Oracle Database Discussions

As some of you may know, Oracle and InfoWorld announced today a bug in Oracle's SCN mechanism [http://www.infoworld.com/d/security/fundamental-oracle-flaw-revealed-184163-0?page=0,1] which could, in very, very rare circumstances crash an instance and leave it unusable for a period of time. The underlying factors appear to be (1) accessing a remote database results in the SCN of both databases being set to the greater of the two, and (2) a bug in the ALTER DATABASE BEGIN BACKUP code, whereby the SCN increases much faster than normal while in hot backup mode, and (3) this rapid increase in SCN growth continues even after ALTER DATABASE END BACKUP. The risk is that if lots of instances on a large network were all growing rapidly due to the hot backup bug, they could reinforce each other and run out of SCN headroom.

I had two reactions after reading this: (A) I did not know that remote transactions caused both instances to adjust their SCN (although it makes good sense now that I think of it) and (B) Who in their right mind uses ALTER DATABASE BEGIN BACKUP any more. Then I read that it is still useful for BCV switches.

So, my questions, since I can't test these for myself, are

i) Does RMAN bump the SCN the same way ALTER DATABASE BEGIN BACKUP does?

ii) Just how much faster than normal does the SCN increase during a hot backup.