Azure SQL

Database & SQL

DW
Bruno Basto
DPSA
bbasto@microsoft.com
Data Platform Airlift
21 de Outubro \\ Microsoft Lisbon Experience
SQL Server de 1.0 a 2016
Magic Quadrant for Operational Database
Management Systems

http://www.gartner.c
om/technology/repri
nts.do?id=1-
2PMFPEN&ct=15101
3&st=sb
How we develop SQL

Cloud-first but not Cloud Only

Using Azure SQL DB to improve core SQL Server features & cadence
Many interesting and compelling on-premises  Cloud scenarios
How we develop SQL
• Predictable performance
• Scalability with no
downtime
• Business continuity
and data protection

• all with near-zero

administration


SQL DB
 V12








Better database manageability Programmability and Insight
• Table partitioning • Analytic Window Functions
• Compression (page, row) • CLR Support (SAFE, bytestream only)
• Large index rebuild • XML index and schema
• Online index build/rebuild with LOB • Heap Support (and SELECT INTO)
column • Application role
• 2GB transaction size limit is removed • Change Tracking
• CHECKPOINT support • Sequences
• Extended Alter Database support • UTF-16 support
• Non-admin DBCC commands, e.g. • Column-store indexes
CHECKDB • Extended DMV set


SQL DB
 V12



 DataMasking
 TDE
 Azure AD
Max per-database burst level
Elastic Database Tools
Single tenant per database
 Each tenant’s data is stored in a different database DB1 DB2
 Better isolation of tenants as compared to multi-tenant model
Multiple tenants per database
Customer Customer
1 2

 Multiple tenants share the same database

 Less isolation of tenants as compared to single tenant model
 Typically more cost-effective than the single tenant model
Hybrid model
 Some tenants share databases, others get their own database
 E.g., premium or paying customers get their own databases, while free tier
customers share databases
Temporal model DB1 DB2
 Sharding based on date/time Customer 1 Customer 4
 Most recent shard is constantly loaded with newly arriving data Customer 2 Customer 5
 New shards added when current most recent shard nears capacity Customer 3 Customer 6

 See guidance from the Azure CAT team on

sharding:
 MSDN: https://msdn.microsoft.com/en-us/library/azure/dn764977.aspx
Client .NET APIs Management Services
Shard map management Fleet Management
• Define groups of shards for your application • Orchestrate administrative tasks across
many databases
• Manage mapping of routing keys to shards
• Same script executed on all databases
Data dependent routing • Ability to save merged results
• Route incoming requests to the correct
shard, e.g., given a customer ID Split/Merge
• Ensure correct routing as tenants move • Grow or shrink capacity by adding or
removing databases
• Cache routing information for efficiency
• Re-balance data among shards
Multi-shard query • Isolate hotspots
• Interactive processing across several shards
• Same statement executed on all shards with
Shard Elasticity
UNION all semantics • Dynamically adjust scale factor of database
• Trigger adjustment through policies
 Client Cross-shard Cross-shard
library operations Cross-Shard Capabilities operations
Cross-
shard
extensions

Elastic
Application
Elastic Scale Admin/
Developer Scale app Manage- DevOps
ability

Shard-local Shard-local
Client … … …
operations shard1 shardi shardj shardn operations
library

Grow/shrink capacity
Admin/
DevOps

Note that your application needs to be shardable to benefit from Elastic Scale. If the application
does not partition well, Elastic Scale (or any other sharding approach) is likely not going to be a good fit.
Vertical: Scale-up vs. scale-down
 Change service-tiers for a given database as capacity needs fluctuate

Horizontal: Scale-out vs. scale-in

 Add or remove databases as more or less capacity is needed
Scale up/down

Premium Premium
Standard Standard
Standard Standard Standard Standard Standard Standard Standard

Scale out/in
Business
continuity &
data protection
Programmatic “oops recovery” of data deletion
or alteration; up to 35 days in Premium
Point-in-time restore
Programmatic “oops recovery” of data deletion or alteration
Auto backups, transactional logs every 5 min
Backups in Azure Storage and geo-replicated
Creates a side-by-side copy, non-disruptive SQL Database Azure Storage
Backups Geo- replicated
sabcp01bl21
REST API, PowerShell or Azure Portal sabcp01bl21

Backups retention policy:

• Basic, up to 7 days
Restore from backup
• Standard, up to 14 days
• Premium, up to 35 days
Automated export of logical backups for
long-term backup protection
Geo-restore
Emergency data recovery when you need it most

Available in Basic, Standard and Premium SQL Database Azure Storage

Backups Geo- replicated sabcp01bl21
sabcp01bl21
Built on geo-redundant Azure Storage

Recover to any Azure region

Restore to any
Azure region
Standard geo-replication
Opt-in business continuity for greater geo-redundancies

Opt-in for Standard & Premium databases

Geo- replicated

Creates a stand-by secondary

Replicate to pre-paired Azure region

Automatic data replication, asynchronous

Opt-in via REST API, PowerShell or Azure Portal

Microsoft-managed, RTO<24h, RPO<1 hr

Active geo-replication
Mission-critical business continuity on your terms, via programmatic APIs
Self-service activation in Premium

Create up to 4 readable secondaries

Replicate to any Azure region

Automatic data replication, asynchronous

REST API, PowerShell or Azure Portal

RTO<1h, RPO<5m, you choose when to failover

Up to 4
secondaries
Hybrid
Apps

27
Aviva implemented a hybrid-cloud solution based on Windows Azure
that brings together mobile devices, on-prem apps, and social media

Benefits One of the things that we

Up and
running in
10
Minutes
loved about Windows Azure
is that we could develop code
with Visual Studio 2010,
Speeds development and implementation upload it, and have it running
by storing data in SQL Database 10 minutes later.
Enables greater flexibility and innovation Jason Steele
Solutions Designer, Aviva

Offers a highly scalable,

fully-managed infrastructure
29
Shift
existing
apps

31
Samsung Electronics migrated its diverse Smart TV systems
to Windows Azure in two months

Benefits Because the infrastructure is

entirely taken care of, we only

80%
Reduction in hardware
& maintenance costs need to deal with software and
compared to on-prem services. For us, Windows Azure
has a competitive edge,
Supports firmware updates and especially its compatibility with
registrations for 20 million smart TVs existing systems and its
reliability.
Chunun Kang
Achieved greater competitive edge Head of Engineers, Visual Display Business Division
through increased focus on core business
32
Website &
mobile
apps

37
3M Informatics hosted its new Visual Attention Service on
Windows Azure so managers can focus on business, not infrastructure

Benefits Because management time and

50%
cost is so low with Windows

$ Revenue Growth
Azure, we focus on growing the
business, not on managing red
tape. As a result, we estimate
Minimizes management time and cost that we’re bringing in fifty
percent more revenue.
Enables cost-effective scalability William K. Smyth
Global Business Manager, 3M VAS

Instills customer confidence

39
BI and analytics

Data enrichment and federated query

Data management and processing

 Azure
Azure

Saas

Azure

Office 365
Public
Cloud
Elastic Scale
Spin up for heavy workloads, cycle down for daily activity
Buy time to insight based on what you need, when you need it
Choose the combo of compute and storage that meets your needs
Pause
Data remains in place – no reloading / restoring of data
When paused, cloud-scale storage is min cost
Automate via PowerShell/REST API

$$$$
 SQL Server Compatibility
Mature enterprise-ready SQL for sophisticated DW scenarios
Existing SQL Server scripts and tools just work
Continuous enhancements on language surface

Modular programming Easier maintenance of Allows proper use and Mature Column-Store
(write once, execute multiple large tables comparisons of technology for best-
times)
characters in different in-class DW query
Improves performance languages performance
Faster code execution
Enhanced scalability
Encapsulated and availability
programming logic
Engine 1 1 1 1 1 1 1 1 1 1 1 1
Nodes
Worker 1 2 3 4 5 6 10 12 15 20 30 60
Nodes
Total # of 60 60 60 60 60 60 60 60 60 60 60 60
distributions

# of 60 30 20 15 12 10 6 5 4 3 2 1
distributions
per node

Concurrency 4 8 12 16 20 24 32 32 32 32 32 32
Slots
http://aka.ms/sqldw
 Azure Storage Blob(s)

D1 D2 D3 D4 D5 D6 D7 D8 D9 D10

D11 D12 D13 D14 D15 D16 D17 D18 D19 D20
Azure SQL Data
Warehouse D21 D22 D23 D24 D25 D26 D27 D28 D29 D30

D31 D32 D33 D34 D35 D36 D37 D38 D39 D40

D41 D42 D43 D44 D45 D46 D47 D48 D49 D50

D51 D52 D53 D54 D55 D56 D57 D58 D59 D60
Query unstructured data via PolyBase/T-SQL
PolyBase for querying & managing non-relational
Hadoop and relational data

• Enables query capabilities across common Hadoop SQL DW Instance

distributions (HDP & Cloudera) and Hadoop file Hadoop VMs /
formats in Azure storage. PolyBase Azure Storage

• Allows leveraging existing SQL skills and BI tools Scale out compute

• Supports multiple non-relational file formats

• Improved time-to-insights & simplified ETL
 Try SQL Server
2016 CTP2
http://aka.ms/trysql2016

Free Azure
Trial
http://aka.ms/tryazure
Use Power BI for Free
http://powerbi.microsoft.com