Electrical computers and digital processing systems: memory – Storage accessing and control – Specific memory composition
Reexamination Certificate
2001-09-20
2003-05-13
Yoo, Do Hyun (Department: 2187)
Electrical computers and digital processing systems: memory
Storage accessing and control
Specific memory composition
C711S147000, C711S154000, C711S167000
Reexamination Certificate
active
06564295
ABSTRACT:
BACKGROUND OF THE INVENTION
1. Field of the Invention
The present invention relates to a data storage array apparatus which is formed by arranging a plurality of data storage parts, a method of controlling access to a data storage array apparatus, and a program and a medium for a data storage array apparatus.
2. Related Art of the Invention
A data storage array apparatus which is formed by arranging a plurality of data storage parts has been developed over recent years for an improved reliability and put into an actual use. In general, hard disk devices are used as the data storage parts, and such a data storage array apparatus is often called “a disk array device” or “RAID (Redundant Arrays of Inexpensive Disks).”
In the following, a conventional data storage array apparatus will be described with reference to associated drawings. This is the “server system which stores and supplies data” described in Japanese Patent Application Laid-Open Gazette No. 9-26854. The entire disclosure of Japanese Patent Application Laid-Open Gazette No. 9-26854 are incorporated herein by reference in its entirety.
FIG. 6
is a configuration diagram of a conventional server system, and
FIG. 7
is a flow chart showing operations in the conventional server system.
As shown in
FIG. 6
, k+1 pieces of disk drives
50
are connected to a buffer memory
51
, and the buffer memory
51
is connected to a control unit
52
and a reconstruction unit
53
. The first one through the k-th one store video data, while the k+1-th one stores parity data. The disk drives
50
are hard disk devices.
Shown in
FIG. 6
is a server device which transmits video data, and basic operations are as described below.
For reading data, a read instruction is queued in each disk drive
50
. The control part
52
identifies, out of the k pieces of disk drives
50
holding chunks of data which express video and a title, one disk drive
50
which holds the longest queue. When the length of the longest queue exceeds a predetermined value, data are read out from other disk drive and also from the k+1-th disk drive which holds parity instead of reading from this disk drive.
The chunk of data which are to be read but failed to be read out from the disk drive from is reconstructed by the reconstruction unit
53
using an error correction coding scheme.
In this manner, it is possible to bypass the most crowded disk drive without temporarily using the most crowded disk drive.
FIG. 7
shows this in the form of a flow chart.
K pieces of disk drives are read when the longest queue (QLlong) is smaller than a predetermined value QL(k+1)+TH. Otherwise, the reconstruction mentioned above is carried out.
As the longest queue is avoided, it is possible to avoid temporary confusion. Further, a new demand is not added to the longest queue, and therefore, a possibility of such temporary confusion is reduced remarkably low.
In short, how much the disk drives can be crowded is managed by means of queues. When a queue in a certain disk drive is long and the length of the queue is equal to or more than the predetermined value, reading from this disk drive is not performed.
However, controlling by means of the statuses of the queues regarding issuance of instruction, the configuration described above has the following problem.
That is, after receiving an instruction, a disk drive transmits data in a certain predetermined period of time. If this delay is approximately constant, the length of a queue regarding issuance of instruction, i.e., the queue being long or short, corresponds to the speed of a response time, i.e., an increased response time or a shortened response time, without any problem.
However, a characteristic of a disk drive is that a response is sometimes extremely slow. For instance, while a usual response is from 50 up to 100 ms, a response may be as extremely slow as 1 sec or more.
In other words, with only information which tells whether an instruction queue is long or short, it is not possible to accurately determine whether a response time increases or decreases.
Further, in the conventional configuration, a condition of an operation that data are restored avoiding a long queue is called “degeneration” during which data can not be restored if a disk drive other than a bypassed disk drive described above has a breakdown.
For a better reliability, therefore, degeneration should be avoided as much as possible. Despite this, the conventional configuration tends to invoke degeneration.
SUMMARY OF THE INVENTION
Accordingly, in view of the problems with the conventional techniques, the present invention aims to provide a data storage array apparatus which bears a low probability of invoking degeneration, a method of controlling access to such a data storage array apparatus, and a program and a medium for a data storage array apparatus.
One aspect of the present invention is a data storage array apparatus which comprises a plurality of data storage parts which store transferred data in accordance with a predetermined rule, an access issuance part, a group response monitoring part, an access response part, a continuous response monitoring part and an array operation control part, comprising:
said access issuance part, by means of a predetermined algorithm, groups first access requests from a sender of said data into a group which includes second access requests directed to said more than one data storage parts, treats said group as a second access requesting group; and issues said second access requests included in said second access requesting group to said data storage parts based on access issuance control provided to said data storage parts by said array operation control part,
said group response monitoring part individually manages each said second access requesting group, monitors processing responses, in said data storage parts, of said second access requesting group until said second access requesting group satisfies a predetermined condition, and issues a notice of completion of monitoring when said condition is satisfied,
said access response part notifies said sender of said data of completion of accessing in response to a predetermined situation,
said continuous response monitoring part classifies, for each data storage part, said second access requests of said second access requesting group which were met with no processing response from the data storage parts and further continuously monitors whether there are corresponding processing responses based on said notice of completion of monitoring thus issued, and issues a result of the monitoring as continuous response monitoring information,
said array operation control part provides access issuance control for said data storage parts to said access issuance part based on said continuous response monitoring information thus issued, so that an operation mode of said data storage array apparatus as a whole is controlled.
Another aspect of the present invention is the data storage array apparatus, wherein said access response part notifies said sender of said data of completion of accessing based on said notice of completion of monitoring from said group response monitoring part.
Still another aspect of the present invention is the data storage array apparatus, wherein said access response part notifies said sender of said data of completion of accessing based on a notice of issuance from said access issuance part.
Yet still another aspect of the present invention is the data storage array apparatus, wherein during reading, based on said notice of completion of monitoring from said group response monitoring part, said access response part restores, using redundancy, data (1) for which there was no processing response, in said data storage parts, of said second access requesting group, or (2) for which there was an error response, or (3) for which issuance of second access requests were suppressed in said access issuance part, and said access response part notifies said sender of said data of completion of accessing.
Still yet another aspect of the present inve
Migita Manabu
Okabayashi Ichiro
Chace Christian P.
RatnerPrestia
Yoo Do Hyun
LandOfFree
Data storage array apparatus, method of controlling access... does not yet have a rating. At this time, there are no reviews or comments for this patent.
If you have personal experience with Data storage array apparatus, method of controlling access..., we encourage you to share that experience with our LandOfFree.com community. Your opinion is very important and Data storage array apparatus, method of controlling access... will most certainly appreciate the feedback.
Profile ID: LFUS-PAI-O-3005671