Highlights
|
Slides
My Goal
- Honest self-examination
- Set DCE on a track for success
- See distributed computing technology deployed in
real applications
My Perspective
- Technologist - 20 years experience in R&D
- ISV - products supporting DCE for 2 years
- User - use DCE and DFS for development
- Developer - created DCE applications
- Advocate - supported DCE around the globe
- Observer - worked with numerous customers
deploying DCE
My Biases
- I helped design and develop DCE
- I represent an ISV that supports the use o
- DCE, but also provides alternatives
- I'm not a Unix bigot
- I have a strong aversion to marketing hype
- I have to face commercial reality
Retrospective
- DCE planning started in 1989
- Development started late 1990
- DCE 1.0 shipped in 1992
- DCE 1.0.3 widely available by 1994
- Most vendors today supporting DCE 1.1, many
bundling runtimes
The Origins of DCE
- Unix, C design center
- Complex APIs for systems programmers
- Command line interfaces for management
- Large footprint
- No subsetting
- Token (no) support for PC desktops
- Little attention paid to management, usability
What's Changed?
- RAD development tools: Visual Basic,
PowerBuilder, etc.
- Success of Windows, NT
- Survival of MVS - "Save the Mainframe"
- For many people - Nothing!
What's Changed?
- Distributed Objects - competition for mindshare
with OLE, CORBA
- Message Queuing - defensive posturing by MOMA
- WorldWide Web - rapidly becoming ubiquitous,
bottom-up approach to distributed computing
- Public Key Encryption
The Real World - Languages
The Real World - Platforms
- 100 Million Microsoft desktops
- Unix growing at 17%
- NT growing at 3-4 times the rate of Unix
- Corporate mainframes everywhere
The Real World - Applications
- The vast majority of client/server applications
are based on simple, 2-tier database access
- 3-tier client/server is growing at 4 times the
rate of 2-tier, but development tools are only
beginning to support this model
The Real World - Deployment
- Different people do development and management
- Managing cells is hard (relatively)
- Lack of DCE management expertise
- People don't want to invest in infrastructure
- Good planning is the exception, not the rule
The Real World - Marketing
- DCE is a strategic sale
- DCE is a solution sale
- DCE is a technical sale
- DCE is a long sale
- For close to 90% of our customers, DCE is No Sale
The Real World - Marketing
- DCE has not gained popular support
- The press and analysts are not on board
- Two ways to succeed:
- Have so much money they can't hear anyone
else talking
- Be so useful it doesn't matter
Is OSF too slow?
- DCE may be too much, too soon
- Emphasis in the market is not on scale,
robustness, security
- Emphasis is on productivity, speed of
development, ease of use
- DCE needs to learn from "just good
enough" philosophy - not be driven by early
adopter market of technical sophisticates
What's Missing?
- Application development tools
- DCE management
- ISV support
- Marketing and press
- Case studies
- Advertising
- High-level advocates
What's Needed?
- Integrate with OLE, CORBA, messaging
- Full support of a manageable security model for
the Internet
- Ubiquitous (and free) availability of DCE runtime
- Lower cost of entry to DCE infrastructure
Summary
- DCE is still the most advanced distributed
computing technology on the market
- DCE is the only possible solution for:
- Security
- Scalability
- Interoperability
- DCE is fighting an uphill battle for recognition
and acceptance
|
Presented
August 6, 1996 
Boston, MA 
