Loading a PDW Region in APS - SQL Server Integration Services Design Patterns - page 230

especially ones that need to scale to hundreds of terabytes—or even petabytes!—of
storage. There are other MPP appliance vendors available besides Microsoft; however,
the ability to easily join unstructured data in HDInsight with structured data in PDW
via PolyBase, close integration with the Microsoft business intelligence stack (SQL
Server, Integration Services, Analysis Services, and Reporting Services), and a compel-
ling cost-per-terabyte makes APS a natural progression for organizations that need to
take their SQL Server data warehouse to the next level. In this chapter, we will walk
through how to efficiently load data into a PDW region in an APS appliance using In-
tegration Services. But before we do that, we will first explore the architecture of APS.
APS could easily consume a book in its own right, so we will only be covering the
most pertinent parts to ensure that you have the foundation necessary for efficiently
loading data into PDW.
Tip Learn more about Microsoft's Analytics Platform System (APS), Parallel Data
Warehouse (PDW), and HDInsight at http://microsoft.com/aps .
APS Appliance Overview
Generally, ETL developers and data integration engineers need not worry about the
hardware specifications of the database systems they interact with. This isn't the case
when you are loading MPP systems. Because MPP systems are designed to take ad-
vantage of distributed data and parallelized workloads, you will find tremendous per-
formance benefits in designing ETL solutions that take advantage of the parallelization
units of an MPP system. For that reason, we will briefly review general hardware and
software architectures for Microsoft's APS appliance before we dive into load patterns.
Hardware Architecture
Microsoft has partnered and closely collaborated with Dell, HP, and Quanta to provide
customers with a choice in hardware vendors. The hardware specifications of an APS
appliance will vary depending on your chosen hardware vendor and capacity require-
ments, although there are some consistencies across the vendors.