Optimize Oracle Applications Performance while Lowering Costs Presenters- Kevin Barry

Download Report

Transcript Optimize Oracle Applications Performance while Lowering Costs Presenters- Kevin Barry

Optimize Oracle Applications
Performance while Lowering Costs
An Agilent Case Study
PresentersKevin Barry
Kevin O’Malley
OuterBay Technologies, Inc.
Agenda
• Introduction to Agilent Corporation
• The Data Growth Problem
• OuterBay ADM Suite
• The Agilent Solution
• Results & Benefits
•Q &A
OuterBay Confidential
About Agilent
• Global technology leader in communications,
electronics, life sciences and chemical analysis
• Revenue
- $7 Billion in Revenues
• Employees
- 29,000 World Wide
• Locations
- Headquartered in Palo Alto, CA
- 30 Facilities worldwide
• Primary businesses
- Test and Measurement
- Automated Test
- Semiconductor Products
- Life Sciences and Chemical Analysis
OuterBay Confidential
Agilent’s Oracle E-Business Environment
•
Oracle E-Business Suite 11i
•
•
•
•
•
•
•
Modules: Entire ERP Suite
Single global instance
Consistent, real-time view across all business units
Common business processes
HP Superdome (64 CPUs)
Go-live in July 2002
Application OLTP production Data Growth
• 92 GB/month
• OLTP instance copied ~ 26 times for
test/development
• Growth expected to increase with additional plants
coming online
OuterBay Confidential
Agenda
• Introduction to Agilent
• The Data Growth Problem
• OuterBay ADM Suite
• The Agilent Solution
• Results & Benefits
•Q &A
OuterBay Confidential
The Problem
Performance issues in production
 Strong relationship between DB size and OLTP / Batch
performance
Storage capacity required was explosive
Data growth was more than Oracle predicted
Total disk capacity requirements to exceed 26 TB in one
year
Business continuity challenges
Increase in backup and recovery times
Higher DR/HA costs
OuterBay Confidential
Data Growth vs. Performance
1300
Data Growth Impact
1200
1200
Application Data
1100
1100
1000
900
25,000
800
700
600
15,000
700
600
0
-1
9
04
-0
20
8
04
-0
7
-0
04
-0
6
400
5
400
04
-0
800
500
20
4
04
-0
20
3
04
-0
20
2
04
-0
OuterBay Confidential
20
1
04
-0
20
2
04
-1
20
1
03
20
-1
0
03
-1
20
9
03
-0
20
8
03
-0
20
7
03
-0
20
03
20
20
03
-0
6
10,000
900
500
20
Batch Run Time
1000
20
20,000
04
30,000
20
35,000
Application Data (GB)
1300
Application Data (GB)
Total Batch Run Time (Hrs/Month)
40,000
Data Growth Alternatives
• Ask Oracle for assistance
Services engagement
• Custom/in-house solution
Major project – year+ to deploy
• Archive vendor search
Selected approach
OuterBay Confidential
Agilent - Key Requirements
No disruption to the business
 Uncompromised end-user reporting
 No change to business processes
 Guaranteed data and transaction integrity
 Aggressive Data Retention Policies
 True 24 x 7 Operations – Zero downtime allowed
 Automation
 Oracle Certification
OuterBay Confidential
Agilent Selects OuterBay
• Suite of proven solutions
• Highly automated
• Oracle certification
• Online/transaction integrity
• Experience
• Ready to Implement
OuterBay Confidential
Agenda
• Introduction to Agilent
• The Data Growth Problem
• OuterBay ADM Suite
• The Agilent Solution
• Results & Benefits
•Q &A
OuterBay Confidential
Application Data Management
Support & Projects
Production
Integration Test
Disaster Recovery
Unit Test
Local Standby
Development
Disk Backup
Archive
Reporting
X(n)
Production
Parallel Initiatives
Application Resource Monitor
Policy & Configuration Console
LiveArchive
Encapsulated
Archive
Developers Edition
OuterBay Confidential
Instance Generator
Data Life Cycle Example
Data Retention Policies
6 months-2 Years
7 – 25+ Years
2-7 Years
Application Transparency
3rd Party Reporting Tools
Encapsulated Archive
.XSD
.XML
Production
History
• Active OLTP
• Online
• Sustained growth & perf.
• Historical transactions
• Online
• Application dependent
• Archive flat file
• XML Query access
• Open / Independent
Storage Class
High Cost
High Availability
Secondary
OuterBay Confidential
Least Expensive
Applications Resource Monitor
• Discover areas of
high and low data
growth
• Decision Support
tool for data
retention policies
• Use across all
instances and
database
applications
DatabaseXtender
OuterBay Confidential
Data Growth Analysis
TTE
125
Relocation
100
Relocation
Relocation
Relocation
Space 75
(GB) 50
25
0
Jan May Sep Jan May Sep
2002 2002 2002 2003 2003 2003
Jan May Sep
2004 2004 2004
• Identify active vs. inactive data
• Build data relocation into the business
process
• Implement based on predictable
storage requirements
OuterBay Confidential
Jan May Sep
2005 2005 2005
Eligible
Eligible Forecast
Ineligible
Ineligible Forecast
Key Features
• Combined View Reporting – native reports/queries
• Data Growth and Process Monitoring
• Data Parity Support
• Database Reorganization Support
• Online Operations (users stay on the system)
• Concurrent Manager integration (or 3rd party
scheduling tools)
• Fully Recoverable/Restartable
• Full audit trail – repository-based
• Reload Support
OuterBay Confidential
OuterBay Platform
LiveArchive
Encapsulated
Archive
Active
Repository
Application Resource
Monitor
Instance
Generator
Developer
Edition
OuterBay Confidential
Agenda
• Introduction to Agilent
• The Data Growth Problem
• OuterBay ADM Suite
• The Agilent Solution
• Results & Benefits
•Q &A
OuterBay Confidential
Agilent Approach
Two-tiered Solution
• Test and Development Environments
- OuterBay Instance Generator: Relationally intact subsetting
- 11 Modules implemented
» INV, COST, WIP, AP/PO, GL, AR, Quotes, Workflow, Cash
Mgmt, Order Mgmt, Cycle Count
• Production Environment
- OuterBay LiveArchive: Online Archiving
- 9 Modules implemented
» INV, COST, WIP, AP/PO, GL, Quotes, Workflow, Cash Mgmt,
Supplier Schedules
OuterBay Confidential
Agilent Approach – Instance Generator
Test and Development Environments
Patch
• Policies selected by time
Policy Sets
Production
OLTP Copy
• Create fully functional subset
databases for development,
test, UAT, training, and demo
Training
Dev
OuterBay Confidential
Agilent Approach - LiveArchive
Best Practices in Data Retention Policies
Archive Module
Data Retention in Production
Archiving
Frequency
Inventory
2 Months + Current Month
Monthly
Cost History
2 Months + Current Month
Monthly
WIP
2 Months + Current Month
Monthly
Supplier Schedules
2 Months + Current Month
Monthly
AP/Procurement
2-3 Months after PO gets finally closed
Monthly
Cash Management
6 Months + Current Month
Monthly
GL
12+2 Months of adjustment period, after a
period is permanently closed
Monthly
Quote
Expired Quotes, older than 9 months will get
deleted
Monthly
Workflow
Delete workflows older than 15 days as part
of regular Oracle workflow purge
Monthly
OuterBay Confidential
Agenda
• Introduction to Agilent
• The Data Growth Problem
• OuterBay ADM Suite
• The Agilent Solution
• Results & Benefits
•Q &A
OuterBay Confidential
The Results
Delivered by Instance Generator
• $ Savings for Storage
- Subsets for Test, Dev, and Training = 30% of Production
» 500 GB from 1.8 TB
- Current usage
» Patch, Training
- Planned usage
» 11.5.10 Upgrade testing: 10+ copies
• IT Process Streamlined
- 17% reduction in subset creation times
OuterBay Confidential
The Results
Expected Benefits
• Defer investments in hardware upgrades
• Decrease TCO for E-Business Suite
• Minimal user impact/end user training
OuterBay Confidential
The Results
Delivered by OuterBay LiveArchive
• $ Savings for Storage: 2.5% of total 2005 IT
budget
-
Number of rows archived = ~ 500 Million
500 GB in storage savings in Production
Total storage savings = 13 TB
Agilent can now defer disk spending for 1 year
• Related Savings
- HA, Backup & Recovery
OuterBay Confidential
The Results
Unexpected Benefits
• Data growth management
 stable, predictable performance management
- Take guesswork out of capacity planning
- Free up resources for business initiatives
- Accelerate critical business processes
OuterBay Confidential
Performance Benefits
1300
1200
35,000
Data Growth
Impact
1100
30,000
1000
900
25,000
Phase II
Tuning
800
Sustained
Predictable
Performance
20,000
15,000
600
500
LiveArchive
OuterBay Confidential
0
-1
9
04
-0
20
8
04
-0
20
7
04
-0
20
6
04
-0
20
5
04
-0
20
4
04
-0
20
3
04
-0
20
2
04
-0
20
1
04
-0
20
2
04
-1
20
1
03
-1
20
03
-1
20
9
03
-0
20
8
03
-0
20
7
03
-0
20
03
-0
20
03
Phase I
Tuning
0
400
6
10,000
20
700
Application Data (GB)
Total Batch Run Time (Hrs/Month)
40,000
Business Benefits Example: Shipping
2,000
1,800
1,600
1,400
1,200
1,000
800
600
400
200
-
12
10
8
Sustained
Predictable
6
Performance
4
2
20
03
-0
4
20
03
-0
6
20
03
-0
8
20
03
-1
0
20
03
-1
2
20
04
-0
2
20
04
-0
4
20
04
-0
6
20
04
-0
8
20
04
-1
0
0
OuterBay Confidential
Average Run time (Min)
Total Batch Run Time (Hrs/Month)
Pick Selection List Generation
Business Benefits Example: OM
1,600
6
1,400
5
1,200
4
Sustained
Predictable3
Performance
1,000
800
600
2
400
1
200
0
20
03
-0
4
20
03
-0
6
20
03
-0
8
20
03
-1
0
20
03
-1
2
20
04
-0
2
20
04
-0
4
20
04
-0
6
20
04
-0
8
20
04
-1
0
-
OuterBay Confidential
Average Run time (Min)
Total Batch Run Time (Hrs/Month)
Order Import
Business Benefits Example: AP
700
10
9
8
7
Sustained
6
Predictable
5
Performance
4
3
2
1
0
600
500
400
300
200
100
20
03
-0
5
20
03
-0
7
20
03
-0
9
20
03
-1
1
20
04
-0
1
20
04
-0
3
20
04
-0
5
20
04
-0
7
20
04
-0
9
20
04
-1
1
-
OuterBay Confidential
Average Run time (Min)
Total Batch Run Time (Hrs/Month)
Payables Approval
Agilent Approach
Lessons Learned – Implementation Project
• You can NEVER START THIS TOO EARLY!!!
• Set the plan by largest “wins” first and get buy-off
immediately
• Get message out that data is NOT being purged; just
relocated
- Users retain transparent access and combined reporting
• Engage key business groups early, establish as a
critical business project: Finance, Audit, etc…
• Strong team work between IT and Business teams
required
• Incremental rollout enabled the project team to
- Refine process
- Enforce accountability
OuterBay Confidential
• “OuterBay Software has provided Agilent
with a long-term positive business impact
with an extraordinary, immediate ROI !”
- Naresh Shanker, Agilent
OuterBay Confidential