gs08 modernize your data platform with sql technologies wash dc
TRANSCRIPT
Modernize your Data Platform with SQL TechnologiesBob WardPrincipal ArchitectData Group, Tiger Team
[email protected]@bobwardms, #bobsql
Deck on slideshare.net
Demos on github
• Disparate systems and processes
• Multiple tools and skillsets
• High cost of ownership
• Siloed insights on disconnected data
• Costly data movement and delay
Challenges of the Modern Data PlatformFragmented architectures are inefficient
Analytics
Visualization
Relational in the cloud
Analytics
Visualization
Hadoop
{ }NoSQL Data
Lake
Analytics
Visualization
Relational DBAppliance
Analytics
Visualization
Hadoop Appliance
Clou
dOn
-pre
mise
s
Relational Non-relational
Azure SQL DatabaseAzure SQL Data WarehouseSQL Server 2016 on VM
Analytics Platform System
Azure Data Lake
SQL Server 2016Analytics Platform System
SQL
Relational Non-relational
On-p
rem
ises
Clou
dMicrosoft Data PlatformTransforming data into intelligent action
• Manage all your data in a mission critical, scalable, secure way
• Gain deep insights across all of your data without data movement
• Utilize existing skills and investments
• Consistent on-premises, cloud and hybrid experiences
Federated queryBusiness intelligenceMachine learning & advanced analytics
Insights
Microsoft SQL Server – On-premises & cloud
SQL Server SQL Server in Azure VM
Azure SQL Database
Dev ops tools Identity VirtualizationManagement tools
Azure SQL Data WarehouseAnalytics Platform System
Common tools for development and managementCommon T-SQL surface area
Simple cloud migration
Consistent experience is everything
Continuous Innovation
End-to-end mobile BI
Advanced Analytics
Enterprise-grade DW
Mission critical OLTP
The bestSQL Server release in history
SQL Server 2016
SpeedAgility
ProvenFeedback
SQL DB SQL DB SQL DB SQL DB
New innovations
Across customer base
Cloud-First Approach + DevOps
Continuous enhancemen
ts
Hybrid cloud
1/12th the cost of Oracle
National Institute of Standards and Technology Comprehensive Vulnerability Database update 12/2016.
Best price/performance
Everything built-in
Highest performing data warehouse
Most secure database
Best TCO Most consistent experience
On-premises to cloud #1 performance
TPC-H
Oracle
is #5#2
SQL Server
#1
SQL Server
#3
SQL Server
The Power of SQL ServerEverything built-in
June 2016SQL Server 2016
TPC-E
6 years in a rowleast vulnerable
2010 2011 2012 2013 2014 2015 20160
10
20
30
40
50
60
70
80
0 1 40 0 3 0
3429
2215
5
2216
6
43
20
69
18
49
74
3
SQL Server
Oracle
MySQL2
SAP HANA
1/12
Only data solution to encrypt your data at rest and in motion
Connect your relational data to big data with PolyBase
Real-time operational analytics without impacting performance
Up to 30x faster transactions, 100x faster queries with InMemory
Unparalleled choice for developer tools and languages
1 T-SQLJavaC/C++C#/VB.NET
PHPNode.jsPythonRuby
For all your applicationsInnovations across all editions
Available nowSQL Server 2016 SP1
On the platform of your choiceSQL Server v.Next
Targeting CY2017SQL Server v.Next GA*
*SQL Server v.Next Public Preview available now on Linux, Windows, and Docker.
Here is How SQL Server 2016 Just Runs FasterCore Engine ScalabilityAutomatic Soft NUMADynamic Memory ObjectsSOS_RWLockFair and Balanced SchedulingParallel INSERT..SELECTParallel Redo
DBCCDBCC ScalabilityDBCC Extended Checks
TempDBGoodbye Trace FlagsSetup and Automatic Configuration of FilesOptimistic Latching
I/OInstant File Initialization is No Longer HiddenMultiple Log WritersIndirect Checkpoint Default Just Makes SenseLog I/O at the Speed of Memory
SpatialNative ImplementationsTVP and Index Improvements
ColumnstoreBatch Mode and Window Functions
Always On Availability GroupsTurbochargedBetter Compression and Encryption
and there is more
Persisted Log Buffer
The evolution of storage
HDD SSD (ms)PCI NVMe SSD (μs)
Tired of WRITELOG waits?Along comes NVDIMM(ns)
Windows Server 2016 supports block storage (standard I/O path)A new interface for DirectAccess (DAX)
2x speeds over PCI NVMe
Persistent Memory
(PM)
Watch these videos
Channel 9 on SQL and PMM
NVDIMM on Win 2016 from \\build
Format your NTFS volume with /dax on
Windows Server 2016
Create a 2nd tlog file on this new volume on SQL
Server 2016 SP1
Tail of the log is now a
“memcpy” so commit is fast
WRITELOG waits = 0 ms Gory Details here
Mission critical platform Deeper Insights Hyperscale
In-Memory OLTPGreater T-SQL surface area, terabytes of memory supported, and higher number of parallel CPUs
Query Store A flight data recorder for your database
Always EncryptedSensitive data remains encrypted at all times, with ability to query
Row-Level SecurityFine-grained access control for table rows
Enhanced Always OnDistributed Availability Groups, Basic Availability Groups, Automatic Failover
The richest set of diagnostics in the industry.
Read more here
High-Performance Execution Engines
ColumnStore + Batch Mode Execution• Columnar compression to save
space in memory and on disk• New execution engine that
optimizes instructions/row and minimizes CPU data bus I/O stalls
• Often gives speedups of 10x• Allows you to store, manage,
and gain value out of significantly more data
In-Memory OLTP Engine Model• Lock-free algorithms• Optimistic concurrency model• Full transactional support• Compiled procedures• Can give significant speedups
(2x-50x) on high-end OLTP applications
Capturing technical breakthroughs in two areas
These enable you to write SQL and leverage ever-increasing performance for your
applications
Mission Critical
Platform
Secure Code
•Secure development lifecycle•Least vulnerable last 6 years
Most Secure DatabaseSecure engine
Least vulnerable last 6 years
2006 2007 2008 2009 2010 2011 2012 2013 2014 20150
1020304050607080
0 211 8
0 1 4 0 0 3
69
53 53 55
3429
2215
5
22
5
28 25 2921
613
3 06
16 169 8 6
43
20
69
18
49
3
SQL Server Oracle DB2 MySQL SAP HANA
* National Institute of Standards and Technology Comprehensive Vulnerability Database update 10/2015
while being most utilized
Mission Critical
Platform
Most Secure DatabaseSurrounded by layers of protection
Secure Code
•Secure development lifecycle•Least vulnerable last 6 years
•SQL Threat Detection •SQL Server Auditing
•Row-level Security•Dynamic Data Masking
•Always Encrypted•Transparent Data Encryption•Encryption-in-flight
Control Access
•SQL Authentication•Windows Authentication• Azure Active Directory Auth.
Monitor activity
Control access
Protect data
Cloud-only
Come to the instructor lead security lab at 4:45
pm today!
Mission Critical
Platform
trust boundary
"SELECT Name FROM Customers WHERE SSN = @SSN","111-22-3333"
Always EncryptedProtect data at rest and in motion, on-premises and in the cloud
NameWayne Jefferson
ADO .NET Name0x19ca706fbd9a
Result SetResult Set
Client
Name SSN Country0x19ca706fbd9a
0x7ff654ae6d USA
SQL Server or SQL Database"SELECT Name FROM Customers WHERE SSN = @SSN",0x7ff654ae6d
ciphertext
Encrypted sensitive data and corresponding keys are never seen in plaintext in SQL Server
dbo.Customers
ciphertext
Security
Mission Critical
Platform
Optimizing query performance and costsQuery Store = flight data recorder for databases
Built-in query monitoring• Negligible perf & storage overhead• Survives failover, memory
pressure, …• Data ready when that incident
happens
T-SQL interface and SSMS UX
Top insights into your workload
• Query and resource stats • Query plans
Developer
App User
Azure
application On-premise
Developer needs to find and fix the underlying problem, ASAP
Customer reports the issue(app is slow/unresponsive)
Mission Critical
Platform
Improved Always On Availability Groups• 3 Replicas for Automatic
Failover• Database Failure Detection
Higher Availability
• Support for Column-store Indexes
• Load Balancing of Read Workloads
More Powerful Readable
Secondaries
• 4X Synchronization ThroughputScalability
• Different Windows Clusters• Different Windows DomainsFlexibility
• Basic Availability Groups in Standard EditionLicensing
Always On Turbocharged
Mission Critical
Platform
Mission critical platform Deeper Insights Hyperscale
Real-time operational analyticsRunning of analytical workloads concurrently with OLTP workloads
R IntegrationInsights from data across SQL Server and Hadoop with the simplicity of T-SQL
PolyBaseInsights from data across SQL Server and Hadoop with the simplicity of T-SQL
Mobile BIMobile reports on SSRS with Mobile Report Publisher
Temporal Tables, JSONTrack change history, integrate with modern service easily
0100101010110
SQL Server OLTP
SQL Serverdata
warehouse
ETL
In-memory ColumnStore
In-memory OLTP
Real-time frauddetection
Operational analyticsSingle system for OLTP and real-time analytics
CapabilityIn-memory Columnstore index over in-memory/disk based OLTP tables
Benefits You gain analytics on fresh data
& 30x faster transactions & 100x faster analytics queries
In-memory for more applications
Fraud detected
2-24hrs
Deeper Insights
Data ScientistInteract directly with data
Built-in to SQL Server
Data Developer/DBAManage data and analytics together
Built-in advanced analyticsRunning R algorithms at massive data scale
Example Solutions• Sales forecasting• Warehouse
efficiency• Predictive
maintenance
Relational Data
Analytic Library
T-SQL Interface
Extensibility
?RR
Integration
010010
100100
010101
Microsoft Azure Marketplace
New R scripts01001
010010
001010
1
010010
100100
010101
010010
100100
010101
010010
100100
010101
010010
100100
010101
• Credit risk protection
Deeper Insights
PolyBaseQuery relational and non-relational data with T-SQL
T-SQL query
CapabilityT-SQL query across SQL Server and Hadoop
Benefits New insights across “data
lake” Leverage existing skillsets
and BI tools Faster time to insights,
simple ETL
SQL Server
Hadoop
Quote:****************************************************************************************************************
$658.39
Jim GrayName
11/13/58
DOBWAState
Ann Smith
04/29/76
ME
Deeper Insights
Support for Temporal DataTrack historical data
Deeper Insights
• Data changes over time (!)
• Databases naturally provide a current view
• But an historical perspective is often critical
Time travel
Data audit
Predictive analytics
Repair record-level corruptions
Mission critical platform Deeper Insights Hyperscale
Stretch DatabaseStretch operational tables in a secure manner into Azure for cost-effective historic data availability.
Enhanced backup to AzureFaster restore times, 50% reduction in storage. Larger DBs
Order history Name SSN DateJane Doe cm61ba906f
d2/28/200
5Jim Gray ox7ff654ae6
d3/18/200
5John Smith i2y36cg776r
g4/10/200
5Bill Brown nx290pldo9
0l4/27/200
5Sue Daniels ypo85ba616
rj5/12/200
5Sarah Jones bns51ra806f
d5/22/200
5Jake Marks mci12hh906
fj6/07/200
5Eric Mears utb76b916gi 6/18/201
4Rachel Hogan px61hi9306f
j 7/1/2014
Sam Johnson ol43bi506gd 7/12/2014
David Simon tx83hal916fi 7/29/2014
Michelle Burns nb95re926gi 8/10/2014
Reed Dean vc61ira536fe
8/23/2014
Order history Name SSN DateJane Doe cm61ba906f
d2/28/200
5Jim Gray ox7ff654ae6
d3/18/200
5John Smith i2y36cg776r
g4/10/200
5Bill Brown nx290pldo9
0l4/27/200
5
Customer data
Product data
Order History
Stretch to cloud
Stretch SQL Server into AzureStretch cold data to Azure with remote query processing
CapabilityStretch cold data from on-premises SQL Server tables to Azure with remote query processing
Benefits Cost effective historical data Entire table is online and remains
queryable from on-premises apps
Transparent to applications Supports Always Encrypted &
Row Level SecuritySQL Server App
Query
Microsoft Azure
Jim Gray ox7ff654ae6d
3/18/2005
Hyperscale Cloud
Fully managed Database Service …So you can focus on your Business
• Built for SaaS and Enterprise applications• Predictable performance & Pricing • Elastic database pool for unpredictable SaaS
workloads• 99.99% availability built-in• Geo-replication and restore services for data
protection• Secure and compliant for your sensitive data• Fully compatible with SQL Server 2016
databases
Predictable workloadsSingle databases or partitioned data across multiple
databases; scale between service tiers and performance levels as capacity needs fluctuate.
Scal
e da
taba
ses
up
as n
eede
d
Scale out/in the pool
reads/writes
reads/writes
read
s/writ
es
reads/
writes
…
Single database or partitioned databases
Customer1
Customer2
Customer3 Customer
#N…
Unpredictable workloadsFor large numbers of databases with unpredictable performance demands; pool resources to be shared between these databases.
Elastic Database Pool
Data
base
s co
nsum
e re
sour
ces
as n
eede
d
Managing large numbers of Databases
SQL Database Service Tiers (single DB model)
*The 99.99% availability SLA does not apply to the existing Web and Business editions, which will continue to be supported at 99.9% availability.
Built For
Available SLA
Database Max Size
Business Continuity
Security
Performance Objectives
Database Transaction Units (DTUs)
Available Tiers ($/Month) and GA
Price
Point-in-time Restore (“oops” Recovery)
BASIC PREMIUMSTANDARD
P1S0
Light transactional workloads
Medium transactional workloads Heavy Transactional Workloads
99.99%*
2 GB 250 GB 500 – 1,000 GB (more soon)
Any point within 7 days Any point within 14 days Any point within 35 days
Geo-restore Geo-replication, standby secondary
Active geo-replication, up to four readable secondary backups
Auditing, row-level security, dynamic data masking, encryption
Transactions per hour Transactions per minute Transactions per second
5
$4.99
S1 S2 S3 P2 P4 P6 P11
10 20 50 100
$15 $30 $75 $150
125 250 500 1,000 1,750
$465
$930
$1,860
$3,720
$7,001
P15 …
4,000
$16,003
Azure SQL Data Warehouse• A relational data-warehouse-as-a-service, fully managed by Microsoft. • Industries first elastic cloud data warehouse with proven SQL Server capabilities.• Support your smallest to your largest data storage needs.
Elastic scale & performance
Scales to petabytes of dataMassively Parallel ProcessingInstant-on compute scales in secondsQuery Relational / Non-Relational
Saas
Azure
PublicCloud
Office 365Office 365
Get started in minutesIntegrated with Azure ML, PowerBI & ADF
Powered by the Cloud
Market Leading Price & Performance
Simple billing compute & storagePay for what you need, when you need it with dynamic pause
AzureAzure
Migration to Meet your Needs
Upgrade MigrationOther
DatabaseSource
Target
Windows Azure Windows Azure
Tool(s) DMA DMA SSMA
Legacy
Modern
ComparisonReports
A’SQL 2008 Instance
Extensive tracing
StatisticalAnalysis
Analyze Results
BSQL 2016 Instance
Extensive tracing
capture prod workload
DEA
replay
BMW Migrations by Industry
Financial Services 139
Discrete Manufacturing 122
Retail & Consumer Goods 113
Government 105
Health 94
Professional Services 93
Process Manufacturing & Resources
59
Hospitality & Travel 44
Telecommunications 36
Power & Utilities 31
Other Industries 119
TOTAL 955
Azure SQL DatabaseAzure SQL Data WarehouseSQL Server 2016 on VM
Analytics Platform System
Azure Data Lake
SQL Server 2016Analytics Platform System
SQL
Relational Non-relational
On-p
rem
ises
Clou
dMicrosoft Data PlatformTransforming data into intelligent action
• Manage all your data in a mission critical, scalable, secure way
• Gain deep insights across all of your data without data movement
• Utilize existing skills and investments
• Consistent on-premises, cloud and hybrid experiences
Federated queryBusiness intelligenceMachine learning & advanced analytics
Insights
• SQL Server 2016 Home Page• Free E-Book on SQL Server 2016 • SQL Server Blog• MSSQL Tiger Team Blog• SQL Server 2016: It Just Runs Faster
Blog Series
Resources
[email protected], github
@bobwardms, #bobsql
Industry ObservationsWorkload-Specific, Optimized Engines Major Performance Enhancements vs. Traditional B+-Tree-based on-disk technologies Leveraging Knowledge of Workload to Improve Performance by Orders of Magnitude In-Memory OLTP and In-Memory + ColumnStore address these within SQL
Rise of Clouds (Public and Private) Have a VM or database in minutes Shut it off when you don’t want it anymore to stop paying for it Simplified Management Experiences, PaaS/IaaS
Accelerated Value Delivery To Customers Customers want value more quickly - Engineering model evolving to go faster Increase Feedback During Development Cycles to Improve Features Make it faster and easier for developers and DBAs to solve problems through T-SQL
In-Memory Transactional Processing
SQL Server Integration
• Same manageability, administration & development experience
• Integrated queries & transactions
• Integrated HA and backup/restore
In-Memory Data Access
• Indexes (hash and range) exist only in memory
• Records exist only in memory (pointers from indexes)
• No buffer pool, B-trees
• Stream-based storage
T-SQL Compiled to Machine Code• T-SQL compiled to
machine code via C code generator
• Invoking a procedure is just a DLL entry-point
• Aggressive optimizations @ compile-time
Declining memory price
Many-core processors
Stalling CPU clock rate Lower TCO
Hardware trends Business
Integrated experience
High performance
data operations
Efficient business-logic
processingCust
ome
r Be
nefit
sAr
chit
ectu
ral P
illar
sD
rive
rs
High Concurrency
• Core engine uses lock-free algorithms
• No lock manager, latches or spinlocks
• Multi-version optimistic concurrency control with full ACID support
Frictionless scale-up
45
Column-Store Indexes
Improved compression:Data from same domain
compress better
Reduced I/O:Fetch only columns needed
…
Row-store indexes Column-store indexes
Ideal for OLTP Efficient operation on small set of rows
C5C1 C2 C3 C4
Improved Performance:More data fits in memoryOptimized for CPU utilization
Ideal for DW Workload
rowgroup
segment
Safe Upgrades to SQL 2016 with Query StoreSafe Upgrade Workflow:
1. Upgrade to SQL 2016, but do not enable new features yet (keep the old DB Compatibility Level)
2. Turn Query Store ON, record your workload in Query Store
3. Enable new Compatibility Level (130)
4. Use Query Store to find any and fix performance regressions
Temporal TablesTracking history is important but challengingCommon to healthcare, pharmaceutical, financial services, social networks…
SQL Server makes it super easyTrack data changes with zero code in your appZoom-in data analysis on any point in time
Time TravelData Audit
Row-level error correction
• Interchange data with apps and services
• Exploit agility of NoSQL to easily extend your app
Agile development with JSON
SQL ServerWeb app, service
CREATE TABLE Departments ( DeptId int NOT NULL PRIMARY KEY CLUSTERED , DeptName varchar(50) NOT NULL , ManagerId int NULL) ;
How Temporal Tables WorkEnd-to-end m
obile BI
CREATE TABLE Departments ( DeptId int NOT NULL PRIMARY KEY CLUSTERED , DeptName varchar(50) NOT NULL , ManagerId int NULL , SysStartTime datetime2 GENERATED ALWAYS AS ROW START NOT NULL , SysEndTime datetime2 GENERATED ALWAYS AS ROW END NOT NULL , PERIOD FOR SYSTEM_TIME (SysStartTime, SysEndTime) ) WITH (SYSTEM_VERSIONING = ON (HISTORY_TABLE = DepartmentHistory)) ;
How Temporal Tables WorkEnd-to-end m
obile BI
How Temporal Tables Work
Insert
prior version Start EndDeptId DeptName ManagerId
DepartmentsHistory
End-to-end mobile
BI
Update / Delete
Departments (temporal)
DepartmentsStart EndDeptId DeptName ManagerId
How Temporal Tables Work
Start EndDeptId DeptName ManagerId
DepartmentsHistory
End-to-end mobile
BI
Normal querySELECT * FROM Departments
GO
Departments (temporal)
DepartmentsStart EndDeptId DeptName ManagerId
How Temporal Tables Work
Start EndDeptId DeptName ManagerId
Departments DepartmentsHistory
End-to-end mobile
BI
Temporal query SELECT * FROM Departments
FOR SYSTEM_TIME AS OF '2015.01.01'
GO
Departments (temporal)
Start EndDeptId DeptName ManagerId
Paginated reportsDesign beautiful documents using updated tools and new features
Mobile reportsCreate responsive, interactive reports optimized for mobile devices
Modern web portal Consume both types of reports in one web portal using modern browsers
SQL Server Reporting ServicesModern enterprise-ready on-premises reporting platform
SQL Server Analysis Services 2016Enterprise scale models with in-memory technology built-in
Use SQL Server Analysis Services asa semantic model without storing dataHigh performance DirectQuery for Tabular
Make it easier to create powerful models Tabular modeling capabilities and performance
Improve manageability and securityProvide additional tools for administrators
Strengthen multidimensionalPerformance and other updates
Easily create models
IT pro
Increase adoption
Business analyst
Share insights faster
Faster time to insight
BI consumers
Move your SQL data tier to CLOUD
Why move? → TCO, Agility, Elasticity, SimplicityElastic resources, on-demand, pay-as-you-goDramatically less management burden on youRich ecosystem of services, to build rich appsMigration assistants to help you move
OLTP
Industry leading TCO
$648K+ $120
per user for Power BI
SQL Server 2016
ADVANCED ANALYTICS
BUSINESS INTELLIGENCE
OLTP
DATA WAREHOUSING
3.4x
20x per user
$2.2M+
$2,230 per user for BI
Note: For OLTP and DW scenario, the price comparison is based on a server with 2 proc, 8 cores each
Built-in with SQL Server vs. expensive add-ons with Oracle
Complete mobile BI
In-memory
End-to-end security
Advanced Analytics
built-in
built-in
built-in
built-in
DATA WAREHOUSING
BUSINESS INTELLIGENCE
ADVANCED ANALYTICS