Electrical computers and digital processing systems: memory – Storage accessing and control – Specific memory composition
Reexamination Certificate
2001-10-01
2004-03-09
Nguyen, Hiep T. (Department: 2187)
Electrical computers and digital processing systems: memory
Storage accessing and control
Specific memory composition
C711S162000, C714S006130, C714S010000, C714S011000
Reexamination Certificate
active
06704839
ABSTRACT:
FIELD OF THE INVENTION
This invention relates generally to a data storage system and a method of storing data and, more particularly, to a system and method implementing a log structured array in a storage subsystem with at least two storage controller processors controlling a shared set of direct access storage devices.
BACKGROUND OF INVENTION
A data storage subsystem having multiple direct access storage devices (DASDs) may store data and other information in an arrangement called a log structured array (LSA).
Log structured arrays combine the approach of the log structured file system architecture as described in “The Design and Implementation of a Log Structured File System” by M. Rosenblum and J. K. Ousterhout, ACM Transactions on Computer Systems, Vol. 10 No. 1, February 1992, pages 26-52 with a disk array architecture such as the well-known RAID (redundant arrays of inexpensive disks) architecture which has a parity technique to improve reliability and availability. RAID architecture is described in “A Case for Redundant Arrays of Inexpensive Disks (RAID)”, Report No. UCBICSD 87/391, December 1987, Computer Sciences Division, University of California, Berkeley, Calif. “A Performance Comparison of RAID 5 and Log Structured Arrays”, Proceedings of the Fourth IEEE International Symposium on High Performance Distributed Computing, 1995, pages 167-178 gives a comparison between LSA and RAID 5 architectures.
An LSA stores data to an array of DASDs in a sequential structure called a log. New information is not updated in place, instead it is written to a new location to reduce seek activity. The data is written in strides or stripes distributed across the array and there may be a form of check data to provide reliability of the data. For example, the check data may be in the form of a parity check as used in the RAID 5 architecture which is rotated across the strides in the array.
An LSA generally consists of a controller and N+M physical DASDs. The storage space of N DASDs is available for storage of data. The storage space of the M DASDs is available for the check data. M could be equal to zero in which case there would not be any check data. If M=1 the system would be a RAID 5 system in which an exclusive-OR parity is rotated through all the DASDs. If M=2 the system would be a known RAID 6 arrangement.
The LSA controller manages the data storage and writes updated data into new DASD locations rather than writing new data in place. The LSA controller keeps an LSA directory which it uses to locate data items in the array.
As an illustration of the N+M physical DASDs, an LSA can be considered as consisting of a group of DASDs. Each DASD is divided into large consecutive areas called segment-columns. If the DASDs are in the form of disks, a segment-column is typically as large as a physical cylinder on the disk. Corresponding segment-columns from the N+M devices constitute a segment. The array has as many segments as there are segment-columns on a single DASD in the array. One or more of the segment-columns of a segment may contain the check data or parity of the remaining segment-columns of the segment. For performance reasons, the check data or parity segment-columns are not usually all on the same DASD, but are rotated among the DASDs.
Logical devices are mapped and stored in the LSA. A logical track is a set of data records to be stored. The data may be compressed or may be in an uncompressed form. Many logical tracks can be stored in the same segment. The location of a logical track in an LSA changes over time. The LSA directory indicates the current location of each logical track. The LSA directory is usually maintained in paged virtual memory.
Whether an LSA stores information according to a variable length format such as a count-key-data (CKD) architecture or according to fixed block architecture, the LSA storage format of segments is mapped onto the physical storage space in the DASDs so that a logical track of the LSA is stored within a single segment.
Reading and writing into an LSA occurs under management of the LSA controller. An LSA controller can include resident microcode that emulates logical devices such as CKD or fixed block DASDs. In this way, the physical nature of the external storage subsystem can be transparent to the operating system and to the applications executing on the computer processor accessing the LSA. Thus, read and write commands sent by the computer processor to the external information storage system would be interpreted by the LSA controller and mapped to the appropriate DASD storage locations in a manner not known to the computer processor. This comprises a mapping of the LSA logical devices onto the actual DASDs of the LSA.
In an LSA, updated data is written into new logical block locations instead of being written in place. Large amounts of updated data are collected as tracks in controller memory and destaged together to a contiguous area of DASD address space called a segment. A segment is usually an integral number of stripes of a parity system such as RAID 5. As data is rewritten into new segments, the old location of the data in previously written segments becomes unreferenced. This unreferenced data is sometimes known as “garbage”. If this were allowed to continue without taking any action, the entire address space would eventually be filled with segments which would contain a mixture of valid (referenced) data and garbage. At this point it would be impossible to destage any more data into the LSA because no free log segments would exist into which to destage data.
To avoid this problem, a process known as “Free Space Collection” (FSC) or “Garbage Collection” must operate upon the old segments. FSC collects together the valid data from partially used segments to produce completely used segments and completely free segments. The completely free segments can then be used to destage new data. In order to perform free space collection, data structures must be maintained which count the number of garbage and referenced tracks in each segment and potentially also statistics which indicate the relative rate of garbage accumulation in a segment. (See “An Age Threshold Scheme for Garbage Collection in a Log Structured Array” Jai Menon, Larry J Stockmeyer. IBM Research Journal 10120.)
Snapshot copy is a facility that is commonly supported by LSA subsystems. Snapshot copy describes a system by which the LSA directory is manipulated so as to map multiple areas of the logical address space onto the same set of physical data on DASDs. This operation is performed as an “atomic event” in the subsystem by means of locking. Either copy of the data can subsequently be written to without affecting the other copy of the data (a facility known as copy on write).
Snapshot copy has several benefits to the customer: (1) It allows the capture of a consistent image of a data set at a point in time. This is useful in many ways including backup and application testing and restart of failing batch runs. (2) It allows multiple copies of the same data to be made and individually modified without allocating storage for the set of data which is common between the copies.
In existing storage subsystems, a redundant storage subsystem is often constructed from a pair of storage controller processors which share a common pool of DASDs to which they are both connected and the pair of controllers support a same set of logical upstream devices. Each storage controller processor typically comprises the following components. (a) An upstream communication channel to the host computer(s). (b) A non-volatile memory into which data written from the host computer may be stored between the time that completion status for the write is given to the host computer and the time that the data is committed to a DASD for long term storage. (c) Some stored programs which operate upon host data so as to transform and or replicate it in some way. Examples are RAID modules, LSA modules, compression modules. (d) Connections to a pool of DASDs used for the long t
Butterworth Henry Esmond
Nicholson Robert Bruce
Turner Douglas
Butler, Esq. Bryan W.
Harrington & Smith ,LLP
Nguyen Hiep T.
LandOfFree
Data storage system and method of storing data 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 system and method of storing data, we encourage you to share that experience with our LandOfFree.com community. Your opinion is very important and Data storage system and method of storing data will most certainly appreciate the feedback.
Profile ID: LFUS-PAI-O-3264969