Richard Seltzer's
home page Publishing home
Articles
about DEC
mgmt memo
.
Volume 7, #6
August, 1988
Digital
Supports
Open Software
Foundation
The
Meaning
Of 'Transportable' Software by Ken Olsen, president
Trade,
Public
Policy And A Global Digital by Cliff Clarke, manager,
International Trade and Policy
A
Systems Vision Of Data Storage Management by Grant
Saviers, vice president, Storage Systems Group
Fernando
Colon
Osorio Named Corporate Consulting Engineer
Personnel
Policies
And Procedures Now On-Line
Infinite
Voyage
Discovers Ancient America
When talking to the press, there are three
areas of our product strategy that often come into question: the
VAX family, desktop systems and Unix*.
We started the VAX family of products over a
decade ago. Today, there are a number of other vendors who claim
that they offer better cost/performance, or offer features that in
some narrow dimension are better than a VAX system. That’s how new
vendors get started — by findÂing some area that is different and
focusing on it. That’s the nature of business. Today, VAX
computers are the standard against which everybody measures
themselves, and everybody tries to do better.
As for overall system capability - the whole
set of things that it does - the VAX family solves customer
problems better than any other system. That’s why we continue to
be very enthusiastic about it, and support it and develop it.
Also, today, the cost of customer solutions is
not driven primarily by hardware — it is driven by software and
systems. Faster, less expensive hardware may not help customers at
all. They end up paying far more for development of software and
system capabilities — costs they would not incur if they started
with a VAX system.
In the area of desktop systems, Digital has a
major program to create leadership 32-bit desktop systems, which
build on our existing VAX, VMS and ULTRIX architecture strengths.
In addition, sales of our terminals are increasing rapidly, and
terminal-based systems will remain important to Digital for the
foreseeable future. Terminal-based systems have the advantage of
lower per-user costs, no management responsibility on the part of
the user, and having data centralÂized in a minicomputer or larger
computer, as opposed to being scattered on everybody’s desk.
Through our Network Application Services (NAS)
program, we also help customers who have made major investments in
the computing architectures of multiple vendors to integrate all
those architectures into their enterprise-wide networks. For
instance, customers with MS-DOS*, OS/2* and Apple-Macintosh*
personal computers will be able to access applications from
anywhere on the network, communicate with other users on the
network and share such resources as printers, files and databases.
(Cont’d on page 2)
Some reporters are puzzled that Digital has two
major software products - a VMS product and a Unix product. The
answer is obvious - most companies have a multiplicity of products
to serve the multiple needs and desires of their customers. Ford
makes Escorts and also Lincoln ContinÂentals. They don’t have a
problem being enthusiastic about both of them. Likewise we can be
enthusiastic about ULTRIX (our Unix product) and our VMS software.
We are in business to provide solutions for
customers. We believe that VMS software by virtue of its
underlying functionality will be a preferred base for a large
percentage of our customÂers for the indefinite future. We also
recognize that some customers want the attributes of Unix, or want
applications that run on Unix, or are more interested in having a
"standard" software system than they are in the underlying
functionality.
Digital’s operating system strategy is to
develop both VMS and ULTRIX software within the framework of a
common, single-system architecture. Currently, this common
architecture inÂcludes the VAX hardware, the DECnet communication
architecture and the Application Integration Architecture (AlA).
We are also working with multiple vendors to insure that Unix
interfaces are true industry standard and not a proprietary
interface of a single vendor.
We expect to see a large number of common
applications across both VMS and ULTRIX software. The fact that
Digital has the two most important computing environments — VMS
and Unix — and that we have common applications which run on both,
will be a very important message that is unique to Digital.
Overall, Digital’s mission is to be the leading
supplier of enterprise-wide information sysÂtems. We are
developing our wide-area networking infrastructure and product
set, as well as network management tools, to help our customers
manage large, complex networks. Also, since many enterprise-wide
applications require large, high-performance databases and the
transaction processing style of computing, Digital has initiated a
major program to enhance its capabiliÂties in this area.
As the term "enterprise-wide" implies,
customers today are less interested in isolated products and more
interested in total solutions. Increasingly, the capabilities of
an individual piece of hardware or software are far less important
than how all of the pieces can work together as a single expanding
computing environment.
As the technology evolves, so does our
business. In its beginnings, Digital sold modules. Later those
modules became the basis for computers. Today computers, software
and support are coming to be seen as components of enterprise-wide
information systems.
First Digital sold only hardware; later, it
sold hardware, software and services. Now Digital is in a position
to take full responsibility for enterprise-wide systems — from the
basic engineering of the systems through integration, testing,
installing, maintaining and evolving those systems.
Digital’s unique position and competitive
advantage will result from its broad range of capabiÂlities. We
will build on our current strengths to ensure that we have the
necessary products and support to help customers who need
computing environments that are both distributed and closely
integrated. And we will do so in ways that are both profitable for
Digital and cost- effective for customers. (Cont’d on page 3)
*Unix is a trademark of AT&T Bell
Laboratories.
*OS/2 is a trademark of International Business
Machines Corporation.
*MS-DOS is a trademark of Microsoft
Corporation.
* Apple and Macintosh are trademarks of Apple
Computer, Inc.
Seven leading computer companies - Digital,
Apollo, Groupe Bull, Hewlett-Packard, IBM, Nixdorf and Siemens —
recently announced an international foundation to develop and
provide an open software environment to make it easier for
customers to use computers and software from many vendors.
The Open Software Foundation (OSF) will define
specifications and promote an open, portable application
environment. Its software environment will include application
interfaces, advancÂed system extensions and a new operating
system, using X/Open* and POSIX specifications as the starting
point. POSIX is an operating system standard, closely related to
the Unix system, which specifies how software should be written to
run on computers from different vendors.
OSF is incorporated as a non-profit,
industry-supported research and development organization. It has a
funding commitment in excess of $90 million to begin immediate
operations. Its iniÂtial development will be based on technologies
offered by the members and its own research, to be carried out
worldwide. Digital was one of the key initiators of the OSF
concept. Henry Crouse, Digital’s vice president of Strategic
Relations, is the acting president of OSF, and Don McInnis,
Digital’s Engineering Systems Group vice president, is on the OSF
board of direcÂtors.
To support its application environment, the
foundation will provide software that makes it easier for users to
mix and match computers and applications from different suppliers
by addÂressing the following needs:
o portability - the ability to use application
software on computers from multiple vendors; o interoperability —
the ability to have computers from different vendors work
together;
o scalability - the ability to use the same
software environment on many classes of computÂers, from personal
computers to supercomputers.
The foundation will follow a direction
consistent with international standards. Where standÂards do not
exist, the foundation will work with standards groups to help
define them. To achieve maximum acceptance for the new software
environment, the foundation will provide all members early and
equal access to the development process.
The foundation’s software environment includes
a set of application programming interfaces to make it easier to
write applications for a variety of systems. The initial set of
interfaces will support POSIX and X/Open specifications, and will
be extended to include areas such as distributed computing,
graphics, and user interfaces. To encourage its widespread use,
the software will run on a wide range of single- and
multi-processor computers. Vendors will be able to add value
through compatible extensions.
"OSF will help Digital respond to customer
needs and requirements for open software systems," explains Bill
Strecker, vice president, Product Strategy and Architecture.
"Prior to the formation of OSF, the de facto ’open’ software
system was Unix, which is really a proprietary system, carefully
controlled and licensed by AT&T. Through OSF, Digital can now
participate along with other vendors in an open process for
defining the open system. We’ll all have a level playing field.
Then we’ll all seek to add unique value to our implementations of
the standard. Standards provide a common denominator that
everybody agrees to provide; then each vendor typically strives to
add value beyond that.
"OSF will define a set of software
specifications so that customers and third parties will know the
interfaces to which they can write their software. That will be a
big help to many people. Programmers who write to these interfaces
can be relatively sure that multiple vendors will support these
interfaces. Customers will be better able to buy software that has
longevity, and suppliers of software will be able to go after
larger markets.
"But from the systems engineering point of
view, the benefits of this effort are not yet clear," notes Bill.
"There is no automatic guarantee that if everybody follows the
specificaÂtions, pieces from different vendors will be able to be
plugged together and work as a single complex system. Having
common interfaces helps a lot, but it doesn’t finish the job,
because there are so many aspects to designing and building
complex systems. Many issues around testÂing and discipline and so
forth go well beyond the ability to capture all of that within
simple specifications. So it’s not at all clear that OSF can solve
that problem. Very likely, the vendor will remain responsible for
building complex systems, making them work and guaranteeing that
they continue to work for the customer."
*X/Open is a trademark of X/Open Co., Ltd.
The Unix market is big business at Digital. It
represents a minimum of 10% of the VAX system revenues. In fact,
independent market research firms, such as IDC and Dataquest, rate
Digital as number one in terms of worldwide system revenues for
Unix systems. We have been number one in the UNIX market for 19
years.
There has been a sharp increase in the demand
for Unix over the last few years, and Unix growth is projected to
be very strong for the next three to four years.
In the 1970s, early interest in Unix was
generated in universities and some of the research community. Now
that base has broadened. In particular, the U.S. government and
other free- world governments, as well as some leading commercial
and engineering customers, are now interested in Unix as well. And
based on AT&T’s experience, regional Bell operating companies,
and other telecommunications firms, such as Siemens, have adopted
Unix.
Digital has been in the Unix marketplace from
its inception 19 years ago and has done much to promote the
success of Unix. In the 1970s Digital distributed Unix software
for PDP-11 compuÂters to universities, and helped fund Unix
development activities at the Massachusetts InstiÂtute of
Technology and at the University of California at Berkeley. Today,
Digital has the widest range of fully compatible Unix systems
available in the marketplace — from the deskÂtop to a
million-dollar VAX system. (Cont’d on page 5)
Unix is a general term, referring to many
different flavors of software. There is no single, unified
standard for Unix. We offer ULTRIX software, a version of Unix
designed to take advantage of the capabilities of our VAX
computers and networking capabilities and that also complies with
key industry standards. We also offer a direct copy of AT&T’s
5.3 UNIX to the Telecommunications industry.
Over the years, we have enhanced ULTRIX
software to support all the popular features developed at Berkeley
and from AT&T System 5. And we are adding features to comply
with the POSIX speciÂfications, soon to be sanctioned by the IEEE
(Institute of Electrical and Electronics EnginÂeers). The POSIX
IEEE 1003 standards effort will unify the various derivatives of
UNIX.
The POSIX standards have been endorsed by all
the major UNIX vendors and by OSF. Our next release of ULTRIX
(3.0), to be released in the next few months, is compliant with
the current POSIX standards and the OSF level 0 portability guide.
In other words, ULTRIX software merges a number of industry
standards, and adds to the capabilities unique to Digital.
ULTRIX compliance with standards and our
support for the Open Software Foundation [see related article in
this issue] are a continuation of Digital’s long-standing support
for industry standards, for instance in networking and languages.
Today, ULTRIX and Digital’s VMS operatÂing system use many common
products, such as DECnet capabilities and languages. Through our
DECwindows program, we will provide users of both ULTRIX and VMS
software with the same user interface. And, over time, we plan to
converge more parts of the VMS and ULTRIX environments, as part of
our single-system architecture. When you have the same networking
protocols, the same user interface, the same compilers and some of
the same applications available on both operating systems, then
the differences between ULTRIX and VMS software become less of an
issue for us in sales situations.
It is important to note that the increasing
emphasis on Unix does not mean that Digital is deemphasizing its
VMS software. VMS software has a very wide customer base and is
selling very well. We’ll continue to enhance it and market it
across the whole spectrum. Basically, Digital is now big enough,
and VMS software is successful enough, that we can
enthusiastically support both systems.
We all need to realize that ULTRIX is a
critical part of Digital’s strategy. Key customers are starting to
require it. It is important that all parts of the company make the
necessary changes, and move to respond to this demand. We need to
position ourselves to get in front of this growing opportunity to
optimize our business.
Software can only be truly "transportable" if
it is generated and maintained with discipline. Software often
grows over many years. It is patched and patched again, with
patches on top of patches. This is often done without discipline,
without documentation or careful record-keepÂing. In time, the
people who did the patching and correcting move on, and what is
left is incomprehensible to anyone else. People only know that the
software works on the particular equipment that it was built and
grown in. (Cont’d on page 6)
Nothing will replace discipline in the writing
and maintaining of software. It has to be done in a planned and
organized way. It has to be tested, maintained, improved, fixed,
and patched with discipline and careful documentation and careful
record-keeping.
Simple programs for personal computers and
workstations do well-defined jobs for one user and can often get
by without such discipline. But when a job grows to include a
number of users and complex software, and needs to be connected to
different parts of the organization, disciÂpline is essential.
Unix was written 19 years ago for people who
wanted to build small, simple systems without the discipline that
goes with big systems. It was purposely unplanned and
undisciplined. It made a contribution to computing by allowing
people to write casual software for small computers and
workstations. Now, many people would like to maintain some of that
freedom, but add the disciÂpline necessary to make Unix software
truly transportable. Digital has joined in that effort, by
supporting the Open Software Foundation.
Software is normally written in one of about
two dozen languages, such as FORTRAN, COBOL or C. The languages
are designed to make it easy for people to write down what they
want a computer to do. The operating system takes these
instructions and converts them to machine language — a form that
is optimized for efficiency and speed of computing, but is very
awkward for a human. The operating system also does many things
automatically so that the software writer does not have to think
about them or even know about them.
Each language is standardized, documented,
disciplined, updated and maintained by an open, public
organization. Everyone has a right to input, and everyone receives
changes and updates at the same time. If someone follows the rules
of the language, then any operating system that accepts that
language should be able to use the software written in that
language. In other words, it should be completely transportable
between different computers with different operaÂting systems.
Often, this does not work. Problems arise
because people do not limit themselves to the stanÂdard language,
but instead, take advantage of unique features in the operating
system to make their software more efficient. As a result, the
software becomes dependent on a particular operating system and
the computer that goes with it, and cannot be readily transported
to another.
Unix has probably suffered more than others
with this problem because of the informality with which it was
originally written. Therefore, IEEE in the U.S. and the X Open
organization in Europe have set about to write standards to
isolate software from the characteristics of the operating system.
For Unix-type operating systems they call this set of standards
"POSIX" for "portable operating system interface." POSIX defines
and limits the calls that a piece of software can make on the
operating system. This means that if someone writes in a language
and is limited by the calls or openings to the operating system,
that are defined by POSIX, and if the operating system allows all
the calls or openings of POSIX, software written with these
limitations will be transportable between operating systems that
follow POSIX.
The transportability of POSIX software will
only be possible if software writers limit their software to the
POSIX calls on the system. But, today, every manufacturer has its
own variaÂtion of Unix with its own set of features, and every
manufacturer is trying to make its featÂures better than the
competitors’ features. The whole justification for investing in
writing Unix software and generating hardware that use it is the
belief that their features will be so much better than others that
customers will buy their system. But when people try to take
advantage of these features and go beyond the calls allowed in
POSIX, they lose all claim to transportable software.
So customers should not presume that they can
buy Unix systems today from several different vendors, plug them
all together and they will work instantly. The label "Unix" is not
a magic word.
They should also keep in mind that large
installations, particularly those which are based on networks, are
very complex. The necessary discipline and testing and effort
required to make them actually work is very hard. You cannot
assume that you can build a large network simply by plugging
together a set of pieces that were acquired from companies that
may never have talked to one another. Claims that you can do that
are unrealistic. I sometimes compare that kind of marketing to
"selling snake oil."
Unix is a strategic product for Digital. We
like it, and we try to do well at it. But we don't want our
customers to choose it because they have unrealistic expectations.
Our VAX family of computers and our VMS
operating system are very popular because software written ten
years ago will play on the newest computers designed today; and
software written today will play on new machines many years from
now. Also, software written for the very smallest desktop machine
will work in a giant cluster of VAX computers. These two forms of
transportability are accomplished by the careful discipline we
follow in the standard interface between the operating system and
the CPU. This Application Binary Interface (ABI) very careÂfully
defines and limits the calls an operating system can make on the
CPU. So, if a CPU, regardless of size or date of manufacture,
follows the VAX ABI standard, the software is interÂchangeable
without recompiling. So much work goes into building a software
system and a comÂputer, and maintaining that kind of discipline,
that only a company with a proprietary interest in them would make
that investment.
In other words, today, Unix software is less
transportable than people commonly presume, and VMS software is
transportable in ways that are often overlooked and undervalued.
For the long term, Digital is a strong
supporter of the kinds of standards that will make software far
more transportable in the future — for instance, for networking,
human interfaces and distributed systems. When communications
within a building or to anywhere in the world are part of the
software system, the protocols have to be standardized. We are
incorporating the OSI networking standards in our DECnet software.
For simple software, the hardest part of
changing software from one system to another is usualÂly in the
human interface, including the keyboard, the monitor and the
windowing system. Every system an organization uses should have
the same human interfaces. Digital and other major major
manufacturers in the Open Software Foundation (OSF) have accepted
the MIT X-Window sysÂtem. If followed, that will be a big step
toward making software for workstations and other small computer
systems transportable. (Cont’d on page 8)
For distributed systems, it is important that
access to data bases be standardized both in how one has access
and how access is restricted. OSF has picked the SQL systems,
which came from IBM. If all manufacturers pick that, computers by
different manufacturers with different operating systems will be
able to operate on the same database.
These efforts will eventually lead to
significant results. But, for today, customers should beware of
exaggerated claims of the transportability of software.
Today we do business in nearly 70 countries.
The number is up significantly from only three years ago. All of
the countries Digital has entered recently have undergone major
changes in a relatively short time. Similarly, countries in which
we have had subsidiaries for years have experienced dramatic
changes inside their major internal markets. In planning our
internationÂal marketing strategies, we have had to better
understand the dynamics underlying these changes in order to meet
our goals.
All of our analysis points to a critical
observation and that is the importance of public policy to our
success.
Collectively, governments make up what is
called the public sector. The public sector purchasÂes computers
for most of the same reasons that the private sector does. Many
opportunities that would be classified as private sector in the
U.S. are public sector activities in other countries. In almost
all the countries in which we do business the public sector buys
directly and indirectly at least half of the computer systems sold
in those countries. Globally, the public sector represents an
enormous market.
The rules are different in marketing to the
public sector as compared to the private sector, and this
difference is crucial. In the public sector, since purchases are
made with public funds, it is expected that the government will
insist on particular terms and conditions in order to further
certain public policies. In practical terms, the notion of fair
and open trading applies to the private sector within countries
but not the public sector.
During our industry’s history, most important
products and technologies have come from the U.S. and more
recently Japan. Because of the great strategic importance of
computer and telecommunÂications technology however, governments
everywhere have targeted the transfer of this technolÂogy into
their country as a high public policy priority.
As we grow in a country, we quickly learn that
our success in selling to the public sector becomes directly tied
to and in many cases limited by public policy. We find great
pressure is applied to transfer technology in exchange for access
to the public sector market. At a time when there is a steadily
declining labor content in our hardware products and a scarcity of
qualified vendors outside of the U.S. and Japan, governments are
insisting that we build factÂories and purchase within their
borders as a condition of sale.
While dealing with governments and the public
sector is very much like managing a relationship with a large
commercial account, at the macro level, the kinds of needs that
governments are attempting to satisfy are different. (Cont’d on
page 9)
At the country level, public policy represents
the contemporary goals and priorities of the nation. These goals
represent social, economic and foreign policy considerations. The
governÂmental process for establishing the laws and regulations
and carrying out these imperatives varies greatly from country to
country. Culture and the ebb and flow reflected in political
process often mask the reasonably predictable nature of most
governments’ machinery. To DigiÂtal, what is important to grasp in
all of this is that while governments may place great priorÂity on
technology transfer and exports, like commercial organizations,
they often don’t know how to undertake solutions that will satisfy
these needs. They look to us for information, technical assistance
and guidance on matters which we are perceived to be truly expert.
There is broad interest in a Digital solution. The challenge is to
advance the solution in an approÂpriate and effective manner.
The Digital public sector solution requires
that our strategy address public policy in each country. Because
of the special needs involved however, our offering, selling
style, and selling cycle will be different than in the private
sector.
Programs to forge public policy directions with
governments take considerable time. Further, engineering, R&D,
manufacturing and sourcing are integral to our solution. We have
learned that a good solution must integrate Digital’s needs and
strategies with a careful appraisal of the country’s needs and
public policy. Again, especially regarding computer technology,
often what is requested by governments does not support the stated
public objective. While political considerations are a factor,
limited knowledge of the technology and industry trends account
for most of the disparity. As in a commercial setting, it is up to
us to develop and present alternate ways forward.
The International Industry Marketing groups are
preparing strategies and messages that will enable the countries
to more efficiently pursue these opportunities in the public
sector. Further, there are important regulatory and standards
public policy considerations that must be integrated into our
strategies and plans.
The nature of the global trading environment
has become complex and economically very interdeÂpendent.
Technology and trade initiatives of any country are carefully
scrutinized by all others in the global trading community.
Countries openly admit to being global, economic competitors.
Coordinated and integrated global marketing messages are vital to
our effectively meeting our goals. Likewise, global strategies
carefully crafted and consistently represented in the public
policy arena are critical to our continued success around the
world.
[This article is based on a speech delivered at
the Dataquest Storage Track ’88 Conference, Santa Clara,
California.]
On our way to the 1990s, computers went from
being number-driven to data-oriented or informaÂtion-based. Our
challenges were much simpler when we had to process and save the
simple seÂquence of arithmetic answers to scientific formulas and
accounting calculations.
But data, which today means words, pictures and
ideas as well as numbers, doesn’t come to simple, logical, or
easily managed conclusions. We’ve evolved, say the management
experts, from a computing environment of sequence to one of
synchrony. Overall data management or information management — our
ability to synchronize our corporate thought-process — embodies
much more complex and strategic issues than disk drive design.
Today, we’re in the thoughtÂprocessing business.
Numbers were the domain of a select group of
technical people. Words are the domain of a larger set of
communicators. Graphics have opened our horizons that much more.
Today our computers systems must both emulate and advance the
corporate thinking process.
That’s why today we’re not talking about
64-user systems. We’re talking about 640-user systems, 6,400-user
systems, even 64,000-user systems. That’s already the order of
magnitude of DigiÂtal’s own internal network.
Corporations want computers that help guarantee
worldwide customer service consistencies, that can help the entire
organization work in concert to bring new products to market
faster, more cheaply. They want computer systems that understand
the management complexities of gaining marketshare. They want
comprehensive, corporate-wide systems that can streamline their
disÂtribution channels, shorten the sales cycle, help locate,
manage and integrate merger and acquisition targets. They want
computer systems that not only provide a real-time view of the
company’s balance sheet and income statement, but also help define
and execute the corporaÂtion’s money management opportunities.
Instead, too many corporations have a raging
crisis of incompatibility - canyons of data separations, and
polarized communications strategies often at odds with one another
or at odds with the corporate purpose. In too many instances,
users have accepted single-purpose, technoÂlogy-driven products
full of great ideas but suffering from a terminal case of
interactive speechlessness.
Data storage management technology is the next
strategic frontier. Business success in the 1990s will depend on
managing data - which implies a lot more than simply storing it in
ever- increasing amounts. Tomorrow’s systems will be driven by
their total, comprehensive ability to capture, hold, access and
manipulate data.
Two years ago, only 20% of Digital’s total
installed base of storage products was co-located with IBM
systems. Today, more than 45% of that installed base is IBM
co-located. In the same period, gross sales of Digital storage
products have almost tripled. During roughly the same period
between 1985-86 and 1986-87, IBM’s storage revenues reportedly
declined by 11% percent and 6% respectively.
Our sales and market penetration over the past
few years has not come exclusively because our product
specifications lead the competition. Particular Digital storage
products may be slower in absolute terms than those made by
competitors. But the systems and methodologies required to perform
corporate-wide thought processing are not commodity issues to be
decided solely on the basis of mathematical, product-to-product
comparisons.
At Digital, we view data management as a core
component and major player in the synergy of our systems. Our
storage products, both separately and in the way they
inter-operate and interÂconnect, are becoming major computer
devices in their own right. Intelligent data management servers
are becoming sophisticated computers in and of themselves.
Tomorrow’s solutions will evolve out of the
collective data management achievements we are working on now:
totally integrated hardware and software systems, integrated
components and distributed intelligent servers. The whole can and
must be greater than the sum of its parts.
This is not meant to diminish the importance of
advances in the processing power of central processors, in the
bandwidth and topologies of networking, the streamlining of
operating systÂems, or the strategic advance of applications
interfaces. But all the processing power in the world, all the
networking capacity, the most highly evolved operating systems and
the best applications are worthless if the right data isn’t
available at the right time, from the right source, at the right
price for the right user.
Our storage products will continue to have more
and more intelligence - because the loads these devices will have
to handle will increase by orders of magnitude in the coming
years, and also because host computers would be degraded to the
point of uselessness if all information management functions were
made the responsibility of the CPU.
Our data management products will continue to
have built-in interchangeabilities so they can attach, reattach,
and detach without disruption, not only to and from each other,
but to diffÂerent hosts or multiple hosts.
Through controller interconnectivity and
inter-operability, Digital will continue to promote approaches
that favor long-distance linkages to host systems, because we see
the long-term need for small and large pockets of data to be
located wherever people work.
Along the same line, Digital will continue to
advance its work with storage subsystem architecÂtures to meet the
full range of low- to mid-sized systems requirements, without any
temporary or lasting incompatibilities.
So that we can be more certain that the right
data management performance is available for the right data need,
Digital will also continue its work on storage subsystems that
will match the performance need with the right device at the
best-possible economy.
Digital will continue to leverage the
advantages of the various memory technologies. We have an
immediate and long-term interest in solid state disk technologies.
Optical technology, which for several years has been a very
appealing technology in terms of raw specifications, has now
finally achieved systems viability. Today, we’ve managed to
combine the optical drive’s enormous capacity potential with the
attendant systems needed to make optical storage a practiÂcal
option within the total framework of information management.
In database developments, our goals are simple
and ambitious. By 1991, Digital’s database systems will be the
leading products in terms of ease-of-use, and inter-operability.
They will be the price/performance leaders, too, but that goal
wouldn’t matter much if ease-of-use and inter-operability weren’t
among its leading features.
Because data availability and data integrity
are so immeasurably important, more and more Digital data storage
management products will anticipate faults in advance of
user-noticeable problems. For instance, a recently introduced
tool, VAXsimPLUS, can predict RA Series disk drive failures in
time for volume data shadowing to take over before there’s any
data loss or downtime. This proprietary service technology
effectively increases the overall availability of stored data by
orders of magnitude.
Storage product installation will move more
aggressively toward push-button simplicity, with automatic on-line
reconfigurability, automatic data migration for least cost and
highest perÂformance, error prediction systems, and automated
back-up and shadow recording redundancies for the protection of
valuable data.
Collectively, these advances will define
tomorrow’s systems. Thousands of engineers at seven Digital sites
worldwide are now working on the product-specific problems from
which technoloÂgical breakthroughs are achieved. But they are
equally focused on the overall synergistic requirements of
corporate-wide, nation-wide, and worldwide data management, which
represents the next strategic frontier.
[This article is based on a presentation
delivered at Digital’s State of the Company Meeting, on May 5 in
Merrimack, N.H.J
Research is one of Digital’s "traditional"
markets. In fact, we’ve been selling into Research for so long
that we’ve developed a number of different ideas about what the
customers in this market want and need. There is a large window of
opportunity for Digital in Research, one we cannot afford to
ignore.
Perhaps a good way to illustrate the needs of
the modem research organization — and Digital’s ability to meet
those needs — is to contrast the computing environment of the
1970s with that of today.
Let’s start our comparison where all computer
purchases start, with the company’s key decision maker. In 1975,
this individual — let’s call him Bob Gold — is the manager of the
Research Department. Bob Gold has a Ph.D. in Chemistry. As
manager, he is responsible for lab operaÂtions and quality
control.
Bob’s lab operations center on taking readings
from instruments. These readings and other test results are
manually recorded in a collection of notebooks. Bob’s quality
control operations are, for all practical purposes, non-existent.
Lab work just takes too much of his staff’s time.
Within the corporation, Bob’s operation is
isolated from other departments. His contact with Manufacturing on
new products, for example, is minimal.
Bob’s department computer in 1975 is a
PDP-11/10 with 16K core memory and a VT05 video termiÂnal. The
computer processes test data and even connects to some instruments
that directly collect analog signals. The view into Bob’s
computing environment was through his terminal into his PDP-11
computer. All he could access was his own little world and, in
fact, Bob’s view was really only a mirror of that world. Things
have changed since 1975. Today, because of an increasingly
competitive worldwide economy, Bob’s department and his role in
the corporaÂtion both have gained in stature.
Corporate survival in the 1980s depends on an
ability to provide new products ahead of the competition; products
with unique features; high quality products with differentiation.
Bob’s company realizes that in each of these three areas Research
plays a key role.
With increased status has come increased size
and complexity for the Research group within the corporation.
Today’s modem Research organization incorporates teams of
scientists working at labs throughout the world. Bob Gold is now
vice president of Research with responsibility to improve time to
market, added value and quality. To accomplish these goals, he
must maximize the productivity of the Research organization and
truly integrate it into corporate operations.
To this high level decision maker, implementing
the right computing strategy is critical. He needs the right tools
to provide staff scientist access to information while enhancing
their ability to share ideas and communicate with the rest of the
company.
The computing strategy has evolved with the
Research organization. The PDP-11/10 computer has developed into
two large VAXcluster systems used for timesharing, data analysis,
conferencing, and mail. In addition, the department professionals
have purchased and installed a number of workstations ranging from
a VAXlab system running VMS or ULTRIX software to an IBM PC under
MS-DOS or an Apple Macintosh.
Today’s lab has seen the advent of Ethernet.
Now, with networking, connections to a variety of departments and
systems are possible. Included on the list of connectable systems
of nonÂDigital computers such as a Cray supercomputer or a
corporate-level IBM mainframe.
Vice President Bob Gold’s organization is a
snapshot of the Research organization of the 1980s. Its role
within the company is more significant. Its computing needs are
greater so its reÂsources are larger and more sophisticated.
Networks allow researchers to work with each other and with other
departments. Yet computing resource issues still exist around
productivity and integration.
In 1975, the single PDP-11/10 computer simply
mirrored the department’s operations, but today’s extensive
network should provide a picture window view of the corporate
computing resources. It should, but it doesn’t. The truth is it
only provides a view the size of a keyhole.
Although the network allows the various
departmental systems to connect in order to share data, exchange
messages, access electronic conferences and notes files; it only
does so through great effort and even then, it does not permit
"true" integration of these system resources.
So as the Research organization moves into the
future, it finds itself in an ironic situation. It has growing
influence within the enterprise yet its ability to share
information within the corporate computing environment is
difficult and restricted.
Like trying to look into a room through a
keyhole, it takes the Research Department a lot of effort to get
only a limited view of the corporate computing resources. The key
issue for Research productivity has become making network
connections more usable.
How has this happened? Quite simply, as the
corporate computing environment evolved, incomÂpatible software
applications grew up in each department.
As an example, Manufacturing and Quality
Control must track samples, perform statistical analÂysis and
report on their work, just as the Research department does. At its
discretion, each department either purchases or writes programs to
do this. Each program has its own user interface, data format,
screen format, and commands. As a result, interaction among these
departmental applications requires a "brute force" approach. Data
conversion and interface programs must be custom-written to
integrate the various applications. For every software change and
enhancement, there is a ripple effect of custom programming to
make sure all the software pieces can still communicate.
Ironically, the Research organization’s
productivity, so important to corporate survival, is steadily
eaten up with the programming, reprogramming, and training
required for researchers to keep up with the computer tools
purchased to save them time.
But as the 1980s portend a future of
increasingly complex application requirements to the Research and
Corporate organization, recent Digital products and programs
continue to provide new solutions.
For example, the Digital DECwindows software
brings together a variety of applications into a single,
consistent, easy-to-use interface for the researcher, providing
easy access to even remote system applications on the network. The
resulting cut in programming and training time frees the
researcher for more pure research.
Digital’s Applications Architecture will mean
that future applications will be written with a consistent data
format, common communication mechanisms and calling standards to
truly inteÂgrate systems throughout the enterprise. The
architecture will also make program enhancements painless.
The Integrated Laboratory Automation standards
program will extend laboratory automation even further within the
lab by promoting consistent, standard instrument-to-computer
interfaces.
More packaged applications will be available
for Digital systems through the company’s coopÂerative marketing
programs.
Add these recent offerings to Digital’s
existing systems, software and networking, and research management
has unparalleled access to the rest of the organization. The
groundwork has been laid for highly productive integration with:
o Manufacturing to improve quality control and
cut time to market,
o Corporate Administration to report to
management for new product and marketing planning,
o Legal to speed patent and government
regulation issues,
o Corporate Library to access new information,
and
o Satellite Labs to share information for more
productive research.
In effect, Digital’s recent offerings continue
to improve integration and productivity for the Research
organization, increasing its value to the corporation. In turn,
the productivity and competitive viability of the corporation
improves as well.
Instead of a 1970s mirror of his own
organization or an 1980s limited "keyhole" view of the
corporation, the Research manager in the future will have the
computing resources to gain a broader view. He or she will have
easy access to corporate resources — a true window to other
departments. (Cont’d on page 15)
Finally, the evolving range of Digital
solutions not only creates a window of resources for the Research
manager, but also a window of opportunity for Digital in the
Research market.
Fernando Colon Osorio has been named Corporate
Consulting Engineer, reporting to Bob Glorioso, vice president,
High Performance Systems (HPS). Senior corporate and corporate
consulting engineers are the most advanced technical positions at
Digital. With the naming of Fernando, 13 persons hold those
titles.
As the technical leader and manager of HPS
Systems Research and Engineering and as acting manager of Clusters
and High Availability Systems, Fernando has made major
contributions to systems engineering, availability and
reliability, and has led the development of Digital’s High
Availability Systems strategy. He also was a principal contributor
to the VAX 8600 ProÂgram, through which he developed new
techniques for the software verification and system testÂing
methodology used today in the design of high performance VAX
systems.
Fernando joined Digital in 1976. He was named
consulting engineer in 1981 and senior consulÂting engineer in
1984. He is the author of a textbook on artificial intelligence
entitled "Engineering Intelligent Systems." He received a B.S.
degree in Electrical Engineering from the University of Puerto
Rico and a master’s and doctorate in Electrical and Computer
EngineerÂing from the University of Massachusetts.
Lee Pledger has been appointed U.S.
Human Resource manager for Field Service, reporting to Donna
Blaney, U.S. Field Personnel manager; Joe Gaffney, Human Resource
manager for Corporate Field Service; and Don Zereski, vice
president, U.S. Field Service. During her 11 years at Digital, Lee
has held various Personnel management positions, including Plant
Personnel manager for the Hudson, Massachusetts, Semiconductor
Facility; U.S. Field Employee Relations manager; and Northeast
Area and In-DEC Personnel manager.
Pat Spratt has been named Corporate
Planning manager, reporting to Jim Osterhoff, vice presiÂdent,
Finance. In his new role, Pat will integrate the multi-dimensional
planning process of products, markets and geographies to develop a
revenue, cost and profit plan for each. He will also be a member
of the Finance staff. Pat has held senior Finance and Strategic
Planning positions with Software, Manufacturing, Engineering and
Field Operations. Most recently, he was Finance and Planning
manager for Product Marketing.
Steve Thomas has been appointed to the
newly created position of U.S. Product Sales manager, reporting to
Chick Shue, vice president, U.S. Sales. In this role, he is
responsible for assuring that U.S. Sales meets or exceeds its
product sales commitments in several major proÂduct segments,
including mid-range systems, software products, networks and
worksystems. A Digital employee for 10 years, Steve began his
career as a sales representative in IndianaÂpolis, Indiana. He has
held a number of position in the Sales organization, most recently
as Connecticut District Sales manager.
The Personnel Policies and Procedures Manual
(the orange book) is now available on-line over the Easynet
network in the U.S. Any employee in the U.S. with a VAX account
can access the manual by typing "VTX ORANGEBOOK" at the dollar
sign ($) prompt. The main menu will provide directions to specific
policies, i.e., benefits, compensation, employee relations,
employment, relocation and travel, security, and new/revised
policies. Whereas before, employees having questions about such
policies had to borrow a manual from a manager, now all employees
in the U.S. should be able to access this information readily.
Hard copies of the manual, which are now
distributed to all managers, supervisors and Personnel people,
will still be offered. Currently, there are over 12,000 owners of
manuals in the U.S. Updates to the on-line manual will appear
simultaneously with the hard copy updates.
Contact your local systems manager if you have
any questions or problems accessing the manual.
The fifth episode of Digital’s Infinite Voyage
series - "The Search for Ancient Americans" - will air in the U.S.
nationwide over PBS on September 7, and on local commercial and
cable stations on September 12. This show will deal with the
origin of early Americans, showing recent discoveries of
archaeologists in Florida, New Mexico and Honduras.
This series of programs is underwritten on
public television and sponsored on commercial teleÂvision by
Digital. Future shows will cover "The New Dinosaurs," "The Web of
Life," and "NaturÂal Disasters."