Paul The trouble with your assumption that leads to the "bug" idea is that you are not comparing like with like. If your IP-attached[1] workstation[2] were to be a regular SNA session to TSO using the 3270 data stream[3] where the data link control layer was a Rapid Transport Protocol (RTP) connection mapping precisely to an High Performance Routing (HPR) connection running over the IP network, you would be comparing like with like. Actually this still also introduces another clarification which is that the coax-attached device uses a 3270 controller with an SNA path to TSO.
One point here is that, when using TN3270 where the secondary logical unit (LU), the "terminal", is provided by the TN3270 server and the SNA session path is concatenated to the TELNET connection, you have introduced some complications not present in the case of the pure SNA session. Another point is that TSO as it is today - and I rather suspect always will be - but I could be wrong - insists on using VTAM and SNA to communicate with secondary LUs. If you want to compare TSO-SNA-3270 with something equivalent connected over IP, you should use Unix System Services TELNET server-IP-TELNET client. Chris Mason [1] I'm obliged to drop "TCP" - and I'm sure you don't insist that the TCP protocol is actually used - because Enterprise Extender economises by using UDP. [2] Let's call a "terminal" a "workstation" today. [3] Let's use LU type 2 rather than LU type 0 in order to maximise the function. ----- Original Message ----- From: "Paul Gilmartin" <[EMAIL PROTECTED]> Newsgroups: bit.listserv.ibm-main To: <[email protected]> Sent: Wednesday, 17 January, 2007 4:58 PM Subject: Re: your mail > In a recent note, [EMAIL PROTECTED] said: > > > Date: Tue, 16 Jan 2007 08:08:37 -0500 > > > > >That design is just plain wrong. Whatever timer expires making it > > >impossible to reconnect ought also cancel the session which has > > >become inaccessible and is only an obstacle to logging in again. > > > > No, they should change the code so that it becomes possible to > > reconnect. Is there an outstanding requirement? > > > I'd consider it a bug rather than request for new function: It > works for coax-attached terminals; I'd assume that it's supposed > to work likewise for TCP/IP attached terminals. ---------------------------------------------------------------------- For IBM-MAIN subscribe / signoff / archive access instructions, send email to [EMAIL PROTECTED] with the message: GET IBM-MAIN INFO Search the archives at http://bama.ua.edu/archives/ibm-main.html
