The architecture that involves having multiple data marts sourcing from a central data warehouse is known as the Hub-and-Spoke architecture. In this setup, the central data warehouse serves as the primary repository, while data marts are created for specific business units or departments, allowing for a balance between centralized control and department-specific needs.

The primary purpose of a physical data model is to define database table structures and access methods. It includes details such as table names, column data types, indexes, and keys. It serves as a blueprint for implementing the data design in a database management system.

Metadata in databases contains information about the structure, organization, and description of the data. It includes details about tables, columns, data types, relationships, and constraints, aiding in data management and retrieval.

The primary goal of a Data Warehouse in Business Intelligence is to store historical data from various sources in a central repository. This historical data is then transformed, cleaned, and made available for analysis and reporting, enabling informed decision-making based on historical trends and insights.

In a Star Schema, dimension tables are typically denormalized, meaning that redundant data is stored to improve query performance. Denormalization simplifies data retrieval as it reduces the need for joins between tables, which can be resource-intensive.

Shifting from batch processing to more frequent updates throughout the day can significantly improve data freshness. This means the data in the data warehouse is more up-to-date, which can lead to more accurate and timely decision-making. Real-time or near-real-time data can be crucial in financial institutions for risk management and customer service.

The process of transforming data from source systems into a format suitable for analysis in a data warehouse is known as Data Integration. Data integration involves cleaning, transforming, and loading (ETL) data into the warehouse, ensuring its quality and compatibility.

In a traditional Relational Database Management System (RDBMS), data is stored in tables, which are a structured collection of rows and columns. Each row represents a record, and each column represents a data attribute. Tables are used to organize and store data in a tabular format.

An essential feature for a real-time dashboard that alerts when sensor values go beyond a threshold is an alerting mechanism. This mechanism allows the system to monitor data continuously and trigger alerts when predefined thresholds are exceeded, enabling proactive response to potential issues.

"Data locality" in Distributed Data Warehousing refers to the geographical proximity of data across multiple data centers. This concept is essential for optimizing query performance, as it reduces data transfer latencies and speeds up data access when distributed data is physically closer to where it's needed.