Maybe someone here has some idea/figures

Previous Topic Next Topic
 
classic Classic list List threaded Threaded
3 messages Options
Reply | Threaded
Open this post in threaded view
|

Maybe someone here has some idea/figures

FDominicus
on this question:
http://blogs.law.harvard.edu/philg/2011/01/10/how-many-updates-per-second-can-a-standard-rdbms-process/

See the hardware requirements.

I really am curious what Gemstone makes out of it.

Regards
Friedrich
Reply | Threaded
Open this post in threaded view
|

Re: Maybe someone here has some idea/figures

Dale Henrichs
On 01/14/2011 11:38 PM, Friedrich Dominicus wrote:
> on this question:
> http://blogs.law.harvard.edu/philg/2011/01/10/how-many-updates-per-second-can-a-standard-rdbms-process/
>
> See the hardware requirements.
>
> I really am curious what Gemstone makes out of it.
>
> Regards
> Friedrich

My post about scaling seaside (written 4 years ago) is the best info
available for GLASS:

 
http://gemstonesoup.wordpress.com/2007/10/19/scaling-seaside-with-gemstones/

The best benchmarks numbers from that post for a singe 4 cpu linux box
was 230 commits/second using fairly standard hardware. My focus has been
on making Seaside requests fast and in a later post I was able to get
somewhere around 7000 seaside requests/second (not commits/second):

 
http://gemstonesoup.wordpress.com/2009/10/18/glass-a-share-everything-architecture-for-seaside-esug-2008/ 


I don't recall what the commit rate from that benchmark was...

We have customers who have gotten up over 10,000 commits/second with
some big hardware ...

I have yet to run GemStone benchmarks on SSD drives, but since the main
limiting factor for commit rates ends up being the disk subsystem (ACID
requirements) I would think that we'd get good numbers:

 
http://gemstonesoup.wordpress.com/2009/02/28/approaching-the-speed-of-light-ssd-drives-for-gemstones/

Dale




Reply | Threaded
Open this post in threaded view
|

Re: Maybe someone here has some idea/figures

SeanTAllen
You might want to talk to Hernan Wilkinson about SSD performance.

When we were working together, he mentioned that at a previous job they had benchmarked Gemstone w/ SSD drives.
I don't remember exact numbers but it was much better performance.

On Mon, Jan 17, 2011 at 12:37 PM, Dale Henrichs <[hidden email]> wrote:
On 01/14/2011 11:38 PM, Friedrich Dominicus wrote:
on this question:
http://blogs.law.harvard.edu/philg/2011/01/10/how-many-updates-per-second-can-a-standard-rdbms-process/

See the hardware requirements.

I really am curious what Gemstone makes out of it.

Regards
Friedrich

My post about scaling seaside (written 4 years ago) is the best info available for GLASS:


http://gemstonesoup.wordpress.com/2007/10/19/scaling-seaside-with-gemstones/

The best benchmarks numbers from that post for a singe 4 cpu linux box was 230 commits/second using fairly standard hardware. My focus has been on making Seaside requests fast and in a later post I was able to get somewhere around 7000 seaside requests/second (not commits/second):


http://gemstonesoup.wordpress.com/2009/10/18/glass-a-share-everything-architecture-for-seaside-esug-2008/

I don't recall what the commit rate from that benchmark was...

We have customers who have gotten up over 10,000 commits/second with some big hardware ...

I have yet to run GemStone benchmarks on SSD drives, but since the main limiting factor for commit rates ends up being the disk subsystem (ACID requirements) I would think that we'd get good numbers:


http://gemstonesoup.wordpress.com/2009/02/28/approaching-the-speed-of-light-ssd-drives-for-gemstones/

Dale