Sql Server 2008 Enhancements

What’s new in SQL
Server 2008 ?

Kobi Cohen

Objectives

Familiarization with main SQL Server 2008 DB
engine enhancements.
Reviewing Integration Services
enhancements.
Reviewing Reporting Services enhancements.
Reviewing Analysis Services enhancements.
Upgrading to 2008 recommended method.

SQL Server 2008 Enhancements
DB Engine Main Enhancements

Data Compression (Enterprise Edition)
Compression in row or page format.
Affect tables, indexes or each partition separately.
Stores fix length type as variable length. An int value of 1:
No compression  4 bytes + 2 bytes metadata overhead.
With Compression  1 byte + 4 bits metadata overhead.
Reduce storage requirements and memory usage.
Less number of pages in cache  less number of latches 
Higher buffer hit rates  better memory utilization with a bit
of extra CPU time.
Significant performance improvement for large I/O bound
workload, like DWH.
Main goal  Shrink fact tables (2X to 7X compression
ratio for real DW fact data).
Secondary goal  improve query performance.


Backup Compression (Enterprise Edition)
Require less I/O  Increase backup speed.
Less storage required to keep backups online (expect 4:1
ratio).
Increase CPU usage significantly  can be set to low priority
session by resource governor.
Average Performance for a Set of Customer Databases:

Resource Governor (Enterprise Edition)
Allow defining resource limit & priorities for different
workloads. Govern queries base on group/ pool.
Limits the amount of CPU & memory that incoming
applications requests can use.
Resource limits can be configured in real time with minimal
impact.
Limited to DB engine  cannot be used for
SSAS, SSIS, SSRS.
DWH usage – handling unpredictable
workloads execution, like 2 un-insolated
DWH applications or a mix of OLTP & DWH
applications.

Filtered Indexes & Statistics (Enterprise Edition)
Non clustered index only, that support WHERE condition and
includes only matching rows.
Improved query performance and execution plan quality 
it’s smaller then full table non clustered index and have
smaller and more accurate statistics.
Reduce index maintenance cost  DML statements is not
always affecting the index.
Reduce index storage cost  Replacing a full table non
clustered index with multiple filtered index without
significantly increasing storage requirements.
CREATE INDEX #IX_Filtered
ON Production.WorkOrderRouting (ActualEndDate)
WHERE ActualStartDate > '2002-01-01' AND ActualStartDate < '2002-12-31‘

Parallel Query Processing on Partitioned Objects
(Enterprise Edition)
Multiple processes work simultaneously to process a single
SQL Statement.
In SQL Server 2005  one thread was created per partition,
regardless the number of processing cores.
In SQL Server 2008  all processors can be used to satisfy
the query requirements. If threads number > partitions
number, query processor allocates equal number of threads
number to each partition.

Change Data Capture - CDC (Enterprise Edition)
Capturing the fact that DML changes were made + the actual
metadata that was changed.
Changes are captured and placed in changes tables.
Log-based capture (by jobs) with minimal impact on source.
Load delta only instead of full refresh – Efficient for updates
of data warehouses or similar.
ETL process can call functions to retrieve
changes into DWH or similar.
Work on schema changes as well – changes to
column structure are stored in cdc.ddl _history
table.
Built-in automatic cleanup (by job) – by default
3 days of data is retained.

Change Tracking – CT
Captures the fact that rows in a table were changed, but does
not capture the data that was changed (only PK value is saved).
Lightweight  Tracks time & context of most recent row
change.
Get latest row data directly from user table  limited historical
data  less storage overhead than CDC.
Crawl Transaction Log  Asynchronous  High Performance.
Provide functions to obtain changes info.
Best for Occasionally Connected
Systems like Mobile Sync Apps or
Remote Dial Up Apps.

Auditing (Enterprise Edition)
SQL Server instance & database level audit specifications.
Available auditing destinations:
Binary file.
Windows application log.
Windows security log (not available in Windows XP).
Audit specifications defines what to audit  Login, Logout,
DDL/DML changes, Restarts, etc.
3 categories of actions for audit:
Server level – server operations, such as management changes
and logon and logoff operations.
Database level – DMLs and DDLs operations.
Audit level – include actions in the auditing
process (CREATE AUDIT, ALTER AUDIT etc).
Based on Extended Events architecture –
can be correlated with windows events to
resolve resources problems.

Data Collector
Data warehouse mechanism.
Running all the time or on a user-defined schedule, and
collecting different sets of data.
Collected data stored in the management data warehouse.
System data collection sets including:
Server Activity.
Disk Usage.
Query Statistics.
Supports dynamic tuning for data collection  extensible
through its API.
Integrated reporting & analyzing.


FILESTREAM Storage
Improved performance:
Break 2GB limit of BLOB storage.
NTFS caching & streaming.
Keep BLOBs out of SQL memory pool.
Transaction file I/O over NTFS.
Stores BLOB data on NTFS – for VARBINARY(MAX) only.
It’s recommended for BLOB objects that are larger than 1MB.
BLOB are stored in SQL-managed folder:
Folder managed as a file group.
Nesting of folders not ok.
Limit folder ACLs to SQL service account.

Policy Based Management
Enforce compliance on database objects:
Asserts a condition against one or more targets.
Repair some violations automatically.
Run on schedule or on demand.
Server Group Management
Creating Central Management Server.
Supports targeting multiple servers for:
Execution T-SQL queries.
Managing policies.

Merge Statement
Perform multiple DML (insert/ update/ delete) once and
perform better than separate DMLs.
Operations based on:
WHEN MATCHED.
WHEN [TARGET] NOT MATCHED.
WHEN SOURCE NOT MATCHED.
MERGE statement is transactional - No explicit transaction
required.
Multiple WHEN clauses possible on:
MATCHED and SOURCE NOT MATCHED.
Only one WHEN clause for TARGET NOT MATCHED.
One Pass Through Tables.
At most a full outer join.
Matching rows = when matched.
Left-outer join rows = when target not matched.
Right-outer join rows = when source not matched.

Merge Statement (2)
Usage Example:
MERGE dbo.FactBuyingHabits AS Target
USING (SELECT CustomerID, ProductID, PurchaseDate FROM dbo.Purchases) AS Source
ON (Target.ProductID = Source.ProductID AND Target.CustomerID = Source.CustomerID)
WHEN MATCHED THEN
UPDATE SET Target.LastPurchaseDate = Source.PurchaseDate
WHEN NOT MATCHED BY TARGET THEN
INSERT (CustomerID, ProductID, LastPurchaseDate)
VALUES (Source.CustomerID, Source.ProductID, Source.PurchaseDate)
OUTPUT $action, Inserted.*, Deleted.*;

Programmability Enhancements
New separate DATE & TIME types.
User-Defined table type:
Declared globally.
Used as argument to stored procedures and functions.
Great for INSERT/ UPDATE procedures.
Eliminates connection isolation.
Declaration example:
CREATE TYPE LocationTableType AS TABLE (LocationName varchar(50), CostRate int);
DECLARE @tbl_LocationTableType AS LocationTableType;

Sparse Columns – Null values are not stored  Ideal for
property/ value storage.
HIRARCHYID Data type – Enables DB applications to model
tree structures  Implemented as a CLR UDT.
Grouping Sets – allow multiple GROUP BY clauses in a single
SQL statement.
GROUP BY GROUPING SETS ((), (a), (a, b, c))

Programmability Enhancements (2)
T-SQL Row constructors – allow multiple value inserts within a
single INSERT statement.
INSERT INTO dbo.Departments
VALUES (1, 'Human Resources', 'Margheim'),(2, 'Sales', 'Byham'),
(3, 'Finance', 'Gill'),(4, 'Purchasing', 'Barber'),
(5, 'Manufacturing', 'Brewer');

Compound Operators – SET @x += 2
Query editor enhancements:
Transact-SQL Debugger – similar to the Visual Studio debuggers
including breakpoints, view call stack and watches.
IntelliSense – word completion, error underlining, parameter help,
colorizing, quick Info and more.
Database Engine Error List Window – displays the syntax and
semantic errors.

Transaction Logging Enhancements
Usually Inserts written to disk twice: once to log +
once to database pages.
Minimally logged INSERTs
In SQL Server 2005 – simple or bulk-logged recovery model only.
In SQL Server 2008 – also in full recovery model.
Performance improvement of large scale INSERT
By reducing the number of log records to be written.
By reducing the amount of log space required.
Performance boost – 3X to 5X.

Spatial Data, Types and Indexes
Spatial data  Info about the physical location and shape of
geometric objects – Point, locations or complex objects such
as countries, roads, or lakes.
SELECT * FROM roads
WHERE roads.geom.Intersects(@ms)=1
Spatial data types:
Geometry – Flat earth data (X/Y coordinates).
Geography – Round earth data (GPS latitude & longitude).
Spatial data objects supported by spatial data types 
Point, MultiPoint, LineString, MultiLineString, Poligon,
MultiPolygon, GeometryCollection.
Visual mapping tool  spatial data results can be
viewed in in Query Editor.
Spatial index  extended index that allows
indexing a spatial column  based on B trees.

Mirroring Enhancements
Automatic page repair  get fresh copy of corrupted pages
from mirroring partner.
Log performance enhancements:
Compress outgoing log stream  at least 12.5 percent
compression ratio.
Write-ahead on the incoming log stream on the mirror server 
incoming log records are written to disk asynchronously.
Reduces manual failover downtime  mirror does not restart to
recover uncommitted log.
Enhanced supportability  includes additional performance
counters & DMVs. For example:
sys.dm_db_mirroring_auto_page_repair - tracks last 100
corrupted pages from mirroring sessions.

Service Broker Enhancements
New Conversation Priorities – Ensures that low priority work
does not block higher priority work.
CREATE BROKER PRIORITY PremierCustomer
FOR CONVERSATION
SET (CONTRACT_NAME = premier,
LOCAL_SERVICE_NAME = sender,
REMOTE_SERVICE_NAME = 'receiver',
PRIORITY_LEVEL = 7);

New Diagnostic Utility – ssbdiagnose utility analyzes the
configuration between two Service Broker services.
New system monitor object and counters.

Star Join Improvements (enterprise edition)
Improved recognition of star join pattern.
While joining a fact table with several dimensions.
Improved costing of star join plans.
Significantly quicker query processing for DWH.
While processing a significant fraction of fact table rows.
Based on bitmap filters
Allows eliminate non-qualifying fact table rows from further processing.
Saves a considerable amount of CPU time.

Integration Services Main Enhancements
Lookup Performance
A lookup tests whether each row in a stream of rows has a
matching row in another dataset (like JOIN operation).
In SQL Server 2005:
Each Lookup component would cache its own copy.
Cache is reloaded every time it’s used.
In SQL Server 2008:
Cache can be shared between multiple lookup components in
the same package or different package.
Cache can be saved to virtual memory or permanent file storage
New miss-cache feature – load to cache key values that have no
matching entries in the reference dataset.

Integration Services Main Enhancements
Data Profiling Task and Data Profile Viewer
The information provided by the profile helps you identify
potential problems with data quality:
The distribution of lengths in the column values.
The percentage of null values.
The distribution of values in the column.

Reporting Services Main Enhancements
Report Authoring
Enhanced Chart Data Region.
New Gauge Data Region.
New Tablix Data Region.
New Report Builder 2.0 offers many new features, like
enhanced data layout, visualization and text formatting, and
on-demand rendering.

Reporting Services Main Enhancements
Report Scalability
Can render much larger reports – renders each page of a
report as you view it.
Can create reports with hundreds or thousands of pages.

Server Scalability
Dependency in IIS removed – includes native support for
HTTP.SYS and ASP.NET.
Can manage its own memory and
memory limits.

Analysis Services Main Enhancements
MDX Query Performance
Improve block computation – No time is wasted evaluating
null cells.
Query and Write back Performance
In SQL Server 2005:
SSAS required write back partitions to have ROLAP storage.
In SQL Server 2008:
Allow write back partitions with MOLAP storage.
Compressed MOLAP format is much faster than querying the
relational data source.
Enhanced Backup
less disk I/O is required.
Scalable Shared Database (SSD)
Can scale out your OLAP query workload across
many small servers with just a single copy of the
database.

Recommended Upgrade Method

Support upgrade from:
SQL Server 2000 SP4.
SQL Server 2005 SP2.
Considerations:
Reporting services – IIS was replaced with HTTP.SYS
Upgrade an existing instance and not databases 
divide and rule.
Existing database retain their compatibility level until
application is ready.

Features by edition
Enterprise Standard
Component Group Feature Description
Edition Edition

Database Scalability & Data Compression Reduce storage requirements
Engine Performance

Backup Compression Less storage required to keep
backups online

Filtered Indexes and Index that support WHERE
Statistics conditions

Resource Governor To define resource limits &
priorities for different workloads

Parallel Query Processing Using multiple processors in
on Partitioned Objects parallel to process query

Star join improvements Better plan + better raw query
execution performance

Performance Data Collecting storing and analyzing
Collector performance data

Logging Enhancements Minimally logged INSERT INTO, in
full recovery model DB

Manageability Policy-Based Management Enables efficient management of
multiple instances from a single
location

Server Group Enables execute of T-SQL queries
Management against multiple servers

SQL Server Extended Can be correlated with Windows
Events events to resolve resources
problems

Features by edition (Part 2)
Enterprise Standard
Edition Edition

Database Manageability Auditing Create & manage database engine
Engine events via DDL

Change Data Capture Data changes are captured and
(CDC) placed in changes tables

Change Tracking (CT) Allow applications to obtain
incremental changes to user table

Transact-SQL Debugger Including break points, data values
watches, call stack and more

Database Engine Error Display syntax & semantic errors
List Window

IntelliSense Word completion, error underlining,
parameter help and more

Security Transparent Database The ability to encrypt the entire
Encryption database files

External Key management Storing keys separate from the data

Programmability MERGE Statement Enable INSERT UPDATE & DELETE
data with a single statement

New Date and Time Data Separate DATE and TIME types,
Types DARETINEOFFSET, DATETIME2

hierarchyid Data Type Enables to model tree structure

Enterprise Standard
Edition Edition

Database Programmability Spatial Data Types GEOMETRY for storing flat-earth
Engine data and GEOGRAPHY for round-
earth data such as GPS

User-Defined Table Type Support table structure as a
parameter for stored procedures
and functions

User Defined Types (UDT) Maximum size of UDT has been
increased to 2 GB

Compound Operators for Example: SET @x += 2

Enhanced CONVERT Allow conversions between binary
& characters hexadecimal values

Grouping Sets Enable defining multiple grouping
in the same query

Table-Valued Parameters Provide easier way to define table
type that can be passed to stored
procedures and functions

FILESTREAM Storage Enables storing of documents and
images on the file system

Sparse Columns Optimizing storage format for NULL
values

Spatial Data Storage Improved spatial indexes.
Results can be viewed in query
editor

Enterprise Standard
Edition Edition

Database Programmability Wide Tables A table that has defined column set.
Engine Up to 8019 bytes size.

Integrated Full Text Catalogs integrated into the
Search database instead of file system

Mirroring Automatic Page Repair Get fresh copy of corrupted pages
from mirroring partner

Improved Performance Compress outgoing log stream

Enhanced Supportability Includes additional performance
counters & DMVs

Service Broker New Conversation Priority Ensures that low priority work does
not block higher priority work

New Diagnostic Utility ssbdiagnose utility analyzes
configuration between two service
broker services

Replication Enhanced Support for Enable execution of SWITCH
Partitioned tables PARTITION on publication database

Integration Components Enhanced Performance Includes faster cache loading and
Services and Caching for the more efficient lookup operations
Lookup Transformation

New ADO.NET ADO NET source component & ADO
Components NET destination component

Enterprise Standard
Edition Edition

Integration Components New data profiling task Help identify data quality problems
Services within individual columns and with
column relationships

New integration services Create package that contains
connections project connection information
wizard

New scrip environment BI Development studio integrates
with MS visual studio for
applications (VSTA)

Data Management Enhanced data type Provides additional info & options
handling in Import and related to date type conversions
Export wizard

New Date and Time data DT_DBTIME2, DT_DBTIMESTAMP2
types DT_DBTIMESTAMPOFFSET

Enhanced SQL statements New merge statement, get changes
to data source, Improved
performance of bulk load

Performance Change data capture Data changes are captured and
Troubleshooting placed in changes tables

New debug dump files Creates dump files(.mdmp & .tmp)
to provide info when package runs

Analysis Designers improvements New Aggregation designer. The
Services (cube, dimension, cube wizard has been simplified
aggregation) and enhanced. New Attribute
Relationship designer.

Enterprise Standard
Edition Edition

Analysis Backup and restore Due improved backup compression
Services improved performance less disk I/O is required.

Reporting Report Authoring Enhanced Chart Data Enhanced Chart Data Region that
Services Region includes a redesigned chart control
that supports many new chart
types.

New Gauge Data Region

New Tablix Data Region New Tablix Data Region for Table,
for Table, Matrix, and List Matrix, and List, which combines
table, list, and matrix structures into
a single structure.

New Report Builder 2.0 New Report Builder 2.0 offers many
new features, like enhanced data
layout, visualization and text
formatting, and on-demand
rendering.

Report New and Enhanced New and Enhanced Rendering
Processing and Rendering Extensions Extensions includes new and
Rendering enhanced rendering extensions for
Word, Excel, and CSV.

New On-Demand Report New On-Demand Report Processing
Processing is a processing enhancement that
renders each page of a report as
you view it.

Enterprise Standard
Edition Edition

Reporting Report Server New Report Server Redesigned server architecture that
Services Architecture and Architecture removes the dependency on IIS and
Tools includes native support for
HTTP.SYS and ASP.NET.

Enhanced Toolset for Support for specific tasks have
Report Server been added to some tools and
Configuration and removed from others.
Management

New Support for Data- New Support for Data-driven
driven Subscriptions and Subscriptions and Job Management
Job Management in in SharePoint Integrated Mode by
SharePoint Integrated using the new subscription
Mode definition pages that are provided
by the SSRS Add-in for SharePoint.

Report Server Report Definition New Report Definition
Programmability Customization Extension Customization extension (RDCE)
that you can use to dynamically
customize a report definition before
it is passed to the processing
engine.

New Methods for the New methods for that endpoint
ReportService2006 Class provide support for data-driven
subscriptions and job management.

Sql Server 2008 Enhancements

More Related Content

What's hot (7)

Similar to Sql Server 2008 Enhancements (20)

Recently uploaded (20)

Sql Server 2008 Enhancements