Electrical computers and digital processing systems: memory – Storage accessing and control – Specific memory composition
Reexamination Certificate
2001-07-30
2003-12-02
Verbrugge, Kevin (Department: 2188)
Electrical computers and digital processing systems: memory
Storage accessing and control
Specific memory composition
Reexamination Certificate
active
06658528
ABSTRACT:
BACKGROUND OF THE INVENTION
1. The Field of the Invention
The present invention relates to file systems and storage devices used within digital computer systems. More specifically, the present invention relates to manners of transferring data in file systems to and from storage devices during read and write cycles.
2. The Relevant Art
As the use of computers becomes increasingly prevalent, it has become necessary to store large amounts of data that cannot be directly stored on a personal computer or workstation at remote locations. Local area networks (LANs) and storage area networks (SANs), which typically include one or more file servers acting as storage hubs for data, are often used in large volume remote data storage.
In situations where large volumes of data storage is required, the file servers often employ arrays of direct access storage devices (DASDs) such as hard disk drives to store the data at a central location of the LAN or SAN. Such storage device subsystems are currently characterized by a predominant volume set geometry. The volume set geometry of the storage device affects the manner in which data is stored and may consist of several member disks partitioned into extents. The extents are mapped directly into volume sets, or are mapped into redundancy groups or parts of redundancy groups which are mapped into volume sets.
Each extent is further partitioned into one or more strips. A strip represents a number of contiguous logical blocks that are mapped onto an extent on a member disk in a volume set. A strip size, typically expressed in kilobytes, is associated with a strip. The immediately preceding or subsequent contiguous logical blocks in the volume set are mapped onto another extents on another member disk in the volume set. The strips on each of the extents are collectively grouped into stripes. A stripe consists of a strip on each member disk of the storage device.
As data is written to the device, it may be written in essentially one of two ways. Either the data consumes a full strip or multiple full strips, or it consumes a partial strip. A partial strip access incurs some read/modify/write performance penalty. This penalty occurs when the parity strip associated with the data being written is locked for updating. Any other data that may share the parity strip with the data being written may incur a resource lock penalty while the parity strip is locked during the write access.
This penalty may be avoided by performing an on-boundary full strip write access to the storage device. When an on-boundary full strip write access occurs, all the data sectors in the full stripe are being updated, and the necessary information for updating the parity strip may be generated in-line as opposed to separately. This process is able to update the parity strip at he same time that data is being written to the device so no separate lock is needed to update the parity strip.
Furthermore, within the storage device, an operating method is employed for locking data segments during read/write operation to protect the integrity of the stored data when the data may be accessed through multiple data paths. The size of the lockable data segment may be a fraction of the strip size, equivalent to the strip size, a multiple of the strip size, or a multiple of a fraction of the strip size.
Current operating environment file system implementations range in their ability to be configured at file system creation time for such parameters as file system blocking factor. This blocking factor defines the smallest common unit for allocation of contiguous data sectors to the file system. The blocking factor further defines the smallest data transfer size that will be passed to the Input/Output (I/O) subsystem driver.
When current operating environment file system implementations are created, such parameters as blocking factor and cluster size of the file system of the file system are defined independent of stripe size and locking segment size of the storage device. Since the operating environment file system is unaware of the geometry of the storage device, performance problems often occur during I/O operation between the file systems and the storage system. One of theses types of problems is the read/modify/write penalty associated with the necessity of writing a partial stripe of data because the file system cluster size does not correspond to the stripe size of the storage device.
Another of ten-encountered problem involves segment locking during write operations. During write operations, data is often locked, as described above, to maintain data integrity during the write access. A resource conflict will occur if an attempt is made to write data to a stripe that is already locked by another write process. Current file systems have no knowledge of storage device locking strategies and thus, often attempt to write to a segment that is already locked by another write process.
Accordingly, a need exists in the art for an intelligent geometry engine file system that is aware of the geometry of the storage device and is capable of using that knowledge to perform I/O operations in a more efficient manner.
OBJECTS AND BRIEF SUMMARY OF THE INVENTION
The intelligent geometry engine file system of the present invention has been developed in response to the present state of the art, and in particular, in response to the problems and needs in the art that have not yet been fully solved by currently available intelligent geometry engine file systems. Accordingly, it is an overall object of the present invention to provide an intelligent geometry engine file system that overcomes many or all of the above-discussed shortcomings in the art.
To achieve the foregoing object, and in accordance with the invention as embodied and broadly described herein in the preferred embodiment, an intelligent geometry engine file system is provided. The intelligent geometry engine file system of the present invention provides increased performance and efficiency in the operation of data input/output (I/O) systems. The file system obtains data about the geometry of the storage device either by automatically querying the storage device at file system creation time or by allowing a user to input the storage device parameters. The intelligent geometry engine file system then manages I/O requests to and from the storage device in such a manner that full advantage is taken of the geometry of the storage device.
The geometry of the storage device may include such parameters as the data stripe size of the storage device, the data placement method of the storage device, or the segment locking method of the storage device. The use of the information about the geometry of the storage device ensures that when data is written to the storage device, the data is written in such a manner that write accesses always occur at stripe boundaries. By ensuring that write accesses always occur at stripe boundaries, the number of segment locks decrease, as does the seek time of the write head. This decrease in the number of segment locks and in seek time substantially increases performance of the I/O operations of the data storage system.
Other manners of applying the intelligent geometry include storing volume header information in multiples of a full stripe, making disk/object geometry available in a volume header, accessing and applying configuration parameters during run-time, placing file directory information at the strip/stripe boundaries, strategically placing blocks and clusters according to the disk geometry, and strategically applying knowledge of the locking size to minimize conflicts and store locking data in caches.
REFERENCES:
patent: 5333305 (1994-07-01), Neufeld
patent: 5408644 (1995-04-01), Schneider et al.
patent: 5613088 (1997-03-01), Achiwa et al.
patent: 5657468 (1997-08-01), Stallmo et al.
patent: 5737741 (1998-04-01), Hilditch et al.
patent: 5799140 (1998-08-01), Niijima et al.
patent: 5809224 (1998-09-01), Schultz et al.
patent: 5859965 (1999-01-01), Gittins et al.
patent: 5915094 (1999-06-01), Kouloheris et al.
pa
Archibald, Jr. John E.
McKean Brian D.
Kunzler & Associates
Verbrugge Kevin
LandOfFree
System and method for improving file system transfer through... does not yet have a rating. At this time, there are no reviews or comments for this patent.
If you have personal experience with System and method for improving file system transfer through..., we encourage you to share that experience with our LandOfFree.com community. Your opinion is very important and System and method for improving file system transfer through... will most certainly appreciate the feedback.
Profile ID: LFUS-PAI-O-3175700