The Need for a Data Delivered as a Service to SOA
Applications and business processes interconnected as services for flexibility aren’t worth anything if the data in them is inaccurate, inconsistent, or not available exactly when needed. If the data stuck inside siloed applications is bad, imagine the calamity when, through an application and business process integration technology, the silos disappear and data from many different applications is commingled.

Without a complementary, standards-based, and flexible technology that delivers holistic, accurate, and timely data as a service to an SOA, typical end-user challenges may include:

• Implementers that embark on an SOA path, quickly realize that SOA presents the following data-centric challenges that can’t be handled by their SOA platform providers:
• Varying data volumes and data movement latencies
• Semantic and syntactic inconsistency of data
• Poor data quality that is often difficult to measure or monitor
• Lack of agreement or visibility (single-view) into critical information assets
• Heterogeneous data silos distributed across the enterprise and beyond
• Little to no reuse of data integration logic and skills across projects

Therefore, far from making data issues irrelevant, the underlying IT infrastructure poses an actual risk to enterprises that lack a right-time data integration strategy. SOA provides a flexible framework by breaking down barriers between silos of applications and enabling the orchestration of business processes using flexible and reusable application-level or business services. If SOA’s inherent capabilities for flexibility, responsiveness, and reuse are enhanced with the holistic, accurate, and timely information, business agility can definitely be achieved.

What’s needed is a scalable data services technology built on a sophisticated data integration platform that can provide a standards-based data abstraction layer to an SOA, and deliver holistic, accurate, and timely data as a service (see Figure 2). A data service is a modular, reusable, well-defined, business-relevant service that enables the access, integration, and right-time delivery of enterprise data throughout the enterprise and across corporate firewalls. However, unlike application-level or business services, data services are specialized services that are more granular, enable loose-coupling with data sources, are data-centric, and are purpose-built to enable right-time data integration in an SOA.

The data services technology should also be capable of orchestrating or sequencing atomic data services to enable flexible combinations and reuse of sophisticated data integration tasks across projects. As shown in Figure 3, atomic data services must enable sophisticated data integration tasks such as access, profiling, cleansing, transformation, and delivery of data that can be sequenced in various combinations to support the entire data integration lifecycle in an enterprise.

Ideally, a data services technology needs to deliver these sophisticated capabilities to an SOA:

• Broad access to all complex data contained in any data source in the enterprise
• Easy and rapid creation of sophisticated, scalable, and reusable data services
• Orchestration of atomic data services to cover the entire data integration lifecycle
• Provisioning data in right-time to applications and business processes over a variety of protocols
• Upfront or proactive data-quality treatments to all data across all data sources
• Tools for data lineage to support change management and impact analysis
• Seamless integration with existing IT infrastructure, the cornerstone in an SOA

Finally, as performance and scalability are at the heart of any mission-critical operation, the data services technology should also be able to handle high concurrency and large data volumes that are typical requirements of such demanding environments.

SOA & Integration Competency Centers
An Integration Competency Center or ICC is an infrastructure of people, technology, policies, best practices, and processes focused on rapid, repeatable, and cost-effective deployment of data integration projects critical to meeting organizational objectives. Organizations have found that there’s a direct connection between the caliber of their ICC and their ability to respond quickly to dynamically changing business models, intense competition, and demanding customers. In short, the ICC is the infrastructure responsible to deliver trustworthy data flexibly and at the speed of the business.

Data services technology encourages the creation and management of ICCs by standardizing development processes on common technology standards for greater reuse, supporting a shared services environment, supporting a centralized services environment, and enabling the governance processes of data services and data architecture so that definitions, semantics, and SLAs are maintained by the appropriate parties.

An ICC enabled by a data services technology is thus the perfect enabler for data integration in an SOA since it provides the adaptive architecture and technology foundation required to access and deliver holistic, accurate, and timely data throughout the enterprise (see Figure 4). It also delivers the collaboration tools necessary to align cross-functional teams through a single technology, increasing responsiveness to the changing needs of the business. Finally, it provides a secure, scalable, and responsive platform to guarantee access to mission-critical information to support various projects across the enterprise.

Conclusion
As we’ve seen in this article, a sophisticated and enterprise-grade data services technology can address SOA’s last mile by providing the cornerstone in an SOA for ensuring the availability of holistic and accurate business-critical information at the speed of business.

Resources

• Ivan Chong and Ashutosh Kulkarni. “Enterprise Data Integration, A Critical Piece of a Service-Oriented Architecture.” SOAWorld Magazine. February 24, 2006.
• Service-Oriented Architecture. Informatica CIO2CIO Perspectives.
  
• Informatica’s Service-Oriented Architecture. Informatica.