llVolumeDetect vs llSensorRepeat over large areas

I'm thinking about security sensors at the moment, and the question of the most efficient and least laggy solution is of course important (the last thing you want to do is bog yourself down with your own security devices).

These have to be scalable - I don't have much land myself, but some people who might be interested might have. llSensorRepeat will provide information over a 96m radius sphere, which is a fair size. Would linked prims with the same volume - c. 29,000m3 - with llVolumeDetect have a speed penalty? How well does llVolumeDetect scale up to large sets of linked prims?

I can see that llVolumeDetect would be better at catching fast-moving drones without low scan intervals, and you can have one script for everything without having to have a server arrangement with multiple client drone scripts... which makes it seem ideal. But if it's not scaling well then it could be worse, and I don't have any metrics to go on.

Well linking prims at that distance, especially lots of them becomes a little problematic. But assuming you want to take the time and endure the tedium of hacking them together, it would be a better way of detecting things while not using active scripts.

However for your detection area to cover the same area as a 96 meter radius detection sphere you would need roughly 290 prims. also, prims, even invisible ones, tend to get in the way of clicking dropping and other things you might want to do in that covered area. Also prims, invisible or not, do i believe have an effect on rendering, and prims themselves, even when not active, do have some small effect on the server.

So, my answer, though not necessarily the final word, is that volume detect linked prims are only better for smaller detection areas that don't have to move or need to cover more complicated areas than spheres precisely.

Yes, thinking a bit more about, a whole bunch of invisiprims would really get in the way of any building at ground level - though they might be okay at high altitude.

Linking a whole bunch of prims to cover a wide area would be a pain, but it would only need to be done once. Also, not so much of a pain if you link sequentially - i.e. create one 10x10x10 cube, shift-drag it to four other locations to get a 20x20x20, link those together, shift-drag that to get a 40x40x40 and so on.

It should also be noted that if you Sit Here on an object and pass through the llVolumeDetect'ing prim to get to the seat a collision will not register until you stand up.

good point. Volume detect prims will also not detect non physical objects. an object must be moving physically for it to "collide" with a volume detect prim.