How come Oracle reverse key indexes don't cause slow random writes?

Question

Oracle's reverse key indexes solve the problem of contention for leaf blocks in the right side of a B-tree index. But the flip side is that reverse key indexes will cause random writes. Why is in memory contention worse than disk seeking to do random writes?

edit: Microsoft advises essentially the same thing for SQL Server http://www.microsoft.com/en-us/download/details.aspx?id=26665

Answer 1

The contention stops the execution of the user session, whereas writes (apart from direct path) are asynchronous -- they are carried out by the DBWR process, and the only synchronous disk write is the LGWR writing the redo log entries.

A larger problem might be that you are more likely under some circumstances to need more of the index segment in memory, as "recent" changes are spread over a great many more blocks than on a regular index, so it might be worth keeping an eye on segment statistics, but even if the problem arises it's likely to be of a small magnitude relative to all the other stuff you could be spending your time on.