Electrical computers and digital processing systems: memory – Storage accessing and control – Access timing
Reexamination Certificate
1999-10-05
2002-07-23
Hudspeth, David (Department: 2651)
Electrical computers and digital processing systems: memory
Storage accessing and control
Access timing
C711S120000, C711S162000
Reexamination Certificate
active
06425063
ABSTRACT:
This application claims priority under 35 U.S.C. §§119 and/or 365 to 9803370-7 filed in Sweden on Oct. 5, 1998; the entire content of which is hereby incorporated by reference.
FIELD OF INVENTION
The present invention relates to a method and to an arrangement for memory management. The method and the arrangement are adapted to manage a first memory and a second memory that operate in parallel with each other, for instance two memories that belong to a first and a second parallel operating processor.
BACKGROUND OF THE INVENTION
The use of primary memories has long been known to the art. By primary memory is meant in this document a memory that will only retain its information content when a voltage is applied and which will lose: its information content when the voltage is removed, such memories also being designated Random Access Memories (RAMs).
In the case of a memory that coacts with a processor, it is also known to allow respective memories to be accessed for reading and/or writing purposes, through the medium of a connection referred to here as a memory bus, and in time slots that are correlated within instruction cycles associated with the execution work of the processors.
With respect to high reliability requirements of a system, it is also known to allow two mutually different processors to perform the same tasks in parallel with one another, so as to generate redundancy in the system. In such contexts, these processors can be referred to as an executing processor and a stand-by processor.
It is also known in such contexts that the memory or memories that coacts/coact with one processor shall contain the same information as the memory or memories that coacts/coact with the other processor. Since writing of information into the memories and reading of information therefrom can be at fault, it is also known to check parts of the memory content or the whole of said memory content continuously during operation of the processors, for instance by comparing the memory content on the executing side with the memory content on the stand-by side.
When large memories are concerned, the work involved is relatively capacity demanding, and consequently it is usual to check solely the memory content that is most important with respect to the work carried out by the processors.
It is also known when starting up one processor, for instance after a crash, after a service, or after upgrading one processor, to go through a start-up phase which comprises, inter alia, transferring the content of the memory/memories belonging to the still operating processor to the memory/memories belonging to the start-up processor.
The process of reading from one memory and writing into the other memory is capacity demanding and means that the processor will not be available for ordinary working purposes for some time.
It is also known that certain types of work data require uninterrupted processing, such as certain telecommunications applications.
It is also known to vary the load in telecommunications applications with time, and consequently applications that require uninterrupted processing are often allocated a processor capacity that is adapted to handle the highest load that occurs.
Moreover, the capacity requirement can change with respect to a given application, either: increase or decrease, and it is known to upgrade processor capacity in accordance with the new capacity requirements on such occasions.
SUMMARY OF THE PRESENT INVENTION
Technical Problems
When considering the earlier standpoint of techniques as described above, it will be seen that a technical problem resides in enabling a given memory to be managed without taking the capacity from the ordinary or standard tasks of the processor in the case of parallel operating processors and their respective primary memories
Another technical problem is one of enabling a start-up phase, for instance in conjunction with service, upgrading, or a re-start after a crash of one processor and its associated memory, while the other processor continues to operate with undiminished capacity with respect to its ordinary tasks.
Another technical problem is one of enabling the memory content of one memory belonging to the operating processor to be transferred to another memory belonging to the start-up processor with undiminished capacity of the working processor in handling its ordinary tasks.
Another technical problem is one of handling changes that occur in the memory of the still operating processor in a manner so that the memory belonging to the start-up processor will also include the changed memory content.
In normal parallel operation between the two processors, another technical problem is one of checking continuously that the parallel operating memories have the same memory content without taking capacity from the ordinary working tasks of the processor to this end.
A further technical problem is one of providing the measures and means required to be able to use empty time slots on a memory bus acting between a processor and a memory for tasks other than the ordinary tasks of a processor.
Solution
With the intention of solving one or more of the aforesaid technical problems, the present invention takes as its starting point a memory managing method where a first and a second memory operate in parallel, such as two memories belonging to first and second parallel operating processors, where respective memories are accessed for reading and/or writing purposes via a so-called memory bus and in time slots correlated with instruction cycles associated with the execution work of the processors, and where the second memory shall contain the same information as the first memory.
With a starting point from such a method and with the intention of providing memory managing that does not take the capacity from the ordinary tasks of the processors, it is proposed in accordance with the invention that memory managing includes the detection of empty time slots in the memory bus, and that requisite memory managing is solely carried out in the detected empty time slots.
This means that during a start-up phase, which includes writing the content of the first memory into the second memory in conjunction with a new start or a re-start of the second processor and therewith the second memory, a memory content belonging to a first memory address in the first memory is read, in accordance with the invention, from the first memory and written to a corresponding memory address in the second memory in a first detected empty time slot.
The memory content belonging to a second memory address in the first memory is then read from the first memory and written to a corresponding memory address in the second memory in a second detected empty time slot, and so on until the memory content of the second memory is the same as the memory content in said first memory.
In order to enable the start-up phase to take place during continued operation of the first processor and to enable information to be written into the first memory to be processed under such conditions during the start-up phase, it is proposed in accordance with the invention that writing of information into memory addresses that have already been read and that belong to the first memory also takes place to the second memory during the start-up phase.
With the intention of further simplifying this managing process, it is proposed in accordance with the invention that all writing to a memory address in the first memory also takes place to a corresponding memory address in the second memory during the start-up phase.
When memory managing during normal parallel operation of the first and second processors also includes checking continuously that the whole or parts of the content of the second memory is the same as corresponding parts of the first memory, it is proposed in accordance with the invention that normal operation will also include detection of empty time slots in the memory bus, so as to enable said check to be carried out without using time slots from the ordinary tasks of the memory bus, and to check the memory cont
Hintukainen Kari Anders
Hofverberg Per Tobias
Mattson Björn Axel
LandOfFree
Method and arrangement for memory management 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 arrangement for memory management, we encourage you to share that experience with our LandOfFree.com community. Your opinion is very important and Method and arrangement for memory management will most certainly appreciate the feedback.
Profile ID: LFUS-PAI-O-2844937