On November 10, Informatica made history with the launch of Informatica 9. In my mind, being a SOA enthusiast, another equally significant event transpired – the birth of SOA-based Data Services – transformational SOA data integration that can revive your enterprise architecture.

So, what exactly are SOA-based Data Services and why am I so excited?

Before I answer this question, let’s look at the data-centric challenges that current architectures are facing:

• No way to quickly find the data you need across heterogeneous data sources, including new and time-sensitive data.
• Challenging to easily understand the structure of all the data, regardless of the type of data or the location where it resides.
• Cannot proactively identify issues concerning inconsistency and inaccuracy across all data sources, in real-time.
• Need to effectively define, enforce and centrally manage policies around data freshness, real-time data quality and data privacy.
• Lack of a data abstraction layer that can insulate all consuming applications from changes in the underlying data.
• Data is typically not represented as the business thinks of it, as business entities, making collaboration extremely difficult.
• Cannot quickly make smart decisions about whether data should be physically moved, or federated on the fly, without rework.
• No support for changing data provisioning mechanisms to various applications, such as Web services or SQL, without rebuilding the logic.
• No support for the principles of service-orientation at the data layer – reuse, modularity, standards-based, loose-coupling.

Architectures defined using current technologies are typically riddled with the use of sub-optimal approaches and diverse tools to address the various aspects of these data-centric challenges – there is little or no standardization of technology, let alone reuse of integration logic or skills. If a technology is good at delivering virtual views across heterogeneous data sources, it will not be able to support the need for sophisticated data transformation including data quality. Further, these standalone tools cannot physically materialize the data within the same design and run-time environment when needed. Another separate tool is required to rebuild the same logic as a batch ETL process thereby requiring a different set of skills and additional maintenance. Additionally, when required to deliver a full range of data services in support of service-oriented architectures, these standalone technologies cannot do much more than enable simple data access services. Finally, there is no support for involvement of the business user in the data integration process, leading to miscommunication and rework.

As you can imagine, managing and maintaining data assets across the enterprise becomes a nightmare. What is needed is a flexible architecture and disciplined architectural approach that can deal with enterprise silos, complement existing infrastructure, leverage principles of service-orientation and deliver timely, trusted and relevant data as a service, to all applications. What is needed are SOA-based Data Services.

SOA-Based Data Services provide the means to read and write enterprise data, based on real-world representations (e.g. CUSTOMER, PRODUCT, CLAIM, etc.), and deliver it to applications exactly as needed, in a timely, trusted and manageable way. With SOA-based Data Services, you can standardize on a single platform and single set of skills for all styles of data integration – physical or virtual. Sophisticated data services can be designed once and deployed many times in multiple ways for applications and projects by reusing data integration and data quality logic.

Next up…we will look at the transformational capabilities of SOA-based Data Services in detail.

Let the revival begin…