Ultra reliable disk memory for multi-processor platforms

Details Ultra reliable disk memory for multi-processor platforms Ultra reliable disk memory for multi-processor platforms

2000-07-14

2003-12-30

Beausoliel, Robert (Department: 2184)

Error detection/correction and fault detection/recovery

Data processing system error or fault handling

Reliability and availability

C714S006130

Reexamination Certificate

active

06671823

ABSTRACT:

PROBLEM
In modern data processing or control systems, including telecommunications systems, a disk memory or its equivalent is used for storing information that is currently not being actively processed, but that can be fetched as needed to perform processing. Some of the information that is kept in disk memories is very valuable, in the sense that the loss associated with a loss of this memory is large. One example of this kind of information, is the billing records of a telecommunication system. If these billing records are lost, the revenue associated with the calls whose billing records have been accumulated in the disk is lost. Such loss can be in the tens and even hundreds of thousands of dollars in a large switching system. Accordingly, systems which use disk memory for such purposes require some form of back-up.
A common arrangement has two processors interconnected by a pair of SCSI buses, each SCSI (Small Computer System Interface) bus also passing through a disk platform for providing duplicate disk storage. A problem with this arrangement is that a failure of the SCSI bus may cause both processors to fail. Another problem is that if this arrangement is used to provide multiple copies of disk data, a heavy processing penalty is incurred in the main processors.
Another solution is to provide periodic back-up of the data in the disk by transmitting this data to some other processing system for storage on tape or other mass memory. Typically, this back-up procedure uses a great deal of processing time, and is therefore, performed only once a day during periods of very light load.
In view of the above, a problem of the prior art is that there is no really satisfactory arrangement for providing for ultra reliable storage of disk information in a multi-processor system.
SOLUTION
The above problem is solved and an advance is made over the prior art in accordance with this invention, wherein each processor platform, comprising a processor and its own simplex or duplex disk memory, is connected to one or more separate disk platforms, each comprising duplicate disk drives; each disk platform is connected to two or more processor platforms. Advantageously, such an arrangement provides a separate platform, having its own power rectification equipment for redundant disk memories which greatly enhances the reliability of the system. The system will only fail if either the disk platform fails and the active processor fails, or if one processor platform and both disk drives of the disk drive platform fail, or if two or more processor platforms fail.
In accordance with one preferred implementation of Applicants' invention, the presence of the disk platform is transparent to the individual processes running on each processor, so that information which is sent to the disk of a platform is automatically also sent to both disk drives of the disk platform. Advantageously, no additional processing time is required to update information in the disk platform.
In accordance with one preferred implementation of Applicants' invention, duplex disk drives are provided on the disk platforms. Advantageously, this greatly increases the reliability of the systems in which the weakest link is usually the disk drive.
In accordance with one feature of Applicants' invention, when all of the disk drive(s) of a processor platform fail, the processor automatically uses the information stored in the disk drives of the disk drive platform. This operation is transparent to the individual processes running on the processor, since disk access software automatically steers a disk read or write request to a working disk drive whether that disk drive is part of the processor platform or a part of the disk drive platform.


REFERENCES:
patent: 5761403 (1998-06-01), Yamagishi
patent: 5933653 (1999-08-01), Ofek
patent: 6035415 (2000-03-01), Fleming
patent: 6035417 (2000-03-01), Kanazawa
patent: 6286110 (2001-09-01), Klein et al.

Affiliated with

Also associated with

Rating

Ultra reliable disk memory for multi-processor platforms does not yet have a rating. At this time, there are no reviews or comments for this patent.

If you have personal experience with Ultra reliable disk memory for multi-processor platforms, we encourage you to share that experience with our LandOfFree.com community. Your opinion is very important and Ultra reliable disk memory for multi-processor platforms will most certainly appreciate the feedback.

Rate now

Comments { 0 }

Profile ID: LFUS-PAI-O-3171677