摘要:
In a data processing system having a multiple write command and a masked write command, a plurality of signal groups can be transferred from a data processing subsystem to a memory unit on consecutive system cycles. Associated with each signal group and applied to lines used to transfer mask signals are control signals that designate when the associated signal group is to stored in the memory unit. When the multiple write command is issued, the apparatus coupled to the mask signal lines is enabled and the control signals can be identified. When the control signals are identified, the operation storing the associated signal group is inhibited.
摘要:
Apparatus is disclosed for selecting a group of address signals to be applied to a memory unit array and for applying the address signals to the memory unit array to permit the activity associated with the address signals to be completed. The apparatus generates a multiplicity of signals controlling a latch-type buffer storage unit. The first generated signal insures that the signal controlling the buffer storage unit is active during application of the address signals to the system bus. The second generated signal overlaps the first generated signal and extends the signal controlling the buffer storage unit a small amount. The third generated signal overlaps the second generated signal and extends the signal controlling the buffer storage unit for the period of time necessary to utilize the memory unit array.
摘要:
A database architecture providing high availability and scalability is provided. According to embodiments of the present invention, multiple data sources are supported by multiple source databases. Partial copies of data stored in one or more of the source databases are provided to a consolidated database in response to requests for data received by the consolidated database from a data client. Multiple consolidated databases may support multiple data clients and data clients that are geographically dispersed. A sparse replication component operates to retrieve data from the source databases for storage in consolidated databases. The sparse data replication component can also monitor source databases for updates to requested data, and retrieve such data for storage in a consolidated database. In addition, each consolidated database may be associated with a data rationalization component to ensure that the data stored in the associated consolidated database does not include pathological data redundancies.
摘要:
For use in a data processing system, a main memory subsystem includes a plurality of memory boards for storing groups of logic signals. Each memory board includes an plurality of array units. Each array unit is adapted to store a group of logic signals that is equivalent in size to the field of data logic signals transferred on the system bus and has an address structure so that each addressable data signal group can be stored in a single array. The address field of each array unit is further adapted so that the probability of interfering activity in each array is low. The arrays are adapted process data signal groups independently, thus, activity involving several arrays can take place simultaneously. The memory subsystem is structured to provide a pipeline types of overlapping activity so that activity involving several array units can be in progress simultaneously. Because the manipulation of the storage cells requires the most amounts of time in the memory unit, and because the arrays are performing this activity independently for each signal group, then the memory unit can be adapted to process the signal groups applied sequentially to the system without delay in nonexceptional circumstances, the most general exceptional circumstance being the masked write operation.