How to Fix Your

Integration Problem
AND

Accelerate Digital
Transformation
B Y V I J AY T E L L A
ntegratio Table of Contents
INTRODUCTION: WHAT IS DIGITAL
TRANSFORMATION?
WHAT’S HOLDING BACK DIGITAL
TRANSFORMATION?
The Technology Gap
3

Customer data is more fragmented than

ever.

Integrations are now much more dynamic

4

and pervasive. 9
Digital customer journeys and experiences
require more than integration. 10
They help automate digital customer
journeys to deliver a modern, engaged CX.11

They deliver smarter, faster process

automation with AI & machine learning. 12

They put your employees in charge with the

help of intelligent workbots. 15

The Social Gap 16

Going digital requires ‘digital literacy’ far
outside of IT. 18

DIGITAL PLATFORMS MUST WORK

DIFFERENTLY! 21

Automations at scale must be

community-powered. 23

So what does an intelligent

automation platform look like? 24

CONCLUSION 26

REFERENCES 27
 3
Introduction:
What Is Digital
Transformation?
Digital transformation is one of the most popular
business trends. But what is it, and more importantly,
how does your business achieve it? In short, digital
transformation is the application of digital capabilities
to processes, products, and assets to improve efficiency,
enhance customer value, manage risk, and uncover new
monetization opportunities.1 These transformations
are driven primarily by changing technology, as well as
by evolving markets and consumer demand.2 Research
shows that two-thirds of Global 2000 companies will, by
the end of 2017, have placed digital transformation at the
core of their corporate strategies.3

Research also reveals that as many as two-thirds of

digital transformation projects fail.4

Let that sink in for a moment. Businesses feel pressure

to implement digital transformation initiatives, but up
1 http://www.cio.com/article/3199030/analytics/what-is-digital-transformation.html
2 https://www.theguardian.com/media-network/media-network-blog/2013/nov/21/
digital-transformation
3 https://www.coxblue.com/21-jaw-dropping-statistics-shaping-digital-transforma-
tion-fortunes/
4 http://www.consultancy.uk/news/2656/two-thirds-of-digital-transformation-proj-
ects-fail
 4
to 66% of those projects fail. For such a popular trend,
those are strikingly depressing odds. Why is the failure
rate so high, and how can businesses ensure that their
digital transformation projects are successful?

What’s holding
back digital
transformation?
Whether you realize it or not, you have an integration
problem. VP of Constellation Research, Holger Mueller
explains:

“The problem is, when you build something new, outside of the
scope of what you have integrated, you by definition have an
integration problem. With the multiple new things happening in
the public cloud, you either have an integration problem between
existing and new technology, or with multiple clouds, which you
have to link together because you have fragmented automation.
In other words, if there’s innovation happening, then by definition
you have an integration problem - and there should always be
innovation happening.

“This creates the need to have an integration tool. The integration

tool and process needs to be faster than the scope at which you are
 5
building new systems, which is what creates integration problems.
If your integration tool is not able to move fast enough, you can
never catch up!”5

Much of this failure is related to the inadequacy of

integration platforms to serve a rapidly evolving digital
business. The first integration platform appeared in the

5 https://www.workato.com/blog/2017/08/holger-mueller-digital-transformation/

Digital Transformation by Company Size

45% Source: “Thriving in the Digital Economy”, IDC, February 2016

41%
40%

35%
33%
30% 29%

25% 25% 25%

22%
21%
20% 19% 19%

15% 15%
13%
12% 12%
10% 10%

6%
5%

Little to Beginning to Coordinating Well underway Have gone beyond

no progress automate and automating in applying integration to
functions activities technology optimize processes

10-99 Employees 100-499 Employees 500-999 Employees

 6
late 1980s. Known as The Information Bus (TIB), it was
designed to be the software analog of a hardware bus. Its
goal was similar: getting disparate, complex (in this case,
software) components from different vendors to work
together seamlessly. TIB prompted a wave of middleware
technologies and public companies like Informatica,
BEA, IBM, Webmethods, and Oracle, which in turn
evolved into more cloud-y versions of those products like
Mulesoft, Boomi, Snaplogic, and Jitterbit.

For a long time, the tech industry assumed that they’d

solved The Integration Problem, so they moved on to
other projects like SaaS, consumerization, and big
data. There was a sense that middleware products and
technologies had matured and become full-featured.
But the urgent and universal digital transformation
imperative—and the explosion of cloud apps and APIs
unleashed by it—has changed everything. It has become
both the biggest opportunity and an existential threat for
companies to transform to dynamic digital companies in
the mold of innovators like Amazon, Salesforce, Netflix,
GE, Starbucks, and Walmart and avoid the fates of a
Blockbuster, a Borders, or a Kmart.

All of this has meant an incredibly high rate of change

across the board:
 7
• In the business strategy as new threats and opportunities
emerge.
• In the number of apps and cloud, big data and AI
technologies being adopted.6
• In the markets and competitive landscape.
• And even in their own employee and customer
expectations, which are set by the revolution that’s
happening in the consumer tech in their lives.

Such high level of dynamism has put tremendous

stresses on business and IT groups across the company
to speed up the painfully slow transformation programs.
There are two major problems holding back companies
today: one is a technology gap in the technical
capabilities required to for modern integrations and
a social gap in the form of Business-IT divide. A new
paradigm, intelligent automation, and supporting
tooling based on a digital native approach is required
to address these gaps and unlock the power of digital
transformation in your business.

THE TECHNOLOGY GAP

Traditional integration tools—which were created
over a decade ago—are inadequate for both handling
this level of dynamism and for supporting digital
6 https://newsroom.cisco.com/press-release-content?articleId=1714059
 8
customer journeys and workflows that cut across your
apps and extend into your customers’ and partners’
systems. In fact, a survey shows that businesses see
current IT systems as the third biggest hurdle to digital
transformation.7

1. Customer data is more

fragmented than ever.
The average large company now uses close to 1300
applications.8 Customer data is splintered across
hundreds of apps across your IT and business groups.
At best, this fragmentation means poor visibility into
customers, inconsistent or incorrect info about them
stored in different places, and incomplete or irrelevant
customer info in the hands of your workers. At worst,
the uncontrolled data sprawl increases business and
compliance risks due to poor transparency and lack
of governance around critical customer information.9
Either way, it results in subpar or counter-productive
digital engagement with your customers—the antithesis
of digital transformation. In 2016, an estimated 90%
of businesses are working under this fragmented
approach.10
7 http://ebooks.capgemini-consulting.com/The-Digital-Advantage/files/assets/ba-
sic-html/page1.html
8 https://blogs.cisco.com/cloud/shadow-it-rampant-pervasive-and-explosive
9 https://www.coxblue.com/21-jaw-dropping-statistics-shaping-digital-transforma-
tion-fortunes/
10 https://blogs.cisco.com/cloud/shadow-it-rampant-pervasive-and-explosive
 9
2. Integrations are now much
more dynamic and pervasive.
It used to be that integrations were between a few
departmental System-of-Record apps, like ERP, CRM,
marketing, customer support, HR, and ITSM. Once
implemented, no one expected these integrations to
change much. The traditional integration tools were
created for use by IT and specialists to solve this smaller
set of integrations between systems-of-record.

The Transformation of Application Integration

Point-to-Point Hub-and-Spoke Bus Middleware Microservices

With point-to-point Hub-and-Spoke offers The bus method of app Buses can serve an When it comes to cloud
integration, a piece of a centralized hub that integration allows enterprise on their computing,
data goes from one connects the data and functions to own as an app microservices are the
application, gets applications, travel and systems to integration service, emerging standard
translated and reformats and speak to each other but middleware can, application
reformatted, and then translates the data, quickly, and without too. Middleware suites architecture.
travels to the next and then decides what human interference. can contain bus or Microservices, if
application, and the system to send to next. This is done through a hub-and-spoke properly done, fix
next, and so on. While While faster for data, set of defined (messaging) app potential data traffic
this route is hub-and-spoke standards that allows integration tools, also. jams by allowing data
straightforward, integration requires any new app to accept to travel to
things get quickly development teams to and emit data applications both in
complicated with more step in and manually (provided it fits the the cloud and on
applications and data. direct the data with standards). premises.
The larger the system, runtime components.
the slower the path.
 10
Integrations now must also cover hundreds of Systems-
of-Engagement (like Slack, Email, Twilio, Zoom, and
social media), Systems-of-Insight (like Redshift, Tableau,
and Splunk), and Productivity apps (like JIRA, the
Google Suite, and Trello) across departments. This
scenario is significantly different from what traditional
tools were designed for11, because:

• The scale of integrations to be done is much greater and

growing rapidly.
• Integrations must happen fast—in hours to days, not
weeks.
• Integrations need to change frequently as business needs
evolve and the set of apps change.
• Integrations need to be created by business users—not
just IT—as it is the business that’s driving the growth of
apps.
• Integrations still need to governed, secure, and compliant.
Otherwise, they’re just even more shadow IT.

3. Digital customer journeys and

experiences require more than integration.
An intelligent automation platform has to do a lot more

11 http://www.itbusinessedge.com/blogs/integration/just-the-stats-cloud-and-saas-inte-
gration.html
 11
than back-end integrations of a few big departmental
apps. It must be able to:

• Automate digital customer journeys for better CX.

• Deliver smarter, faster process automation with AI and
machine learning.
• Empower employees to better serve customers by
delivering them timely, relevant information, assist in
the automation of their personal workflows across all of
their apps.

A) THEY HELP AUTOMATE DIGITAL CUSTOMER

JOURNEYS TO DELIVER A MODERN, ENGAGED CX.

Digital transformation, at its heart, is about delivering

new products and services to customers in new ways, via
new channels, based on new business models to engage
them better. Transforming the customer journey with
things like streamlined lead prospecting, enhanced
customer onboarding, preemptive churn reduction,
and accelerated quote-to-cash is at the core of digital
transformation.

Workflow automation is essential to defining and

orchestrating detailed business processes across all
your systems, people, partners, and customers. From a
technical requirements standpoint, it means:
 12
• Complex logic and process automation.
• Across thousands of apps, APIs and technologies.
• Human workflows and approvals.
• Transactional workflows: processing events exactly once,
in sequence.
• Rolling back busted transactions via compensating
transactions.
• Re-entrant processes to support long running workflows
• Google Docs-like auto-versioning of workflows to make it
easy to go forward or roll back.
• Automated error and exception management.
• The ability to process millions of transactions per day and
scaling up on demand.

B) THEY DELIVER SMARTER, FASTER PROCESS

AUTOMATION WITH AI & MACHINE LEARNING.

Even with a great workflow automation tool, re-

inventing the wheel to solve the same problems from
scratch that thousands of other companies have
already solved is not very smart. As a company with
an aggressive timeline, you want to start with the best
practices and then apply your secret sauce or point-of-
view.
 13
What would it mean for your business if you could access
smart, customizable workflows automatically created
for you based on proven results from thousands of other
companies doing similar things? An intelligent process
automation platform gets you to smarter processes more
quickly. It changes the digital transformation game
because:

• It applies machine learning technology to a rich corpus

of popular business workflows being used by thousands
of other innovative companies. Such technology would
automatically create customizable workflows for you that
are smarter and more well-thought-out than you would
likely get by starting from scratch.12
• It weaves in rich, third party AI technologies like Watson,
Cortana, and Einstein to create smarter workflows than
before.13 When a customer support request comes in, for
example, having Watson analyze the tone, urgency, and
emotion of the message can help you dictate smarter
automations surrounding how that request is handled.
• It helps you vastly simplify B2B and business process
outsourcing. Business workflows must extend to engage
thousands of companies and individuals in your customer

12 https://www.accenture.com/t20160125T111718Z__w__/us-en/_acnmedia/Accen-
ture/Omobono/TechnologyVision/pdf/Intelligent-Automation-Technology-Vi-
sion-2016.pdfla=en#zoom=50
13 http://www.teletech.com/resources/articles/machine-learning-next-generation-in-
sight#.WYSYsNPyuRs
 14
and partner ecosystems. Brute force, one-at-a-time
workflow execution is expensive and error-prone, and
it lacks digital engagement. Workflow-as-a-service can
drastically simplify operations, reduce costs, and improve
engagement.14

14 http://www.computerworld.com/article/3083264/it-careers/automa-
tion-not-cheap-labor-is-reshaping-outsourcing.html

The Marketing Funnel & Workflow

Newsletter
Social
Email Ads

Social
Online PPC Reviews Networks
Ads
Website Community Blog
Blog Forum

Awareness Interest Consideration Purchase Retention Advocacy

Word Media Store Promotions

of Mouth
Email Ecommerce
FAQ
Knowledge Base

PR Radio
TV
Print
 15
C) THEY PUT YOUR EMPLOYEES IN CHARGE WITH
THE HELP OF INTELLIGENT WORKBOTS.

The fragmentation of apps and splintering of customer

context is not just a back-end or a plumbing problem
for IT. It impacts every employee in the company as
they work with dozens of business apps to get their jobs
done. The customer context needed by a support rep is
spread across a dozen different apps—not just in, say,
Salesforce. Employees’ individual workflows and tasks
are fragmented across these apps, with tens of tabs open
in their browsers or their mobiles. They can miss time-
critical business updates, must constantly app-hop to
gather context, and then must open even more apps to
take actions before repeating the process all over again.
All of this is is incredibly frustrating and productivity-
killing. Today’s employees and customers expect a
smarter, more fulfilling work experience.

Enterprise chat products like Slack, Microsoft Teams,

and Facebook@Work are vying to become the single
place from which employees can collaborate with
coworkers and get most of their work done. But for such
platforms to realize these ambitions, they need a digital
assistant that:

• Watches over their apps and prompts them with the

important updates across their apps that are relevant to
them and their jobs.
 16
• Provides 360° customer context or relevant historical
data around those alerts.
• Helps employees take actions in their apps without
leaving the chat console.
• Does all of this respecting the privacy and security
of data an employee is authorized for in a shared,
collaborative work setting—like Slack or MS Teams.

A smart digital assistant empowers employees to be

like mini-CEOs, with access to everything they need to
get their jobs done quickly and well to enable next-gen
customer journeys and experiences.

THE SOCIAL GAP

While fragmented data and the digital platform
technology gap are significant hurdles for
transformation, perhaps the most important challenge
to overcome is the Social Gap or divide between business
and IT. True transformation requires that you harness
the full power of your entire organization.

Enterprise-grade tools are accessible only by IT and

developers, while the digital business initiatives are
being driven by business users across the company. Since
IT can’t keep up with all the integration requests from
the business15, shadow IT has exploded. According to
15 https://www.accenture.com/t20160919T031307Z__w__/us-en/_acnmedia/Accen-
ture/Conversion-Assets/DotCom/Documents/Global/PDF/Technology_10/Accen-
 17
Gartner, simply managing shadow IT accounted for 35%
of IT expenditures in 2016.16 The proliferation of rogue
IT leads to poor visibility, lack of governance, high
corporate risk, and an ever-expanding divide between
business and IT, with as few as 8% of IT teams having
strong knowledge of their companies’ unmanaged cloud
apps.17

IT and business teams using different tools and working

at different speeds, on different wavelengths, for
different purposes is the biggest obstacle to companies
becoming dynamic digital businesses.18 Bi-modal IT
(control OR productivity) is a Faustian bargain; your
company needs both to succeed, but 81% of executives
believe IT cannot effectively support multiple objectives
at once.19

More than ever, businesses need an intelligent

automation tool that bridges this divide, aligns the
organization, and makes it viable for business users to
create robust automations in collaboration with IT for
governance, compliance, transparency, and security.
Business users should also be able to operate these
ture-Multi-Speed-IT-PoV.pdf#zoom=50
16 http://www.itproportal.com/2015/11/20/why-shutting-down-shadow-it-stifles-inno-
vation/
17 http://www.itproportal.com/2015/11/20/why-shutting-down-shadow-it-stifles-inno-
vation/
18 http://sloanreview.mit.edu/projects/strategy-drives-digital-transformation/
19 https://www.accenture.com/us-en/insight-calibrating-multi-speed-it
 18
integrations themselves. An intelligent automation
platform must be accessible to all and work differently,
drive teaming and governance between business and
IT, and leverage the power of business community to
execute faster and smarter.

Going digital requires ‘digital

literacy’ far outside of IT.
Business users must be able to automate workflows
themselves or maintain those created by IT so they can
make ongoing changes as needed. “Citizen developers”
and “citizen integrators” are among the most misused
terms20 when it comes to integration platforms, because:

• The majority of all “citizen developers” come from a

strong technical background21, but a normal Salesforce
admin or a marketing analyst simply cannot use the
“citizen integrator” tools out there. These analysts and
admins know their apps, but not things like XML, XSLT,
REST, SOAP, or deployment concepts that traditional
tools require them to know. Intelligent automation tools
must present an entirely different, business-friendly level
of abstraction around these technologies.
• An intelligent automation tool must present a rational,
humanized API-level interface so an average business

20 https://www.forbes.com/sites/jasonbloomberg/2016/05/16/citizen-develop-
ers-low-code-is-now-enterprise-class/#3cd13b723b4f
21 https://devops.com/meet-citizen-developer/
 19
user can create workflows and maintain them. With the 22

existing integration tools, implementing an integration

or a workflow that is logically ten steps when written
down on paper often devolves into 100 micro-steps with a
ton of look-ups and low level API calls in order to get the
job done. Such integrations are hard to create, impossible
to understand, and painful for IT to maintain— let alone
business users.
• Exceptions are the norm in enterprise automations.
Handling these should not force work back onto IT.
Business users must be able to operate the integrations
themselves in a governed, secure manner. A digital native
approach means automating the automations to recover
quickly from data or process errors. It supports an
intuitive ability to review errors, fix the data or process,
and re-run them. An intelligent automation platform
lets you set custom recovery and notification policies to
proactively detect and resolve the inevitable data, process,
or other exceptions.

22 https://www.ibm.com/developerworks/websphere/library/techarticles/1503_
clark/1305_clark.html
 20

Practical and Critical Thinking

E-safety
Functional Skills and Evaluation

Digital
Literacy
Cultural & Social
Creativity
Understanding

01101001
01101110
01100110
01101111

Proficient Curate
Collaboration
Communicator Information
 21
Digital platforms must
work differently!
You don’t go to your Instagram account and “provision”
it for certain number of photos, videos, shares or
comments; you don’t think about setting it up for “peak
loads”; you don’t configure (or be forced pay for!) “high
availability” so the service is Always On for you; and
you certainly don’t operate the sharing your posts and
feedback! However, today’s cloud iPaaS and ESB tools
work exactly this way. This is incredibly backwards in
today’s enterprise platforms!

An intelligent automation tool—which enables speed,

agility, and dynamism with beautiful experiences—must
be based on a digital native architecture and operate very
differently in 2017. It must work more like an Instagram
account (where everything but taking the pictures is
automatic) than like a camcorder, where you have to do
literally everything.

An intelligent automation platform, one that is based on

a digital native architecture:

• Is highly productive, fun, and delivers instant feedback

and quick gratification.
 22
• Has instantaneous “deployment” of workflows—not a
long, technical process. No need to think about or set up
nodes, servers, atoms, or molecules!
• Scales automatically. No guesswork needed around
“provisioning” (and paying for!) sufficiently overhead
server capacity that sits unused 99% of the time.
• Is Always On, just like your Instagram. Redundancy and
high-availability are built in, not options you can opt into
and be forced to pay for.
• Recovers automatically from common errors like API
rate limits, network disconnect, API going down, and
expired authentication tokens.
• Drives secure, collaborative, fine-grained IT governance.
There will not be real empowerment without security
and compliance to corporate standards. Security and
empowerment are two sides of the same coin.
• Supports a role-based collaborative integration project
lifecycle across the different different roles and orgs.
• Detects automatically, anomalous data traffic patterns
(via ML) and prompts admins with courses of action.
• Operates in real time with built-in analytics.
• Is open (composable) to enable complex business
workflows from simpler ones.
 23
AUTOMATIONS AT SCALE MUST
BE COMMUNITY-POWERED.
Automating workflows mainly consists of two things:

• Thinking about the automation challenge and designing a

solution for how you will automate.
• Transcribing this design into a workflow automation tool
to actually create and run workflows.

The traditional approach to automations is that of a lone-

wolf working on their own on the integrations. They
(typically an integration specialist) rethinks integrations
from scratch, implements them from scratch, deploys
them from scratch, and then hands them over to
business or IT. This is how developers worked, too, until
GitHub and the open source movement came along. Now
developers are vastly more productive by collaborating
with other developers and leveraging the work that
thousands of others have done. GitHub solves not just
the productivity problem around writing code; it also
solves the thinking problem around designing what code
to write!

Similarly, intelligent automation for digital businesses

demands a collaborative community of business
users within your company, across companies and
industries—a GitHub for automations. Community-
powered integrations:
 24
• Have more thought behind them and are more whetted
than lone-wolf integrations.
• Are at least 10x faster to get done and function better out
of the gate.
• Are secure and protect your unique value differentiation
via private integrations and automations that cannot be
shared back to the community.

SO WHAT DOES AN
INTELLIGENT AUTOMATION
PLATFORM LOOK LIKE?
Intelligent automation, then, is necessary for digital
transformation. But choosing an integration solution can
seem daunting, especially to the less technically inclined.
How do you know whether a platform is intelligent or
not? When you’re evaluating an integration platform,
ask:

• Can it connect all your cloud apps, ground apps,

databases, processes, and APIs?
• Is it enterprise-grade? Can everyone, both business and
IT, use it in harmony?
• Does it use machine learning to creates better
integrations, faster?
 25
• Can it create smarter automations with AI tools like
Watson and Einstein?
• Does it auto-scale? Is it always on? Does it instant deploy,
and are there absolutely no DevOps required?
• Does it feature automated error repair and rollback, with
no data loss or duplicates?
• Will it deliver relevant info and better workflows with AI-
powered chatbots?
 26
Conclusion
The opportunities for businesses to grow through
intelligent automation are huge. When lines-of-business
users are empowered to work more efficiently with
integrations that are governed, secure, and easy to
implement, the productivity gains offer a significant
return on investment.

Before investing in digital transformation, organizations

need to determine whether their integration solution
meets the requirements of a dynamic digital business.
Can business users create integrations to fit their rapidly
changing needs? Are those integrations still centrally
governable by IT? Does the platform require strong
technical knowledge, or is its UX user-friendly to app
admins and non-specialists? Is there a robust community
pool of replicable integrations? If the answer to these
questions is “no,” it’s not a an intelligent platform. And
without intelligent automation, a digital transformation
project will likely fail.
 27
References
1. http://www.cio.com/article/3199030/analytics/what-is-digital-transformation.html
2. https://www.theguardian.com/media-network/media-network-blog/2013/nov/21/
digital-transformation
3. https://www.coxblue.com/21-jaw-dropping-statistics-shaping-digital-
transformation-fortunes/
4. https://www.workato.com/blog/2017/08/holger-mueller-digital-transformation/
5. http://www.consultancy.uk/news/2656/two-thirds-of-digital-transformation-
projects-fail
6. https://newsroom.cisco.com/press-release-content?articleId=1714059
7. http://ebooks.capgemini-consulting.com/The-Digital-Advantage/files/assets/basic-
html/page1.html
8. https://blogs.cisco.com/cloud/shadow-it-rampant-pervasive-and-explosive
9. https://www.coxblue.com/21-jaw-dropping-statistics-shaping-digital-
transformation-fortunes/
10. https://blogs.cisco.com/cloud/shadow-it-rampant-pervasive-and-explosive
11. http://www.itbusinessedge.com/blogs/integration/just-the-stats-cloud-and-saas-
integration.html
12. https://www.accenture.com/t20160125T111718Z__w__/us-en/_acnmedia/Accenture/
Omobono/TechnologyVision/pdf/Intelligent-Automation-Technology-Vision-2016.
pdfla=en#zoom=50
13. http://www.teletech.com/resources/articles/machine-learning-next-generation-
insight#.WYSYsNPyuRs
14. http://www.computerworld.com/article/3083264/it-careers/automation-not-cheap-
labor-is-reshaping-outsourcing.html
15. https://www.accenture.com/t20160919T031307Z__w__/us-en/_acnmedia/Accenture/
Conversion-Assets/DotCom/Documents/Global/PDF/Technology_10/Accenture-
Multi-Speed-IT-PoV.pdf#zoom=50
16. http://www.itproportal.com/2015/11/20/why-shutting-down-shadow-it-stifles-
innovation/
17. http://www.itproportal.com/2015/11/20/why-shutting-down-shadow-it-stifles-
innovation/
18. http://sloanreview.mit.edu/projects/strategy-drives-digital-transformation/
19. https://www.accenture.com/us-en/insight-calibrating-multi-speed-it
20. https://www.forbes.com/sites/jasonbloomberg/2016/05/16/citizen-developers-low-
code-is-now-enterprise-class/#3cd13b723b4f
21. https://devops.com/meet-citizen-developer/
22. https://www.ibm.com/developerworks/websphere/library/techarticles/1503_
clark/1305_clark.html