1. SQL Server Fast Track & Project Madison – SQL MPP
Roger Moore – Data Warehouse SSP

2. Agenda Microsoft Data Warehouse Strategy SQL DW & BI
SQL Server Fast Track
Madison Overview – SQL MPP (DATAllegro)
Hub and Spoke
Multi-Temperature
MTP – Technology Preview (PoC)
Summary
Microsoft Confidential

3. Our Integrated BI-DW Offering
SharePoint Server
SEARCH
DELIVERY
Reports
Dashboards
Excel
Workbooks
Analytic
Views
Scorecards
Plans
CONTENT MANAGEMENT
COLLABORATION
END USER TOOLS & PERFORMANCE MANAGEMENT APPS
Excel
PerformancePoint Server
BI PLATFORM
SQL Server
Reporting Services
Analysis Services
SQL Server DBMS
SQL Server Integration Services

4. Microsoft Is Serious About Data Warehousing
2005
2008
Futures
Data Warehouse Scale
10s of TB Warehouses
Parallel partitioning
Data compression
New Reference Architectures
Policy Based Admin.
DB Resource Governance
High Perf. Connectors (Oracle, Teradata, SAP BW)
Data Profiling
Policy based auditing
Multi TB Warehouses
Enterprise scalability
DW Reference Architectures
Unified manageability
Enterprise class ETL tool
Data Cleansing (Fuzzy lookup/matching)
Data Protection
& Tracing
PB Warehouses
>64 Core Processing
Scale out through MPP
Perf. Management Tools
BI Resource Governance
Improved Predictability
Mixed workload support
Continuous Loading
Integrated DQ Services
(Zoomix)
Master Data Management (Stratature Integration)
Rights Management
Data Warehouse Management
Heterogeneous Connectivity & Workloads
Data Integrity & Quality
Compliance & Security

5. The Appliance Model for Data Warehousing
Building a traditional DW
Time consuming
Expensive
Performance varies
Scalability issues
Potential bottlenecks in standard DW architecture
The DW appliance model
Pre-built & tuned h/w + s/w
Views entire stack holistically
Known performance & scalability
Encapsulates best practices
Leverages Sequential I/O
Lower TCO
Faster deployment
Better performance
Minimised DBA time
Benefits

6. What is SQL Server Fast Track Data Warehouse?
An appliance approach to SMP data
warehouse reference architectures
Pre-built & tuned h/w + s/w
Views entire stack holistically
Known performance & scalability
Encapsulates best practices
Leverages Sequential I/O
Seven distinct reference architectures
Delivered with SI Partners –
QuickStart assessments
Solution templates
Helping Customers & Partners Accelerate Their Data Warehouse Deployments

7. 4/1/2017 4:22 AM
Fast Track Data Warehouse Components Key Principle 1: Tight Specifications
Software:
SQL Server Enterprise
Windows Server 2008
Configuration guidelines:
Physical table structures
Indexes
Compression
SQL Server settings
Windows Server settings
Loading
Hardware:
Tight specifications for servers, storage and networking
‘Per core’ building block
<Session Name> Microsoft NDA-only
© 2004 Microsoft Corporation. All rights reserved.
This presentation is for informational purposes only. Microsoft makes no warranties, express or implied, in this summary.

8. Key Principle 2: Balanced Across All Components
SQL Server 2008 Potential Performance Bottlenecks
SERVER
CACHE
SQL SERVER
WINDOWS
CPU CORES
FC SWITCH A B
DISK
LUN FC
HBA A B
STORAGE
CONTROLLER A B
CACHE
Balanced for data warehousing workload
Provide calculations for sizing one’s workload
Base system on CPU and its’ ability to grind data regardless of workload
I/O provides data required to keep CPU performing at max ability
CPU Feed Rate
SQL Server
Read Ahead Rate
HBA Port Rate
Switch Port Rate
SP Port Rate
LUN Read Rate
Disk Feed Rate

9. Key Principle 3: Sequential I/O
Random I/O
Ideal for data warehousing
Scalable, predictable performance
Large reads & writes
Requires 1/3 or fewer drives for same performance
Ideal for OLTP
Not as predictable & scalable for data warehousing
Small reads and writes
Requires large number of drives
Best practices focus on preserving the
sequential order of data

10. SQL Server Fast Track Data Warehouse for HP
2 Processor Configuration
Server: HP ProLiant DL385 G5p with 2 Quad-core AMD Opteron processors
Storage server: EMC or MSA Storage
Scalability: up to 8 TB
4 Processor Configuration
Server: HP ProLiant DL 585 G5 with 4 Quad-core AMD Opteron processors
Scalability: 4 – 16 TB
8 Processor Configuration
Server: HP ProLiant DL 785 G5 with 8 Quad-core AMD
Opteron processors
Scalability: 16 – 32 TB
Note Compression assumes 2.5:1

11. SQL Server Fast Track Data Warehouse for DELL
2 Processor Configuration
Server: Dell Power Edge 2950 MLK with 2 Quad-core Intel Xeon processors
Storage server: EMC CX4-240 & AX4
Scalability: up to 8 TB
4 Processor Configuration
Server: Dell Power Edge R900 with 4 6-core Intel Xeon processors
Scalability: 12 – 24 TB
Note Compression assumes 2.5:1
- Fully loaded only adds drives to minimum HW required
- Data space can be increased by using 450GB drives

12. Fast Track Case Study - Environment
Current Environment
Teradata 4-node (5450 model) with 6TB of user data
BI: Business Objects
ETL: Informatica and BTEQ scripts
Proposed Microsoft Platform
SQL Server Fast Track Data Warehouse
HP DL580 Server - 4 Quadcore Processors  (16 core total)
256 GB Memory
SAN Storage: MSA 2000 (Qty 4) – 8TB User Data Capacity
ETL: SQL Server and SSIS

13. Fast Track Case Study - Results
Teradata
SQL Server
Fast Track DW
Comparison
Loading –
Subject Area 1
5:10:21 total time
51:31 total time
R SQL Server 6x
faster
Subject Area 2
4:36:08 total time
1:50.01 total time
R SQL Server 2.5x
Query times –
3:03 avg query time
(using 9 benchmark queries)
0:15 avg query time
R SQL Server 12x
56:44 avg query time
(using 4 benchmark queries)
8:09 avg query time
R SQL Server 7x

14. Fast Track Benefits Summary
4/1/2017 4:22 AM
Fast Track Benefits Summary
Appliance-like time to value
Reduces DBA effort; fewer indexes, much higher level of sequential I/O
Choice of HW Platforms
Dell, HP, Bull – more in future
Low TCO Through
Commodity Hardware and value pricing; Lower storage costs.
Appliance-like time to value
Reduces DBA effort by tightly specifying and managing the DBMS to HW interface
Very good out of the box performance
Fewer indexes
Much higher level of sequential I/O
Choice of HW platforms
Dell, HP, Bull – more in future
Low TCO through
Commodity Hardware and value pricing. Price/TB of new Reference Architectures will be less than half of Netezza’s and under 1/3rd of Oracle’s Exadata
Easier to achieve and maintain good performance
Lower storage costs
High scale
New reference architectures scale up to 32 TB (assuming 2.5X Compression), allowing customers to scale their DWs on SQL Server
Reduced risk
Tested by Microsoft; better choice of hardware; application of Best Practices and industry solutions
High Scale
New reference architectures scale up to 32 TB (assuming 2.5x compression)
Reduced Risk
Tested by Microsoft; better choice of hardware; application of Best Practice
<Session Name> Microsoft NDA-only
© 2004 Microsoft Corporation. All rights reserved.
This presentation is for informational purposes only. Microsoft makes no warranties, express or implied, in this summary.

15. The Bridge to Project "Madison"
Fast Track offers appliance-like ease of deployment, scalability and performance for SMP
Madison to offer massively parallel (MPP) scale and performance
Madison hub-and-spoke architecture to include support for SMP spokes

16. Fast Track Data Warehouses
Scaling SQL Server 2008
INDUSTRY STANDARD
SERVERS
Fast Track Data Warehouses
Scale Up
INDUSTRY STANDARD
NETWORKING
INDUSTRY STANDARD
STORAGE
Scale Out
Project
“Madison”

17. MPP (Madison) Overview
INDUSTRY STANDARD
NETWORKING
SERVERS
Reference Hardware Platforms
Project
Madison
STORAGE

18. Madison – SQL MPP Architecture Sample
Compute Nodes
Compute Nodes
Storage Node
Control Nodes
Active / Passive
Dual Infiniband
Landing Zone
Backup Node
Storage Servers
Spare Compute Node

19. Project “Madison” Demonstration Architecture
TPCDS – 150+ Terabytes
Date Dim
D_date_sk
D_date_id
D_date
D_month …
Store Sales
Ss_sold_date_sk
Ss_item_sk
Ss_customer_sk
Ss_cdemo_sk
Ss_store_sk
Ss_promo_sk
Ss_quantity
Promotion
P_promo_sk
P_promo_id
P_start_date_sk
P_end_date_sk
Customer
C-Customer_sk
C_customer_id
C_current_addr
Item
i_item_sk
i_item_id
i_rec_start_date
i_item_desc
Store
S_store_sk
S_store_id
S_rec_start_date
S_rec_end_date
S_store_name
Demographics
Cd_demo_sk
Cd_gender
Cd_marital_status
Cd_education
73, 049
100 Million
502, 000 1
Trillion
Rows
2, 500
1.92 Million
1, 902

20. Distributed & Replicated Tables
Data Distribution with Replication
Database
Distributed & Replicated Tables
Date Dim
D_date_sk
D_date_id
D_date
D_month … C I D CD S P
Customer
C-Customer_sk
C_customer_id
C_current_addr … SS
Item
i_item_sk
i_item_id
i_rec_start_date
i_item_desc …
Store Sales
Ss_sold_date_sk
Ss_item_sk
Ss_customer_sk
Ss_cdemo_sk
Ss_store_sk
Ss_promo_sk
Ss_quantity … C I D CD S P SS C I D CD S P SS C I D CD S P SS
Promotion
P_promo_sk
P_promo_id
P_start_date_sk
P_end_date_sk …
Customer
Demographics
Cd_demo_sk
Cd_gender
Cd_marital_status
Cd_education … C I D CD S P SS C I D CD S P SS
Store
S_STORE_SK
S_STORE_ID
S_REC_START_DATE
S_REC_END_DATE
S_STORE_NAME … C I D CD S P SS C I D CD S P SS

21. Processor Utilization
4/1/2017 4:22 AM
Processor Utilization
MICROSOFT CONFIDENTIAL
© 2006 Microsoft Corporation. All rights reserved. Microsoft, Windows, Windows Vista and other product names are or may be registered trademarks and/or trademarks in the U.S. and/or other countries.
The information herein is for informational purposes only and represents the current view of Microsoft Corporation as of the date of this presentation. Because Microsoft must respond to changing market conditions, it should not be interpreted to be a commitment on the part of Microsoft, and Microsoft cannot guarantee the accuracy of any information provided after the date of this presentation. MICROSOFT MAKES NO WARRANTIES, EXPRESS, IMPLIED OR STATUTORY, AS TO THE INFORMATION IN THIS PRESENTATION.

22. Madison and Fast Track Hub and Spoke
4/1/2017 4:22 AM
Madison and Fast Track Hub and Spoke
Central EDW Hub
Regional Reporting
Departmental Reporting
ETL Tools
High Performance HQ Reporting
Each business unit has own Data Marts
More responsive to business needs
Fits budget realities
Hub provides centralized data governance platform
Node-to-node data movement
Parallel over Infiniband or 10 Gig Networks
<500GB per min with minimal overhead
1) Simplification of complex EDW workloads through distribution of workload across multiple appliances.
2) Centralized platform for data governance –
A single platform for ETL and ELTL processes
Simplified governance through publishing spokes versus complex ETL processes
One version of the truth… a single golden copy of data… resolves issue with dependent
data marts
Flexible to support customer unique data governance processes and policies.
3) Node to node data movement –
Current plan is to support 10Gige and infiniband for high speed data movement.
True data movement at the database level. No overhead for exporting and loading of
data.
Transfer processes in Madison are in development. Transfer rates will vary based on
size of the source and destination systems.
<Session Name> Microsoft NDA-only
© 2004 Microsoft Corporation. All rights reserved.
This presentation is for informational purposes only. Microsoft makes no warranties, express or implied, in this summary.

23. Madison Multi-Temperature
Auto Publish
FRESH DATA LOADING
Most Recent -
3 Months
2 Years
7 Years
User Queries
BI Server
Queries
User Data
Hot -> Warm -> Cold
Stage -> ODS -> Prod
Back-up / Archive
Data structure in synch
Fast response to users
Easy Data Movement
High Availability
Grid Flexibility – architectural options
Multi-temperature user data
Staging -> ODS -> Production
ODS -> Production
Hot -> Warm -> Cold data
Multi-temperature back-up or archive data – Not commonly queried by users
Enables ability to keep archive data structures in synch with production data structures
Fast response to user data requests
Better trend analysis
Easy data movement between Development, Test and Production systems
High availability
Spokes or Hubs can back each other up

24. Case study: Tier 1 Carrier - CDR Architecture including Multi Temperature Archive
220TB
ARCHIVE DW
120 TB HIGH CAPACITY
‘WARM’ CDRs
ROLL OFF TO
ARCHIVE
UP TO 500M
ROWS/DAY
HIGH-SPEED
PARALLEL
UPDATES
MARGIN
ANALYSIS
Key messages:
High capacity 300+TB installation
Appliance left - High volume of data coming in and being transformed / enriched (CDR mediation and cleansing)
Appliance Middle - Reporting infrastructure / Data Marts –
Appliance Right - Archive infrastructure – Appliance Right
- Grid replication for replicating data at high speed between appliances.
Multi-temperature strategy with very recent data on the left, reporting data in the middle, and archive on the right.
Right-sized appliance for each use-case
Archive appliance lets them store up to 2 years of CDR call history
FRAUD
DETECTION
COST
MGT
BILLING
60 TB HIGH PERFORMANCE
FOR MEDIATION & AUGMENTATION USING ETL TOOLS
REVENUE
ASSURANCE

25. "DW Appliance" Experience
All hardware from a single vendor
Multiple vendors to chose from
Orderable at the rack or cluster
Vendor will
Assemble appliances
Image appliances with OS, SQL Server and Madison software
Appliance installed in less than a day
Support –
Vendor provides hardware support
Microsoft provides software support

26. Madison Beta Programs Two Programs Requirements
MTP – Madison Technology Preview
15-20 participants
Duration of 4 to 6 weeks
TAP – Beta production implementation
4-6 customers
First iteration 9 to 12 weeks
Requirements
Focus on EDW and large data marts
Migration projects, not green field
Open to customers & prospects

27. DW QuickStart – Data Warehouse Roadmap Service
Requirements
Existing DW
Volume of end-user data 1TB+
Considering change to BI or DW infrastructure
On site survey
Interview of key stake holders in Data Warehouse environment
Performed by Microsoft Architect
Service also available from selected Microsoft partners with deep Data Warehouse expertise
2-5 days duration
Deliverables
Presentation of key findings
Report detailing findings
Results delivered approximately 10 days after survey

28. Summary Microsoft has a compelling EDW vision
BI, ETL, scale up and out
Hub & Spoke architecture
Fast Track available today
Up to 30TB
Scale up today with SMP, scale out tomorrow with MPP
MTP and TAP for Madison in June 2009
Scales up SQL Server to >1PB
Sets a new bar in appliance pricing and performance
Hub-and-Spoke will integrate Fast Track with Madison

29. Our Integrated BI-DW Offering
SharePoint Server
SEARCH
DELIVERY
Reports
Dashboards
Excel
Workbooks
Analytic
Views
Scorecards
Plans
CONTENT MANAGEMENT
COLLABORATION
END USER TOOLS & PERFORMANCE MANAGEMENT APPS
Excel
PerformancePoint Server
BI PLATFORM
SQL Server
Reporting Services
Analysis Services
SQL Server DBMS
SQL Server Integration Services

30. 4/1/2017 4:22 AM
© 2009 Microsoft Corporation. All rights reserved. Microsoft, Windows, Windows Vista and other product names are or may be registered trademarks and/or trademarks in the U.S. and/or other countries.
The information herein is for informational purposes only and represents the current view of Microsoft Corporation as of the date of this presentation. Because Microsoft must respond to changing market conditions, it should not be interpreted to be a commitment on the part of Microsoft, and Microsoft cannot guarantee the accuracy of any information provided after the date of this presentation. MICROSOFT MAKES NO WARRANTIES, EXPRESS, IMPLIED OR STATUTORY, AS TO THE INFORMATION IN THIS PRESENTATION.
© 2007 Microsoft Corporation. All rights reserved. Microsoft, Windows, Windows Vista and other product names are or may be registered trademarks and/or trademarks in the U.S. and/or other countries.
The information herein is for informational purposes only and represents the current view of Microsoft Corporation as of the date of this presentation. Because Microsoft must respond to changing market conditions, it should not be interpreted to be a commitment on the part of Microsoft, and Microsoft cannot guarantee the accuracy of any information provided after the date of this presentation. MICROSOFT MAKES NO WARRANTIES, EXPRESS, IMPLIED OR STATUTORY, AS TO THE INFORMATION IN THIS PRESENTATION.