Mainframe Linux Mythbusting Was: Using Java in batch on

in hsdt

i had one effort to ship a peer-to-peer networking operation that had ability to emulate 37xx (hardware-software) around the edges. it had originally been implemented on S-1 boxes by one of the RBOCs. i was going to move it to ha-cmp rs-6000 configurations.

the trivial use of real (peer-to-peer) network provided the ability to did a large number of things for sna configurations that were impossible in real sna operation.

dominant issue was that it would have severely impacted the communication groups 37xx revenue.

i accidently once got invented to give a presentation on the project at an SNA ARB (architecture review board) meeting in raleigh. afterwards, the person running ARB was very agitated and wanted to know how it was that I had been invited ... I think he wanted to make sure it never happened again. part of the problem was that it seemed that most of the people (or at least most of the technical people) in the audience wanted to stop work on what they were doing and come work on what I was doing. That, of course, wasn't allowed to last long.

old thread discussing some amount of that effort:

the post

includes part of the presentation that I gave at the ARB meeting. it gives both a functional and a business cost analysis for compareable configurations (based on series-1 hardware ... not rs-6000 hardware).

disclaimer: as an undergraduate ... there was some things that I wanted to do with 2702 telecommunication controller ... that it turned out couldn't be done. somewhat as a result, the university started a project that reversed engineered the mainframe channel interface and built a channel interface board for an interdata-3 ... which was programmed to emulate a telecommunication controller. there was some write-up blaiming four of us for spawning the clone controller business