Packet Generator

yes tiamat. i have taken a look at their site already. and guess what, this was their recent venture :
"Calabasas, CA, October 27, 2005 - Ixia (Nasdaq: XXIA), a leading, global provider of IP network testing solutions, is pleased to announce the opening of its new offices in Kolkata, India. Housed in the Infinity Tower II in the hub of the city's IT sector, this new Ixia facility will be a local center for software product development, designed to respond to the high demand for the Company's innovative network testing products. "
that is where i am from, and Infinity Tower is exactly 10 minutes from where i lived!!!

the protocol is designed to provide both kind of services, connection oriented and connection-less, yes both!!!
it has congestion control, error checking, leaf node processing, and many more interesting things which was not possible earlier. next step is to code the OS for a router which can deal with such a protocol

What he said.

have found a nearly workable approach towards planning a packet structure. its half the size of the conventional packets structures used nowadays, and it still carries more information than what is being done now. how i made it possible was to use the available bits to alternatively carry information with some preconditions. if its option A, then it carries a specific info, if its option B, then the same sequence of bits will carry some other option dependent info. helps to keep the header length same. and this is designed for the connectionless services, for connection oriented services, the header length is even smaller. but still ideally it would carry enough info to manage a resource hogging application at both ends.

I knew that, I was just waiting for you to say that.

using basic C++ routines to generate a packet and then another routine to process it. so for now its just one packet at a time. the packets are not randomly generated. i am deciding what packets to throw at the processing module, and let it find out what kind of information it can gather and in what time. next step would be generating random packets, but still one packet at a time. another step forward would be generating random packets automatically, and see how fast and efficiently the algorithm can process the packets and forward them onto the respective outgoing interfaces