Better, Faster, Cheaper: Optimizing Application Performance and Availability

1
Better, Faster, Cheaper Optimizing Application
Performance and Availability

• Part I Infrastructure

2
Agenda

• Business Drivers and Pain Points
• Oracle Solution
• Oracle Virtual Machine
• Clustering
• ASM
• RAC
• Coherence
• Partitioning
• Active Data Guard
• Advanced Compression
• Summary/Contact Info

3
IT Challenges Quality of Service

• Data centers are out of power, space and cooling
• People are expensive and skills are hard to find
• IT pressured to do more with less
• Much of the infrastructure is underutilized
• IT challenged to keep pace with rapid business
  change

4
Infrastructure Trends

• Infrastructure trends are shaping ITs response
  to these challenges
• Increasingly powerful low-cost commodity servers
• Virtualization
• Server Virtualization
• Virtual Machines
• Server Partitioning
• Server Pooling
• Scaling workloads across multiple servers

5
2009 CIO Deployment Priorities
Source Morgan Stanley CIO Survey, 12/18/2008
6
VirtualizationTwo Approaches
Server Pooling Aggregating many physical
servers to appear like a single logical server
Server Virtualization Disaggregating a single
physical server into multiple logical servers
(VMs) or partitions
Virtual Machines
Virtualization Layer
Hardware Platform
7
Server Virtualization Technologies

• Can provide electrical as well
• as fault, OS, and resource
• isolation
• Available on most Risc
• platforms
• Evolution from static to
• dynamic resource allocation.
• Supports both dedicated and
• shared resource models
• Runs at native speed

• Can provide fault, OS, and
• resource isolation
• Mostly focused on x86
• platforms
• Fully dynamic environment
• can support cross server
• resource sharing.
• Based on a shared resource
• model, can support pinning
• of resources
• Operations overhead can be
• significant.

• Can provide process level
• resource isolation as well as
• some fault isolation
• Extension of some operating
• systems
• Enables efficient resources
• sharing for OS environments at
• the same level.
• Shared resource model
• Requires all environments to be
• at same OS level.
• Minimal overhead.

8
Overhead varies widely

• Platform
• CPU generation
• Server platform
• Virtualization technique
• Application
• CPU intensive
• Context switching
• Memory usage
• I/O access
• Workloads
• Variation in usage
• Resource over-commitment
• CPU
• Memory
• I/O

DB on Oracle VM (Linux)
Nirvana
DB on VMware
Acceleration
Overhead
-10
-95
10
Memory Intensive
I/O Intensive
CPU Intensive
Memory, CPU, I/O Over-commitment
9
Introducing Oracle High Performance and
Availability Infrastructure
10
Oracle VM

• Free product based on Xen 3.1
• Oracle tested and supported server virtualization
  technology
• Maximizes consolidation of Linux and Windows
  servers, saves on power, cooling and space
• Virtual Machine templates for automated deployment

• Live migration included at no additional cost
• Integrated, browser-based management console
• Downloadable pre-built images for Oracle products
• Enterprise-quality support at low annual cost

11
Oracle Product Certification with Oracle VM

• Oracle Database
• Oracle Application Server
• Oracle Enterprise Manager
• Oracle Berkeley DB
• Oracle TimesTen
• Oracle E-Business Suite
• Oracle PeopleSoft
• Oracle Siebel
• Oracle Hyperion
• More information on Metalink Note 464754.1

12
VirtualizationTwo Approaches
Server Pooling Aggregating many physical
servers to appear like a single logical server
Server Virtualization Disaggregating a single
physical server into multiple logical servers
(VMs) or partitions
Virtual Machines
Virtualization Layer
Hardware Platform
13
Evolution of Server Pooling Standalone SMP to
Grid Computing
RACClustersforAvailability
14
Cluster TechnologyFoundation for Server Pooling
Cluster Benefits Workload Balancing Workload
Failover Optimal Capacity Planning Lower
Hardware Costs Increased I/O Throughput Reduced
administrative costs and downtime (e.g., rolling
upgrades)
15
ASM Storage Grid

• Oracle Automatic Storage Manager (ASM)
• Provisions storage capacity automatically to
  Oracle 10g as needed
• Stripes and Balances I/O
• Mirrors Immune to disk failure
• Oracle Automatic Backup and Recovery
• Single backup areafor all Grid databases
• Archive to tape

16
RAC The Database Grid
Network
Users
Centralized Management Console
Interconnect
No Single Point Of Failure
High Speed Switch or Interconnect
ClusteredDatabase Servers
Shared Cache
Hub or Switch Fabric
Storage Area Network
Drive and Exploit Industry Advances in Clustering
Mirrored Disk Subsystem
17
Before Clusters - Higher Costs

• Poor Resource Utilization
• Built for peak periods
• Gartner estimates average server utilization rate
  at 5-10!
• Standby hardware and software costs virtually
  double the investment and further reduce useful
  utilization
• Management happens in silos
• Uneven process maturity across managed silos
• Availability, security, performance
• Increased staff
• Proliferation of tools that have overlapping
  capabilities
• Software patching/testing, upgrade tasks are
  multiplied
• Increases information complexity and lowers
  agility
• More data movement required (i.e. increased
  latency, increased storage, increased integrity
  issues)

18
Proven for Production 8 RAC Node Customers

• SAIC
• Fairmont Hotels
• ADESLAS
• Evite.com
• Quelle AG
• Telstra
• Gas Natural
• MyTravel
• Thomson
• AOL
• Vivo
• Sagawa Kyubin

• Citigroup
• Burlington Coat Factory
• J2 Global Communications
• Genworth Financial
• Amazon.com
• MSDS
• Mercado Libre
• Yahoo! Overture
• Babcock Engineering
• Ordnance Survey
• Dell
• Yahoo!

19
Rolling Patch Upgrades With RAC
20
Oracle Clustering - Not Just RAC!
Web Tier
Application Tier
Database Tier
In-Memory Cache
Web Cache
Web Servers
Application Servers
Coherence Data Grid
RAC
HTML Data Structures in Memory
Java Data Structures in Memory
SQL Data Structures in Memory
Offload Web Servers, Improve Network Performance
via Compression
Cache Java Structures in Memory Very Fast Access
to Java Data in Memory across Mid-Tier Grid
Provide Scalability to Database Data improving
Query Transaction Write Performance
21
Oracle Coherence Data Grid ServiceDistributed,
In Memory
Distributed In-Memory Data Management Provides a
reliable data tier with a single, consistent view
of data Enables dynamic data capacity including
fault tolerance and load balancing Ensures that
data capacity scales with processing capacity
22
Summary Virtualization Server
PoolingComprehensive Perf/HA Architecture

• Optimize within a server and across servers
• Server Virtualization benefits
• Easy consolidation of underutilized servers
• Reduced floor space, power, and cooling
  requirements
• Easy provisioning for test and dev environments
• Server Pooling benefits
• True business continuity
• Scalability across servers
• Higher performance (high overhead with VMs)
• A reduction in management burden (app/db
  consolidation)
• Server virtualization best for small workloads,
  test, dev, and non-critical applications
• Server pooling (RAC) best for mid-to-large and
  business critical applications

23
Oracle Partitioning A History of Innovation
24
New in 11g - Interval Partitioning

• Partitioning is key-enabling functionality for
  managing large volumes of data
• One logical object for application transparency
• Multiple physical segments for Administration
• Improves Manageability, Availability, and
  Performance

• BUT
• Physical segmentation requires additional data
  management overhead
• E.g. new partitions must be created on-time for
  new data

Solution Automate partition management
25
New in 11g - REF Partitioning
Table ORDERS
...
...
Jan 2006
Feb 2006

• PARTITION BY REFERENCE
• Partitioning key inherited through PK-FK
  relationship

• Related tables benefit from same partitioning
  strategy
• Eliminates data and maintenance overhead
• Intuitive modelling
• Enhanced Performance and Manageability

Table LINEITEMS
...
...
Jan 2006
Feb 2006
26
Partitioning-based Rolling Window Operations
Order Table (partitioned by quarter)
Q107
27
Unlocking the Value of Standby DBs
Standbyfor OnlineUpgrade,Auto Failover
Standbyfor Testing,ReadablePhysical
Standbyfor DRand Backup
Logical Standbyfor RealtimeQuery
28
Disaster Recovery ChallengeInvestment in
Disaster Recovery only

• Applications, backups, reports run on production
  only

Real-time Queries
Standby Database
Production Database
29
With Oracle Active Data GuardOffload production
reporting to standby

• Simultaneously available in read and recovery
  mode

Real-time Queries
Standby Database
Production Database
30
Active Data Guard Real-time Query
Concurrent Real-Time Query
Continuous Redo Shipment and Apply
Primary Database
Physical Standby Database

• Read-only queries on physical standby concurrent
  with redo apply
• Supports RAC on primary and/or standby
• Queries see transactionally consistent results
• Handles all data types, but not as flexible as
  logical standby

31
With Oracle Active Data GuardOffload database
backups to standby

• Complete database and fast incremental backups

Standby Database
Production Database
32
With Oracle Data GuardTest changes
Production Database
Standby Database
33
Rolling Release Upgrades w/Data Guard
34
Introducing Advanced Compression

• Advanced compression in Oracle Database 11g
• Structured data compression
• Unstructured data compression
• Compression for backup data
• Network transport compression
• Reduces resource requirements and costs
• Storage System
• Network Bandwidth
• Memory Usage

35
Advanced CompressionNew in Oracle Database 11g

• Extends table compression for OLTP data
• Support for conventional DML Operations (INSERT
  and UPDATE)
• Block level compression dictionary is dynamic,
  making it adaptive to frequent data changes in
  OLTP environments
• New algorithm significantly reduces write
  overhead
• Batched compression ensures no impact for most
  transactions
• Makes compression feasible for OLTP systems as
  performance is not compromised for DML operations
• Available with new Advanced Compression option
• Table Compression for bulk load operations
  continues to be available as a feature at no
  extra charge

36
Real World Compression Results10 Largest ERP
Database Tables
Data Storage
MB

• 3x Savings

37
Unstructured Data CompressionOracle Database 11g
- SecureFiles

• SecureFiles is a new database feature designed to
  break the performance barrier keeping file data
  out of databases
• Similar to LOBs but much faster, and with more
  capabilities
• Transparent encryption, compression,
  deduplication, etc.
• Preserves the security, reliability, and
  scalability of database
• Superset of LOB interfaces allows easy migration
  from LOBs
• Enables consolidation of file and relational data
• Single security model
• Single view of data
• Single management of data

38
SecureFiles Compression

• 2-3x compression for typical files (doc, pdf,
  xml)
• Industry standard compression algorithms
• Can be specified at a partition level
• Automatically detects if SecureFile data is
  compressible
• Skips compression for already compressed data
• Auto-turn off compression when space savings are
  minimal or zero
• Two levels of compression provide different
  compression ratios
• Compression Levels MEDIUM (default), HIGH
• Higher the degree, higher the latency and CPU
  overhead incurred
• Compression is independent of table or index
  compression
• Part of Advanced Compression option

39
SecureFiles Deduplication

• Enables storage of a single physical image for
  duplicate data
• Significantly reduces space consumption
• Dramatically improves writes and copy operations
• No adverse impact on read operations
• Duplicate detection happens within a table,
  partition or sub-partition
• Ideal for content management, email and data
  archival applications
• Part of the Advanced Compression Option

40
Why Compress Files Inside Database?Several
advantages over storing zipped files

• Enables reads and writes to compressed data
• without decompressing the complete file
• Indexing and search can be done on compressed
  data
• Does not require manual intervention to zip and
  unzip files
• Compression is completely transparent to the
  application
• De-duplication is not feasible outside the
  database

41
Advanced Compression Summary

• Compress Large Application Tables
• Transaction processing, data warehousing
• Compress All Data Types
• Structured, unstructured, backup data
• Typical Compression of 2-4X
• Cascade storage savings throughout data center

42
TUSC Trusted Oracle Expertise Across Techology
and Applications
Fusion Middleware
Information Age Applications
Database and Grid Computing

• Oracle E-Business Suite
• PeopleSoft Enterprise
• Siebel CRM
• JD Edwards EnterpriseOne
• JD Edwards World
• Oracle Retail
• i-flex
• Communications Billing
• ProfitLogic
• G-Log

• Application Server
• Integration / SOA
• Hot-Pluggable
• Business Intelligence
• Identity Management
• Data Hubs
• Collaboration Services
• Process Orchestration
• Java Development Tools

• Database
• Real Application Clusters (RAC)
• Enterprise Manager
• Partitioning
• OLAP
• Security
• Lite
• Times Ten

43
Contact Us

• West Brian Decker, deckerb_at_tusc.com, (626)
  836-9574
• South/Central Lisa DiNitto, dinittol_at_tusc.com,
  [phone]
• East/Central Mike Margulies, mjm_at_tusc.com, (203)
  293-4422
• For additional information and consultation
• Oracle Investment Value Analysis
• Review of existing Oracle topology and
  architecture, including deployment growth and
  capacity analysis
• Review of existing Oracle licenses ownership and
  license surplus/exposure analysis
• License optimization recommendations, including
  leveraging maximum available discounts and
  financing options
• Solutions Requirements Assessments
• Performance/HA Architecture healthcheck and
  high-level roadmap
• Quickstart options