I can hold my own in the basic tech world but have little knowledge of asterisk and so my inquiry. I have a text to speech app running a couple year old version of asterisk on Centos on a server in my phone room. There is a separate Windows server running some coldfusion apps that are a part of the whole. Is it possible to combine and move all of this in to the cloud in a cost efficient manner…? I have heard of latency issues and the like and don’t know what to believe so thought I would pose the question here. My apologies for not being more detailed.
I have Asterisk running on a VMware server so it works in a virtual environment. If you find a provider who uses VMware it should be fine. There are several Physical to Virtual conversion tools out there as well that might help you move your system as they are unless you fancy a complete re-install to bring things up to date.
Running Asterisk in a virtual environment reliably requires that you have control of the host machine. You must ensure that the machine is only lightly loaded and you need to set the parameter that controls the maximum CPU scheduling delay, preferably to make it significantly less than 20ms.
Getting the former wrong will cause audio breakup, particularly in music on hold and announcements. Getting the latter one wrong will result in the Asterisk system disproportionately degrading the jitter on the audio stream. (The phones may compensate by increasing the latency.)
I rather suspect that nearly all VM installs of Asterisk have poor jitter characteristics.
In general, VM systems are not designed for applications with the real time constrains associated with VoIP, even though a lot of people report using them successfully.