Electrical computers and digital processing systems: support – Digital data processing system initialization or configuration – Loading initialization program
Reexamination Certificate
2000-11-13
2003-11-18
Heckler, Thomas M. (Department: 2185)
Electrical computers and digital processing systems: support
Digital data processing system initialization or configuration
Loading initialization program
Reexamination Certificate
active
06651165
ABSTRACT:
This invention relates to computer systems and methods which involve data storage and memory media organized into a redundant array of independent disks (RAID). More particularly, the present invention relates to a new and improved method and apparatus for directly booting or establishing functionality of the computer operating system (OS) with a RAID storage media volume as the primary OS memory, rather than a single disk or storage device, without requiring the use of a dedicated RAID controller. As a result of the RAID volume being the primary OS memory, system performance is substantially enhanced by the OS having the capability to directly send and receive large numbers of input output operations (I/Os) to the RAID volume without encountering the delay and complexity of converting single memory device I/Os into RAID I/Os. Greater reliability is also achieved by storing the OS on the RAID volume.
BACKGROUND OF THE INVENTION
One type of data storage and memory media which is well known and used in many modern computer systems is a redundant array of independent disks (RAID). RAID media use a plurality of separate disk drives which are connected in a group or array, rather than a single disk drive or storage media. RAID systems offer numerous advantages, particularly in high-performance computer systems which require high data throughput performance and high reliability against the accidental loss of data.
To achieve higher data throughput, the RAID volume is formed by distributing or striping the data across a number of the RAID disk drives to form a data volume. A data volume is the collection of physical storage spaces or capacity which is reserved on a multiplicity of the RAID disk drives in which a single integral collection of data is distributed. This type of RAID volume is referred to as “RAID 0.” By distributing the data volume throughout a number of disk drives in the array, input/output operations (I/Os) to multiple disk drives can be performed approximately simultaneously. The time to perform I/Os constitutes a substantial restriction in computer performance, so by performing the I/Os simultaneously on all of the disk drives, the time consumed for the I/Os to access and retrieve the data is substantially reduced. The performance of the computer system is substantially increased in terms of throughput of information.
Reliability against accidental data failures is achieved in a type of RAID media known as “RAID 1.” Any disk drive is susceptible to failure and a possible loss of the data and information recorded on it. However, in a RAID 1 configuration, identical copies of the same data and information are recorded on different disk drives. The multiple disk drives which receive the identical copies of the data are referred to as “mirrored.” The failure of one of the mirrored disks does not prevent access to the data because the other mirrored disk drive is probably functional and available for I/Os without any significant loss of performance or downtime in computer functionality.
RAID media may also make use of combinations of both RAID 0 and RAID 1 to achieve both high-performance and redundancy. Such combinations are known as “RAID 0+1,” “RAID 1+0” or “RAID 10.” Moreover, parity codes which allow the reconstruction of lost data are also typically used in RAID media as a further assurance against accidental destruction of data resulting from causes not involving the failure of disk drives. Because of both performance enhancements and gains in reliability against the unintended loss of data, RAID media offer substantial improvements in computer systems.
Despite the advantages of RAID storage media compared to single storage media, most computer software operating systems (OSs) have been written to interact with a single storage media as the primary OS memory media. This fact is primarily a historical artifact of the earlier and more primitive OS architectures which were originally developed for earlier versions of personal computers. However, because of the large number of personal computer systems in use, because contemporary personal computers offer great computing capacity for relatively low cost, and because contemporary personal computers are being used more predominantly in high-capacity business computing applications, attempting to change the OS from the basic original single storage media architecture would cause enormous problems of backward compatibility for use with earlier versions of personal computers. Indeed, none of the contemporary personal computers are believed to offer the capability of the OS interacting directly with a RAID media as the primary OS memory.
Instead, to use RAID media with contemporary computers, it has been necessary to interpose RAID controllers between the RAID media and the computer system to perform RAID algorithms between the conventional, single storage media I/Os generated by the conventional OS and the multiple different I/Os which are involved in communicating with the multiple disk drives of the RAID storage media. Such RAID controllers are dedicated, single purpose devices whose sole function is to perform these RAID algorithms. In addition, certain types of hardware adapters have been used as RAID controllers by programming the hardware adapters to perform the RAID algorithms. Because of the similarity in dedicated, sole purpose functionality, both the dedicated RAID controllers and the hardware adapters which have been programmed to perform only the RAID algorithms are referred to herein as RAID controllers.
RAID controllers hide the physical properties of the RAID volume from the OS. The OS performs I/O instructions as though a single memory media or disk drive was the mass storage element for the computer system. The RAID controller prevents the OS from directly interacting with the physical properties of the RAID volume. While RAID controllers are effective, they add complexity and cost to a high-performance computer system. The complexity arises because the RAID volume I/O processing is off-loaded from the OS to the RAID controller, and because of the conversion required between I/Os generated by a conventional, single storage media OS and the I/Os delivered to and received from the RAID storage media. The conversions create additional possibilities for malfunction, failure and diminished performance of the computer system. The cost of a dedicated RAID controller or a RAID hardware adapter increases the cost of the computer system, because the cost of these items is generally considerably higher than the cost of more generic components use for a variety of different purposes in a computer system.
In response to the cost and complexity of RAID controllers, RAID emulation software programs have been written which directly interact with the conventional, single storage media I/Os generated by the OS and convert those instructions into I/Os suitable for interaction with the RAID volume. Such RAID emulation software programs are effective, but they must rely on the operation of the conventional single storage media OS as a basis for their own functionality in implementing the RAID algorithms. The conventional single storage media OS must be booted and fully operative before the RAID emulation program can be loaded and used.
These previous RAID controller and emulation techniques have made it costly, complex, impossible or difficult to distribute the OS on the RAID volume and to boot the OS from the RAID volume. Consequently, the benefits of redundancy and reliability against accidental data loss offered by RAID storage systems has not generally been available to the OS. Furthermore, in order to use the previous RAID controllers or RAID emulation software programs, it was necessary to have the conventional OS operating, and operation of the conventional OS meant that a single storage media was the primary OS memory. Consequently, the operation of the OS with its primary OS memory as a single storage media could not be interrupted to replace the primary OS memory with a RAID volume. It was impossible to d
Heckler Thomas M.
Ley, L.L.C. John R.
LSI Logic Corporation
LandOfFree
Method and apparatus for directly booting a RAID volume as... does not yet have a rating. At this time, there are no reviews or comments for this patent.
If you have personal experience with Method and apparatus for directly booting a RAID volume as..., we encourage you to share that experience with our LandOfFree.com community. Your opinion is very important and Method and apparatus for directly booting a RAID volume as... will most certainly appreciate the feedback.
Profile ID: LFUS-PAI-O-3126856