Patent
1993-12-01
1996-04-16
Beausoliel, Jr., Robert W.
39518216, 39518202, 395650, G06F 1540
Patent
active
055091180
DESCRIPTION:
BRIEF SUMMARY
FIELD OF THE INVENTION
The present invention relates to a fault tolerant change distribution method in a distributed database system comprising at least two separate intercommunicating databases, at least part of the data stored therein having identical contents in the different databases.
A distributed database system refers typically to a system with two or more separate intercommunicating databases. At least part of the stored data is identical in two or more databases. Therefore, when common data is changed in one of these databases, the same change must be made in all the other databases containing the same data in order that the required uniformity of the data would be maintained in the database system. Under normal circumstances, a database in which certain data has been changed first simply transmits the change to the other databases through interconnecting links. Problems arise, however, when faults occurring either in the databases or in the interconnecting links prevent the transmission of a change to all or part of the databases.
Conventionally, this problem has been solved by storing the changes in special buffers in the databases until they have been transmitted. This method has, however, serious drawbacks: 1) separate buffers may require a significant amount of additional memory capacity, 2) the updating of buffers requires extra computer resources, and 3) buffer overflows may occur. In case of a buffer overflow, some data is always lost, and thus the validity of the data suffers. Processes preventing overflows and processes used for recovering from overflows are complicated, and compromises must often be made in practical applications.
SUMMARY OF THE INVENTION
The object of the invention is to provide such a fault tolerant change distribution method for a distributed database system that solves the above-mentioned problems.
This and other objects and advantages of the invention are achieved with a fault tolerant change distribution method of the type described in the introduction, which is characterized in that it comprises the steps of state field of said data item in a state which indicates that the change should be transmitted to another database, database, and the changed data item in a state which indicates that the change has been transmitted.
The basic idea of the method of the invention is to add a state information field to each data item in a distributed database system instead of using separate buffers for storing changes. The state information field indicates how far the distribution of a data item/change has advanced in the distributed database system. According to the invention, the distribution of a data item/change is controlled by acknowledgement messages. The distribution/transmission is repeated periodically until it succeeds.
The method of the present invention requires little additional memory, and there are no problems of overflow as it is the original file, where the data items merely have different states, that functions as a "buffer". In addition, the method renders it easy to withdraw such deletion of a data item even after the deletion has been effected, as the data item does not have to be deleted physically. Instead, a state information field is merely set to indicate that this data item has been deleted. This renders it even possible to delete logically interconnected data items in a controlled manner.
According to an embodiment of the invention, the capacity of a very large database can be improved by using a compact transmission list for rapid finding of those data items whose transmission/distribution has not yet succeeded. The purpose of the transmission list is to render it unnecessary to go through large databases in order for such data items to be found. When transmission lists are employed, similar overflows may occur as in separate buffers, but in such cases the use of a transmission list can be easily prevented until it is operational again. When the transmission list is out of use, the data items to be transmitted are searched for directly from t
REFERENCES:
patent: 5212788 (1993-05-01), Lomet et al.
patent: 5247664 (1993-09-01), Thompson et al.
Fernardez et al, Database Security and Integrity, 1981, pp. 280-285.
Alewine et al, Branch Recovery with Complier Assisted-Multiple Instruction Retry, Fault-Tolerant Computing, 1992 Intl'l Symposium, at 66.
Pentikainen Jukka
Tuulos Matti
Beausoliel, Jr. Robert W.
Fisch Alan M.
Nokia Telecommunications Oy
LandOfFree
Fault tolerant change distribution method in a distributed datab does not yet have a rating. At this time, there are no reviews or comments for this patent.
If you have personal experience with Fault tolerant change distribution method in a distributed datab, we encourage you to share that experience with our LandOfFree.com community. Your opinion is very important and Fault tolerant change distribution method in a distributed datab will most certainly appreciate the feedback.
Profile ID: LFUS-PAI-O-331996