The Case for Coordinated EDM and SOA Strategies

Strategies for enterprise data management (EDM) and service-oriented architecture (SOA) are often pursued as separate, disparate programs and initiatives within organizations, both as a business requirement as well as an IT implementation perspective. However, there are important overlapping and interdependent components, processes, and quality checkpoints for each in which coordination is necessary to ensure the success of either strategy.

Furthermore, by coordinating these strategies, organizations will realize opportunities to optimize:

There are several reasons driving the coordination of EDM and SOA strategies, including asset value, organizational efficiency, profitability, and cost optimization.

The CIO Perspective on Data and SOA Dependencies
The need for coordinated data and services, under the guidance of coordinated EDM-SOA programs, has recently been raised to the executive office. As shown in Figure 1, CIOs are focusing on data-centric initiatives such as Customer Data Integration (CDI) and Master Data Management (MDM) as drivers for service-oriented architecture projects. Even traditional non-SOA business analytics or knowledge management programs are now key drivers for SOA, according to surveyed CIOs.

This survey also demonstrates the realization by most CIOs that SOA creates interdependencies between systems requiring high-quality data, which further suggests that the full benefit of a SOA program cannot be gained without coordinated EDM strategy components.

The "Perfect Storm" for EDM and SOA
Interrelated factors across industries are creating an environment ripe for organizations to develop EDM and SOA strategies/programs in parallel timeframes and a coordinated fashion. Figure 2 lists major contributing industry trends resulting in corporate initiatives driving requirements for coordinated EDM and SOA capabilities, including joint governance programs.

EDM Framework and Component Considerations
Let's take a look at the EDM framework in Figure 3 to see which components have potential dependencies and synergies with SOA strategies/components. This will help identify where we should focus strategic efforts for coordinated EDM-SOA strategies.

Enterprise Data Management initiatives generally consist of activities addressing several of these components, simultaneously or at least in coordination. When considering which EDM components have a significant impact on SOA strategies, and which are significantly impacted by SOA, we see that some have direct major considerations (primary coordination points) in joint EDM-SOA strategies, while others have a smaller impact (secondary).

Organizations should start with identified primary (coordination) EDM components and SOA strategy impacts, dependencies, and synergies for emphasis, then determine which secondary components are pertinent for coordination within that particular EDM-SOA environments and initiatives.

Organizations that fail to coordinate EDM and SOA strategies appropriately will inherently cause enterprise data and services to evolve disparately rather than synergistically as part of a well-managed enterprise architecture, and should ask:

These are important issues from the EDM perspective that may go unmanaged when organizational data and service strategies, governance, architecture, and development go uncoordinated.

The Enterprise Information Architecture (EIA) component, a primary coordination point, is relatively recent to EDM organizations. It:

Figure 4 shows how EIA works with organizational business processes and the SOA, under the guidance of data/SOA governance. It is directly leveraged by the SOA and includes the enterprise data model that service designs will leverage. Hence, the EIA works with and contains the enterprise-level aspects of MDM and metadata management. It coordinates pertinent data integration and quality issues, best practices, and tools between EDM and SOA strategies as a primary EDM component for coordination with SOA strategies.

SOA Framework and Component Considerations
Looking similarly at a SOA framework, it becomes clearer where dependencies and synergies between EDM and SOA strategies exist.

SOA initiatives generally address several of the components shown in Figure 5, simultaneously or at least in coordination. Considering how EDM components are impacted in a SOA environment, most SOA components play some role in coordinated EDM-SOA strategies/programs.

Key cross-impacts of EDM and SOA components in a SOA environment are:

SOA Best Practice Considerations
Another way to consider SOA strategy impacts on and synergies with EDM is in addressing how organizations achieve SOA best practices. Significant dependencies between an organization's EDM and SOA strategies will surely reveal themselves and must be taken into account when pursuing SOA best practices.

The following are among the key best practices for SOA strategies, and most/all can also be applied as EDM best practices.

1. When thinking about services, don't forget to consider the data
Systematically designing a service model is like designing a data model. For either, its impact should be considered long term, and the level of normalization of designed components, services, or data is considered a sign of quality and maturity.

Figure 6 shows service-data normalization from immature to mature organizations:

Most organizations pursuing services-data normalization have progressed to ownership/stewardship levels, yet need to reach encapsulation before realizing major benefits in efficiencies, maintenance costs, and asset business value.

The highest level of service-data normalization, object, may not make sense for some organizations, especially where master data or business services change frequently. Depending on their stability, the more possible an object level may be. However, cost/benefit analysis may make encapsulation preferred for some organizations.

Transitioning to advanced service-data normalization is a process of increasing organizational maturity toward coordinated EDM-SOA strategies.

This is facilitated through coordinated:

2. Focus on avoiding the proliferation of services that can't be shared
SOA strategies have little business value if their enterprise services aren't shared (i.e., reused) among multiple user groups and business domains in the enterprise and sometimes outside the immediate enterprise.

Not coordinating data with SOA initiatives (e.g., via governance), services may inadvertently propagate non-"gold standard" data to service consumers when developed by the initiatives. Such services become, or should be, unable to be shared.

Worse, the absence of coordinated data and services may tempt developers to create their own data stores/marts to support their services' domain, causing unnecessary propagation of potentially unmanaged databases. This will damage both your EDM and SOA strategies.

Avoiding services that can't be shared is facilitated by coordinated:

3. Reward both reusability and reuse
Reusability and reuse applies to services and data in development, deployment, and consumption cycles. Services and data should be normalized for reuse (see SOA Best Practice #2), and developers and consumers of either should be rewarded.

This is a delicate balance that should be managed by data-SOA governance and processes to ensure appropriate reuse when it makes business sense (i.e., unless service/data requirements are new or existing designs/implementations are obsolete).

SOA governance will sometimes advocate development of new or improved services when it makes sense. Similarly, data governance will almost always advocate reuse of existing master data or managed metadata, but decreasingly over time as managed data stabilizes, there may be reasons to extend or change "gold standard" data.

Reuse and reusability should also be enforced, and best practices established, by coordinated data and SOA governance programs. Governance should include the identification of which data and services can potentially be reused for a given initiative, and the criteria for when new data/services should be created or modified.

Rewarding reusability/reuse is facilitated by coordinated:

4. When establishing governance, stay away from dictatorships
As shown in Figure 7, there are three general governance approaches for organizations:

The goal is not to progress to dictatorial governance. While "Wild West" is a problem with lack of control/coordination, dictatorships swing the pendulum too far toward centralized control of all decisions regarding enterprise-wide and domain-specific assets (data or services).

Establishing balanced governance is facilitated by coordinated:

5. Establish a Center of Excellence (COE) to provide guidance, governance, and coordination
A well-organized and managed EDM or SOA environment should establish a COE to:

Mature EDM/SOA strategies include development and management of a supporting COE. However, there is the danger of organizations developing disparate EDM/SOA COEs when pursuing both strategies, despite overlapping roles, dependencies, and synergies.

As Figure 8 demonstrates, an Integration Competency Center (ICC) does more than a traditional COE by providing a design, development, prototyping, and testing sandbox for data and services integration.

An ICC provides processes/tools for EDM and SOA environments, and facilitates coordinated services and information architecture and development to achieve, among other things, service-data normalization.

An ICC coordinates with data-SOA governance and the EIA, as well as overall enterprise architecture concerns. It identifies key components managed within governance for which instances will reside in the ICC and/or production implementations.

Establishing a services-data COE/ICC for stakeholders and interdependent projects is facilitated by coordinated:

6. Start with the right program size in the right area of emphasis
Starting too big with EDM or SOA can lead to mistakes. Think strategically, but act tactically with an emphasis on launching a realistic program that can become successful initially and grow. Develop a coordinated long-term vision for your EDM/SOA programs, and implement these incrementally in support of each other. Coordinated data-SOA governance programs, and a supporting COE/ICC that addresses data and services (see Best Practice # 5), can support the scoping and launch of a "right-sized" program.

For example, you can design sets of services around selected business and data models. The data model can be used to encapsulate business data, and the business model can link further business processes of applications with its software implementation (i.e., services). Business services typically consume data services that are usually not exposed outside the enterprise.

Alternatively, the data model may become primary design criteria for an application; the data model would be the best choice on which to base application services, user interfaces, and business process designs.

Implementing an appropriately sized and focused EDM-SOA program is facilitated by coordinated:

7. Invest in systematically designed core services initially, allowing for rapid opportunistic extensions later
An SOA program may be launched tactically with flexibility and scalability to systematically incorporate program scope under the guidance of governance and an ICC.

Key intermediary services that intercept and handle operations common across services that should be reused include: authentication, logging, monitoring, and routing.

A common data services layer:

SOA strategies cause organizations to implement an EIA and infrastructure in support of services, including common data services capable of supporting producers and consumers with timely and consistent information.

The main categories of data services include:

Within other services, and across the data services layer, common infrastructure methods for search, access, creation, update, and deletion functionality are made available.

Systematically/progressively designing services is facilitated by coordinated:

Conclusion
This article demonstrates the case for coordinated EDM and SOA strategies and capabilities in organizations. It further shows what strategic EDM and SOA components require attention to facilitate appropriate coordination.

This article shows that organizations should develop a coordinated data-SOA governance program, as coordination needs to be established at leadership levels to enable the optimal value of their services and data. This is the highest priority toward coordinated EDM-SOA capabilities.

Initial data/SOA governance models should be coordinated for processes, checkpoints, and ownership. We rarely develop these strategies and governance models in coordination from scratch though hopefully this article will spark this. Hence, it's important to adapt appropriate processes and checkpoints between separate models and (re)define roles to support coordinated data-SOA governance.

Organizations should scale progressive EDM-SOA initiatives with shared data-SOA governance, processes, and communications complemented by educating EDM/SOA resources/stakeholder to effectively leverage each other in joint data-services development.

Ensure business- and data-modeling analysts are involved in services design (e.g., not just services analysts), so business functionalities are reflected rather than technical partitioning of software/data. An appropriate COE/ICC will facilitate diverse stakeholders working proactively to drive phases of coordinated EDM-SOA initiatives. Last, promote a culture of collaboration as an underpinning of successful EDM, SOA, and coordinated programs.

Part 2 of this series on EDM and SOA will introduce the coordinated service-oriented data architecture framework and the capability maturity model (CMM), including how to use the C-SODA framework and CMM to drive assessments and improvement roadmaps for organizational maturity.

© 2008 SYS-CON Media