Evolving an important theme I've been working on and presenting all year, this new deck summarizes how enterprise architecture and large scale technology-based business solutions must transform to be more effective in the 21st century.
Contains material on a hypothesis for what's wrong with today's EA as well as potential solutions of merit such as emergent architecture, WOA, enterprise REST, open supply chains (APIs), mashups, and other models.
Presented this week in Oslo Norway to Bouvet's enterprise architecture council.
4. Software architecture
– The definition of the fundamental structure
and properties of software systems:
Components, resilience, scalability,
adaptability, reliability, changeability,
maintainability, extensibility, security,
technologies, standards, and other key
constraints.
6. Many sophisticated architectural
frameworks exist today
• 4+1
• Department of Defense Architecture Framework (DODAF)
• UK Ministry of Defence Architectural Framework (MODAF)
• The Open Group Architecture Framework (TOGAF)
• Zachman framework
• Federal Enterprise Architecture
• Reference Model of Open Distributed Processing (RM-ODP)
• Service-Oriented Modeling Framework (SOMF)
7.
8. Areas where traditional
EA models often struggle
• Don’t respond to change quickly enough
• Aren’t aligned with current business reality
• Lack of focus on driving consumption (or network
effects)
• Too centralized and isolated
• Expensive and resource-intensive
• Overengineered in the wrong places. Excessively
constraining.
11. Today’s Software Architectures
Are Also Extremely Sophisticated
• Highly distributed and federated
• Often have a social architecture
• Built from cutting edge ingredients
Example: http://clickatell.com
• Have to scale globally
• Set with expectations that are very high for
Integrating with 3rd party
functionality and low for the cost to
suppliers live on the Web
develop/own new solutions as well as being a 3rd party
supplier is the name of the
• Increasingly created with productivity-oriented game circa-2009
design & development platforms
12. There’s A Lot To
Master Today To
Architect Credible
Solutions:
13. • Some of this is around
what we call “2.0”
• Peer production and crowdsourcing
• Owning your classes of data on the network
• Using new distribution models to leverage
the Web as your platform
• Social systems
• Open supply chains and 3rd party sourcing
(http://programmableweb.com)
15. But existing integration models
have been challenged
• Most SOA initiatives are delivering low ROI to the business
• The reasons are many but boil down to:
– Lack of engagement: Focus on technology instead of business
problems.
– Slow adaptation: Top-down enterprise architecture moves slower than
the environment changes.
– Low levels of use: Important avenues of SOA consumption and
production points are often excluded from participation.
16. The results of a large new
SOA effectiveness study:
•“It has become clear to me
that SOA is not working in
most organizations.”
– Anne Thomas Manes,
Burton Group
17. Demand for Breadth
Integration
• “48 percent of the
CIOs we surveyed
said that they plan to
implement service-
oriented
architectures for
integration with
external trading
partners this year.” –
McKinsey & Co.
18. And we now have real-world experience with
traditional means of connecting to our data
• Traditional Web services
was a good first try but has
a long list of challenges for
the outcomes we desire
today.
• The model of the Web has
continued to teach us about
how to structure
information and services.
19.
20. Strange Attractors: Similarities
between Web 2.0 and SOA
• Web 2.0 • SOA
– Software as a service – Software as services
– Interoperability based on
– Interoperability based on Web heavyweight standards
principles
– Applications as platforms
– Applications as platforms – Permits unintended uses
– Encourages unintended uses – Composite Apps
– Mashups – Little user interface guidance
– Little prescription of user
– Rich user interfaces participation
– Architecture of Participation
21.
22. One Emergent Solution:
Web-Oriented Architecture
distribution and
composition Open APIs identity
OpenID and
Data Mashups security
WOA OAuth
Widgets Core
SSL
HMAC-SHA-1
REST
WADL
XML URIs
data formats BitTorrent
ATOM
and description
protocols and
WOA IXMLHTTPRequest
interfaces
JSON
Full
23. Enabling New
Consumption Scenarios
• Cut-and-Paste deployment
anywhere on the Intranet
• Consumption of the SOA in
any application that can use
a URL
• Discovery of data via search
• Integration moves out of
the spreadsheet
24. Recent technological
innovations coming primarily
from the online world
• Cloud computing
• Utility/grid/Platform-as-a-service
• Non-relational databases
• S3, CouchDB, GAE Datastore, Drizzle, etc.
• New “productivity-oriented” platforms
• RIA: Flex/AIR, JavaFX
• Stacks: Rails, CakePHP, Grails, GAE, iPhone, etc.
• Web-Oriented Architecture
25. Changes to the processes
that create architecture
• Increasing move to assembly and integration
over development of new code
• Perpetual Beta and “extreme” agile
• Community-based development and
“commercial source”
• Product Development 2.0
28. Benefits
• Dynamic response and adaptation to
change
• Architecture supported and driven widely
by local users
• Less waste
• More access to opportunity
• Better fit to business needs
31. Motivations for
Open Supply Chains
• Increase reach and head off
competition
• Tap into innovation
• Grow external investment
• Cost-effectively scale business
relationships
• Going from 10s to thousands of
integrated partners
32. Example: Amazon
• 1st Gen. Product: E-commerce store
– No differentiation
– Scaling of a single site
– Single site
• 2nd Gen. Product: E-commerce platform
– 55,000 partners using their e-commerce APIs live
– Scaling of the Web
• 3rd Gen. Product: A series of Web platforms
– Simple Storage Service (S3)
– Elastic Compute Cloud (EC2) EC2 S3
– Mechanical Turk (Mturk)
– Many others
– 300K businesses build on top of what they’ve produced
• 2nd and 3rd generation platforms generate large net revenue
34. The Market Share
Opportunity
• The vast majority of Internet user activity is
elsewhere, on 3rd party Web sites and applications
• If firms could reach this traffic, the growth potential is
as large as the Web itself
• Reaching this traffic before competitors do can
result in successful marketshare “lock-out”
• Businesses able to cost-effectively integrate with a
large number of partners to grow
• Access and offer value to existing ecosystems of
customers
35. Opportunity:
Going To the Customer
and Open Web APIs
Tens of Thousands of Dynamic Web Partners
Partner Partner Partner Partner Partner Partner Partner
New Business
Partner Partner Partner Partner Partner Partner Division:
Partner Partner Partner Partner Partner
Interact Additional
Partner Partner Partner Partner Revenue via
Usage Fees,
Live Web + Advertising, etc.
Integration $$$
Open API Monetization
Boundary
+
Consumer or
Business Interact
Online Business Direct Revenue
36. Platforms vs. Applications
Distribution Models Target Audiences
Consumers
Native App
existing
Web Application Small Businesses
Medium-Sized Business
Open Widgets
Power/Web Saavy Users
Facebook/Open Social
Developers
Web API
SDK, Developer Community, SLA,
Billing Businesses
37. Platforms vs. Applications
Distribution
Distribution Models Order of Magnitude Method
Native App 10M Users Push
Desktop Client API 10M Users Pull
Open Widgets
10-20M Users Pull
Facebook/Open Social
Open Web API
SDK, Developer Community, SLA, 100M+ Users Pull
Billing
38. Key API Goals
• Leveraging existing investments as much as
possible (reduce rework in design and architecture)
• Protect intellectual property around proprietary
capabilities
• Select API model that will result in 1) the most
developer uptake and 2) access to the largest
possible audience
• Selecting a discriminating factor (rich vs. reach)
• Scope: Graduated capability vs. full initial API
39. Long-term future usage
breakdown w/API
Other Apps
Embedded Apps • Reach every distribution
Web Mobile Apps channel possible
Web Widget Apps • Leverage 3rd party customer
iPhone Apps
bases
Open Social Apps • Cut off competitor’s growth
OPPORTUNITIES
3rd Party Web Apps • Ride the MAXIMUM
POTENTIAL growth curve
Facebook Apps (driving consumption)
• Harness innovation of
Existing Web Site or hundreds and thousands of
Application 3rd party developers
40.
41. Reasons Developers Select
APIs
Key to initial adoption Key to long-term adoption
• Provides access to • Reliable, well-known, scalable
functionality not possible provider that is trusted
to develop internally
• Developers can get answers to
• Easy to use and integrate
questions, support, and
with
problems fixed when bugs are
• Good documentation and found
easy to get started
• Strong user base for 3rd party
developers to tap
42. “Platforming” Your
Business
• Requires opening the server-side to 3rd party developers
• Allowing the construction of widgets and Web apps
offering some or of all of your functionality by external
partners
• Harnessing the innovation on the network
• Generating the greatest potential reach, competitive
lock-out, market share, and revenue
• Warning: Must maintain control of hard-to-recreate data
43. Open API Challenges
• Foreign business model for traditional companies
• Requires full-spectrum support from the business
(marketing, sales, customer service, technical
support, etc.)
• Successful monetization strategies vary greatly
• The biggest successes are firms which create a
well-funded dedicated business division
44. Open Supply Chains:
The bottom line
• Good repeatability
• Can be costly
• Unproven in some
industries (yet)
• Proven ROI
(example: $300M+
net revenue)
Strategic
Industry Play
45. High Velocity Processes:
The Web’s Version of Agile
• Shadow Apps for real-
time feedback
• Customer-Sampling
and Live Testing
• Granular Versions
(constant evolution)
• Daily, even hourly,
releases
49. 2.0 models are beginning
to transform everything
• Product Development
• Marketing and Advertising
• Operations
• Customer Service
50. The network is consistently
proving to be the best
solution for many classes of
problems
51. So how do we
re-imagine our
software
architecture for
the 21st century?
52. Challenges to Transitioning to
New Architectural Modes
• Innovator’s Dilemma
• “How do we disrupt ourselves
before our competition does?”
• Not-Invented Here
• Overly fearful of failure
• Deeply ingrained classical software culture
• Low level of 2.0 literacy
53. What we often see in
the marketplace today
• Too many copy-cat methods
• Failure of imagination and courage
• New architectural concepts as an after-
thought. Or tacked on as a “checklist” item.
• Companies that pay lip service to
innovation but are having trouble or
unwilling to make the necessary changes
54. Key Lesson:
We now have a
fundamentally new and
better set of lenses through
which to look at leveraging
value on the network:
55. • Push to pull systems
• Web 2.0 design patterns and business
models
• New modes of software, platforms, and
architectures
• Productivity-Oriented Platforms
• Web-Oriented & Emergent
Architecture
• New Distribution Models
56. It’s time to change
our DNA
• Moving from the 20th century towards
21st century businesses
• Deeply understanding the network and its
profound potential for creating growth and
building value
• Putting proven new models into the core of
our lines of business and enterprise
architecture