Library system storing removable storage media

Details Library system storing removable storage media Library system storing removable storage media

2001-04-19

2002-12-31

Le, Dieu-Minh (Department: 2184)

Error detection/correction and fault detection/recovery

Data processing system error or fault handling

Reliability and availability

C714S006130

Reexamination Certificate

active

06502204

ABSTRACT:

BACKGROUND OF THE INVENTION
The present invention relates generally to storage device systems with increased usability adaptable for use with hand-carriable storage media. More particularly, but not exclusively, the invention relates to memory device systems with redundancy at respective constituent elements for improvement of usability.
In computer systems a document known as Patterson's paper has been known as one of the prior art documents most pertinent to the invention.
A.C.M. SIGMOD Conference Proceeding, “A Case for Redundant Arrays of Inexpensive Disks (RAID)” by D. Patterson et al., Chicago, Ill., Jun. 1-3, 1988 at pp. 109-116.
The Patterson paper discloses therein one technique relating to data distribution/layout on disk arrays.
Disk array is a mechanism for attainment of high performance with increased reliability of disk systems. In such disk array, enhanced performance may be attainable by forcing processor devices to recognize a plurality of physical disks as a single “virtual” disk device. On the other hand, for achievement of higher reliability, redundant data for use in recovering data as accidentally destructed upon occurrence of failure or malfunction in data storage disk devices is prestored in a separate disk device(s). Generally, one segment of data used as a unit for read/write operations of disk devices is called the “record” among those skilled in the art; in this respect, the Patterson paper has proposed several record layout methods. Note here that where disk arrays are employed, it can happen in some cases that the “record” acting as the read/write unit when viewed from the host computer is different in data length from the “record” as actually recorded in disk devices, namely, a record as stored at a single sector of disks. The former will be referred to as “logical record,” whist the latter as “physical record” hereinafter. Now, record layout methods as disclosed in Patterson's paper will be explained below.
One typical record layout scheme is to store logical records—i.e. those records as viewed from the processor device side—in a way such that these are subdivided for storage into a preselected number (say “m”; here m is the integer greater than or equal to 1) of physical records on disk devices. This approach will be called the “divisional layout” method. With such a divisional layout scheme, the same effect may be obtainable as would be attained when the data transfer rate is virtually increased by m times because of the capability of transferring a single logical record between or among m disk devices operatively associated therewith. Then, a redundant data preparation method in the divisional record layout scheme will be explained as follows. In the divisional layout, for m physical records as divided from a logical record, a plurality of—“n” where n is the integer more than 1—redundant data items are prepared for storage in respective disk devices as a single physical record with respect to one disk device (n records as a whole). Hereinafter, those physical records storing therein certain data directly accessible by the processor device for reading/writing will be referred to as “data records” whereas other physical records storing redundant data as “parity records.” Further, a combination of m data records and n parity records as organized together into a group will be called the “parity group.” Typically, if n parity records are within the parity group, data of such parity group may be recovered even upon occurrence of operation failures or disturbances in disk devices as far as the number thereof is less than or equal to n.
Another Patterson's record layout scheme proposed is to store a logical record acting as the read/write unit as looked at from the processor device functionally supervising disk devices as a single physical record, i.e. a single data record. This approach will be called the “non-divisional layout” scheme hereinafter. Accordingly, logical records are equivalent to data records. (Since respective physical records are assigned with either data records or parity records, the physical records will not always remain exactly equal to logical records. In other words, while a single logical record is one physical record, it is not always true that one physical is a single logical record and can be a parity record in several cases.) One noticeable feature of the non-divisional layout scheme lies in that read/write processings are executable for every one of respective disk devices. (With the divisional layout scheme, it should be required that plural disk devices be exclusively dedicated or “slaved” during execution of read/write operations.) As a consequence, with the non-divisional layout scheme, it becomes possible to improve the multiplexibility or “multi-tasking” offerability of read/write processings to be executed within disk arrays, which in turn leads to capability of achieving enhanced performance. With this non-divisional layout scheme also, n parity records are prepared from m data records for storage in disk devices. Note however that while the divisional layout scheme is designed to use a collection of data records within a parity group to form a single logical record as viewed from the processor device, the non-divisional layout scheme treats a respective one of data records as if it were a complete independent logical record when looking at from the processor device.
In computer systems, magnetic tape drives or optical storage drives or equivalents thereto are frequently employable as data storage devices other than the disk devices. Especially in recent years, digital versatile disks (dvds) are becoming more important in the manufacture of advanced computer systems. One significant feature of these storage devices of the types mentioned above is that storage media or record carrier bodies are separated in structure from read/write (R/W) devices operatively associated therewith, and that one storage medium is loaded into any desired R/W device for permitting execution of reading data therefrom or writing data thereinto. These media are generally known as hand-carriable media. In large-scaled computer systems a library unit is introduced in order to readily accomplish management of an extremely great number of carriable storage media. The library may include, in addition to storage media and R/W devices, a containment or “rack” structure for housing therein an increased number of storage media, and a computer-controlled robot module for carrying and delivering storage media between the rack and R/Ws. In the computer systems with such library architecture, an appropriate library management software unit is typically provided on a supervisory or “host” computer used. The library management in this case is for managing the system to monitor or “watch” which type of storage medium is present to store what kind of information. To this end, an ordinary approach upon loading a new storage medium or media is to let the host computer become aware of occurrence of such new media loading event by sending a corresponding notice thereto.
Recently, data to be processed by computer systems increases in scale more and more; thus, achievement of its usability and maintenance flexibility—this may also be called “availability” among those skilled in the computer art—is required more strictly. Therefore, in storage device systems including the aforesaid carriable recording media also, it remains effective to attain enhanced usability by incorporating therein the concepts as proposed by the Patterson paper discussed supra.
One prior known architecture applying such concepts to carriable storage media is described, for example, in Alan E. Bell (IBM Research Division), DVD Applications, COMDEX '96, Nov. 20, 1996. This Bell document has proposed redundant arrays of inexpensive libraries (RAIL) with redundancy employing a combination of plural sets of currently available standard libraries including dvds, R/W devices, robot modules, and others.
With regard to data recovery/restore technology, one typical prior art t

Affiliated with

Also associated with

Rating

Library system storing removable storage media does not yet have a rating. At this time, there are no reviews or comments for this patent.

If you have personal experience with Library system storing removable storage media, we encourage you to share that experience with our LandOfFree.com community. Your opinion is very important and Library system storing removable storage media will most certainly appreciate the feedback.

Rate now

Comments { 0 }

Profile ID: LFUS-PAI-O-2973469