Jerry Kara Solutions Architect Dallas Technology Group [email protected].
-
Upload
dominick-hicks -
Category
Documents
-
view
215 -
download
1
Transcript of Jerry Kara Solutions Architect Dallas Technology Group [email protected].
Enterprise Applications
Product Family– Contracts, Corporate Performance Management, Customer
Relationship Management, Financials, Human Resources, Interaction Center, Intelligence, Maintenance Management, manufacturing, Marketing, Order Management, Procurement, Product Lifecycle Management, projects, Sales, Service, Supply Chain Planning, Supply Chain Management
Oracle Ebusiness Suiteonline document available
A Brief look at Oracle’s Database History
Oracle 7– Oracle Parallel Server
Oracle 8i– Transition to Real Application Clusters begins
Oracle 9i– Integrated Data Warehousing – Real Application Clusters
Oracle 10g– New grid capable features
Federated Database
A logical unification of distinct databases
Independent servers No shared resources
(including disk) Connected by LAN
Shared Disk Architecture
Equal access to all disks
No notion of data ownership on any node
Managed by Oracle’s Real Application Clusters (RAC)
Shared Disk Architecture: Scalable
Nodes can be dynamically added and removed from the grid
Adding and removing nodes does not require data to be redistributed
Multi-version read consistency model Based upon Real Application Clusters
Multi-Version Read Consistency
Writers never block readers and readers never block writers
Data is tracked by a System Change Number (SCN) Always provides consistent and accurate results Non-escalating row-level locking
– Locks are contained at the row level
Shared Disk Architecture: Availability
The loss of a cluster node does not cause a loss of data
The loss of a cluster node does not cause another node to “take over” data
Processes can be redistributed to other nodes
Traditional Shared-Disk Clustered Databases
Maintaining data coherency is a hard problem
– Need to synchronize updates to shared data– The disk is the only medium for data sharing
Disk I/O latencies appear in the critical path when multiple nodes access shared data
Disk-based coherency is the main bottleneck to achieving a scalable shared disk cluster
– Only synthetic fully partitioned workloads scale!
Oracle Real Application Clusters (RAC)
An application transparent clustered database
– single node applications run and scale with no changes
Cluster interconnect fabric replaces the disk as the medium for inter-node data sharing
Cache Fusion protocol for data sharing results in a scalable cluster for OLTP and DSS workloads
Oracle RACArchitecture
ClusteredDatabase Instances
Mirrored Disk Subsystem
High Speed Switch or Interconnect
Hub or Switch Fabric
Application Servers/ Network
Centralized Management Console
Storage Area Network
Low Latency Interconnect
Drive and Exploit Industry Advances in Clustering
Users
No SinglePoint Of Failure
Shared Cache
What is Cache Fusion?
The underlying technology that enables RAC
Protocol that allows instances to combine their data caches into a shared global cache
– Global Cache Service (GCS) coordinates sharing
Key features are– Direct sharing of volatile buffer caches– Efficient inter-node messaging framework– Fast recovery from node failures using
cache and CPU resources from all surviving nodes
Read Sharing for Queries– query needs to read a data block that is
currently in another instance’s buffer cache.
Write Sharing for Updates– update needs to modify a data block that is
currently in another instance’s buffer cache.
With Cache Fusion, a disk read is performed only if the block is not already in the global shared cache
Data Sharing Problem
Cache Fusion Read Sharing
Uses Oracle’s Consistent Read (CR) scheme
– undo is applied to make a block transactionally consistent to a System Change Number (SCN).
– a CR copy is shipped to the requesting instance
Instance A Instance B
225 Data BlockData BlockQuerySCN 200
1
CR CopyCR Copy200
2
3
Cache Fusion Write Sharing
Multiple dirty copies of a data block can exist in the global cache, but only one is current
The current copy can move between instances without first being written to disk
– Changes are logged if not already on disk
Non-current dirty copies can directly service queries from any node and instance recovery
Cache Fusion Write Sharing
Instance C
1 2
Instance A Instance B
GCS
Current225
UpdateBlock 10
Requester Holder
200200
Copy225
Copy225
3
4
Master
Efficient Inter-Node Messaging
Messaging cost independent of cluster size– At most 3 nodes involved in a request
requester, holder and master (directory) number of messages to service a request is
bounded
Inter-Node Message Latency– exploits high performance interconnect substrates
so that on-the-wire message transmission times are minimal
Frequency of Inter-Node Synchronization– adaptive directory migration based on access
patterns– fast reconfiguration of resources when a node
joins/leaves
Oracle Grid Computing Goals
Half the Cost Highest Quality of Service Easiest to Manage
StorageGrid
DatabaseGrid
ApplicationGrid
GridControl
Problem with Today’s IT Infrastructure
Islands of Computing Resource
– Some are melting down
– Some are almost idle High Costs
– Hardware– Labor– Software
Hard to Align with Business Priorities
ERP
DW
CRM
Solution: Computing On Demand
Partitioning of one large server Built with high quality,
high cost parts Complete, integrated software High quality of service at
high cost
Coordinated use of many small servers
Built with low cost, standard, modular parts
Open, Complete, integrated software
High quality of service at low cost
Mainframe Model Grid Computing Model
Early Grid Computing
Desktop Grids – SETI@home
Scientific Research Grids– Globus Toolkit, CERN
1111
What about Grids for the Enterprise?
Requirements for Enterprise Grids
Run all your applications Stores all your information Highly scalable, available, reliable Secure Easy to manage
– Make individual systems self-managing– Manage thousands of servers at once
Introducing Oracle 10G
10G10G
Runs all your applications Stores all your information Highly scalable, available,
reliable Secure Easy to manage
– Make individual systems self-managing
– Manage thousands of servers at once
Orders of MAGNITUDE…
~354 NEW 10g features * @5 slides / feature * @2 minutes /slide = 4.5 DAYs [ 9 hour/day ]
Oracle Database 10G – Database for the Grid
Storage for All Your Data Real Application Clusters High Availability Ease of Management Data Provisioning across Grid Business Intelligence on the Grid Application Development on the Grid
Oracle Database 10G – Database for the Grid
Storage for All Your Data Real Application Clusters High Availability Ease of Management Data Provisioning across Grid Business Intelligence on the Grid Application Development on the Grid
Oracle Database 10G – Align Storage with the Business
Problem: Islands of storage
– “My storage is 50% utilized and growing 30% a year”
ERP
DW
CRM
Oracle Database 10G – Align Storage with the Business
Problem: Islands of storage– “My storage is 50% utilized
and growing 30% a year”
Solution: Groups of standard, modular disks
– Consolidated,Networked Storage
– Dynamically Provision
Oracle Grid Storage ManagementHighly Available and Low Cost
Oracle Automatic Storage Management
– Manages disk storage groups– Mirrors, Stripes, Re-balances– Huge reduction in labor cost
Built-In Data Protection at Lowest Cost
– Disk Based Backup and Recovery
– Flashback in Time– Site Recovery
Automatic StorageManagement
Disk BasedBackup &Recovery
Automatic Storage Management – Dynamically Provision and Tune Storage
Portable, high performance file system
Eliminates need for conventional file system and volume manager
Automatic mirroring Automatic I/O tuning
– Stripes data across disks to balance load
ASM
The Operational Stack
Disks
Logical Vols
File System
0010 0010 0010 0010 00100010 0010 0010 0010 0010Files
Tablespace
Tables
Disk Group
Logical Vols
File System
Files
Tablespace
Tables
TODAY ASM
“The best way to lower mgmt costs is to remove complexity”
ASM Disk Groups
Disk Group
A pool of disks managed as a logical unit
Partitions total disk space into uniform sized megabyte units
ASM spreads each file evenly across all disks in a disk group
Coarse or fine grain striping based on file type
Disk groups integrated with Oracle Managed Files
ASM Disk Groups
Fast Disk Group
• Administer disk groups not files• Number of disk groups remains
constant while number of files and disks continually increases
Slow Disk Group
Mirroring with ASMFailure Groups
Disk Group 1
Storage resources organized into Disk Groups and Failure Groups
A Disk Group is a collection of disks managed as a logical unit in which ASM files are stored
Disk Groups assigned by DBAs A Failure Group is a set of disks
sharing a common resource whose failure needs to be tolerated
– Redundant copies of an extent are stored in separate Failure Groups
Failure Groups assigned by DBAs or automatically by ASM
Failure Group 1 Failure Group 2
Controller 1 Controller 2
Automatic Storage Management – Lowest Labor Cost
0
1
2
3
4
5
6
7
8
Nb
r S
tep
s
Install Add Storage RemoveStorage
MigrateStorage
Tune I/O ManageSpace
ASM
Traditional
7 76
8
6 6
5
2 2
4
0 0
ASM Delivers Great Performance with no Manual Tuning
0
4000
8000
12000
16000
20000
24000
Tp
mC
ASM (24 Disks) VxVM (24 Disks, 12 s)
22,648
20,590
Automatic Storage Management -Mission Critical, Low Cost Storage for Enterprise Grids
Mission-critical– Integrated mirroring– Add/drop disk online– Best I/O performance
Easy– Automatic I/O Tuning– Up to 50% less DBA/Sys Admin work
Low Cost– Works well with inexpensive disks – Better storage utilization
41
Oracle DB 10G Stores All Your Data
Complete Integrated
Oracle interMediaOracle interMediaMultimedia managementMultimedia management
Oracle LocatorOracle LocatorLocation and Proximity SearchingLocation and Proximity Searching
XML DBXML DBIntegrated Native XML DatabaseIntegrated Native XML Database
Oracle TextOracle TextText management and searchText management and search
RelationalRelationalCharacters, Numbers and DatesCharacters, Numbers and Dates
OCS FilesOCS FilesLow-cost and EasyLow-cost and Easy
Oracle DB 10G ULDB - Stores ALL Your Data
Database size limit raised to millions of Terabytes (8 Exabytes)
Ultra Large Data Files – Terabytes in single file
Unlimited size LOB columns – Terabytes Automatic Storage Management removes
file system limits
Oracle Database 10G – Database for the Grid
Storage for All Your Data Real Application Clusters High Availability Ease of Management Data Provisioning across Grid Business Intelligence on the Grid Application Development on the Grid
Align Processing with the Business
Problem: Islands of computation
– “My servers are 15% utilized“
ERP
DW
CRM
Align Processing with the Business
Problem: Islands of computation
– “My servers are 15% utilized“
Solution: Groups of industry standard servers
– Lowest cost, highest performance
– Capacity on demand– Not Self-healing,
Disposable
Oracle Database 10G RACCapacity on Demand for the Grid
Only Cluster Database that runs Real Applications and Scales
Highly Available Runs on standard,
low cost servers Proven
– Hundreds of customers
– Certified with leading apps
ERP CRM DW
Oracle Database 10G RAC -Complete, Integrated Clusterware
Complete Oracle cluster software solution
Single-vendor support Low Cost
– No need to purchase additional software
– Easy to install, manage High quality and functionality
across all platforms Common event and
management API’s Support for third-party
clusterware
ConnectivityMessaging and Locking
Cluster Control/RecoveryServices Framework
Benefits of Integrated Clusterware
Lower software cost Easy
– Deploys in hours not days– Single integrated target for cluster
management provides improved single system image
Less Risk – Same high functionality and quality across
platforms– Single vendor support
Automatic Server Provisioning
Push-button add/drop server to cluster Hands-free allocation and re-allocation of
servers to services – Automatic routing of service requests to
appropriate server with lowest load– On server failure, automatic re-allocation of
surviving servers to services
Works across all platforms
Oracle RAC 10G - Only Solution for Enterprise Grids
Mission Critical Quality of Service on Industry Standard, Low Cost Servers
Complete, Integrated Oracle Clusterware Runs all Applications Provides Capacity on Demand RAC Made Easy for Everyone Automatic Server Provisioning
Oracle Database 10G – Database for the Grid
Storage for All Your Data Real Application Clusters High Availability Ease of Management Data Provisioning across Grid Business Intelligence on the Grid Application Development on the Grid
Low Cost Fault Tolerance
Computer Computer FailuresFailures
Data Data FailuresFailures
System System ChangesChanges
Data Data ChangesChanges
UnplannedUnplannedDowntimeDowntime
PlannedPlannedDowntimeDowntime
Low Cost Fault Tolerance
Oracle10G RAC Integrated
clusterware– Integrated, less
moving parts, better tested
Faster failover between servers
– Single digit seconds
Computer Computer FailuresFailures
Data Data FailuresFailures
System System ChangesChanges
Data Data ChangesChanges
UnplannedUnplannedDowntimeDowntime
PlannedPlannedDowntimeDowntime
56
New Economics forData Protection & Recovery
Disk-Based RecoveryDisk-Based Recovery
Computer Computer FailuresFailures
Data Data FailuresFailures
System System ChangesChanges
Data Data ChangesChanges
UnplannedUnplannedDowntimeDowntime
PlannedPlannedDowntimeDowntime
Trade cheap disk spacefor expensive downtime
Flashback Log
Flashback Database
New strategy for point-in-time recovery
Flashback Log captures old versions of changed blocks
– Think of it as a continuous backup
– Replay log to restore DB to time– Restores just changed blocks
It’s fast - recovers in minutes, not hours
It’s easy - single command restore
– SQL> Flashback Database to ‘2:05 PM’
Like a “Rewind” button for the Database
Old Block
VersionDisk Write
New Block
Version
Flash Backup
Fully automatic disk-based backup and recovery
– Set and Forget Nightly incremental
backup rolls forward recovery area backup
– Changed blocks are tracked in production DB
– Dramatically faster (20x)
Use low cost ATA disk array for recovery area
Database Area Recovery Area Tape
Nightly Applyof
IncrementalBackup
WeeklyArchive to
Tape
Oracle Database 10G Highest Data Protection for Lowest Cost
Storage FailureProtection
ASM Mirroring Human Error Protection
Flashback Logs
Flashback
Corruption Protection
On DiskBackups
Flash Backup Site FailureProtection
RedoShipping
Data Guard
60
Allow any change to the system with no downtime
Online ReconfigurationOnline Reconfiguration
Goal
No Cost System Changes
Rolling UpgradesRolling Upgrades
Computer Computer FailuresFailures
Data Data FailuresFailures
System System ChangesChanges
Data Data ChangesChanges
UnplannedUnplannedDowntimeDowntime
PlannedPlannedDowntimeDowntime
Oracle10G High Availability Features
Backup and Recovery– Automated Disk-
Based Backup and Recovery
– Drop Database– RMAN Database
Deregistration– Backup of Standby
Control File– Automated TSPITR
Instantiation– Simplified Recovery
Manager Cataloging of Backup Files
– Automated Channel Failover for Backup and Restore
– Automated File Creation During Recovery
– Simplified Backups to Disk
– Proxy Copy Backup of Archivelogs
Backup and Recovery (Cont)
– Incrementally Updated Backups
– Simplified Recovery Through Resetlogs
– Restore Tolerates Corrupt or Missing Backups
– Full Database Begin Backup Command
– Change-Aware Incremental Backups
Data Guard– Data Guard Broker
Support for RAC– Automated LogMiner
Configuration– Log Miner Support for
Index-Organized Tables
– Secured Redo Transmission
Data Guard (Cont)– Fine-Grained
Supplemental Logging
– Named Data Guard Configurations
– Simplified Zero Data Loss for Data Guard SQL Apply
– Zero Downtime Instantiation for SQL Apply
– Real Time Apply– Redo Data and
Archivelog Compression
Flashback– Flashback Database– Flashback
Reinstantiation– Flashback Standby
Database– Flashback Table– Flashback Row
History– Flashback Drop– Flashback
Transaction History Online Operations
– Enhanced Online Redefinition
– Improved Handling of DDL Locks on Busy Tables
– Signature-Based Dependency Tracking Using Synonyms
– Rolling Upgrades with Data Guard SQL Apply
Oracle Database 10G – Database for the Grid
Storage for All Your Data Real Application Clusters High Availability Ease of Management Data Provisioning across Grid Business Intelligence on the Grid Application Development on the Grid
Oracle Database 10G Ease of Management
Self-Managing Database
OEM Grid Control– Grid-wide
operations– Scale from one to
hundreds of systems with little incremental cost
Business Challenge – How to Scale DBA’s to the Grid ?
Software Maintenance 6%
Ongoing System Management 55%
Installation 6%
Others
15%
Creation & Configuration 12 %
Source: IOUG 2001 DBA Survey
DataLoading 6%
Oracle10G System Management – Built-In Intelligent Infrastructure
Automatic Workload Repository
– Automatically collects statistics, SQL workload, Feature usage
Automatic Maintenance Tasks
– Pre-packaged, resource controlled
Server-Generated Alerts– Out-of-the-box, Push vs.
Pull, Just-in-time
Oracle10G System Management – Automatic SGA Management
Only 2 memory parameters
Advisor helpsset parameters
Automatically resizes with changes in workload
Maximize usage of available memory
SGA Pool
PGA Pool
Large Batch Jobs
sort
Java Pool
SQL Cache
Large Pool
Buffer Cache
SQL Cache
sort
Online
Users
Large Pool
Buffer Cache
Java Pool
Oracle10G System Management – Self Optimizing SQL Management
Oracle Query Optimizer – Proven Technology– Comprehensive set of access paths, adaptive search
strategy– All cost-based transformations– Preconfigured task to collect object and system statistics – If no statistics available, are collected at parse-time
Automatic memory allocations, automatic degree of parallelism
Automatic identification and capture of high-load SQL– Automatic SQL Tuning– Automatic SQL Analysis – SQL Access Advisor
Oracle10G System Management – Self Optimizing SQL Management
Automated capture and identification of high-load SQL
Automatic SQL Tuning
Automatic SQL Analysis
SQL Access Advisor
Well-tuned SQL
Custom Applications
SQL Advice
PackagedApplications
SQLProfile
High-load SQLPackaged
Applications
Optimizer
Auto SQL Tuning
Auto SQL Analysis
Access Advisor
CustomApplications
CustomApplications
Indexes &MVs
Oracle10G System Management – Automatic Database Diagnostic Monitor
Performance expert in a box
Automatically provides database-wide performance diagnosis, including RAC
Pinpoints root cause and non-problem areas
Runs proactively
Central management for the complete platform
Dramatically lower operations costs
Application service level management
Manage one or many systems with little incremental cost
Enterprise Manager 10G Grid ControlManagement that Scales to the Grid
Scalable Software Maintenance – Discover and Analyze Configurations
Complete inventory of all Oracle software– Versions– Patch levels
Oracle product configuration details Search and compare configurations Change tracking Reference configurations
Scalable Software Maintenance - Automated Software Cloning
Reduce manual labor in software life-cycle– From hours to minutes
Automate mass provisioning of reference systems
Update
Inventory
Clone to
Selected Targets
2
3
Select Software (and Instances) to Clone
1
Scalable Software Maintenance - Automated Patch Management
Real-time discovery of new patches – Security patch rapid deployment dramatically reduces
vulnerabilities Automatic staging and application
– From hours to minutes Rolling RAC upgrade
Update
Inventory
Determine
Applicability2
Apply Patch3
4Patch Published1
Oracle10G Manageability Features
Infrastructure– Easy OracleJVM
Monitoring– SYSAUX Tablespace– Enhanced Wait Model– Database Features Usage
Tracking– Enhanced Database Time
Model– Top Consumers
Performance Monitoring & Diagnostic
– JDBC: End-to-end Java/J2EE Tracing
– SQLJ: End-to-end Java/J2EE Tracing
– Automated Routine Administration Tasks
– Proactive Server-based Alerts and Notifications
– Database Statistics and Workload Respository
– Threshold-Based Alerts
Server Configuration– Automated Storage
Management Configuration
– Automated RAC Services Configuration
– Simplified Upgrade for RAC and OPS Databases
– Automated Enterprise Manager Configuration
– Automated Portable Clusterware Installation
– Automated Configuration of Recovery Area
– Out-of-the-box LDAP Configuration
– Simplified Initialization Parameters
– Easy Upgrade– Upgrade Information Tool– Simplified Database
Install– User Default Tablespace
Instance Tuning– User-Initiated Buffer Cache
Flushing– Database Resource Manager
- Adaptive Consumer Group Mapping
– Database Resource Manager - Fixed CPU Quota
– New Performance Overview Charts in Oracle Enterprise Manager
– Improved SQL Reporting Using Oracle Enterprise Manager
– Integration of Resource Manager and Profiles
– Automated Checkpoint Tuning
– Transaction Rollback & Recovery Monitoring
– Database Performance Analyzer
– Redo Logfile Sizing Advisor– Automated SGA Memory
Tuning
Application Tuning– SQLAccess Advisor– Materialized View Tuning
API– SQL Tuning Advisor– Enhanced SQLAccess
Advisor– Automatic Optimizer
Statistics Collection Backup and Recovery
– Enhanced RMAN Reporting
– Backup Compression– Bounded Backup Window– Manage RMAN Scripts– Recovery Area Full Alert
Storage Management – Multiple Default Temporary
Tablespace Support for SQL Operations
– Automated Storage Management
– Rename Tablespace
Oracle10G Manageability Features(continued)
Network Management – Back Up Directory Naming
Entries to Local Naming File
– Dynamic Connection Manager Configuration
– Easy Connect Naming Method
– Easy OCI/JDBC Install– Improved Network Outage
Detection– Automated LDAP
Discovery for Clients– Improved Connection
Manager Access Rules– Automated Shared Server
Configuration– Simplified Shared Server
Configuration Parameters
Space, Object, Transaction Mgmt
– Automatic Undo Retention Tuning
– Segment Shrink Advisor– Segment Size Estimation– Online Segment Shrink– Proactive Tablespace
Management– Undo Advisor
Database Utilities– SQL*Plus - SPOOL
CREATE, REPLACE, and APPEND Options
– SQL*Plus - Show Recycle Bin
– SQL*Plus - DBMS_OUTPUT After SELECT
– SQL*Plus - SET SQLPROMPT Runtime Variable Substitution
– SQL*Plus -COMPATIBILITY Command Line Option
– SQL*Plus - Run glogin.sql and login.sql After Each CONNECT
– iSQL*Plus Input Prompting
Scheduler– Core Scheduler Features– Usability and
Manageability– Monitoring– Recovery– Support for Jobs in a
Clustered Environment– Job Processing– ILMS Support– Increased Throughput– Redo Minimization– Dynamic Slave Pool
Oracle Database 10G – Database for the Grid
Storage for All Your Data Real Application Clusters High Availability Ease of Management Data Provisioning across Grid Business Intelligence on the Grid Application Development on the Grid
Transportable Tablespaces –Fastest Way to Provision Data
Oracle can unplug and plug database data files
– Move and share data without unload and reload
– Plug/unplug performance independent of tablespace size
– Mount read-only tablespaces on multiple databases
New: Now Works Cross-Platform!
2424
Oracle10G Data Pump
High performance import and export – 60% faster than 9i export (single thread) – 15x-20x faster than 9i import (single thread)
Scales with parallel threads Network transfer data between databases
with no intermediate representation
Oracle Streams Oracle
Oracle Streams moves or copiestablespaces faster than FTP
Same platform or cross platform
Detach, move, attachwith one command
Oracle Streams Oracle
Oracle Streams moves or copiestablespaces faster than FTP
Same platform or cross platform
Detach, move, attachwith one command
Oracle Streams Oracle
Oracle Streams moves or copiestablespaces faster than FTP
Same platform or cross platform
Detach, move, attachwith one command
Optionally stream delta updates
Migrate to the Grid withthe self-propelled database
Oracle Database 10G– Database for the Grid
Storage for All Your Data Real Application Clusters High Availability Ease of Management Data Provisioning across Grid Business Intelligence on the Grid Application Development on the Grid
Business Intelligence on the Grid
Lower Cost– Leverage 10G hardware, software, labor savings– Manage millions of terabytes
Intelligence when you need it– Real-time data feeds and integrated BI capabilities
More value from your data– Additional integrated BI capabilities
Managing Large Volumes of Data
Partitioning and parallelism are crucial for VLDB Partitioning provides ‘incremental’ operations for:
– WORK SMARTER– Data loading, Indexing, Referential Integrity– Backup and Recovery– Queries
Parallelism for all operations– WORK HARDER– DBA operations: loading, index-creation, table-creation, data-
modification, backup and recovery– End-user operations: Queries– Unbounded scalability: Real Application Clusters
Partitioning - Minimal Unit of Work
Intelligent usage of partitioning information is key success factor for availability
– Partition Pruning– Backup and Recovery– Read-Write versus Read-Only
order table
may99jan99 feb99 mar99 apr99
Partitioning - Rolling Window Partition level management
– Range, Composite (range + hash), List– On-line addition and removal of partitions– Data management operations (loading, index builds)
Improved availability– Localized disk failures, backup and recovery
order table
Drop
may99
Add
jan99 feb99 mar99 apr99
Local Index
Other data is not affected
Table Compression
Table Compression increases:– I/O-subsystem capacity– I/O throughput– query scan performance– buffer cache capacity
Table Compression:– reduces cost of ownership– is easy to use– requires little or no table definition changes– is transparent to applications
Table Compression: What is it?
Tables can be compressed – Compression can also be specified at the partition level– Indexes and index-organized tables are not compressed
Typical compression ratios are 3:1 - 6:1– Compression is dependent upon the actual data– Compression algorithm based on removing data redundancy
All DDL/DML commands are supported on compressed tables
Data is Partitioned into Granules (block range or partition)
Each Scanner is Assigned Multiple Granules
No two scanners ever contend for the same granule
Granules are assigned so that the Load is Balanced Across all Scanners
Dynamic Granules
Scanner #1
Scanner #2
Scanner #3
Oracle Parallel Query - Scanning
. . .
10G Real Time WarehouseAsynchronous CDC
SQL interface to change data Publish/subscribe paradigm Parallel access to log files, leveraging
Oracle Streams Parallel transformation of data
OLTPDB
Archivedlogfiles
Logical Change DataBased on
Streams
DWTables
SQL, PL/SQL,Java
Transform
Oracle10G Business Intelligence Features
Real Time Data Warehousing
– Asynchronous Change Data Capture
– Cross-Platform Transportable Tablespaces
– Enhanced Table Functions– External Tables Unload– Enhanced MERGE
Functionality– Data Pump Export and
Import Utilities– Parallel Data Pump Export
and Import– SQL*Loader Direct Path
Load Support for ROWID Datatype
– SQL*Loader Direct Path Load Support for VARRAY Datatype
– SQL*Loader Direct Path Load Support for XMLType Tables
VLDB– Bigfile Tablespace– Enhanced Partition
Management in Enterprise Manager
– Global Partitioned Indexes - Hash Partitioning
– Partitioned Index-Organized Tables - List Partitioning
– Partitioned Index-Organized Tables - LOB Support
– Partitioned Index-Organized Tables - Global Index Maintenance
– Enhanced Bitmap Index Performance and Space Management
– Local Partitioned Indexes Manageability Improvements
Summary Management– Summary Management
OLAP Query Performance Improvements
– Summary Management - Enhanced Partition-aware Materialized View Refresh
– Summary Management - EXPLAIN PLAN Shows Materialized View Access
– Summary Management - Enhanced Dimensions
– Summary Management - Nested Materialized View Refresh
– Summary Management - Enhanced PCT Refresh
– Summary Management - Query Rewrite Uses Multiple Materialized
OLAP – Parallel AGGREGATE
Command– Parallel SQL IMPORT
Command– PL/SQL Creation of
Analytic Workspaces– Intra-Dimensional
(Attribute) Aggregation– Enhanced Composite
Dimensions Indexing– PL/SQL OLAP Measure
Calculation Definition– Complete Analytical
Workspace Management– Multi-Writer Support– XML Interface for Analytic
Workspace
Oracle10G Business Intelligence Features (Continued)
Data Mining – Frequent Itemsets– Attribute Importance
Algorithms– Enhanced Adaptive Bayes
Network– JSR-73 RI & TCK - Java
Data Mining– Multi-User Access Control– Complete Oracle Data
Mining Management– Enhanced Model Seeker– PMML Producers /
Consumers– Enhanced Data
Preprocessing– Data Mining Rapid
Application Development
Bio-informatics– DOUBLE and FLOAT
Datatypes– BLAST Queries– Statistical Functions– Document Clustering– Document Clustering Non-
negative Matrix Factorization
– Document Clustering Neural Nets
– Document Named Entity Extraction
– Support Vector Machines– Taxonomy Builder
SQL Analytics– Grouped Table Outer Join– Increased Number of
Aggregates per Query– SQL Interrow Calculations– Upsert Through SQL
Interrow Calculations
Oracle Database 10G – Database for the Grid
Storage for All Your Data Real Application Clusters High Availability Ease of Management Data Provisioning across Grid Business Intelligence on the Grid Application Development on the Grid
Application Development on the Grid
HTML DB– Low Cost, Hosted,
Web Based Development Environment
Improved PL/SQL Performance
Database Web Services
HTML DB and the Grid - Business Problem
Departmental and LOB data sharing and information tracking
Lots and lots of desktop databases
Fragmented Inaccurate Untimely Vulnerable
HTML DB – Perfect for the Grid
Productivity of a Desktop Database
Infrastructure of an Enterprise Database
Built for the Web Leverages the Grid
Oracle Database 10G – Database for the Grid
Storage for All Your Data Real Application Clusters High Availability Ease of Management Data Provisioning across Grid Business Intelligence on the Grid Application Development on the Grid
Oracle Database 10G
Half the Cost– Low cost hardware with high utilization– Integrated software stack– Self-managing
Highest Quality of Service– Highest Performance, Scalability,
Availability, Reliability, Security
Easy to Align with Business Priorities
Oracle Database 10g Performance Metrics
PLSQL up to 2x Table Scans up to 2x SQL Profile Tuning
huge Floating point math up
to 10x Bulk LOB Update up to
3x Import up to 15x Export up to 35%
Incremental backup up 10x Parallel media recovery
2-4x Backup compression 17x Infiniband interconnect
½ latency, 2x block transfer
TPC-C – Best absolute and best on 4-
way Intel– +28% on same hardware