I have been using an ‘intelligent’ Sharp Ethernet NIC wherein the ‘driver’ code resides and runs on a Hitachi RISC processor. My host communicates with it via an on-board dual-port RAM. This smart card eliminates the need to perform the low-level functions that the OSI transport, network, data link and physical layers do. It’s a bit bizzare, perhaps, but the product works. We are presently connected to a POS system that uses UDP/IP as its message protocol. We wish to move to TCP. However, the initialization of the TCP/IP stack is unfamiliar to me. To initialize this card for UDP is simple: I call the Init, Socket_UDP, Bind and then Receive_UDP functions via the DPR. When the Receive_UDP function times out, I simply call it again. My/their application takes care of checksumming and handshaking (retry, etc.). Does anyone know where I can obtain more in-depth info about the TCP-specific functions? We can just allow the user to program a static IP and port# into our device but I’d like to be able to support DHCP. Oh, the device is centered around a NEC V20 running the 8086 subset, written in assembly language. I only have a 27256 PROM to store the code in; hence the need for a ‘smart’ NIC – no room for TCP/IP stack code in the 32k PROM. I’d appreciate some direction or mentoring in the low-level (application side) manipulation of the transport layer (TCP, in this case). Thanks a lot!!!