data warehousing guidelines for bi and bam solutions
DESCRIPTION
Data warehousing guidelines for bi and BAM solutions session at SQL Saturday Italy event on 13/12/2013TRANSCRIPT
Performance Dreams wait for you at SQL Server 2014
At KACST @ Riyadh on 21/12/2013
Data Warehousing Guidelines for BI and BAM solutions
12/12/2013 by Shehap El Nagar
Sponsors & Media Partners
Shehap EL-Nagar
About me
I am MVP,MCTS , MCITP SQL Server, I am DB consultant and Architect for lots of Banking, Telecom ,Ministries and governmental organizations all over Gulf ,also he has deep knowledge about T-SQL performance , HW Performance issues, Data Warehousing solutions , SQL Server Replication, Clustering solutions and Database Designs for different kinds of systems ...
The founder of the biggest SQL Server community all over the middle east http://sqlserver-performance-tuning.net/ , you can watch its success memories at http://www.youtube.com/user/ShehapElNagar
Moderator and author at http://www.sql-server-performance.com ,, the 1st SQL Server Author at MSDN Arabia http://msdn.microsoft.com/ar-sa/library/jj149119.aspx
, Speaker at SQL Saturday Events worldwide , local events at Saudi Arabia , many online events , more than 90 video tutorials and also many private sessions for .net developers and Database Administrators
And also influent participator at Microsoft Forums of SQL Server at http://social.technet.microsoft.com.More about him , you can find him on MVP Microsoft site http://mvp.microsoft.com/en-us/mvp/Shehap%20El-Nagar-5000188 .You can contact him at the below contacts :Mail :[email protected] ….Cellular phone :00966560700733
Agenda and Overview:First :Definitions and benefits of DWH •Definition of Data Warehousing Solutions •Benefits of DWH Solutions•Why RTDWH (Real Time Data Warehousing ) is high necessary..? •Data Warehouse vs. Data Mart•Relational DB vs. Dimensional DB•Dimensional Database vs. Multidimensional Database•Star Schema vs. Snow flake schema •Techniques of DWH solutions
Second : RTDWH for online Reporting•Technique and concepts•Demo
Third :DWH for online Archiving•Technique and concepts•Demo
Fourth :DWH for online ETL•Technique and concepts•Demo
First :Definitions and Benefits of DWH
Definition of Data Warehousing
6
Data Warehouse Engine
Optimized Loader
DataCleansing
De-normalize Data
Metadata Repository
RelationalDatabase 1
RelationalDatabase 3
RelationalDatabase 2
RelationalDatabase 4
Benefits of Data warehousing:Data Consolidation & organization
Data standardization for different attributes such as Collation
Support numerous RDBM sources flexibly like SQL Server , Oracle , TeraData , Informix , SAP BI , Sybase, Access , CSV files , Excel…etc
Scale up reports either SSRS or SSAS reports (OLAP Reports)
Speed up reports performance
Why Real Time Data Warehousing..? Active decision support
Business activity monitoring (BAM)
Alerting
Efficiently execute business strategy
Relational DB vs. Dimensional DB:
Relational DB represents a normalized DB for OLTP transactions purposes.
More normalization >>>Less no of columns >>> less possibility of indexes >>> Less IO cost of cluster indexes while using them for insert /update /delete of OLTP transactions Dimensional DB represents a de-normalized DB for OLAP purposes
More number of interrelated columns in one table >>> Less possibility for joins >>> More covering compound indexes
Data Warehouse vs. Data Mart:
•Data warehouse is a global repository for wide scale of business
•Data mart is a smaller repository for specific business scope
Therefore, we could say a Data Mart solution is sub set of a bigger Data warehousing solution
11
Data Warehouse vs. Data Mart
Data Marts
Data Sources
Data Warehouse
Dimensional Database vs. Multidimensional Database:
Dimensional DBs could be used as staging DBs for SSAS reports or they could be used directly for SSRS reports Multidimensional DB represent SSAS DBs composed of cubes which are formed basically of : •Facts tables which Contain business process core where aggregative columns called measures could be found there.
•Dimension tables which Contain Lookup details relevant to these aggregative data
(DWH DB)Dimensional DB
OLAP Service
Decision Support Client
DB Service Presentation Layer
(OLAP DB)Multi-Dimensional DB
Star Schema vs. Snow flake schema :
Snow flake schema close much the design of star schema design but the first one is trying to break down schema design more into smaller tables to avoid more redundancy of columns.
Snow flake schema isn’t recommended for neither OLAP transaction nor OLTP transaction
Star Schema
Telephone
Fact Table
date, custno, prodno, cityname,Region ...
Gender
Name
Marital status
Snow flake Schema
Telephone
Fact Table
date, custno, prodno, cityname,Region ...
Gender
Name
Marital status
Marital status lookup table
Gender lookup table
Data warehousing techniques• • Old 2005 codes (Select /insert/Update /Delete)
• New 2008 codes “Merge” which could replace more efficiently all of above commands in one statement
• DTS (Data transformation Service) and SSIS Packages
• Enterprise platform solutions for LDWH(Large DWH)
Fast Data tracking solution
Sybase IQ
Red Brick Warehouse
IBM
DB2 MVS
Universal Server
IBM Data Warehousing
TeradataInformix
Online Dynamic ServerXPS --Extended Parallel ServerUniversal Server for object relational applications
Enterprise Platform Solutions
Second :RTDWH for online Reporting
Technique of DWH Solutions used for Online Reporting
• Creating 2 tables (One Temp table and the second is DWH table itself)• • Making all DML transactions on a Temp table.
• Then compare Temp table results with DWH table.
• If not match for any record /column, then Bulk Merge command from Temp table to DWH Table
• You can use now this DWH Table for your online Reports
Concepts of DWH Solution used for Online Reporting
1- Set xact_abort on : To ensure the highest transactional status for group of DML transactions to commit all if all succeeded and rollback all if any of them failed 2- Set nocount on :To speed up queries by avoiding counting no of records each time of run 3- Set deadlock_priority low; To avoid any impact on end users transactions while this online data warehousing. 4- Try /Catch commands : To capture any possible errors and report them by mail.
5- Bulk Logged mode :To save efficiently more storage capacity while bulk merge
6-Using Read committed snapshot isolation level using row versioning is recommended to avoid heavy locks/deadlcoksd
Demo
Third :Data warehousing
For online Archiving solution
Techniques of DWH Solutions for Archiving
•Bulk insert the old data from a Source table to an Archived table
•Bulk delete from source table after success of 1st step
•Bulk delete should be split into smaller patches with small
no of records like 1000 delay of 5-30 sec between each patch and another to avoid any tangible locks or deadlocks
Concepts of DWH Solutions used for Archiving
1 - Bulk Logged mode :To save efficiently more storage capacity while bulk merge as we are going more to show by next workshops
2- Use WAITFOR DELAY '00:00:30'doesn’t mark for risky waits here, but just a normal wait command like service broker wait. 3- Bulk Insert and bulk delete phases could be conducted in different transactions in different time intervals without any risks
4- You could validate that also using output commands
Demo
Fourth :Data warehousing
For online ETL solutions
Technique of DWH Solutions used for ETL
Run your ETL process in parallel with end users activities but to a different table rather than online tables
Once finish, start to scan all mismatches between the 2 tables through the 3 data warehousing statements
Concepts of DWH Solutions of DWH used for ETL
Scanning any new inserted data entity within the source tables to be inserted to the target tables
Scanning any updated data entity through scanning any records shared between the 2 tables for PK values but different for any other columns.Scanning any deleted data through using except commands
•The 3 phases could be undertaken asynchronously without any risk at all
Q & APost your questions at:
http://www.sqlserver-performance-tuning.net/forums/
Performance Dreams wait for you at SQL Server 2014
At KACST @ Riyadh on 21/12/2013
12/13/2013
Thank you ..See you again
Sponsors & Media Partners