* Maintenance is generally easier. Patches can be tested in a separate instance, then swapped in to take over a production instance.
* Rapid development time, new software can be tested without having to reboot the kernel.
* More persistence in general, if one instance goes hay-wire, it is often possible to substitute it with an operational mirror
* Rapid development time, new software can be tested without having to reboot the kernel.
* More persistence in general, if one instance goes hay-wire, it is often possible to substitute it with an operational mirror