The reason I created this is to:
1. Tackle my challenge of scalability.
2. I love Minecraft and have seen it being non-scalable for a long time, without people doing anything about it.
3. I want to remove the hassle of you having to setup your own server, with all the installation, and plugins, etc.. Maybe you and your family can just hop in and play, and things will be kept for as long as Minecraft is alive.
4. The already setup version of Minecraftly has a disadvantage that it is limited in terms of plugins. I only choose cosmetics, fun plugins to install in this project, so actually you can't choose plugins, unless you use the open source code and host it by yourself. It's a little like Wordpress.com's add-ons vs wordpress.org self hosted version
That is how I view it. :)
To avoid crazy costs I setup a bit of infrastructure: I managed to pipe the active / logged in user count of the minecraft server to cloudwatch: 30 mins of server in-activity and I had the server shutdown (to avoid paying for it to run 24x7). Additionally I setup a webpage that he could login to (only used basic auth to be honest) and click a "start server" button. This sent the ec2 start instance command to AWS. Granted he didn't have instant play capabilities (had to wait for server boot up) but the costs were quite low.
I personally installed any mods or plugins he wanted.
I'd highly recommend rolling your own setup like this to anyone interested :)