Oracle Exadata for SAP

Report
Oracle Exadata for SAP
http://www.sdn.sap.com/irj/sdn/ora
Copyright © 2010, Oracle Corporation and/or its affiliates
– 33
–
World’s Most Popular Database
Oracle Database 11g
•
•
•
•
•
#1 in RDBMS market share
#1 for enterprise applications
#1 for data warehousing
#1 for SAP applications
#1 for SAP Business Warehouse
© 2011 Oracle Corporation
4
Oracle Development for SAP Team
Located at SAP HQ in Walldorf, Germany
• Long-established working relationship
• Enables technology transfers between Oracle and SAP
• Optimization of SAP products for Oracle technology
‒ Implementation & testing new Oracle products on SAP
• Worldwide customer & pilot support
‒ Walldorf
‒ Palo Alto
‒ Tokyo
• SAP is Oracle’s largest reseller
‒ Multi year Partners agreements in place
© 2011 Oracle Corporation
5
Enabling SAP Applications
Years of continuous Oracle innovation
Oracle Exalogic
Oracle Exadata
Oracle Linux
Oracle Database 11g
2011
2010
2009
Oracle Database 10g
2007
2006
Automatic Storage Management
Real Application Testing
OLTP Compression
Oracle Database Vault
Transparent Data Encryption
Read-only table compression
Partitioning
2003 Real Application Clusters
1998 First SAP/BW release on Oracle
1992
First SAP/R3 release on Oracle
© 2011 Oracle Corporation
Oracle9i Database
6
Why SAP Customers Choose Oracle
Highest Quality of Service for Business Users
Best
Performance
& Scalability
Best
Deployment
Flexibility
Best
Availability
& Reliability
Best
Support for
Very Large DBs
Best
Database
Security
Best
Database
Manageability
© 2011 Oracle Corporation
7
SAP Customers Rely on Oracle
© 2011 Oracle Corporation
8
Why Oracle Exadata?
© 2011 Oracle Corporation
9
Oracle Exadata Database Machine
One architecture for…
• Online Transaction Processing
• Data Warehousing
• Database Consolidation
© 2011 Oracle Corporation
10
Oracle Exadata Database Machine
One architecture for…
• SAP ERP 6.0
• SAP Business Warehouse
• Database Consolidation
© 2011 Oracle Corporation
11
Extreme Performance Platform
Ideal for Consolidating SAP Application Databases
• Database Server Pool
– Oracle Database 11g Release 2
– Oracle Real Application Clusters
– Automatic Storage Management
• Storage Server Pool
– Up to 336 TB disk
– 5 TB flash storage
– Oracle Exadata Storage Software
• InfiniBand Network
– 40 Gb/sec redundant switches
© 2011 Oracle Corporation
12
Intelligent Storage Innovations
Oracle Exadata Storage Software
• Smart Scans
‒ Offload query processing
‒ Scale-out storage
+
+
• Hybrid Columnar Compression1)
Efficient compression increases effective
storage capacity and increases user data
scan bandwidths by a factor of up to 10X
+
Uncompressed
• Smart Flash Cache
– Accelerates I/Os up to 30x
– Doubles data scan rate
primary backup
Benefits
Multiply
test
standby
dev’t
Compressed
1) Hybrid Columnar Compression can be used after SAP approval
© 2011 Oracle Corporation
13
Exadata Smart Flash Cache
Fastest Flash with Best Database Integration
• Exadata has 5 TB of PCI flash
‒ Flash PCI cards avoid disk controller bottlenecks
‒ Store up to 15 TB using Advanced compression
• Intelligently manages flash storage
‒ Smart Flash Cache holds hot data
‒ Gives speed of flash, cost of disk
• Exadata flash cache achieves:
• Over 1.5 million IO/sec from SQL (8K)
• Sub-millisecond response times
© 2011 Oracle Corporation
14
Oracle Exalogic Elastic Cloud X2-2
Integrated Compute, I/O, Networking and Storage
Integrated Storage
•
•
•
•
•
Shared storage for applications
Clustered for HA
40 TB SAS disk
4 TB read cache
72 GB write cache
InfiniBand I/O Fabric and 10GbE/GbE
• 40 Gb/sec links
• 10 Gigabit Ethernet
connectivity to datacenter
Compute Nodes
•
•
•
•
30 x86 compute nodes
360 Xeon cores (2.93 GHz)
2.8 TB DRAM
960 GB SSD
© 2011 Oracle Corporation
15
Standardized Configuration
Deploy in days not months
• All Database Machines are the same
‒
‒
‒
‒
Delivered tested and ready-to-run
Highly optimized
Highly supportable
No unique configuration issues
• Runs existing OLTP and DW applications
‒ Over 30 years of Oracle Database capabilities
• Leverages Oracle ecosystem
‒ Skills, knowledge base, people, partners
© 2011 Oracle Corporation
16
Consolidating Databases on Exadata
• Serves as farm/cloud for databases
‒ Lowers costs
• Extreme performance for complex mix of
OLTP, DW, batch & reporting workloads
‒ Increases utilization
• Automated I/O and CPU resource
management isolates workloads
‒ Reduces management
© 2011 Oracle Corporation
17
Workload Resource Management
For mixed Workloads and Multi-Databases
• Ensure databases get correct amount of
I/O and CPU resources
‒ ERP:
‒ BW:
33% of resources
67% of resources
• Ensure ERP tasks get correct resources
‒ Interactive gets priority before Batch
• Ensure BW tasks get correct resources
‒ Reporting: 70% of resources
‒ ETL:
30% of resources
© 2011 Oracle Corporation
18
SAP Customer Case Study
• Business Background
‒ European B-2-B stationery and office supply retailer
‒ Operates in 30+ countries
‒ 30% growth expected within next 4 years
• Challenges
‒ Long-running invoicing batch run times
‒ Cost and effort to administer large number of SAP databases
‒ New B-2-B process couldn‘t be supported on current system
© 2011 Oracle Corporation
19
Customer’s Requirements
• Goals
‒ Expose “real“ SAP workloads to an Exadata environment
‒ Measure performance increase (throughput & response times)
‒ Show how Oracle technology can help minimize the impact of
hardware and software failures for the business
• Conditions
‒ Use standardized Oracle Exadata configuration
‒ No manual tuning or configuration changes
‒ All tests administered by customer's staff
© 2011 Oracle Corporation
20
Conducting Proof of Concept
at Oracle Solution Centre in UK
• Preparation
‒
‒
‒
‒
Installation of the entire IT infrastructure for SAP
Upgrade 6.5 TB data from Oracle Database 10g to Exadata
Generate simulation of online users in a web shop
Generate batch job load for invoicing
• Execution Environment
‒ Oracle Exadata Database Machine X2-2 Half-Rack
‒ 5 x Sun Blade 6270 M2 running Solaris10 for Application Tier
‒ All tests administered by customer's staff
© 2011 Oracle Corporation
21
Results from Proof of Concept
“Faster performance out-of-the-box”
• Online User Activity
‒ Simulated over 6,000 concurrent users
‒ Response times of ~0.9s with 17% CPU utilization across 4 nodes
‒ Unable to “saturate“ system
• Batch Job Results
‒ Simulated over 900 batch jobs running in parallel
‒ Reduced batch run times by 2/3rds (11:30 to 3:45 hours)
‒ Only 2 Database Server nodes used for batch jobs
© 2011 Oracle Corporation
22
Complete Family Of Database Machines
From quarter-rack to multi-rack configurations
Oracle Exadata X2-2
Oracle Exadata X2-8
2 to 8 (12 core) database nodes
2 (64 core) database nodes
Quarter
Half
Full, Multi-Rack
Full-Rack
Multi-Rack
© 2011 Oracle Corporation
23
Exadata Storage Expansion Rack
Big Data for less
• Pre-built storage-only rack
• Connects to anyExadata Database Machine
• For applications that need more space but
not more compute
–
–
–
–
On-disk backups (27 TB/hour)
Historical or archive data
File data, LOBs, XML data, documents
Images, and other large unstructured data
• Attractive pricing ( ~ $7,000 per TB )
© 2011 Oracle Corporation
24
Oracle Exalogic and Oracle Exadata
Easiest Path to Extreme Performance
SOA
E20
IDM
SAP ERP
Data
Custom
App.
SAP BW
Data
SAP ERP
App.
InfiniBand
SOA
Dehydration
Java
EE
App.
Data
Warehouse
SAP BW
App.
Exalogic
Data
Mart
Exadata
Application Tier
Database Tier
© 2011 Oracle Corporation
25
Oracle Exadata for SAP
A major leap in performance, capacity and value
•
•
•
•
•
•
Redundant, secure, scale-out architecture
Eliminate systems integration trial-and-error
Run existing SAP databases unchanged
Utilize existing personnel, skills, Oracle licenses
Improve performance of SAP applications
Consolidate platforms, databases, power and cooling
© 2011 Oracle Corporation
26
For More Information
oracle.com/sap
oracle.com/exadata
sdn.sap.com/irj/sdn/ora
© 2011 Oracle Corporation
27
28

similar documents