To take maximum advantage of Big Data and changing environments, enterprises are re-thinking their IT architecture for collecting and analyzing data. Traditional data warehouse and ETL approaches require high up-front labor intensive data set up effort. These approaches are not well suited for unstructured data and fast evolving environments where searching, analyzing and integrating the data are iterative processes and are best done by the user. Having IT involved in this iterative process using traditional data warehouse and ETL approaches creates long lead times and user frustration. As unstructured content grows, there is a growing need for a new approach.

Cheap storage and fast processing are allowing IT departments to deploy a Data Lake IT strategy. In this approach, IT first makes sure that all relevant data is captured and stored as data becomes available for further mining and integration. On the front-end, relevant data from varied sources are auto tagged, profiled and stored along with metadata in a Big Data archival system. Semantic middleware is used for managing the metadata, tags and profiles and for easy access and data migration on the backend. More extensive content mining and text processing are done on the back-end, as business needs are identified or when data is needed. Using the Semantic middleware, users are provided increased self-serve analytics capabilities and user-driven automated & semi-automated approaches for extracting, curating and integrating relevant extracted data with data from internal systems

In this session we will discuss:

• Introduction to the key elements of this strategy
• The early learnings of who is considering such an approach and for what types of use cases
• The implication to the existing IT architecture
• Technology and tools being deployed
• The value this approach seeks to provide

And the challenges it creates.