Robert Haas <robertmhaas@gmail.com> writes:
> If we switched from per-tuple MVCC based on XIDs to per-page MVCC
> based on LSNs and a rollback segment, all of this stuff would go out
> the window. Hint bits, gone. Anti-wraparound VACUUM, gone. CRCs,
> feasible. Visibility map... we might still need that, but the
> page-level bits go away.
> Of course, it would also create new problems.
Yup, we've seen that proposal before. It's called Oracle. There's
no good reason to believe that we'd have a net win after we were done
switching over ... not to mention the likelihood that they hold a ton
of patents about particular aspects of doing things that way.
regards, tom lane