Careers: Interviews
Roger Sessions, an Expert in High-End Distributed Software
Architectures
This week,
Stephen Ibaraki, ISP, has an exclusive interview with Roger
Sessions, considered by many as the world's foremost expert in
high-end distributed software architectures, author of many books,
magazine articles, his own ObjectWatch Newsletter plus he heads up
ObjectWatch Inc.
Roger is a highly regarded conference speaker. ObjectWatch is an
information transfer company located in Austin Texas, specializing
in courses geared towards Software Architects and CxOs. They focus
on Microsoft's Windows Server Platform and Java's J2EE
architectures, since these architectures offer companies the
opportunity to build high throughput (100,000,000+ transactions per
day) commerce systems with low cost per transactions.
From 1990-1995 Roger Sessions worked at IBM involved in the CORBA
effort. He spent a year as a lead architect for the CORBA
persistence service and four years as one of the lead architects for
the IBM implementation of CORBA's persistence technologies. Roger
Sessions left IBM in 1995 to start ObjectWatch, Inc., a company
dedicated to offering training and consulting services in the field
of high scalability, component architectures. He started out
focusing on the CORBA technologies, but then turned his attention to
Microsoft's distributed technologies, including Java, COM, DCOM, MTS,
MSCS, and MSMQ.
His earlier books include: “COM+ and the Battle for the Middle
Tier”, “COM and DCOM; Microsoft’s Vision for Distributed Objects”,
“Reusable Data Structures for C”, “Class Construction in C and C++”,
and “Object Persistence: Beyond Object-Oriented Databases.”
Roger’s most recent book, Software Fortresses; Modeling Enterprise
Architectures, is published by Addison Wesley and provides an
essential roadmap to all aspects of software fortresses including
their design, and implementation. Roger is the originator of the
Software Fortress Model where enterprise systems are treated as a
series of self contained, mutually suspicious, but cooperating
software fortresses—matching the real world of J2EE and the
Microsoft Windows Platform found in many enterprises. Using Roger’s
Software Fortresses, you don't try to choose between enterprise
platforms, you can use them all by designing and implementing the
unifying architecture that recognizes the strengths and weaknesses
of each platform. Each fortress makes its own choices as to software
platform; data storage mechanisms; and how it interacts with other
fortresses through carefully crafted treaties.
Discussion:
Q: Roger, you have a particularly demanding schedule with your
expertise requested worldwide. We are most fortunate to meet up with
you. Thank you for sharing your years of experience by agreeing to
this interview.
A: As always, a pleasure.
Q: You have such a remarkable history starting with your studies at
Bard College in biology; working as a research scientist at the US
National Institute for Health; a stint as a VisiCalc developer at
Software Arts; developing software for Prime Computer; then onto IBM
and CORBA. Can you share some stories and lessons from these
multiple careers?
A: Let’s see. My most important lessons from each of these stages...
From my career in science, I learned the importance of thinking
about a problem logically. This means forming a hypothesis, creating
experiments to challenge that hypothesis, running those experiments
in a carefully controlled manner, and then interpreting the
experimental results with a healthy dose of skepticism.
From my work at Software Arts, I learned the importance of
predicting future trends. Software Arts created the original
spreadsheet program, VisiCalc, in 1979 and bet heavily on the Apple
IIe computer. VisiCalc was by far the most successful business
application of its time, fueling not only huge profits for Software
Arts but driving much of Apple’s business. IBM came out with the PC
in 1981, but Software Arts largely ignored this new machine and
continued its focus on the Apple IIe. Lotus Corporation bet
differently. It decided that the IBM PC would become the dominant
machine and wrote a new spreadsheet, Lotus 1-2-3, for that machine.
Lotus Corporation made the right call and within two more years,
Software Arts had gone from being the most important company in the
software industry to complete obscurity, all because they waited two
years too long to recognize what would become a decade defining
industry trend.
From my work at Prime, I learned what “mission-critical” means. I
was project lead for Oracle on Prime. One of Prime’s customers for
Oracle was Soloman Brothers. Soloman Brothers was a major finance
institution of its time and probably still is. Soloman Brothers
wanted to make sure that I, as project lead for one of their
dependent technologies, understood the nature of their critical
dependency on my software. They gave me personal tours of their
stock trading and computer operations. I saw first hand what it
means to a finance institution for a machine or application to go
down and how quickly downtime piles up into huge amounts of lost
income.
From my work at IBM, I learned both negative and positive lessons.
On the negative side, I learned how stifling politics can be in a
large bureaucracy. On the positive side, I learned how important it
is to manage customer relationships.
From my work on CORBA, at IBM, I learned the difference between
standards that succeed and those that fail. I learned that standards
that focus on portability fail while those that focus on
interoperability often succeed. CORBA was a huge standards activity.
It could be divided into two areas. The first was the hugely complex
CORBA API, which focused on portability. This easily accounted for
95% of the CORBA activity. The second was IIOP, which was about
interoperability. It accounted for at most 5% of the CORBA activity.
Although the CORBA API received almost all of the early hype, it was
ultimately the lowly IIOP that was the only part of CORBA that had
any real success.
Q: As a world renowned expert in distributed architectures, can you
share your views on the major competing technologies today, the
nature of these technologies, similarities and differences, their
strengths and weaknesses, market penetration, and where you see them
in the two year and five year time frame?
A: In the enterprise space, the major competing technologies are
IBM’s WebSphere and Microsoft’s Windows Sever Platform. BEA still
plays a role in the enterprise, but it seems to be diminishing.
WebSphere and the Windows Server Platform are both enterprise
platforms, that is, platforms on which you can build enterprise
systems. An enterprise platform is more than just an operating
system; it must also provide various cross-machine capabilities such
as:
• high reliability through clustering across redundant machines
• asynchronous messaging across the enterprise
• loosely coupled transactional support
• security services
• workflow management
• interoperability with other enterprise platforms
In general, the WebSphere technologies have two strengths. First,
they have better consulting support, primarily from IBM Global
Services. Second, they have support for most operating systems, in
particular, Linux.
In general, the Windows technologies also have two strengths. First,
because of their close operating system ties, they have better
performance than do the WebSphere technologies. Second, they are
less expensive.
Both systems have enough advantages so that both will survive.
The biggest surprise to many over the next two years will be the
dwindling influence of J2EE and Sun on IBM. IBM is already
deemphasizing its J2EE heritage. As Sun pushes the Java licensing
terms harder and harder, IBM will increasingly distance itself from
Sun. Within five years I predict that the term J2EE will disappear
from IBM marketing literature.
Q: Describe the evolution of your “Fairies, Gnomes, Wizards,
Dragons, and all kinds of other interesting and imaginative
figures?”
A: My figures are driven by my own boredom. I have sat through
countless presentations. I find it horribly tedious to look at slide
after slide showing boxes layered on boxes with more boxes within
boxes and yet more boxes connected to even more boxes with perhaps
an occasional circle thrown as a trinket for the masses. I have made
a conscious decision that I will not create more boring slides.
There are already more than enough boring slides to satisfy the
universe’s needs for eons to come.
Q: Considering the evolution of the enterprise, your latest book is
a must read for all IT professionals and I highly recommend it. We
would be fortunate if you could share some highlights from the book
and important lessons. Also if you could expand on the different
fortresses including: Presentation, Web Service, Line of Business,
Legacy, and Treaty Management.
A: The most important lesson is that architecture is not the same as
technology. Architecture is about how to think about systems.
Technology is about how to implement systems. A good architecture
can be implemented with any number of technologies. The technology
becomes an implementation detail of the architecture. If there is
one critical lesson for IT organizations, it is this: Architecture
Comes First!
Enterprise architectural models, such as the Software Fortress
Model, need to focus on enterprise level issues. These issues
include:
• How one thinks about individual systems and how they relate to
other systems in the enterprise.
• How systems should protect themselves from the riff-raff.
• How systems interoperate with other systems.
• How systems identify themselves to each other.
• How transactions should flow through an enterprise.
The Software Fortress Model considers an enterprise to be made up of
a number of self contained systems. I call these systems “software
fortresses”, because they have the following fortress-like
characteristics:
• They are self-contained.
• They form a trust boundary.
• They interact with other software fortresses through well defined
relationships.
• They are mutually suspicious of each other.
• They are each responsible for their own security, although they
may use other fortresses for credentials management.
Interestingly, as one goes through the different types of enterprise
systems and starts to understand them within the context of the
Software Fortress Model, one notices certain recurring
characteristics and patterns. This leads to a natural classification
of systems. I believe there are six basic types of systems, or
software fortresses. These types are as follows:
• Presentation fortresses, which interact with browsers.
• Web service fortresses, which interact with collaborating systems
over the Internet.
• Business application fortresses, which process mission critical
business logic.
• Service fortresses, which provide enterprise level services such
as credential management.
• Legacy fortresses, which wrap legacy systems.
• Treaty management fortresses, which manage complex workflow.
Q: What do you see on the horizon that businesses and IT
professionals “must” be aware of to be competitive?
A: A difficult but important skill is the ability to compare and
contrast technologies. That seems simple. Take IBM and Microsoft,
for example. IBM’s WebSphere compares to Microsoft’s Windows Server
Platform. Doesn’t it? In reality, WebSphere consists of more than 70
technologies, and the Windows Server Platform consists of more than
15 technologies. You can and should build enterprises systems that
choose from among all of these. But which of these technologies are
comparable to which? This is very difficult to determine, for these
reasons:
• Most people don’t have a standard model for how to build
enterprise systems in the first place. Therefore it isn’t even clear
which of these technologies you really need and where exactly you
need them.
• Both Microsoft and IBM have overlap in their own technologies. You
can usually use more than one technology to accomplish the same
goal. A good example of this is database stored procedures and
middle-tier components. Both can be used to write business logic.
Which is the preferred technology? You won’t get much help from
either IBM or Microsoft figuring this out.
• Both Microsoft and IBM are constantly renaming, repackaging, and
repricing their technologies. Even if you did manage to figure out
how they compared at one moment in time, things would quickly
change.
I have been working on a standard process for conducting these types
of evaluations that I call the Software Fortress Enterprise Planning
Process (SF-EPP). My hope is that this will go a long way to helping
people compare and contrast technologies in a meaningful way.
The SF-EPP is broken into several steps. The first step involves
doing a software fortress model for the enterprise. The second step
is defining the categories of systems needed in a particular
enterprise. The third step is conducting a requirements analysis for
those categories. The fourth step is determining candidate
configurations that meet those requirements. The final step is
choosing from the different candidate configurations. I hope that
the SF-EPP will help business decision makers and CxO types to
better understand how to make these choices. For those readers
interested, I have an extensive white paper and accompanying
spreadsheet that takes you through this process. It is titled
“Modeling Software Architectures and Enterprise Choices”. It is
available free at the ObjectWatch web site (www.objectwatch.com)
under ObjectWatch White Papers.
Q: What do you feel are the top five hottest topics of interest to
both businesses and IT professionals today and what will be the
topics in two years and in five years?
A: The trends that I think will be the most important over the next
few years are these:
• The continued displacement of big expensive machines by little
cheap machines.
• Industry standards for security based on PKI.
• Continued work in the area of inter-enterprise messaging standards
and implementation of those standards.
• Intellectual models for enterprise architectures and tools to
support those models.
• The almost total movement to browser-based interfaces.
Q: Who/what do you think are the winners and losers in IT in next
five years? [This could be companies, technologies, …and so on.]
What advice would you give to enterprises in their adoption of
technologies in the next five years?
A: Sun Microsystems is rapidly heading for the trash bin. BEA will
struggle along for a while. All other J2EE vendors with the
exception of IBM will either disappear or be marginalized. IBM and
Microsoft will be the big winners in the enterprise space. IBM will
focus more and more on consulting. Microsoft will focus more and
more on cost and performance. My advice is simple. If you are an
enterprise and you want a Java or a non-Windows solution, go to IBM.
If you want low cost, high performance, and scalability, go to
Microsoft. If you go to anybody else, you are taking a risk.
Q: What would be your recommended top references for the business
professional? And what would be your recommended top references for
IT professionals?
A: For the business professional:
• On Writing Well by William K. Zinsser
• Elements of Style by William Strunk Jr., and E.B. White (of
Charlotte’s Web Fame)
• The Chicago Manual of Style by The University of Chicago Press
• Roget’s 21st Century Thesauras
For the IT professional:
• Of course, Software Fortresses; Modeling Enterprise Architectures,
by yours truly.
• UML Distilled, by Martin Fowler with Kendall Scott
• The CERT Guide to System and Network Security Practices by Julia
H. Allen
• Transaction Processing: Concepts and Techniques by Jim Gray and
Andreas Reuter (oldie but goldie)
• Using CRC Cards by Nancy M. Wilkinson (another oldie but goldie)
• And, last but not least, Applied Cryptography by Bruce Schneier.
For everybody:
• Harry Potter and the Order of the Phoenix by J. K. Rowling,
because it is nice to be able to talk to your kids and even nicer to
be one.
• A good book of poetry, because what is the point of life without
poetry?
Q:What are the top challenges facing businesses and IT departments
in the next five years and what are your recommendations to
meet/overcome these challenges? Please provide specifics…
A: The top challenges are to integrate, interoperate, and innovate.
The companies that win over the next few years will be those that
can integrate their own systems together, interoperate with other
companies, and come up with innovative ways of leveraging
information and relationships.
Q: Can you comment more about the Open Source Movement—its current
position, its philosophy, the major innovations, and where it’s
going?
A: Ultimately, I don’t expect the Open Source Movement to go
anywhere. It is primarily fueled by the “Stop Microsoft” sentiment,
as was CORBA and J2EE before, the former of which clearly failed and
the later of which is well on its way to failure. Open Source is
still in the honeymoon phase, but the harsh reality of daily life
will rapidly dull its luster. Here are the primary reasons I don’t
expect it to go anyplace:
• There is no way to ensure that open source code doesn’t violate
intellectual property laws. Anybody can put in code from anyplace.
There is no way to know where the code came from or what legal
agreements may have violated along the way. The current SCO lawsuit
against IBM is just the tip of the coming iceberg, in my opinion.
• There is no business model for its ongoing success.
• There is no quality control.
I know people will hate me for saying this, but Open Source is a
dead-end for the enterprise. I make this prediction with great
confidence: ultimately, the only successful Open Source projects
will be those that IBM takes over and proprietarizes.
The only winner from Open Source will be IBM who will use Linux as
an excuse to discontinue support for operating systems it no longer
wishes to support; to drive the final nail in Sun’s coffin; and to
strike fear in the heart of Microsoft. No wonder IBM loves Open
Source! Open Source is IBM’s dream come true.
Q: What additional books are you planning in the near and far term?
A: I would like to do a book on Software Fortresses for developers
(my last book was really for architects). I would like to do a book
on case studies of the Software Fortress Model. I would love to do a
book of poetry. Beyond that, who knows?
Q: From a context of past, present and future, what drives you to do
what you do?
A: The fun, the challenge, and an endless supply of doppio
macchiatos from Starbucks.
Q: What are ObjectWatch’s current vision, mission and roles,
strategies and values and how will they evolve over the next ten
years?
A: We are an independent information transfer company specializing
in information that is critical to CxOs and Architects. I don’t see
that changing. What will change is the information that we transfer
and the technologies that we use to accomplish that transfer.
Q: Your list of accomplishments are staggering. Which ones standout
foremost in your mind and what lessons can you share with our
audience?
A: I still get my greatest kick out of writing the ObjectWatch
Newsletter (available at www.objectwatch.com). I love it when
somebody comes up to me and starts talking about an article I wrote
years ago and how it helped them understand some major technical
issue for the first time. If you are interested in technical
writing, my best advice is to remember that technical writing is,
first and foremost, writing. Treat it as non-fiction writing about
technical topics. That means that you must strive to write well,
write clearly, write succinctly, and write in an interesting and
engaging style that defines you as a writer.
Q: If you were doing this interview, what other question would you
ask of someone in your position and what would be your answer?
A: I would ask this question: What keeps you up at night? What
development in the IT industry do you fear the most? My answer is
that somebody, someplace, will discover and publish an algorithm
that can be used to easily determine a private PKI-type key given a
public key. That would be the biggest disaster I can think of that
could befall our industry.
Q: Roger, we are very appreciative of the many contributions you
have made to the world, providing shape and a clarity for so many.
Thank you for coming in to share your views with our audience. We
look forward to reading your insightful books, articles, and
benefiting from your wisdom in your global seminars. Roger, you have
a most illustrious career. I asked this question before, “If you had
to do it over again ….?”
A: Next time, I think I’ll order chocolate.
|