On 04/10/2011 11:49 AM, Rafael Martinez wrote:
> On Sun, 2011-04-10 at 09:34 -0500, Joshua Berkus wrote:
>> ... which is coming up, or is already here. Can anyone find the date of the orignial POSTGRES paper in 1986?
> In the "The POSTGRES Data Model" paper , there is a reference to "The
> Design of POSTGRES" paper, it was presented in a conference in june
> "M. R. Stonebraker and L. A. Rowe, ‘‘The Design of POSTGRES’’, Proc.
> 1986 ACM-ACM-SIGMOD Conf. on Management of Data Int. Conf. on the Mgt.
> of Data,June 1986"
The paper is available at
http://citeseerx.ist.psu.edu/viewdoc/summary?doi=10.1.1.48.2909 ; click
on the little PDF logo underneath of "CACHED:" on the right hand side.
Fun trip back into the 80's. "The major cost of the storage system is
the requirement to push dirty pages of data to stable storage at commit
time"; some things never change. I'm also amused how VLSI acceleration
is treated exactly the same way as GPU offloading is now: as a
potential accelerator for something, but exactly what isn't clear.
The 1987 "The Design Of The Postgres Storage System" is a more
interesting read to me. Much of the transaction model and even VACUUM
all go back to there. Also amusing to see them bash on WAL based
approaches given how things evolved...
Greg Smith 2ndQuadrant US greg(at)2ndQuadrant(dot)com Baltimore, MD
PostgreSQL Training, Services, and 24x7 Support www.2ndQuadrant.us