lookout limitations and selection of the computer system for lookout

[nkrsaini]

This is regarding the limitation of the lookout i.e. upto how much IO lookout is reccomended to use.

or it is on the basis of the l4p or lks file size basis.

Do you have some criteria for the selection of the computer system for lookout on the basis of the IO's.

is there any criteria by which we can check the execution cycle time for the lookout process.

Is there any NI controller also avaiable for lookout applications?

[Mike@Forshock]

From what I have observed, there is no standard.  The best thing to do is to create test setups.

Its entirely possible that a system with 100 IO may require more processing/memory than a 5000 IO system.  Database, archiving, reporting, analytics, etc.

It is variable.

Just for reference on our systems, we dont use the IO count for performance curves, we go based on the site configurations for the system.

A site could be 1 pump, 6 pumps and a tank, 1 pump and a tank, 4 tanks, etc.  Consider than a tank may only take 5 IO and a well upwards of 50 IO each, we could hit 150-200 IO per site.

Smaller systems (<10 sites) we have run very successfully on Atom based servers (not netbooks...).

10-40 sites run well on basic dual-core workstations with 2GB or more of RAM

40-100 we recommend quad-core servers with 4GB or more of RAM.

(100+ gets a little more intricate)

We try to make sure that the system is not more than 15% loaded (memory and/or CPU) by services at normal work.  Lots of room left over for future growth or unexpected workloads and backup routines.