Re: Implementation Proposal For Add Free Behind Capability

Amit Kumar Khare wrote:
> --- Neil Padgett <npadgett(at)redhat(dot)com> wrote:
> > On Sat, 2002-02-23 at 08:30, Amit Kumar Khare wrote:
> >
> > > Considering thesesuggestions as our baseline we
> > have decided to
> > implement Most Recently Used(MRU) cache replacement
> > policy. Our task
> > will be to identify a large sequentialscan and then
> > advice the buffer
> > manager to apply MRU for cache replacement.
>
> Neil wrote:
> > This could lead to poor buffer performance in the
> > case of multiple
> > backends simultaneously executing queries. How will
> > you handle this?
>
> Certainly it's cause of concern. I am looking for
> literature regarding it. I kindly request you all to
> inform me if any body finds some thing related to it.

My guess is that we should test MRU for large sequenatial scan vs. LRU-K
and see how the compare in various tests.

--
Bruce Momjian | http://candle.pha.pa.us
pgman(at)candle(dot)pha(dot)pa(dot)us | (610) 853-3000
+ If your life is a hard drive, | 830 Blythe Avenue
+ Christ can be your backup. | Drexel Hill, Pennsylvania 19026