157
CHAPTER 6 | Databases
C H A P T E R
6
Databases
A persistent data store is at the heart of many applications. As you migrate
existing applications to the Azure cloud or
create new applications, you will
likely find yourself needing to interact with a database. The Azure platform
provides several options from which to choose. You can choose from
relational database offerings such as Azure SQL Database,
SQL Server
running in Azure Virtual Machines, or non-Microsoft databases such as
Oracle or MySQL. If a non-relational (or NoSQL) database fits your
application needs better, services such as DocumentDB and
Azure Table
Storage might be a good fit. Furthermore, with Azure Virtual Machines you
can install a wide range of database platforms (see Chapter 3, “Azure
Virtual Machines,” for more information).
When the time comes to determine a data storage approach for your
application, the Azure platform offers a variety of database choices,
enabling you to balance reduced friction and management with fully
customizable virtual machines.
Azure SQL Database
Azure SQL Database provides a relational database as a service, targeted at online transaction
processing (OLTP; that is, data entry and retrieval transactions) workloads. This falls firmly in the
platform as a service (PaaS) category of cloud computing. Using SQL Database enables
you to give up
the physical management responsibilities of a database server but retain the vast majority of logical
management and administrative responsibilities. SQL Database provides many attractive features,
such as elastic scale, predictable performance,
business continuity, near-zero maintenance, and the
use of familiar development languages and tools.
158
CHAPTER 6 | Databases
It’s important to understand that with SQL Database you do not get a physical server that you can
manage. Because SQL Database is a database as a service, the underlying physical implementation
details are outside your control. There are still logical services
when working with SQL Database, but
these are not the same as the Microsoft SQL Servers you may be used to working with on-premises.
More will be included later in this chapter about SQL Database servers.
SQL Database is available in two different, yet similar, models: elastic database pools and single
databases. Elastic database pools enable you to manage multiple databases in a pool,
scaling
performance up and down as demand changes while maintaining a predictable budget. One of the
key features of the elastic pool model is the ability to share performance across many databases in the
pool. Alternatively, if you have
only a handful of databases, the single database model might be more
appropriate. Both models allow you to adjust performance as necessary with no downtime and
provide a 99.99 percent service level agreement (SLA).
Both the elastic database model and single database model are available in three service tiers: Basic,
Standard, and Premium. Within these tiers, performance is expressed in database throughput units
(DTUs). A DTU is a synthetic measure that allows a quick comparison of the relative performance of
the various database tiers. Within each tier, there are also performance levels (for Standard, they are
S0, S1, S2, and S3). These performance levels provide a way to increase or decrease the DTUs available
within the tier. The maximum database size will vary
across tiers, ranging from 2 GB to 1 TB. Table 6-1
lists some of the pertinent details for each of the database tiers.
Table 6-1 SQL Database tiers and performance levels
Do'stlaringiz bilan baham: