On Wed, 15 May 2019 14:59:00 +0200, R.S. wrote: >(this is mostly historical question)
It might be helpful to you to get a copy of the System/360 Principles of Operation, which describes this a little bit. You can find it on bitsavers. http://bitsavers.trailing-edge.com/pdf/ibm/360/princOps/A22-6821-7_360PrincOpsDec67.pdf The OEMI night also be of interest. http://bitsavers.trailing-edge.com/pdf/ibm/360/A22-6843-3_360channelOEM.pdf >I'm trying to understand the role of old boxes in DASD connectivity. > From IODF/HCD point of view we have >CPC-channels-CU-some_cable-DEVice In those days, there was no IODF or HCD. There was only IOGEN. There was no channel subsystem until System/370 Extended Architecture. There were no subchannels as we know them, and no HSA. XA made a radical departure in the channel architecture, but the interface between the channel and the control unit remained the same, both physically and logically. >Nowadays in real world we have >CPC-ficon-DASDbox >and DASDbox is both CU+DEV (emulated). > >In the old days there was a Storage Control Unit, i.e. 3830 and disk >controller within disk cabinet, i.e. 3350-A2 > >So, we have CPC-cable1-3830-cable2-3350A2controller-internal_cable3-disk. > >I'm trying to understand separation of duties between 3830 and 3350A2 >controller. Electronics were much larger in those days. Integrated circuits were in their infancy. As a result, the logic in the control unit took up a considerable amount of space. >What was defined as CU - it was 3830 or controller within 3350 cabinet? The control unit is the device that the channel cable connects to. It is a well defined interface from the point of view of the channel, and often did not concern itself with the details needed to access the device. The channel would, for example, send a seek command to the control unit. The control unit would communicate with the disk controller in the dasd box, which would handle the details of switching on and off hydraulic pumps and valves, etc. for the purpose of actually moving the heads. The control unit typically didn't deal with those details. >Which cable was a channel (Bus&Tag)? I guess it is "cable1" connecting >CPC and 3830. Bus and tag cables are the cables used to connect a channel to a control unit. That standard allowed other vendors to supply compatible devices with some assurance that they would work on any processor. The interface between control unit and device was not similarly standardized in general. As a result, you couldn't connect one vendor's DASD to another vendor's control unit. IIRC, the cables between control unit and dasd looked similar to bus and tag cables. I don't know if any vendors used the same cables for that purpose. >Not to mention that some old reference manual's diagram shows yet >another box between CPC and 3830 SCU. That might be a channel switch, such as a 2914. It would be connected to some number of channels, typically on different processors and some number of control units. remember that there were no LPARs in those days. The switch allowed devices to be switched to different processors. It was a manual process to switch. The device(s) had to be varied offline, then the connection was switched by turning dials, then varied online to the other processor. To further complicate things, on some early processors, the channels were located in a separate box from the CPU. -- Tom Marchant ---------------------------------------------------------------------- For IBM-MAIN subscribe / signoff / archive access instructions, send email to lists...@listserv.ua.edu with the message: INFO IBM-MAIN
