| Article: 38379
Newsgroups: comp.unix.questions,comp.sys.dec,comp.arch,comp.os.vms
From: [email protected] (Mitch Wagner)
Subject: net.views -- dec and alpha -- responses
Organization: Open Systems Today
Date: Wed, 03 Feb 93 23:25:32 GMT
Following are the responses to a previous net.views question. Phone numbers
have been deleted, as have personal messages to the editor (that means you,
Pope Clifton :-) ). But complaints about OST have been left in.
Thans for the participation of those who participated, and the forebearance
of those who didn't. Enjoy!
Is Alpha going to revitalize DEC?
We're asking this question to collect opinions for the "net.views" column in
By responding to that address, you'll be granting permission to OST to publish
your response.
In order to use your response, we'd like to know who you are. Please tell us
our real, full name, your job title, your employer, and location---or,
alternately, if you're a student, your class standing (freshman, sophomore,
junior, senior, grad, etc.), college and university, and its location. Please
also include a telephone number where you can be reached during daytime hours
in North America.
Please limit responses to about two full screens (24x80) of text, not including
headers and other administrivia.
Responses to net.views will be posted to this space in about two weeks.
Thanks for participating, and we look forward to hearing from you!
From: Steve Suttles <[email protected]>
I'm a long-standing DEC customer--19 years using DEC, 15 professionally,
10 on a wide variety of VAXen. Currently, I'm involved in O/S level
development, and my company has an ISV agreement with DEC. I have worked
at virtually every technical level, with virtually every every PDP-11 and
VAX model, and had decided for and against DEC and DEC products for about
the last 14 years. I own a PDP-11, but I use a PC clone at home.
My call is no. They have to change their outlook. Techies such as
myself are going to replace management eventually--fewer IS managers
lack hands-on experience than in even the very recent past.
A common complaint of the next crop of managers is not that the product
line stops too short of the top. This is a recurring chorus in the industry,
but it is not the current complaint. The current complaint is the lack of
support. For a long time, DEC grew so fast it couldn't keep up with itself.
We should all have to solve problems like that. However, now that DEC is
not growing as fast, that excuse doesn't wash. DEC has abominally slow
response to problems when support agreements are in place, and the red tape
to put an agreement in place is horrible (in terms of time and money).
My first act as honorary CEO is to issue a directive to change the focus.
DEC needs to be a service-oriented company, not a hardware sales company.
There is an installed base, and while upgrades and new sales will continue,
it is very important not to drive existing customers away. From where I
sit, on the outside looking in, the systems sales force is twice as big
as it should be. The software sales force is about right, and the support
and development teams should be doubled. A software bug report should
be acknowledged upon reciept, and ones returned "will be fixed in a future
release" should account for substantially less than one quarter of those that
are deemed to need "fixing". When a decision is made that there will be no
immediate action on a bug report, that decision should occur within a week or
two of the reciept of the bug report, not six months plus.
Getting on hardware maintenance should not require two customer man-days to
arrange, nor should there be punitive fees for telling the system sales force
that the hardware and its maintenance couldn't be purchased together. Field
service staff should be required to intern at a DEC facility rather than take
their semester final at a customer site by phone. Field service should use
computers (or something) to ensure that a maintenance customer with a tape
drive of designation X has a maintenance office staffed with someone who has
seen one within a given range. Field service could realize cost savings and
simultaneously gain goodwill by installing hot spares at sites willing to
do so when it becomes necessary to inform the customer that the nearest
experience with that type of hardware is in a different time zone, for example.
"Once our customers like us, they may even decide they need more computers or
software! They may even tell other people to buy our stuff, instead of
listening to them and buying those fruity things or those blue boxes! And
remember--our real money comes from service, not sales. We only sell it once,
and the service income lasts for as long as they are happy. If they are happy,
we'll sell more, and service more, and so on. If they are unhappy, we only
sell it once."
I am: Steve Suttles
a Senior Systems Programmer
working at Cross Access Corporation
2900 Gordon Avenue, Suite 100
Santa Clara, CA 95051
My office phone number is (408)735-7545
I grant permission to publish this or a derivative, and specifically encourage
this information to be given to any and every one at DEC who can make use of it.
--
Steve Suttles Internet: [email protected] Dr. DCL is IN!
CROSS ACCESS Corporation UUCP: {uunet,sgiblab}!troi!steve Yo speako TECO!
2900 Gordon Ave, Suite 100 Talk data to me!
Santa Clara, CA 95051-0718 HA! It's under 4 lines NOW!
From: John LoSecco <[email protected]>
Organization: High Energy Physics Notre Dame
No
--
John M. LoSecco
Physics Department HEPnet/SPAN: UNDHEP::LOSECCO or 47344::LOSECCO
University of Notre Dame Internet: [email protected]
Notre Dame, Indiana 46556 USA Bitnet: losecco@undhep
From ross!cs.utexas.edu!ross.com!tony Mon Jan 11 11:59:30 1993
From: Tony Hurson <[email protected]>
I am a pessimist about DEC and Alpha. The new chip and the series of
machines it will go into might stabilize the company, but I can't see
it returning to the heady days of the early 'eighties, when DEC ruled
the minicomputer world.
The main reason for my gloomy outlook is that DEC is late to the game.
The dominant architectures of the workstation business - Sun, HP,
SGI and IBM - were created years ago and have a huge installed base
that will be hard to eat into. Granted, Alpha has a VMS backward
compatibility, but I sense that users and MIS managers are moving from
proprietary operating systems like VMS in droves. Think of it
individually: if you were just graduating from college with a computer
science degree, would you consider a career in VMS-based software or
systems a smart move?
With VMS fading, DEC will be left to battle with more nimble competitors
in the crowded "open systems" (mostly Unix) arena.
The possibilities of Alpha with Windows-NT are harder to call, since the
latter is not yet on the market, but my guess is that Microsoft will
target its news OS at the widest possible swath of business users, which
means Intel-based PC machines. A Pentium-based multiprocessing machine
running Windows-NT sounds like a very adequate and price competitive
high end choice for business applications in 1994, say.
[Respondent details are in the signature box below, except for the fact
that I am an integrated circuit design engineer at Ross Technology.]
+------------------------------------------------------------------------+
| Disclaimer: Any opinions above have nothing to do with Ross Technology |
+------------------------------------+-----------------------------------+
| Tony Hurson | Email: [email protected] (work) |
| | [email protected] (home) |
+------------------------------------+-----------------------------------+
| Ross Technology Inc., |
| 5316 Hwy 290 West Suite 500, |
| Austin, Texas 78735, USA. |
+------------------------------------------------------------------------+
From ross!cs.utexas.edu!ross.com!tony Mon Jan 11 15:21:50 1993
From: Tony Hurson <[email protected]>
No, neither myself nor Ross Technology have ANY business relationships
with DEC, direct or indirect.
- Tony Hurson ([email protected]).
From: "Charles T. Smith, Jr." <[email protected]>
Few things are sure in todays market, but Alpha appears to be a pretty good
bet. It will take time; even though the first systems are out, the software
is not up to the point where its ready to deploy mission critical applications,
especally where co-existing with current hardware is important. Yet, these
issues are being addressed in a fashion that builds confidence. There does
not seem to be any of the loose ends that were evident with the 9000 systems;
the entire alpha process seems well thought out.
Another way in which Dec seems to be getting its house in order is its
attitude toward third parties. The "Not Invented Here" syndrome is finally
giving way to the reality that DEC cannot - and should not - try to be
everything to everyone. A lot of this shows externally, in the selling of
some assets, such as manufactoring circuit board assemblies. And Dec is
taking proactive steps to undo much of the damage done by driving away
the third parties in the '80s.
Second source for the Alpha chip is still important - back in the PDP
days, DEC was friendly to the third party market, too. Second sourcing
alpha will provide firm proof of Dec's intentions.
It won't be an overnight happening; it will take time. Digital does have
world class hardware, looks like it has a good handle on the software side,
and is making the changes to its internal processes and business practices
that will allow it to suceed in the years ahead. And if the rumors about the
next generation of ALpha are anywhere close to true - 1000mhz - Dec has the
stuff in the pipeline to remain a competitive player.
Current customer, former OEM, owner of three personal microVaxen home
systems.
Charles T. Smith, Jr.
Senior Systems Programmer
Sun America Financial
11 Executive Park
Atlanta, GA 30329
From: Rick Westerman <[email protected]>
Current involvement: 11 VAXstations/MicroVaxen. Plus knowledge of other
VMS activity on the campus (we are mainly a UNIX, non-DEC campus).
Revitalize might be too strong of a word. "Keeping DEC from going under" is
how I would phrase it. With the advent of the model 90s (VAXstation and
MicroVax), the newer 4000s and the Alpha machines, we now plan to stay
with VMS instead of replacing our few remaining VMS systems with Unix-based
systems. However I don't expect any of our Unix users to reverse-migrate
to VMS. As for buying Alpha machines to run OSF/Unix, we may buy some, but
probably not too many. After all, a Unix box is a Unix box is a Unix box.
Sun, HP, DEC, IBM; in many ways they are all the same and we've been burnt
too many times with DEC to want to put many of our eggs back into their basket.
-- Rick
Rick Westerman System Manager of the AIDS Center Laboratory
[email protected] for Computational Biochemistry (ACLCB), BCHM
bldg., Purdue University, W. Lafayette, IN 47907
From: Daniel Packman <[email protected]>
Organization: Ntl Center for Atmospheric Research - Atmospheric Chemistry Div
Current DEC VAX/VMS customer.
Daniel Packman
Programmer III
National Center for Atmospheric Research
Boulder, CO 80307-3000
Within the context of the computing industry, downsizing and competition
will remain fierce in the forseeable future. These factors will affect
all large computer makers including DEC. The Alpha is just another product
to allow DEC to compete. It is not a generation ahead of the competition.
Indeed, if we examine the SPECmark efficiency (SPEC/Mhz) for several
architectures, we see that the Alpha lags at an effiency of about 1 but
other vendors (IBM and HP) have achieved a value close to 1.5. To the
extent that DEC can retain a lead in clock speed alone, they can produce
a comparable product to the competition.
DEC is ahead of some with a 64bit product but behind others in that it
does not support multiple cpus. DEC has a slight lead with its Futurebus,
but almost comparable bus speeds are available in the competition. I see
the Alpha product as one that might stop DEC's fall but certainly not
revitalize it.
--
Dan Packman NCAR INTERNET: [email protected]
7 P.O. Box 3000
Boulder, CO 80307-3000 NSI/DECNET: 9583::PACK
Phill Hallam-Baker
From: [email protected]
Ya.
From: VAX System Manager <[email protected]>
I hope so. This all depends on DEC, I think. They really should have done this
(i.e. coming out with a RISC-Y VMS machine) a long time ago. The only way it
will work, however, is if they COMPETATIVELY price the machines. Actually,
they need to blow away the other manufacturers. If they do it strictly on a
per-vup basis, I don't know if that will work. If they come out with AFFORDABLE
machines (i.e. servers in the under $10K range), then they have a chance.
I've taken a lot of abuse from people who manage UNIX machines over the
performance of VAX/VMS. Some of it is deserved, but a lot isn't. I'd much
rather work with VMS than UNIX anyday... it's a much more user-friendly
environment. Of course, when you start getting into X-Windows applications,
some of the underlying elements of the operating system no longer apply, but
as soon as you shell out to the operating system, I'd much rather see the
familiar "$"-VMS prompt than a CShell prompt.
Warren Kring ABAM Consulting Engineers
Computer Systems Manager A Member of the Berger Group
33301 Ninth Avenue South
Email: kring%[email protected] Federal Way, WA 98003-6395
From: Simon Townsend <[email protected]>
Maybe if you didnt now charge for your mag, people would be more
willing to give you useful opinions for free too!
From: "Andrew C. Burnette" <[email protected]>
Andrew C. Burnette, Network Manager, Computer Systems Lab,
box 7911
Raleigh, NC, 27695
You asked.......
Alpha looks like a great microprocessor, but for DEC to be successful hinges on several
items which DEC is trying to address internally.
Dec has dug a very deep hole which won't be easy to crawl out of.
1- You can't develope an installed base today without a good OS.
OSF-1 looks like a party which no-one else will attend. True, it is a 64 bit
OS, and the beta release looks clean, and hopefully will be more competant
than Ultrix.
Dec does not plan to offer OSF-1 on their larger systems, and will not have SMP
support for OSF-1 for quite some time. HP on the other hand has made clear it's single OS
stratagy for HP-UX accross it's entire line of systems. Sun(now that they have a
mirange offering) is doing the same thing. Perhaps lessons from PROFITABLE
computer companies would help.
2- You can't develope an installed base today without a market presence in the
$5k arena. Given the high clock rates (I have looked inside an alpha box) and
the heavy duty cooling needs of the CPU, it will be difficult for Digital to penetrate
this area. Sun does make money in the under $5k arena. HP sells low end(fast as hell)
in the $5k area, even though they don't make money on them. Both realize
there is a need for the inexpensive desktop.
3- Without any installed base or significant applications for Alpha, it
will be difficult to sell these boxes.
4- We have almost 1000 MIPs based decstations. Since we've been dec'ed more than
once, how can we expect better treatment? So DEC can't produce a $5k alpha box yet,
why don't we have any r4000 based decstations yet?
5- tradeup prgrams at digital suck. While announcing new systems, dec sells older
systems, and charges up to 30% premium to upgrade to a newer box. Sun has realized
this pisses customers off, and does something like a cost plus upgrade,
based on current prices. It at least looks fair. I don't know what HP's policies are.
Later,
--
******************************************************************************
Andrew C. Burnette [email protected]
Electrical and Computer Engineering
From: Victor Arnold <[email protected]>
Vic Arnold
Director MIS
Stanford University Clinic
Suite 225
The Stanford Barn
700 Welch Road
Palo Alto California
The clinic is the group practice of the Stanford School Of Medicine
We use VAX platforms with VMS for business applications: Patient
Billing, Managed Care Plans, Patient Appointment Scheduling,
Patient Registration, etc.
These machines are integrated into the larger university WAN
environment (5,000+ nodes with additional internet connections)
We also share real time application and data exchange with Stanford
Hospital IBM Big Blue Iron via SNA-CT gateway and LU6.2
Alpha provides the needed migration path for us to more MIPS at
a much more reasonable price. The OpenVMS to (hopefully) a standard
UNIIX environment (again hopefully not Ultrix 4.X but OSF/1 or
something like it). Because we have a large investment of money
and resources in the VMS environment we can't just toss the gear
for some other vendor's offerings. Alpha appears to help that.
The extra MIPS too should help us migrate some applications to
a client/server environment (end user reporting with windows/
Xterm based GUIs for example) w/ the alpha boxes as servers and
possibly clients.
My 2 cents worth...
Vic Arnold
From watzman.quest.sub.org!slice!udo Fri Jan 8 16:40:13 1993
From: Udo Klimaschewski <[email protected]>
This is from:
Udo Klimaschewski
Programmer & Sysadmin
QUEST systems GmbH
Dortmund, Germany
Mitch,
From my point of view, a revitalisation is not depending on any
hardware. The Risc machines from DEC are an excelent hardware,
so is ULTRIX. ULTRIX was and is one of the most stable, powerful
and easy-to-upgrade systems today. If you take a look at SOLARIS
or AIX, you'll see that most upgrades are not upgrades, but rewrites.
The Alpha series from DEC introduce some kind of a next generation
architecture, leaving the discussion about the OS open.
And this is, not only for DEC, the main problem.
Customers and foremost ISVs are much more interested in the OS.
DEC has to work more on the ISV and "political" frontier, not on
the HW side. (Take a look at the IBM 6150/6151 desaster)
If we (as an ISV) would get a good, stable and "promoting" support
from DEC, along with a long term strategy, on which you can count on,
people may concern using DEC instead of Sun,HP or IBM.
A good car does not make a new image, but a good image may sell a bad car.
kind regards (& sorry about my ignorant english)
--udo
--
From: [email protected]
I have used several VAX machines over the years but not too much
recently. There are others around me with P-max, so I am a bit
familiar with those too.
-- Vince
From: William Bardwell <[email protected]>
Yes, assuming DEC can do some things they have had trouble with
before. They need to price their systems competitive with their
competitors. They need to make a real low end (priced with low
end of their competitors). And they need to do one thing that
they shouldn't have trouble with, provide good, continuously
improving OSs
William Bardwell
wbardwel+@[cs.]cmu.edu
Junior
Carnegie Mellon Univ.
Pittsburgh, PA
From: Pope Clifton <[email protected]>
Organization: Inst. for Epistemological Pathology
X-Newsreader: TIN [version 1.1 PL8]
Mitch,
I think it is too early to say whether Alpha will turn DEC around. At
present they are facing an extremely tough market, with SGI's R4000 based
high performance workstations coming out, and HP driving its PA-RISC
architecture to higher and higher clock rates. The main driver for Alpha,
if workstations built on it live up to DEC's claims, will be how fast DEC
can get popular commercial CAD/CAM/CASE packages and popular public domain
software ported to it. The HP 700s (Snakes) faced a sales lag of over a
year after they were the best performer on the market, because there just
were not enough popular software packages available for them.
-- Clifton
--
[email protected]
Clifton Royston, Pope of the Church of the Subgenius in Paradise
- Dissecting personal psychopathology at the edge of the 20th century -
From: [email protected]
Yes. Big time. Micro-based parallel processors such as the Alpha-7000
and Alpha-10000 will replace multi-chip-cpu mainframes. Shared
machines tend to bring in more money for a given performance as a
higher percentage of the CPU cycles are actually used (more efficient
use). Because the costs/perf for manufacturing is now much lower (due
to microprocessor) profit margins should be good. Even as DEC
undercuts the older multi-chip-cmu mainframes to take away their
market, DEC will make lots of money. IBM, Amdahl, Control Data, Unisys
(or who ever it is that has been making multi-chip-cpu mainframes) will
suffer more than ever.
It also looks like the $3,500 PC version of Alpha will be twice as fast
as the Pentium/586 for about the same price. This is a fantastic
advantage in such a competitive market as PCs. As windows/NT takes
off, so will Alpha. The Intel X86 architecture has only 8 registers -
a limitation that will keep it from ever being able to compete in
performance with the Alpha. As the Alpha-PC and windows/NT become
cheap, the X86 will become a thing of the past. PC clone makers will
switch over to the Alpha chip, giving DEC the position Intel used to
have of making money off the sale of most PCs.
Also, the Alpha is used in the Cray MPP, so DEC will be getting some
money from the supercomputer market as well. So DEC is positioned well
for supercomputers, large shared machines, workstations, and PCs.
Yes, the Alpha will revitalize DEC.
-- Vince Cate
Computer Science Grad Student
Carnegie Mellon
From: Marcus J Ranum <[email protected]>
The question is "which Digital?" Digital, in this case, can be
thought of as several corporations - it certainly doesn't react to the
market like a single unified corporation with a coherent strategy. The
part of Digital that we might call "VMS, Inc" will probably do fairly
well from Alpha, as the processor's performance slows the erosion of
VMS towards UNIX, and helps stave off UNIX workstation vendors like
Sun that traditionally made inroads into VMS installations based on
cost/performance. Digital's "UNIX, Inc" probably will continue to
flounder helplessly, continuing to be chronically short of resources
and behind the operating systems power curve.
Digital's problems in the UNIX arena aren't hardware and never
were, so there's no reason to assume that faster iron is going to solve
them. In the last few years Digital has kept fairly well abreast in
cost/performance with other UNIX vendors, and has consistently lost
by being the only vendor still selling 4.2BSD with enhancements, by
having a sales force you still don't know their own UNIX products,
and by having corporate executives that cut funding for UNIX
projects to feed VMS projects. Faster hardware doesn't improve any
of these situations, and in fact makes life worse for Digital
because it forces them to transition their installed base to a new
architecture. Digital's inability to ship OSF/1 at initial Alpha
release does not bode well for their ability to get all-important
third party applications moved over. By the time Digital has a
"commercial quality" UNIX environment on Alpha, HP and SGI/MIPS
will be right behind them on the hardware performance, and will
still be way ahead in the operating systems arena.
Digital will survive. Parts of Digital will prosper thanks
to Alpha and parts of Digital will continue to be as badly off as
they were before. If Alpha performance had been available for UNIX
a year ago, its positive impact might have been tremendous. Digital
still has the advantage of making the world's fastest hardware for
running VMS.
From: "Phillip M. Hallam-Baker" <[email protected]>
Will Alpha revive DEC? Nobody knows of course, but without Alpha DEC would be
finished.
The Alpha seems well placed for companies upgrading from the PC world in a few
years time. In terms of commercial success this will be the key market. If
Windows/NT has not managed to provide multiuser functionality by then, VMS will
be a useful card to play.
The main question is of cost. In terms of price performance, alpha is cheap.
However it is not cheap enough for many people afford one to have at home, nor
is it cheap enough to give to undergraduates on a one per student basis. If DEC
can sell an Alpha system offering a 3000/500 level of performance at VAX VLC,
DECstation 25 prices within a year they will have a winner. Otherwise they will
find that the competition has caught up on them.
--
Phill Hallam-Baker
From: Marcus J Ranum <[email protected]>
Organization: Trusted Information Systems, Inc.
Now *THAT* is an amusing question!
From: "Ehud Gavron " <[email protected]>
From: [email protected] (Ehud Gavron 602-570-2000 x. 2546)
Organization: ACES Consulting Inc.
Please go read the NSFNET usage guidelines to better understand
how your posting violates them.
--
Ehud Gavron (EG76)
[email protected]
From: "David B. Horvath, CDP" <[email protected]>
Organization: Hidden - I don't speak for them.
Yes, I think Alpha can revitalize DEC. With the chip being
"open" - that is, not just installed in DEC machines or running
DEC operating systems, there will be an increased demand.
Microsoft is porting Windows NT to the Alpha chip. Can't get an
Intel chip fast enough for you but you want to keep the same
user interface? Go to Alpha! Want to develop applications for
Intel PC's and other platforms? Go with Windows NT.
How strongly do I feel about this? I just bought DEC stock last
week because I believe in DEC & Alpha.
Who I am:
David B. Horvath, CDP. Email after 1/31/93:
[email protected]
title(s): Independent Consultant, Adjunct Instructor,
Graduate student (beginning thesis).
affiliations: CoOperative Business Solutions (various clients),
Delaware County (PA) & Camden County Colleges (part-time),
University of Pennsylvania, Dynamics of Organization
DEC experience: user, developer, system management, DECUS Seminar
presenter.
Standard disclaimer: these opinions are my own, they do not represent
the client who owns *this* machine, anyone who employs me, or DECUS.
From: "Benjamin Z. Goldsteen" <[email protected]>
Work for OU HSC which has one VAX 6000-610.
Don't like VMS. Don't really like DEC.
From what I have heard and read the Alpha is a very impressive
offering that will surely attract VAX/VMS users who have been thinking
of abandoning ship to RISC systems. However, certain attitudes of
DEC, which are very attractive to commerical data centers,
are often lacking in the general UNIX market. Recent rumors have it
that OpenVMS actually works better then DEC claims it will -- I would
have a hard time saying this about most other system vendors. Although,
the performance of the AXP systems may not live up to expectation.
For a UNIX site, however, there is some question as to OSF/1.
Support for OSF doesn't appear to be there. Is OSF falls apart,
would DEC continue to offer and improve it? What about WinNT --
can DEC support WinNT, VMS, and OSF? Rumors have it that upper DEC management
is partial to WinNT over OSF.
--
Benjamin Z. Goldsteen
|