Re: [twsocket] OT rather big project question

2006-01-03 Thread Dod
ttp://www.overbyte.be FP> - Original Message - FP> From: "Dod" <[EMAIL PROTECTED]> FP> To: "ICS support mailing" FP> Sent: Tuesday, January 03, 2006 10:41 AM FP> Subject: [twsocket] OT rather big project question >> Hello, >> >&g

Re: [twsocket] OT rather big project question

2006-01-03 Thread Francois Piette
t;ICS support mailing" Sent: Tuesday, January 03, 2006 10:41 AM Subject: [twsocket] OT rather big project question > Hello, > > I am starting a re-write one of my ICS-based server to be able to > handle 15.000 permanent cnx smoothly. The datas Clients will > send

Re: [twsocket] OT rather big project question

2006-01-03 Thread Dod
Hello Wilfried, Yes, I also thought that concentrator-to-server could be direct TCP-IP or any other kind of protocol (HTTP tunneling for example), also I found other advantage, I could make concentrator-to-server connection initiator, or server-to-concentrator way so this would be nice whenever th

Re: [twsocket] OT rather big project question

2006-01-03 Thread Wilfried Mestdagh
Hello Dod, At first tought I should go for first solution because of a few advantages. - server updates as you say yourself - application is splitted separate according the job it does. The concentrators dealing with the clients and the proto, between concentrator and main server you can exchang

[twsocket] OT rather big project question

2006-01-03 Thread Dod
Hello, I am starting a re-write one of my ICS-based server to be able to handle 15.000 permanent cnx smoothly. The datas Clients will send/receive will be only 2KB packets time to time. As Windows OS is limited with sockets and handle (whenever you can do some tricky changes in re