Netherlands: Software

Introductie van Micorosoft SQL Server 2016

Issue link:

Contents of this Issue


Page 188 of 212

176 C H A P T E R 9 | Introducing Azure SQL Data Warehouse change. Whereas SQL Database is designed for OLTP systems and small organizational data marts, SQL Data Warehouse is built and optimized for systems with at least a terabyte of data, with the ability to scale out far beyond that. It uses the same performance optimizations built into APS that allow queries to scale out and take advantage of the parallel computation of complex, analytic queries. There are four key components in the SQL Data Warehouse architecture, as shown in Figure 9-1: Control node You connect to SQL Data Warehouse through the control node, whether you are using business-intelligence, development, or data-loading tools. The control node is a SQL Database that coordinates the data movement and computational workloads in the system. The control node receives the queries and breaks them into units of work to be split across the compute nodes. Compute nodes The compute nodes are separate SQL Databases that provide the computing power for SQL Data Warehouse. The control node evenly distributes data across compute nodes when loading new data into the warehouse. For other operations, the control node assigns operations to each individual database based on the data it contains. Each compute node returns its partial results back to the control node, which assembles the final results. Storage All SQL Data Warehouse data is stored in Azure Blob Storage. This object-based storage service offers nearly limitless and transparent scale that allows the warehouse to scale storage operations separately from compute operations. An additional benefit is the ability to pause your data warehouse and persist the data in blob storage. That means you pay for computing resources only when they are in use. Furthermore, because blob storage is highly fault tolerant, it provides additional layers of protection for your data. Data movement services Data movement services (DMS) transparently move data between nodes as required to respond to user queries. Although this is a background service, you might notice execution plans containing DMS operations.

Articles in this issue

Archives of this issue

view archives of Netherlands: Software - Introductie van Micorosoft SQL Server 2016