Thanks for the input!
The model isn’t designed to address the issue with Australia, the are other reasons for it, political and practical. Azure isn’t necessarily the answer it’s available and there is funding/support readily available to test it as an approach. Nobody is averse to pivoting/reverting if it proves not to be right/best.
In essence the question is about how it might be possible to deal with a globally distributed user audience when Moodle’s normal setup is a single point of presence.
The latency to Aus is a given, it’s well documented and best case scenario for a round trip from the UK you are looking at getting on for 800ms or so. I think infrastructure may also be an issue as it can also fluctuate up from that.
Hosting in the UK It makes Australia the poor relation and when users see how performant it is from the UK in comparison it’s not great. It’s the TTFB that gives the perception and pages taking an extra second or more to start rendering is the issue. We’re well aware that further optimisation of infrastructure and code may eek away at that.
Unfortunately hosting more centrally, to share the extra latency. may not be an issue due to data residency requirements.
If we discover anything that may be useful it will be shared back. It sounds like nobody has yet uncovered the golden bullet but there are some great techniques being applied to squeeze out every last ounce of improvement!
Thanks again everyone!
Simon