This has nothing to do with a certain database vendor. Paul Tuckfield of (first) PayPal and (now) YouTube fame described a nifty method to speed up MySQL Replication in his keynote at the MySQL Conference.
The pitch is simple: He implemented an oracle algorithm. In more detail: In a MySQL setup, several layers of caches exist. In front of the database sits the query cache that saves the MySQL from doing actual work (query analysis, execution, disk I/O). The next two levels sit between MySQL and a the actual disk. The first is the filesystem cache that the operating system provides. The second is the block-level cache that the RAID controller provides. Considering the last, from a MySQL point of view, data has been written, when it hits the RAID controller.
All these caches serve one purpose: Reduce execution of complex or slow operations such as physically reading data from a harddrive or writing to it. MySQL Replication, in a nutshell, works like this: A master server records all queries it receives that alter data into a logfile. A slave server reads that logfile and serially applies all queries to its data. This happens in a single thread on the client.
Writing the data to disk can be made faster by making sure that the last cache contains the correct data to speed up the writes. One cannot look into the future, but with the serial logfile that a MySQL Replication slave is reading, there is a small amount of future available.
Paul wrote a script that reads from the logfile the queries that are going to be executed moments later. He parses the queries and constructs new select queries that populate the cache with the data that speeds up the upcoming writes. He claims, if I remember correctly, a three to four times speed-increase. Wow.