NetNews Usenet Archive 1992 #19/NN_1992_19.iso/spool/comp/sys/hp/10043

Internet Message Format | 1992-09-03 | 8.5 KB

Path: sparky!uunet!olivea!spool.mu.edu!sdd.hp.com!scd.hp.com!hplextra!hpfcso!hpcuhb!hpcupt3!larryt
From: [e-mail] (Larry Thompson)
Newsgroups: comp.sys.hp
Subject: Re: HP-UX 9.0
Message-ID: <[e-mail]>
Date: 4 Sep 92 00:31:19 GMT
References: <[e-mail]>
Organization: Hewlett Packard, Cupertino
Lines: 149

I don't know if this will help or not. It is a copy of a note that was
posted awhile back (April). I don't think the authors knowledge of where HP
was going with PA-RISC and the 68K product lines was very good, but his
insights into the OSF seem pretty good even today.

This is my personal opinion. Not an offical statement of the Hewlett-Packard
Company.
      ______________________________________________________________
     |           ___                                                |
     |          /  /                          Larry L. Thompson     |
     |  HEWLETT/hp/PACKARD       Commercial Open Systems Lab (COSL) |
     |        /__/                            [e-mail]     |
     |Open Systems Software Div                                     |
     |______________________________________________________________|


    --------------------- cut ------------------------------------------------

From [e-mail] Fri Apr 17 09:28:19 1992
Relay-Version: version Notes 2.8.4  1990/05/09; site hpcupt3.cup.hp.com
From: [e-mail] (Doug Eltoft)
Date: Fri, 17 Apr 1992 17:28:19 GMT
Date-Received: Sat, 18 Apr 1992 20:48:49 GMT
Subject: HP & OSF plans
Message-ID: <[e-mail]>
Organization: InterWorks
Path: hpcupt3!hpcuhb!hpfcso!hplextra!hplabs!hp-cv!ogicse!uwm.edu!zaphod.mps.ohio-state.edu!hobbes.physics.uiowa.edu!ns-mx!icaen.uiowa.edu!daeltoft
Newsgroups: comp.sys.apollo,comp.sys.hp
Sender: [e-mail]
Followup-to: comp.sys.apollo
Lines: 116
Xref: hplextra comp.sys.apollo:374 comp.sys.hp:1388

As a result of the significant number of misunderstandings and
misrepresentations that I read on the net, in the press, and hear from
other workstation users, I felt that it would be a good time to review the
situation with regard to The Open Software Foundation technologies,
Hewlett Packard's Workstation Business Unit, and Hewlett Packard's
operating system plans. Because of my position on the board of directors
of InterWorks (the HP workstation users group) and my position
representing The University of Iowa as a member of the OSF, I have been
able to monitor OSF and HP activities and plans as well as keeping in
touch with many end users. Although I do not represent HP or the OSF or
even other end users, I have tried to get the facts straight.

First, the Open Software Foundation. The OSF is a Not-for-profit
Corporation. OSF takes standards conforming technologies and weaves them
together to produce what they call "technologies." These "technologies"
are licensed to vendors who then have to develop marketable code which can
be sold to end users. Many times in conversations or in the press I hear
or read possibly misleading references to 'having OSF' on a given
computer. When someone refers to Unix, they do not usually mean just the
AT&T kernel. They mean instead a usable environment with all of the
software tools and utilities that over the years have become a part of
what we now commonly refer to as Unix. The OSF environment is the same in
that it has many independent parts from which hardware & OS vendors can
pick and choose. Thus at this early stage it is misleading to use 'OSF' in
the same way we use 'UNIX.' The OSF has delivered several OSF technologies
including:

Motif - a graphical user interface based upon X-windows. It has a style
guide for software designers, a library of graphical objects and an
example window manager.

OSF/1 - a fully-threaded, Posix (current state) compliant operating system

kernel based upon AT&T, BSD, and MACH code with some features from other
sources (e.g. B1 security from Secureware, and STREAMS code from Mentat).
It requires only an AT&T 5.2 license and an OSF license. There is a POSIX
compliant OSF application environment specification (AES). OSF is working
on a micro kernel version of OSF/1.

DCE - a distributed computing environment. Implements a user space
p-threads for those systems that do not have kernel-based p-threads,
timing service, naming service, authentication service using access
control lists (ACLs), a distributed file system (DFS) that provides remote
file locking (via tokens) and local data caching, and a network computing
system (NCS) for remote procedure calls (RPC).

Future OSF technologies include:

DME - a distributed management environment. Technologies have been
selected and are now being integrated into a cohesive framework.

ANDF - Architecture neutral distribution format. A two stage compiler made
up of a producer and an installer that allows software to be delivered in
a vendor independent and operating system independent format. This is not
a Unix-only product.

Now for HP. HP currently sells two operating systems, Domain/OS and HP-UX,
for its HP Apollo workstation line.

Domain/OS - acquired when HP purchased Apollo Computer. The current
release is 10.4. The 10.4 release is the first mature commercial operating
system to meet the OSF AES. HP will support it with the maintenance money
from Domain/OS contracts and will supply bug fixes and minor enhancements
as the money allows. This hospice care should continue through the year
2000. HP indicated that they will lobby software suppliers to support
versions of their software for the OSF AES compliant sysV environment of
Domain/OS. 10.4 will only run on the DN3000 and newer series machines and
the 9000 series 4xx models. There is also a 10.4p for the DN10000. The
Domain/OS implementation of NFS has been rebuilt so that it now works
well. HP will provide maintenance customers with a site license for NFS.
NFS will be the only way for you to interconnect the Domain/OS file system
with other operating systems now and in the future. Some DCE technologies
should be ported to Domain/OS but the final determinations have not yet
been made.

HP-UX - After evaluating the OSF/1.0 kernel, HP and all of the other
vendors had several choices. One choice was to fix the OSF/1 code. HP's
attempt at doing this resulted in the HP-OSF/1 technology release. The
work getting OSF/1 ready for the technology release made HP realize that,
as with any new code, it was not as stable as older established code.
Another choice was to replace large sections of OSF/1 code with more
robust older code from their established operating systems. This is the
approach DEC took with their DEC-OSF/1.0 code. HP selected a third option,
to take their existing robust HP-UX kernel and blend in the OSF
technology. They also decided to blend in some Domain/OS features as well.
This allowed them to deliver a stable operating system kernel and maintain
binary compatibility with all of the applications now running on HP-UX
while developing an OSF environment. HP refers to this as their unified
release and currently is targeting late 1993 for their PA RISC version. My
guess would be that the 68K version would not deliver until at least six
months after the PA-RISC version. The 68k version will be for the 9000
series 4xx models *only* and will not be available for older HP hardware,
such as the 3xx models or older Apollo hardware like the DNxxxx models.

Originally, when HP stated their intentions to provide OSF/1 on the
DN5500, they thought that they would be able to fix up and use the OSF/1.0
code. Due to the fact that software vendors told HP that they would *not*
support HP-UX on a DN5500, HP decided not to port the HP-UX based unified
release to the DN5500. They currently are offering to credit half the
price of the 68040 upgrade, a trade in allowance on the DN5500, software
license conversion, and standard discounts toward the purchase of PA RISC
based workstation.

HP will be offering 68K 9000/4xx to PA RISC 9000/7xx motherboard upgrades
in the $6,000 to $8,000 list price range. The case, HP-UX keyboard, hard
disk, HP software licenses, and memory will all transfer.

Each administrator needs to determine the most cost effective method of
protecting system stability while moving along with the hardware and
software technology. It is obvious that all new HP workstation models will
use PA RISC and that at some point in the future the 68k based
workstations will become functionally obsolete. The real question is when
to stop purchasing 68k based workstations and purchase only PA RISC based
workstations.

Doug Eltoft, InterWorks Board Chair 319-335-5751  [e-mail]