In most companies today, there is a real need to run multiple different analytical workloads that go well beyond the capabilities of the traditional data warehouse. This includes analysis of streaming data (e.g. IoT sensor data), analysis of unstructured text (e.g. twitter messages), graph analysis on social network data, deep learning on image data and analysis of huge volumes of machine data such as on-line clickstream from weblogs. These requirements have resulted in new types of data store appearing in the enterprise, including NoSQL Column Family databases, graph databases, Hadoop systems, and streaming analytics platforms. Also, as business demands more and more structured, semi-structured and unstructured information, the number of internal and external data sources is exploding. It’s taking us way beyond the OLTP and master data management systems that feed typical data warehouses. The problem is that the classic architecture of an enterprise data warehouse and dependent data marts can no longer cope with these new requirements. Also, self-service BI users are now challenged because the data they need is now sitting in multiple analytical data stores, making it hard to access. So, what can we do about it? We need a new architecture that separates analytical data stores from the tools and applications that need to access them, that allows us to add new data stores to accommodate these new workloads but that hides the complexity of all this from business users. The answer is a Logical Data Warehouse architecture. This session discusses these challenges and defines the architecture needed to meet these needs. It then discusses how you can implement a Logical Data Warehouse and to transition from a traditional data warehouse to a new, more flexible logical one.
• New data, new analytical workloads and new analytical data stores
• Why is the traditional data warehouse architecture no longer enough?
• What is a Logical Data Warehouse?
• How to implement one using data virtualisation
• Transitioning from a traditional data warehouse to a logical data warehouse
• Simplifying access for self-service and operational BI