From: joel garry <>
Date: Mon, 15 Dec 2008 14:15:40 -0800 (PST)
Message-ID: <>

On Dec 14, 6:31 pm, Michael Austin <> wrote:
> hpuxrac wrote:
> > On Dec 13, 9:26 pm, Michael Austin <> wrote:
> >> Can anyone share performance impressions when using ASM with SVC or even
> >>   9i(yeah we know...) cooked file systems?
> >> We currently have the impression that SVC can only return data in 32k
> >> chunks regardless of I/O setting at the OS (Solaris, AIX and HPUX) and
> >> or storage parameters in Oracle causing additional I/O traffic.
> >> Just curious.
> > Back in my old IBM days SVC was a SuperVisor Call ... what are you
> > talking about exactly?
> > Is this some kind of storage?   ( We are an EMC shop ... so no clue ).
> ahh yes.. more acronyms...
> SVC=SAN Volume Controller.
> I am just trying to figure it out but have been told you can think of it
> as an array agnostic filer - similar to NetApp, but will work with any
> storage.  Essentially this "controller (cluster)" is a gateway between
> your server and your storage whereby you can have any vendors arrays
> behind it and it presents logical LUNS to your server and it handles all
> of the I/O.
> Do a quick google search on IBM SVC architecture.

OK, I did a quick google on that along with transfer size.

I would be bummed if it could do 256K reads but only returned 32K...

And further down:

"So this is just FUD. To re-iterate. This was a generated 70/30 mixed workload. So 70% read, 30% write, with a block size of 4K. This is industry accepted as a pseudo-typical database workload - and has been for many years."

Putting this into context is left as an exercise for the student. As is why someone would test raid-0 because raid-10 isn't done yet, and admit it.

This may be cool stuff, but too much DKBness for my tastes. And besides that, I'd rather know where my data is.


-- is bogus.
"not picking on HP, but we do seem to migrate off a lot of these
boxes" - Barry Whyte
Received on Mon Dec 15 2008 - 16:15:40 CST

Original text of this message