Newsgroups: comp.parallel.pvm,comp.unix.aix
From: woks4000@mailszrz.zrz.TU-Berlin.DE (Wolfgang Ksoll)
Subject: Re: Counting Users
Organization: CompuNet Berlin
Date: 5 Aug 1994 16:55:16 GMT
Message-ID: <31tqtl$bbs@brachio.zrz.TU-Berlin.DE>

In article <31tk1s$nq3@stc06.CTD.ORNL.GOV>, Hank Cochran <hdc@ORNL.GOV> wrote:
>IBM has announced AIX 4.1 with a new pricing structure. The cost depends on the 
>number of simultaneous "users" where "user" is the number of login plus rlogin 
>users. Does a PVM job running on a node utilize an rlogin on that node? 
>
>If so, it seems AIX 4.1 (at $200/box upgrade price) turns our RS/6000s into high 
>priced PCs or else requires an additional $588 per box to upgrade to AIX 4.1 and 
>maintain workstation capability (1 user + root + PVM). Pretty steep!
>
Hank,
the artifact IBM choosed to count users depends obviously on the
login(1) program. This means telnet-sessions, console-sessions and
terminal-sessions (tty) count. Someone mentioned it in comp.unix.aix
that rsh does not count.

So if you are debugging your pvm-session and you have 20 windows open
on twenty nodes with a pvm console running in it (just for example)
you are 21 users if you opened the windows to the nodes with telnet
and 1 user if your sessions were opened with rsh.

If you only run a pvm-session for production you are one user if you
doit interactively, no user in a batch environment :-)
If you use PVM/6000 instead of PVM-PD the same applies because it 
uses a rcmd-call instead of rsh.

Isn't it an easy to understand licensing scheme? Be asured that other 
people count other things when counting users. But I have not seen
that someone counts users :-)

Enjoy,
Wolfgang Ksoll


