Yeah, you are right, if you want to finish your thesis topic and learn something, starting from simple is a wise choice. You may consider pomelo when your project scale up. Hoping the pomelo example and tutorial can still inspire you.
On Friday, January 11, 2013 12:51:15 AM UTC+8, DerDree wrote: > > Thanks for your suggestions. Sure I will give the framework a try but Im > quite forward with my game so I dont want to start at the beginning again, > so I need to figure out how to combine it with my actual version. > Also like I said it is possible that I want to make this project to my > thesis topic in university where own implementations and deep inside are > requested, so a ready to work framework may not be that good for it altough > it seems very appropiate. > > So I think next to trying the framework out I will still try to implement > a simple version of it by my own just to study it. When I fail its not that > bad because of the framework, but I think it can help a lot learn some > things for it. > > As to your statements to my thought of a simpler solution: > I am not that familiar with multiple socket.io processes, but Pomelo as a > framework states that the complex part is connecting the different servers > and that players on different servers can interact in an mmo for example. > In my game totally seperated processes would be okay, so that for example > gameroom 1 - 100 are processed by process1 and the next 100 by the next > process and so on. I thought that once the client is connected to the right > socketserver-process the connection always stays with this process and > there is no need for me to change a server for a client or to communicate > between these because my gamerooms are totaly seperated. So a gameroom is > using one socket-process and when the client connects to this gameroom he > connects to the same socket-process. I hoped that this would need a single > routing to the right process at the beginning and then the connection to > the right process would be established. > > But as I stated I am not that experienced with it. Maybe connecting to the > same process as the gameroom is is not that easy as it sounds first. But if > anyone has some good readings for this I would be very happy going a bit > deeper into this topic. > > > On Thursday, January 10, 2013 10:56:16 AM UTC+1, Charlie Edward wrote: >> >> Well, I do not think the simple version is as simple as you think. Game >> is more complex than web application in some aspects, you have to take >> care of a lot things, especially in multi-process environment. Servers >> communication, route, session management, broadcast, request/response, >> package parser, channel assignment, dynamic server extension, all these >> things will make your simple version game not simple at all, and error >> prone. >> >> Using a framework is a much more reasonable way, actually your >> requirement is really close to pomelo. Take a try, really less code, and >> much more scalable. Github: https://github.com/NetEase/pomelo >> >> On Thursday, January 10, 2013 5:02:26 PM UTC+8, DerDree wrote: >>> >>> Thank you both very much. That are two nice solutions for this kind of >>> problem. >>> >>> I studied the solutions and am I right that the main idea behind these >>> is to instantiate more processes for handling the websocket messages? This >>> was also one of my ideas aboth. Altough I would like to try programming it >>> by my own in a simpler version rather than using a full framework. The main >>> functionality of my game is quite far in development and maybe I will use >>> this whole project for a thesis during my study. But these two projects >>> give a great inside in managing such thinks, that is very usefull. >>> >>> I also think that my particular game can work with a simpler solution >>> just now. I have completly seperated games (with each up to only 8 players >>> max) and there is no need to communicate between the different games. So I >>> think I just could try to open another process at a specific number of >>> games going on and then sending all new games to this new process until the >>> first process is not anymore at the limit. I just hope that the number of >>> games a process can handle will be acceptable so that there doesnt need to >>> be too much processes for a few games. >>> >>> But my concern would still be that, altough there are multiple processes >>> handling the messaging, it still would be only one machine. Are there any >>> studies on how many messages can be send in a second or half using multiple >>> processes on one machine? I could imagine at some point the machine reaches >>> its maximum network load, making more processes useless. Than this really >>> would need some complexer solution where maybe these frameworks would be >>> really handy. >>> >> -- Job Board: http://jobs.nodejs.org/ Posting guidelines: https://github.com/joyent/node/wiki/Mailing-List-Posting-Guidelines You received this message because you are subscribed to the Google Groups "nodejs" group. To post to this group, send email to nodejs@googlegroups.com To unsubscribe from this group, send email to nodejs+unsubscr...@googlegroups.com For more options, visit this group at http://groups.google.com/group/nodejs?hl=en?hl=en
