wraggster
May 9th, 2006, 19:12
Sgstair posted this news:
I thought I might mention that I’ve put some thought into the new DSTunnel protocol, and have written up some basic documentation on how I want the protocol to work, what kind of messages will be present, and how the system should behave. When I get some time I’ll be properly putting this into code, and hopefully getting a bug-free implementation going this time ;) - some of the premises of the protocol have changed since last attempt, this one will act as a cross between a true p2p udp network and a server-client network, based on configuration and network ability… it will determine whether a p2p or client/server system is more effective based on connection speeds and lag times, and configure the network in an attempt to create the best performance for an arbitrary group of peers.
Additionally, since I know a bit more about what I’m up against now, I’ve designed some mechanisms in the protocol that will make managing some aspects of timing a lot easier… no idea if they’ll work as well as I hope, but they should at least work better than last attempt (which uh, really screwed a lot of timing stuff up) - Also I’ve designed the interface between the transport protocol and the hardware interface to be a lot more streamlined, and I will probably introduce a simple plugin system so people can write their own custom servers easily.
All this when I have time….
More info --> http://blog.akkit.org/
I thought I might mention that I’ve put some thought into the new DSTunnel protocol, and have written up some basic documentation on how I want the protocol to work, what kind of messages will be present, and how the system should behave. When I get some time I’ll be properly putting this into code, and hopefully getting a bug-free implementation going this time ;) - some of the premises of the protocol have changed since last attempt, this one will act as a cross between a true p2p udp network and a server-client network, based on configuration and network ability… it will determine whether a p2p or client/server system is more effective based on connection speeds and lag times, and configure the network in an attempt to create the best performance for an arbitrary group of peers.
Additionally, since I know a bit more about what I’m up against now, I’ve designed some mechanisms in the protocol that will make managing some aspects of timing a lot easier… no idea if they’ll work as well as I hope, but they should at least work better than last attempt (which uh, really screwed a lot of timing stuff up) - Also I’ve designed the interface between the transport protocol and the hardware interface to be a lot more streamlined, and I will probably introduce a simple plugin system so people can write their own custom servers easily.
All this when I have time….
More info --> http://blog.akkit.org/