Best conventions for using kryonet

Im setting up a multi server network all managed by one server, the way it will work is the ‘Core Server’ will handle how all other servers communicate and will handle any messages between then, then the lower level servers will connect to the clients…is it bad practice to have a server that is also a client to a larger server?

It depends. :clue:

Well it works, so i guess that’s the only thing that matters xD

That should really be your concern until you start running into problems. Once you get down to more specific cases you can test a few things and see where it gets you.

Well im not going to design a server system to fail on launch xD I’m making a sturdy server system to start of with so i don’t need to worry about re-writing it again and again in the future

You’ll probably not be able to accommodate for a large volume regardless of what you do with your server structure anyway, unless you’re sitting on piles of money. I’m not saying design to fail, I’m saying design within constraints that are workable, don’t aim for the best implementation at launch, because you won’t know where you need to pour resource until later, and you might find that something works better for your specific demands down the road.

As a sort of related example, this is the server structure for a game called Osu!:

They have all that working behind the scenes for a rhythm game with ~20k peak players at best probably.

I’m using a service called Digital Ocean and what they let you do is set up server instances on the fly as you need them (using there api) and take them down when they are not required, and this is handy because the server instances are just ISO snapshots of the current Linux state so when i need a new server, it will already have my Java software running.