The Executive View FEBRUARY 2002

The Executive View is

a monthly letter for
senior executives
that offers insight on
business strategies
and trends.

The End of the Enterprise Application Era

Rod Johnson, Bob Parker, Jim Shepherd, Lance Travis

The current enterprise application model is becoming obsolete. Each applica-

tion was designed to stand alone, with the goal of being the center of their
users’ universe. Nearly all of the enterprise applications were built from the
ground up with proprietary tools, languages, application servers, and integra-
tion frameworks. These applications were built in a different era and on a
technology stack that fails to address the realities of a multivendor world.

Figure 1: Current enterprise application model designed for a single

vendor world

Global ERP Project

Presentation (SAP GUI)

Business Logic (FI/CO, SD, MM, PP)
Workflow/BPM (SAP Workflow)
Integration (BAPI, iDocs)
Reporting (BW)
Application Server (Basis, ABAP4)
System Management (CMMS)

Network Management

Operating System

Database

Hardware
Source: AMR Research, 2002

FEBRUARY 2002 © AMR RESEARCH, INC.

Two Oliver Street, Fifth Floor, Boston, MA 02109 Tel: +1-617-542-6600 Fax: +1-617-542-5670 www.amrresearch.com
Offices in Atlanta, GA, Irvine, CA, and London, United Kingdom
The Executive View 2 of 5

The current model is broken and a browser UI isn’t going to fix it

As we have moved into the era of the Internet, vendors have raced to eliminate all the busi-
ness logic and code on the client in support of a browser-based user interface. However, just
like the screen scraping technology at the front end of the client/server era, user interface
band-aids do not address the fundamental problems of how enterprise applications are built.
And they do nothing to address the problems users face, like the cost of integration, achiev-
ing integration, managing change, data integrity, and poor reporting.

The reasons for these problems stem from the incompatibility between various enterprise
application products. For example, if a company has Siebel, SAP, and PeopleSoft applica-
tions installed, the user must contend with three presentation layers, three workflow engines,
three flavors of Application Programming Interfaces (APIs), three application servers, and
three partial reporting environments (see Figure 2). This equates to three sets of specialized
skills required to manage the applications, expensive app-to-app interfacing, disjointed
processes, and user complexity. It’s time for a reality check. Whether they want to or not,
companies live in a multivendor world, and CIOs and their suppliers must rethink their
approach to architecture, integration, and application deployment.

Figure 2: Current enterprise application model does not work in a multivendor world

Siebel 2000 SAP R/3 PeopleSoft 7.5

SFA, Call Center, Field Service Order-to-Cash Human Resources
Presentation Presentation Presentation
Business Logic Business Logic Business Logic
Workflow/BPM Workflow/BPM Workflow/BPM
Integration Integration Integration
Reporting Reporting Reporting
Application Server Application Server Application Server
System Management System Management System Management

Network Management Network Management Network Management

Operating System Operating System Operating System

Database Database Database

Hardware Hardware Hardware

Source: AMR Research, 2002

Two Oliver Street, Fifth Floor, Boston, MA 02109 Tel: +1-617-542-6600 Fax: +1-617-542-5670 www.amrresearch.com
Offices in Atlanta, GA, Irvine, CA, and London, United Kingdom
© 2002 AMR Research, Inc. All rights reserved. Contents may not be reproduced in whole or in part without written permission of publisher.
The Executive View 3 of 5

No longer the center of the universe, vendors must find their

core competency
At every significant stage in the history of packaged applications, vendors are forced to
reassess what business they are really in and reconsider their core competencies. IBM, for
example, realized that its real skills were in hardware, infrastructure, and services, not busi-
ness applications, so it exited the market. Now it is time for SAP, Oracle, PeopleSoft, Siebel,
and others to recognize that they should no longer be in the business of building application
servers and integration tools. Their real value is in developing, within industry-standard tech-
nology frameworks, the business logic that lets key business processes run more effectively.

Figure 3: The next wave of componentization of the application stack

Presentation

Presentation
Business Logic
Business Logic

Presentation Workflow/BPM Workflow/BPM

Business Logic Integration

Reporting Integration
Workflow/BPM
Integration Application Server
Reporting
Reporting System Management

Application Server Application Server

System Management
Network Management System Management
Network Management
Operating System
Network Management
Database Operating System

Hardware Operating System

Database

Database
Hardware

Hardware

Host-Based Client-Server Web-Based

Hardware-Centric Application-Centric App-Server-Centric

Source: AMR Research, 2021

Industry-standard technology is emerging which, if widely adopted by the application ven-

dors, will let users address many of the costs of maintaining current environments. The most
significant change is the maturing of commercial application servers, based on Java 2
Enterprise Edition (J2EE) or Microsoft .NET. Application servers and the move to Web serv-
ices should change the way applications are built, bought, and deployed, including the com-
ponentization of the currently monolithic enterprise application stacks. To support the real
environment that exists in most businesses, vendors must design their applications to operate
on data they do not own, through a presentation tier they did not design, and most of all, to
run on an application server they did not build.

Two Oliver Street, Fifth Floor, Boston, MA 02109 Tel: +1-617-542-6600 Fax: +1-617-542-5670 www.amrresearch.com
Offices in Atlanta, GA, Irvine, CA, and London, United Kingdom
© 2002 AMR Research, Inc. All rights reserved. Contents may not be reproduced in whole or in part without written permission of publisher.
The Executive View 4 of 5

To survive the change, users and vendors need a common roadmap

AMR Research created the Enterprise Commerce Management (ECM) model to understand
how key advances in technology will alter application and architecture strategy. It lets AMR
Research analysts and our clients assess their current environments and create a pragmatic IT
strategic plan. Whether you are a CIO considering a new architecture for an IT environment
comprised of hundreds of disparate applications or a software vendor CEO thinking about
converting your products to a new application server, the decision is daunting given the cost
and risk.

For users, our first message is to stop buying the integration vision from leading ERP or
CRM vendors and base your selections on the quality of their business logic. The second
message is to heavily favor key infrastructure technologies that make use of your existing
investments. In the near-term, the four best investments to address existing challenges and
prepare for the future are Private Trading Exchange (PTX), single portal framework, integra-
tion bus framework, and single analytical data model.

For vendors, especially the large ones, the shift to an industry-standard application server will
take time, deliberation, and courage. It’s a massive expense to rewrite all existing business
logic and retrain developers. For a vendor like SAP, it will take at least three years, and the
price tag could exceed $1B. And, if successful, its customers can more easily adopt applica-
tions from other vendors, thus eliminating much of the locked-in effect it currently enjoys.
So the question is why would SAP or any of the others make this investment. Well, it’s based
on one key metric that all successful software companies share: the ability to drive sales from
new customers. AMR Research believes that by 2003, user buying requirements will shift to
applications available on these new platforms and away from products built on proprietary
stacks, thus threatening company viability.

Further vendor proliferation, or the rebirth of the IBM shop

Some CIOs may worry that this change will cause applications to proliferate further, but we
actually expect it to lead to consolidation around key infrastructure investments. Leaders are
likely to come from two camps: existing enterprise application leaders able to make the tech-
nology transition and leading general-purpose infrastructure providers.

Today, enterprise application vendors are the only vendors that are accustomed to taking
responsibility for software maintenance, hotline support, and much of the technology infra-
structure. However, significant investments are required to become part of the solution
rather than part of the problem in order to secure their position. In order to retain this
position, they will have to demonstrate that their architectures and use of industry-standard
technology actually facilitate the inclusion and integration of applications and data from
other vendors. They should also have new opportunities to become enterprise-wide suppli-
ers of portals, analytics, and exchanges.

Two Oliver Street, Fifth Floor, Boston, MA 02109 Tel: +1-617-542-6600 Fax: +1-617-542-5670 www.amrresearch.com
Offices in Atlanta, GA, Irvine, CA, and London, United Kingdom
© 2002 AMR Research, Inc. All rights reserved. Contents may not be reproduced in whole or in part without written permission of publisher.
The Executive View 5 of 5

From the infrastructure camp, users may well choose to be “IBM shops” once again, but
their decision will have nothing to do with hardware. IBM appears to be a clear winner due
to the strength of its products across application servers, integration, system management and
services, its emerging products in portals, analytics, and business process management, and
its partnerships at the business logic layer and exchanges. However, IBM must get organized
and step up to take accountability through a maintenance contract of this stack, including
the business logic, if it is to have any hope of regaining account control. Other big winners
are likely to be Microsoft and Oracle, based on overall completeness of their products, but
each has baggage it needs to shed. Microsoft will dominate the mid-market, but it’s unclear if
it can vie for control at the enterprise level; Oracle must learn to partner with other applica-
tion and technology suppliers. Other key players should be BEA Systems, because of its pres-
ence in the application server market, and Sybase, due to the strength of its products across
the ECM technology stack. Sun and HP have potential, but they must establish their appli-
cation server businesses before they will be considered real solution providers.

Time to start planning, not reacting

Like all major technology transitions, the promise is great, but the realization will take
longer, cost more, and deliver less. Application server-based architectures will ease the burden
and cost of application management and integration, however competition between J2EE
and .NET must end, standards on the app servers must mature, and, like RosettaNet, users
and vendors must participate in determining agreed contextual meaning for key transactions,
if vision will ever approach reality.

Two Oliver Street, Fifth Floor, Boston, MA 02109 Tel: +1-617-542-6600 Fax: +1-617-542-5670 www.amrresearch.com
Offices in Atlanta, GA, Irvine, CA, and London, United Kingdom
© 2002 AMR Research, Inc. All rights reserved. Contents may not be reproduced in whole or in part without written permission of publisher.