Subject | RE: [IB-Architect] Super-transactions and Incremental Backup |
---|---|
Author | David Berg |
Post date | 2000-06-23T17:29:42Z |
To add to Jim's comments, my experiences with using memory mapped files is
that they are not the panacea that OS developers try to paint them as. Just
the opposite.
Software that understands that it's dealing with a disk based device that
has a bias towards consecutive, sequential reads and writes can (and will)
_substantially_ outperform software that leaves such 'details' to the
operating system.
I've never seen memory mapped file access beat traditional file access, but
I've seen traditional file access outperform memory maps by an order of
magnitude or more.
Dave Berg
-----Original Message-----
From: Jim Starkey [mailto:jas@...]
Sent: Thursday, June 22, 2000 7:58 AM
To: IB-Architect@egroups.com
Subject: Re: [IB-Architect] Super-transactions and Incremental Backup
At 06:22 PM 6/21/00 -0700, you wrote:
structure) is center around fixed length pages which are read and
written atomically. Database pages and operation system pages are
quite different things.
disk I/O, only mapped files. Rather than simplifying database
implementation, in fact it made it more complex. From time to
time operation system geeks come up with the recurring idea of
a single level store -- no files, no directories, just a very
large, flat virtual space. Horrible concept.
Jim Starkey
------------------------------------------------------------------------
SALESFORCE.COM MAKES SOFTWARE OBSOLETE
Secure, online sales force automation with 5 users FREE for 1 year!
http://click.egroups.com/1/2658/6/_/830676/_/961686006/
------------------------------------------------------------------------
To unsubscribe from this group, send an email to:
IB-Architect-unsubscribe@onelist.com
that they are not the panacea that OS developers try to paint them as. Just
the opposite.
Software that understands that it's dealing with a disk based device that
has a bias towards consecutive, sequential reads and writes can (and will)
_substantially_ outperform software that leaves such 'details' to the
operating system.
I've never seen memory mapped file access beat traditional file access, but
I've seen traditional file access outperform memory maps by an order of
magnitude or more.
Dave Berg
-----Original Message-----
From: Jim Starkey [mailto:jas@...]
Sent: Thursday, June 22, 2000 7:58 AM
To: IB-Architect@egroups.com
Subject: Re: [IB-Architect] Super-transactions and Incremental Backup
At 06:22 PM 6/21/00 -0700, you wrote:
>Could you elaborate, please?
>In my mind I don't see where deleted and limbo records would be a problem.
>They just seem to fit in with all the rest of the stuff.
>
>the
>Is the paging architecture still valid in todays OS and hardware
>environments? To me at least, I wonder if perhaps operating systems aren't
>advanced enough such that the whole paging scheme can just be left up to
>OS.It has very little to do with OS paging. The database ODS (on disk
>
structure) is center around fixed length pages which are read and
written atomically. Database pages and operation system pages are
quite different things.
>Feel free to flame me but please understand that I do not know OS's at allof
>when programming at these levels. I would be very interested in hearing how
>these layers come into play when dealing with reading and writing
>information in the file system. I would also like to know the feasibility
>creating a database system that off-loaded the complexity of paging totallyThe Apollo operation system, Domain, had no concept whatsoever of
>to the OS.
>
disk I/O, only mapped files. Rather than simplifying database
implementation, in fact it made it more complex. From time to
time operation system geeks come up with the recurring idea of
a single level store -- no files, no directories, just a very
large, flat virtual space. Horrible concept.
>I think perhaps the first thing I am going to hear is that with a crosswaiting
>platform database being able to abstract this layer makes things much
>simpler. Let's put that aside if we can for the sake of discussion. Jim's
>replies would be nice but let's not all get into the habit of simply
>to see what Jim has to say, ok?Works for me.
>
Jim Starkey
------------------------------------------------------------------------
SALESFORCE.COM MAKES SOFTWARE OBSOLETE
Secure, online sales force automation with 5 users FREE for 1 year!
http://click.egroups.com/1/2658/6/_/830676/_/961686006/
------------------------------------------------------------------------
To unsubscribe from this group, send an email to:
IB-Architect-unsubscribe@onelist.com