Interoperable: First, this capability enables users to connect to and ingest data from various types of storage platforms, both on-premises and cloud, remote and local. Second, this is able to operate with MDM (master data management) tools, ETL (extract, transform, and loading), data modeling, data lineage, DevOps, governance, and metadata tools. For example, IBM data stage and Informatica have parallel connectors to modern data warehouses, enabling high-speed data transfer with custom data transformation algorithms executed on data nodes. It is integrable with analytical, ML, an statistical analysis packages.
Interconnectivity: An organization that does not have one but rather a collection of analytical servers, data warehouses, data lakes, data marts, and so on needs to have these be integrated together logically, with metadata and data virtualization (federation), so that it appears as a single logical system for integration by providing interfaces between multiple servers with multiple data integration methods. First, we combine the logic data warehouse with an abstraction layer on top and then automate manual tasks in the LDW and other systems. This provides features that allow the system to provide prescriptive
advice to the users about what data elements to use, which datasets/tables to join, and, additionally, provides a significant degree of ability to self-administer the platform.
Resilient: If a modern data warehouse architecture is resilient it means it has the capability of being highly available, able to recover from disaster scenarios quickly, and able to provide a backup and restore DB from the point of failure. This is very important for highly critical data applications ranging from payment applications in the banking security industry or IOT-based application in automotive industries. All cloud vendors use hundreds of server farms and are designed with built-in redundancy and fail-over backup recovery with high service level agreements by setting up mirror images in geographically distributed data centers across regions, countries, and other parts of the world.
Elastic: With the varying processing demands across time and a range of requirements, modern data warehouses require scalable architecture that adapts to such variations on demand. A positive trend is that all companies providing cloud services platforms (public, private) provide on-demand and auto and on-demand elastic scalability at low prices. Elastic architectures help administrators by their not having to spend time to calibrate capacity, increase or throttle usage if necessary, avoid risk of overspending on hardware, and decrease time to market. Elasticity has many factors to consider that vary, from type and criticality of application to environment of deployment, such as development, QA, test environment setup, and setting up analytic sandboxes.
