Intelligent Enterprise

Better Insight for Business Decisions

RSS
Webcasts
Whitepapers
Subscribe
Home

March 1, 2003

Collaborative Commerce: Not Dead Yet

E-business collaboration is still an imperative — and thanks to new standards, it's also easier and more affordable

by Michael Brown and Derek Sappenfield

Continued from Page 1

Application networks are the result of the convergence of two previous separate areas of IT: application integration and network infrastructure. The model comprises six abstraction layers (see Figure 2); each layer provides common infrastructure services or functions (or both) to the wider community of IT and business systems:

Presentation. The presentation layer manages human-machine and human-human interactions, selecting the appropriate channels and presenting the information in a manner appropriate to that channel. Portal technology, for example, provides a personalized, potentially role-based, interface on multiple systems. Technologies and standards, such as portals and business process modeling notation, give employees, customers, suppliers, and partners a single user interface for accessing and integrating processes, information, applications, and services. These user interfaces rely on standard protocols to provide a screen display that extracts relevant user information from different systems and presents it all as a single, integrated view to the user.

Identification. Concerns about security and trust have made many enterprises migrate toward EDI and other proprietary and exclusive standards. However, standards-based technology makes it possible to extract identity management from the disparate systems that currently reside in the enterprise and manage it from a single, consolidated perspective. Single sign-on and authorization enable a user's identity and profile to be created and managed in one system that interacts with all other business systems. For example, imagine the simplicity and related savings associated with a single password for all systems; these savings are augmented when dealing with external stakeholders. (Example vendors that provide related products and services include Cisco, Check Point, Microsoft, Netegrity Inc., and Sun Microsystems.)

Identity management provides a set of processes and underlying infrastructure to support the creation and maintenance of digital identity including user attributes, entitlements, and credentials. These services allow centralized, delegated, and self-service administration, combined with workflow, to ensure that identity management information is accurate and accessible from authoritative sources. The system can also allow systems to obtain just-in-time identity assertions from third parties.

Data management. Real-time information must be available to stakeholders and systems when they need it so that effective decisions can be made. Most information loses its strategic and operational value over time; in order to convert information into business intelligence, it's vital to process, analyze, and report on corporate data promptly. Unfortunately, most organizations have disparate data repositories, multiple data formats, and multiple data management and analysis tools. This state of affairs makes consolidated views and intelligence an increasingly difficult and costly goal.

To address that challenge, companies must create a common data model that most or all systems can be mapped to — not necessarily a single schema or model to which all systems migrate. (Example vendors in this area include Oracle, Microsoft, and Documentum Inc.) The handling and management of unstructured data — including workflows for their creation and approval and their integration with structured data — has also become critical.

Integration. System integration is fundamental for linking multiple business systems. An information bus or middleware architecture is capable of resolving many internal corporate integration challenges; however, they can't facilitate the integration across enterprises because of the lack of industry standards stipulating how applications communicate — including messaging formats, session management, and data exchange. (Vendors involved here include IBM, Tibco Software Inc., and WebMethods Inc.)

Previous attempts to integrate applications have either been too proprietary or too complex for widespread adoption and true interoperability. But eventually, services-based architectures developed around Web services will make cross-enterprise integration a reality. The rapidly maturing suite of XML-based standards that form the Web services stack have the potential to become as pervasive as the Internet itself, demolishing the barriers to entry for all parties.

Data structures also need to evolve along with business needs. Data models in traditional applications are relatively static, changing only when a new release is available. Applications also overwrite data, destroying the ability to evaluate the past and forecast the future. Rather, standard protocols will define how these services (or applications) can be registered and located, and how sessions will be invoked and which standards have to be adhered to. The result will be a truly distributed processing environment that overcomes the challenge of interenterprise network integration. This architecture will serve as the roadmap for how stakeholders and systems conduct electronic interaction in the future.

Connectivity. Although many enterprises believe they have their physical networks under control, a significant need continues to optimize the flow of network traffic. Thus, the logical architecture must consider differentiation among traffic types, application or service types — and, most important for collaboration, quality-of-service management.

Many networks still don't adequately support business processes. Implementation of dynamic bandwidth management mechanisms and quality of service controls will allow IT departments to do more with potentially less bandwidth, as the networks are aligned to business priorities and integrated with application functionality.

Platform. The platform layer includes computer and storage infrastructure as well as application platforms. It's usually filled with multiple vendor products and older technologies, which tend to be underutilized. However, recent advances in server technology have enabled multiple applications to run on a single processing platform. (Example vendors that provide products and services include EMC, Hewlett-Packard, and Sun Microsystems.) This development in turn has affected storage as well as software and application licensing. Standard technology and protocols have allowed IT to redesign these functions, generating substantial efficiencies and savings in cases where each application has its own server, storage, and application platform.

Right Model, Right Time

The Application Network provides a reference model that allows your IT architecture to finally deliver on the promise of collaborative e-business by facilitating system-to-system, machine-to-machine, application-to-application communication via standards and protocols. As technology building blocks and processes cease to be viewed in isolation, business systems and solutions will ultimately utilize the Application Network architecture as a critical path to business success.

The Application Network, however, involves more than just the evolution of a services-oriented architecture. In addition to making common infrastructure functions operate as network-accessible services, the model is one in which the layers are mutually aware. In other words, instead of optimizing each layer in a vacuum, resulting in an overall suboptimal system, all elements of the IT infrastructure are designed to work together. What's the sense of a highly prioritized application process on a best-effort network, or a Quality of Service-aware network, when your applications can't take advantage of it?

Rate This Article

Comments:

Optional e-mail address:

Technology is only part of the answer, however. Successful e-business collaboration initiatives have an end game defined before you ever get started. For example, when undertaking an SRM initiative, the roadmap is better understood when the project team clearly understands the interdependencies among strategic sourcing, e-sourcing, and e-procurement.

The same perspective applies to the enabling technologies. Using the Application Network Reference Model ensures that your initiatives address all dimensions of the solution, including IT architecture. Beyond defining the one right architecture for collaborative e-business, this architecture must be flexible enough to accommodate "industrial strength" IT architecture for large-scale trading partners as well as "lean" IT architecture for smaller trading partners.

Michael Brown [michael.brown@us.didata.com] is the managing director of the global supply chain practice for Dimension Data. He has spent his career developing operational strategies, reengineering and implementing supply chain solutions for clients in multiple industries.

Derek Sappenfield [derek.sappenfield@us.didata.com] is the senior vice president of global solutions for Dimension Data. He develops and implements business and technology solutions for clients across many industries.