How should browser based muiltiplayer game state be syncronized using WebSockets and a Node.js backend?

Question

I'm in the process of learning how to create browser based html5 games. I'd like to eventually be able to achieve a game with mechanics very similar to that of the game agar.io. I've been observing coding streams and doing a fair amount of reading on the subject of network latency, jitter, and buffering but I'm still having difficulty ironing out an appropriate aproach to coding a node.js implementation to handle these things. I'm not so much as interested in having a solution provided to me as learning how more expierenced developers code tick counts, frames, etc...

Thank you all very much for your time and I very much look forward to any input your willing to provide regarding the subject. :)

Answer 1

Node.js & the event loop :
To prevent flooding the action queue by multiple users playing the same game, I would divide the MMO into game rooms or game regions / zones. Dividing the game server into independent threads that could even run on independent machines.

A login socket server for handling new incoming clients
with a db reader & writer process for db queries & updates
For each game room available a zone socket server (for communicating with clients that are interacting in that specific zone : action requests & action result emission)
with a corresponding forked zone manager process
(for validating requests & result generation)

Maybe you like some personal notes...
(maybe a chance for this model to be reviewed & evaluated by the opinions of other devs)

<client>        Socket.emit('login_request', {credentials}) 
<server>        Receive login request [send to db_rwp]
<db_rwp>        Validate login_request {credentials} :: if valid fetch client (user) data & insert entry in active_sessions db table
<server>        Evaluate user.region : prepare region for client (ipc between procs with the folowwing data) {user:'{userData}', secret:'%ha$h%', csid:'login_handshake'}
<region>        Store @ incomingClients['%ha$h%'] ) = '{userData}'
                Inform to login.server {action:'prepared', region:'region_id', csid:'login_handshake'}
<server>        Inform Database then client 
<db_rwp>        UPDATE sessions WHERE socket_id = login_handshake 
<server>        Now inform the client :: io.to(data.csid).emit('valid_login', user:'{userData}', region:{id:'region_id', port:'region_port', secret:'%ha$h%'})
<client>        Load region splash screen while socket disconnects from <server> and connects to <region> (server:region.port) 
<region>        On client connect, return a client probe (requesting the secret hash)
<client>        Client is probed, return a Socket.emit('region_snapshot_request', '%hash%');
<region>        if incomingClients['%ha$h%'] is found valid
                    return the snapshot to the incomingClient
                    notify the incomingClient to the activeClient[]
                    activeClient['region_handshake'] = incomingClients['%ha$h%']
                    then destroy the incomingClients['%ha$h%']
                    Inform to login.server {action:'completed', region:'region_id', old_csid:'login_handshake', csid:'region_handshake'}
<server>        Inform handshake changes to Database
<db_rwp>        UPDATE sessions SET socket_id=csid WHERE socket_id=old_csid

Here I gave an answer that contains a client model that will only render what the server emits.
https://gamedev.stackexchange.com/a/116531/52482