Re: oracle benchmarks on VMS

Hein van den Heuvel <hein_netscape_at_eps.zko.dec.com> wrote in message news:<3E63C5E9.88569D3F_at_eps.zko.dec.com>...
> Tim Smith wrote:
>
> > I see Oracle publishes some benchmarks at
> > http://www.oracle.com/apps_benchmark/, but notably there is nothing
> > for VMS - is that because Oracle writes to the filesystem, not direct
> > to device file files directly i.e. VMS filesystem is a lot slower than
> > raw devices?
>
> It has nothing to do with (potential) speed.
> It is just a commercial / marketing decision mostly from Oracle.
> They decided there was not enough critical mass to maintain support for
> the Oracle Applicaiton suite on VMS. The database itself is and will be
> supported at one of the higher tier levels. Details are not up to me, but
> it used to be product release on VMS 90 days after first release. Please
> verify with Oracle. They may also choose to skip 'dot' releases. Dunno,
> let's say they had 9.0, skipped 9.1 but released 9.2. Again, this is not a
> statement of support, just a line of thinking. Check with Oracle.
>
> The VMS Filesystem is actually an advantage! The VMS filesystem does NOT
> buffer data.
> On Unix systems the OS tends to waste time and memory buffering Oracle
> data pages which are better managerd by Oracle in its buffer pool (SGA).
> On many Unix implementation, for ultimate Oracle perfromance once has to
> deal with hard-to-manage 'Raw Devices' to avoid said buffering.
> On VMS you have the comfort of a file system for Alloaction, Naming and
> backups yet the speed of a raw device. On HP Tru64 Unix Oracle can (and
> will) use the DIRECT IO feature to get the same effect on single systems
> as well as in clusters.
>
> > If anyone has older benchmarks that include VMS I would like to see
> > them.
>
> It would be nice to see some VMS / Oracle benchmark, but I will not hold
> my breath.
> Benchmarks require major investment which both companies believe is better
> spend on the products itself. VMS will offer comparable (ballpark)
> performance as Unix on the same platform.
> It will not be 2x slower. It might be a little slower or a a little faster
> depending on the application.
> The performance will be close enough to focus on other, more important,
> platform decision factors: Cost-of-ownership, Reliability, Availability,
> Experience, Applications, Installed base,..
>
> Hope this helps some, but it is just an opinion from a guy in the
> sidelines.
> Be sure to contact officials at Oracle and HP for the official positions.
>
> Cheers,
> Hein.

That makes sense, and I appreciate your comments. The only problem I have is that I find that simple I/O operations are much slower on VMS than AIX. I understand that this is a complex topic dependent on disk subsystems, disk speeds, etc etc but take this for example:

I have two systems, one a Alpha 4100 with VMS 7.1 and a RS6000 H-50 both with 2GB RAM and 2 processors. These systems were purchased around the same time for the same purpose and I would term them as comparable. I did simple file copies from the same disk to same disk.  While I do not have the disk speeds, I used an older disk on the AIX (4GB) and a newer disk on VMS (12GB). I don't believe there was any fragmentation to consider.

AIX: 297MB - 2MB/second, if repeated - 2.84MB/second

VMS: 297MB - 1.02MB/second, if repeated - 1.19MB/second

I don't mean to bash VMS, I just am trying to understand if it really comes close for I/O performance. I've seen several places that have to through a lot of hardware at performance bound applications. Any time I've done SQLPlus inserts they have been a lot faster on AIX. I know that many people still prefer VMS so I am trying to figure out why...

Tim Received on Wed Mar 05 2003 - 17:08:19 CST