[Info-vax] ISAM on disk layout. Was: Re: HP wins Oracle Itanium case

[Bob Koehler]

In article <4bp1h9-5s12.ln1 at news1.chingola.ch>, Paul Sture <nospam at sture.ch> writes:
> 
> This highlights a problem with RMS.  On mainframe systems I was used to 
> placing files carefully on disk, aligning by cylinder etc.  I cannot 
> remember whether the compatibility mode RMS utilities (IFL - Initial File 
> Load and friends) could do file placement, but once FDL came along you 
> could.

   This level probably isn't RMS.  It's more likely in the Files-11
   XQP.

   Once upon a time I had a system where I had to specify which blocks
   of a disk I wanted to use when I created a file.  Putting a name on
   it was optional.  I never actually needed that level of control on
   that system, but it was an early to middle 1960s OS and it required 
   the user to figure that out.  We had to know the geometry of the disk.

   We needed that level of control for a couple of systems on VMS.
   Set up the entire disk as one pre-allocated file and did $QIO access
   to blocks in the file.  Skipped RMS.

   In contrast, I've used a few DBMS on PCs (not because I wanted to).
   One of them used block location data in it's internal meta-data.  Had
   to have a backup product that knew that, and could update it on
   restore.

   The amount of time saved in accessing a file by physical location vs.
   mapping a logical location is almost always tiny, so it always seemed
   just a PITA to use ISAM on MVS or DBMS on PCs that could only store
   physical block locations in thier meta-data.