[mapguide-users] RE: Tooltip performance

[Trevor Wekel]

Yes.  I can take the lead on this one.  I should be able to start work
on the RFC next week.
 
Trevor
 
  _____  

From: Jason Birch [mailto:Jason.Birch at nanaimo.ca] 
Sent: Friday, November 17, 2006 12:21 PM
To: dev at mapguide.osgeo.org
Cc: Trevor Wekel
Subject: RE: [mapguide-users] RE: Tooltip performance


Yes, I think so.
 
Are you willing to take the lead on this?  I don't mind setting up the
RFP, but don't have resources to allocate to the change.
 
Jason

  _____  

From: Trevor Wekel [mailto:trevor.wekel at autodesk.com] 
Sent: Friday, November 17, 2006 11:15
To: Jason Birch; users at mapguide.osgeo.org
Subject: RE: [mapguide-users] RE: Tooltip performance


Hi Jason,
 
It looks as though we will need some API changes (HTTP Protocol, MgMap)
to implement this.  Since API changes are involved, should this be
handled as an RFC?
 
Thanks,
Trevor
 
 

  _____  

From: Jason Birch [mailto:Jason.Birch at nanaimo.ca] 
Sent: Friday, November 17, 2006 11:39 AM
To: users at mapguide.osgeo.org
Cc: Trevor Wekel
Subject: RE: [mapguide-users] RE: Tooltip performance


That seems sub-optimal.  
 
- It's spending extra cycles on layers that don't need to be queried.
- What happens if I have a layer which is not selectable, but for which
I want to return a maptip?
 
What would it take to change this?
 
Jason

  _____  

From: Trevor Wekel [mailto:trevor.wekel at autodesk.com] 
Sent: Friday, November 17, 2006 10:34
To: Jason Birch; users at mapguide.osgeo.org
Subject: RE: [mapguide-users] RE: Tooltip performance


Hi Jason,
 
Looks as though we only use the selectability property currently.  The
routine is shared code with selection handling.
 
Trevor

  _____  

From: Jason Birch [mailto:Jason.Birch at nanaimo.ca] 
Sent: Friday, November 17, 2006 11:17 AM
To: users at mapguide.osgeo.org
Subject: RE: [mapguide-users] RE: Tooltip performance


Is it just layers that have had their selectability property cleared
that will be ignored, or will not setting a maptip have the same effect?
 
Jason

  _____  

From: Trevor Wekel [mailto:trevor.wekel at autodesk.com] 
Sent: Friday, November 17, 2006 10:15
To: Campbell, Keith A; users at mapguide.osgeo.org
Subject: [mapguide-users] RE: Tooltip performance


Hi Keith,
 
You can also disable selection for layers which do not require tooltips.
If a layer is not selectable, it is not included in the query.
 
Thanks,
Trevor

  _____  

From: Trevor Wekel 
Sent: Friday, November 17, 2006 10:54 AM
To: 'Campbell, Keith A'; users at mapguide.osgeo.org
Subject: RE: Tooltip performance


Hi Keith,
 
You are correct with the basic flow of events for generating the
tooltips.  The AJAX client sends the map coordinates of the point along
with the map name to the Web Extensions.  The Web Extensions in turn,
forwards this information to the Server.  Once at the server, all of the
layers in the map are queried to see if they have any features
underneath the point.  Since every layer is queried, the performance
will depend on the number of layers in the Map.  Each FeatureSource for
each layer will be queried.  If you have a number of layers from
external databases, the database will be queried for each layer.
 
Feature Service connection pooling and caching may help.  These are
controlled by the following settings in serverconfig.ini and should be
accessible from Server Admin.
 
[FeatureServiceProperties]
DataConnectionPoolSize = 20
DataConnectionPoolExcludedProviders = OSGeo.SDF,OSGeo.SHP
DataConnectionPoolEnabled = 1
DataCacheSize = 100
CacheSize = 100
 
Just as a side note, I think MapGuide 6.x handles tooltip generation
client side.  I don't think it goes back to the server so it will be
less affected by server load.
 
Thanks,
Trevor

  _____  

From: Campbell, Keith A [mailto:keith.campbell at atkinsglobal.com] 
Sent: Friday, November 17, 2006 6:53 AM
To: users at mapguide.osgeo.org
Subject: Tooltip performance



I'm finding that the performance of the tooltips is varying greatly and
I'd like to understand what is going on behind the scenes and therefore
what factors are affecting this.

Presumably when the cursor s moved over the map, a mouseover event is
triggered and this sends the position information and state of the
layers is sent back to the server via an xmlhttprequest. The web
extensions are used to identify what object on what layer is at the map
location, and the tooltip data is queried and returned to the client,
which is then displayed. So factors affecting this process will be what
is happening on the client machine, the network, and the server at the
time the process occurs.

The performance seems to change a great deal from very good to
unacceptably slow (the user would give up). I find this change strange
because other applications (e.g. MapGuide 6.x) run from the same client
to the same server are not varying nearly as much. It is also strange in
that the initial request for the layout and subsequent requests for
layer data are not particularly slow when the tooltips are. Are there
any settings which can be altered to improve this? This kind of
unreliability could almost rule this technology out in my view.

Keith 



This email and any attached files are confidential and copyright
protected. If you are not the addressee, any dissemination of this
communication is strictly prohibited. Unless otherwise expressly agreed
in writing, nothing stated in this communication shall be legally
binding. <http://www.blackspider.com/> 

-------------- next part --------------
An HTML attachment was scrubbed...
URL: http://lists.osgeo.org/pipermail/mapguide-internals/attachments/20061117/5f706510/attachment.html