MapServer performance

[Ed McNierney]

Antti -

Does that mean there are other applications running on the same server, of
which you are unaware?  That's generally a very good hardware setup, unless
you're competing with other apps.

The only other obvious question is to check your overviews against the
images you're requesting.  Are your test WMS requests asking for image
resolutions for which overviews are available, or is MapServer doing a lot
of resampling for you?  Resampling can be very expensive.

     - Ed

-- 
Ed McNierney
President and Chief Mapmaker
Maps a la carte, Inc. / TopoZone.com
73 Princeton Street, Suite 305
North Chelmsford, MA  01863
Phone: (978) 251-4242
Fax: (978) 251-1396
ed at topozone.com

> From: antti siukola <siukola.antti at GMAIL.COM>
> Reply-To: antti siukola <siukola.antti at GMAIL.COM>
> Date: Tue, 27 Mar 2007 13:51:26 +0300
> To: <MAPSERVER-USERS at LISTS.UMN.EDU>
> Subject: Re: [UMN_MAPSERVER-USERS] MapServer performance
> 
> Hi!
> 
> I finally got some specs for our server. We have an vmware Virtual
> infrastructure 3 with fysical hardware of Dell PE 6850 dual dual-core
> cpu with 12 GB of RAM. Hard drives are EMC CX3-20 SAN system. And the
> system allocates 1GB RAM for us.
> 
> Antti S
> 
> 
> On 3/27/07, Romolo Manfredini <romoloman at hotmail.com> wrote:
>> From: "Rahkonen Jukka"
>>> My test system has 3 GHz prosessor and 1 GB of memory.  I have on an
>> external Firewire 400 unit worth 2 TB of imagery, consisting or
>>> individual 300 MB GeoTIFF images.  Those are uncompressed and with
>> overviews, but untiled.  I have made layers out of those by creating
>>> tileindices. When I run test with Jmeter by sending simple, unique WMS
>> request hitting 1-4 original GeoTIFF files the throughput is around 200
>>> requests per minute.  The size of the output image is 500 by 500 pixels
>> and the requests are tuned to suit the original resolution.
>> 
>> You can try performing Jmeter test against a very simple php script that
>> output a map, so you can understand the phpmapscript overhead.
>> I'm not in condition to perform such a test on my high-end server and the
>> test server is really downsized so i would not take my results as reliable,
>> sorry.
>> 
>> Romolo Manfredini
>>