I have been playing around with ArcIMS image overlays in Google Earth.  I like the ease of laying a single image down over the viewing window, but I don't like the loss of continuous panning that I get when I move the globe just a bit.  Whatever happens when the camera stops GE goes out and makes the new request for the image.  There is no buffer.  I can make the image bigger, but that makes the actual image less detailed.  I thought I would explore mutliple network links wrapped in a GE Folder.  Give this a try:  http://www.mapdex.org/GE_KML_LINK_TILE.cfm?servername=hercules.kgs.ku.edu&mapservice=lawrence_Imagery

I have fused together 4 network links and taken advantage of the cookie element on the networklinkcontrol element. 

I have broken up the single overlay into four parts.  It is sort of hack that allows continuous panning (because each image extent goes farther than what GE requests).  I use the cookie tag as a way to use the previous image if the user only pans a little bit.  It still seems a bit tacked on...And I am not sure the extra download time is worth it.  

What do you think? 

FYI:  Here is the same service with the traditional single image overlay.

http://www.mapdex.org/GE_KML_LINK.cfm?servername=hercules.kgs.ku.edu&mapservice=lawrence_Imagery

Jeremy