What Database Best Supports Quick Retrieval Of One Child Table

[Quoted] [Quoted] I have an application that has a parent record and up to ten million child records. I am guaranteed that for this application there will never be relational joins against other tables. In theory, which kind of database would best support these requirements, and which commercial product best instantiates the theory. The requirements are:

[Quoted]
• Must be able to very quickly create a new database and copy the one parent record and its ten million child records into the new database [Quoted]
• Must be save against multiple-thread access (dozens of threads might be active against the same database at any one time) [Quoted]
• Must be very fast for random access for read/write operations against individual child records [Quoted]
• Locks on the child records must not lock the parent record or any other child (i.e., record-level locks, not page locks)

[Quoted] [Quoted] Is a relational database the best kind for this sort of access? Which commercial database product will support these requirements?

--
[Quoted] Will Estes                                  U.S. Computer
Internet: westes_at_usc.com                    POB 3150
                                            Saratoga, CA  95070
                                            FAX: 408-446-1013