> In theory, I see no reason why TCP couldn't be the medium, but > IUCV doesn't support that type of connection.
Yeah, then AVS or TSAF are probably your only other options. There's a nice market niche there for someone to create an IP-based 3088-like device. It'd be hard, but very, very useful. Just thinking out loud, you could do a RTSP-based implementation that would provide the real-time transport component (CTC response is not all that different from isochronous voice traffic, so you could probably borrow some ideas from that, and RTSP lends itself nicely to multistream processing (so you could do multiple subchannels over the same connection). The hard part would be the actual channel interfaces at the mainframe end, although if you assume it's only for 2nd level systems, you could probably do something useful with a VCTC driver in a Linux guest. -- db
