| 摘要 |
The hierarchical data distribution system (HDDS) includes a top level master database system, bottom level client servers, each with its own database copy, and at least one intermediate database level. HDDS is an event driven pusher type system. The entry of update data into the system invokes the distribution process. First, the master database system is updated. Then, the master database system updates several database systems at the first intermediate database level. Each database system at the first intermediate database level then updates several database systems at the next lower database level. This may be a lower level intermediate level or it may be client servers. This process continues until the lowest level database systems, the client servers, have been updated. The distribution process performs any necessary reformatting, data assembly and data view processing before transmitting the update data. Each higher level database system must update fewer lower level servers and overall update performance is improved. If communications to a destination server are not functional, the distribution system detects this immediately because it is unable to establish communications with the destination server. Transitional inconsistency is reduced because the distribution process is invoked for all destination servers at the same time. System resources are not wasted because communications are only established when there is data to be updated.
|
