity group. Instead, such storage apparatuses have a function whereby logical disks are defined in a parity group, logical disks are divided and joined repeatedly across several hierarchical layers and a logical disk at the top hierarchical layer is defined as a volume. In the storage apparatus shown in FIG.
1
(
b
), for example, a logical disk is defined in each of parity groups
1013
and
1014
, and what is obtained by joining the defined logical disks forms a volume.
One of references describing the RAID technology is a document authored by Mark Farley with a title of “Building Storage Networks,” Network Professional's Library, Osborne.
By the way, in recent years, attention is paid to a SAN (Storage Area Network), which is a high-speed network comprising typically fiber channels connecting a plurality of storage apparatuses to a plurality of client computers using the storage apparatuses. In the SAN environment, a client computer is capable of using two or more storage apparatuses at the same time and a storage apparatus can be used by two or more client computers. A typical configuration of the SAN is shown in FIG.
2
. In the SAN configuration shown in
FIG. 2
, a network
2001
connects storage apparatuses
2005
,
2006
and
2007
, client computers
2002
and
2003
using the storage apparatuses
2005
,
2006
and
2007
as well as a management computer
2004
for managing the storage apparatuses
2005
,
2006
and
2007
to each other. The management computer
2004
defines internal configurations of the storage apparatuses
2005
,
2006
and
2007
, which are connected to the network
2001
, and controls permissions of accesses to the storage apparatuses
2005
,
2006
and
2007
. In the SAN environment shown in
FIG. 2
, if the client computer
2002
is given an access permission by the management computer
2004
, the client computer
2002
is capable of making accesses to all the storage apparatuses
2005
,
2006
and
2007
. In addition, the client computer
2002
and the client computer
2003
are capable of sharing the storage apparatus
2005
.
Many storage apparatuses such as the storage apparatuses
2005
,
2006
and
2007
, which are used in the SAN environment like the one shown in
FIG. 2
, adopt the RAID technology described earlier. For example, a plurality of volumes using the same physical disks are created in the storage apparatus
2005
, and each of the volumes can be used by the client computers
2002
and
2003
.
One of references describing the SAN is a document with a title of “Data Storage: Report 2000,” an edition of Nikkei Computopia, 2000.
A storage management program is used for doing work to create and move a parity group and a volume. The storage management program is executed by the management computer connected to the SAN or a SAN terminal provided with a storage apparatus. If physical disks in a storage apparatus being processed have sufficiently large storage capacities and the storage apparatus' necessary resources can all be used, requested operations can always be carried out.
By the way, the storage management program may have a function relevant to management of performance of physical disks and volumes. The storage management program may be provided with a function to monitor an average access time of a volume in physical disks and issues a warning if the monitored access time exceeds the average access time's upper limit determined in advance. The storage management program may have a function to store a performance log and optimize performance by relocation of volumes in accordance with the stored performance log.
SUMMARY OF THE INVENTION
In a case wherein two or more volumes use the same physical disk shown in FIGS.
1
(
a
) and
1
(
b
), if accesses are concentrated on a specific one of the volumes, it is quite within the bounds of possibility that the performance of another volume using the same physical disk deteriorates. Assume that accesses are concentrated on the volume
1006
shown in FIG.
1
(
a
). In this case, the performance of the volumes
1007
and
1008
, which use the same physical disk, deteriorates. From a different point of view, if an already existing volume displays requested performance and a new volume is created to use the same physical disk as the already existing volume or a volume is moved from another physical disk to the same physical disk, it is quite within the bounds of possibility that the already existing volume is no longer capable of satisfying the requested performance. Assume that the volume
1006
shown in FIG.
1
(
a
) currently has an average access time of 18 msec, satisfying a required average access time of 20 msec. Let a new volume be created in the parity group
1005
.
Asano Masayasu
Miyazaki Satoshi
Soejima Ken'ichi
Mattingly Stanger & Malur, P.C.
Moazzami Nasser
LandOfFree
Volume management method and apparatus does not yet have a rating. At this time, there are no reviews or comments for this patent.
If you have personal experience with Volume management method and apparatus, we encourage you to share that experience with our LandOfFree.com community. Your opinion is very important and Volume management method and apparatus will most certainly appreciate the feedback.
Profile ID: LFUS-PAI-O-3332138