Best Practices for Virtualizing Oracle with EMC and VMware <Title>

Today’s Choices in Cloud Computing Title Month Year Today’s Choices in Cloud Computing Verticalization Virtualization Applications Database Operating System Server Storage Middleware Applications Database Operating System Middleware CPU Pool Storage Pool Virtual Infrastructure Network Über-cloud Nearly every IT discussion today focuses to some extend on Cloud Computing and the journey to the cloud. The promise of cloud computing is to provide higher levels of efficiency across all levels of IT, and provide more control of your IT systems through linking all applications and infrastructure together as one cohesive framework. Our belief at EMC is that the cloud should also provide you the benefit of choice, which means the ability to quickly adapt your IT environment for new applications or to deploy new technologies as they become available, without sacrificing efficiency or control. Today the IT industry is becoming polarized around two approaches to cloud computing, verticalization or virtualization. As Oracle customers all of you have seen the change in Oracle the past few years, through the purchase of SUN, in an effort to offer a vertical “Red Stack” of Oracle products, and in some cases bundling everything into one appliance box. Quick show of hands, how many of you here today are Oracle DBAs? For those of you who are DBAs, you understand today, regardless of how it is packaged…. the real value of the Oracle solutions lies in their software, applications, middleware and databases. EMC and VMware have worked with Oracle for many years to tightly integrate our infrastructure with Oracle software to target the specific Oracle software features, map their functionality to our infrastructure, and develop joint solutions which together provide significant measurable benefits. The belief at EMC and VMware is that deploying a virtualized infrastructure integrated with Oracle software provides the best of both worlds….maximum value from your Oracle software investment, with the added efficiency control and choice through virtualization.

The Journey to The Cloud Consolidate x86 Servers Increase Utilization Enhance Data Mobility And High Availability Cross-Application Active-Active use of Infrastructure Reasons Oracle Customers Virtualize IT Production Business Production IT-As-A-Service Lower Costs Improve Quality Of Service Improve Agility % Virtualized 85% 95% The journey to the cloud we believe is centered upon increasing use of virtualization. Virtualization across both servers and storage infrastructure is the fundamental enabling technology which allows all tiers of the Oracle software stack to become seamlessly mobile, continuously available, and much more efficient to operate. Virtualization also allows not just Oracle environments, but ALL environments in the data center to be leveraged across shared infrastructure. Note to presenter: Click for animation In talking with thousands of our customers today, most are at the second phase of the journey today. They started with virtualizing their non-critical environments, such as development and test, which provided them strong cost hardware cost savings. They are now looking to take those same benefits into their production application environments to increase application mobility and high availability. Moving forward over time, some of the new technologies from EMC and VMware can enable active-active application leverage even across multiple data centers. We will review examples of all of these phases and benefits to Oracle customers in this presentation. 50% 30% 15% Platinum Gold

Top Oracle Customer Challenges Increase Oracle performance up to 44% with lower TCO 95%+ Oracle instances match native performance on VMware Increase server and storage utilization up to 80% Consolidate servers by 4x – 20x More Performance and Utilization Faster Provisioning Deploy and manage more instances with less effort 50-70% faster deployment with 90% less App downtime Improve dev/test/QA; new application deployments/migrations Improved backups / Application-aware disaster recovery Backup with less downtime, less storage and faster recovery Recover to application-consistent point in time Enhanced High Availability We speak with many Oracle customers just like you every day and hear consistent themes around areas where there’s a need to increase efficiency around Oracle software and IT infrastructure: Performance and utilization: Maximizing performance of Oracle databases and applications, and sustaining this performance at scale as environments grow. Coupled with this is the need to increase and balance server and storage utilization to ensure high performance does not come at “extreme” cost Provisioning: Every Oracle DBA we talk to today has to create and manage multiple copies of Oracle databases and applications on a consistent basis. This can be for development, test, quality assurance, or for deploying new application environments. EMC Global IT ourselves is one of the largest Oracle deployments in the world, managing hundreds of Oracle databases and we face this challenge everyday High Availability: Backup and recovery of Oracle environments is for most customers the primary source of data protection. The time of backup and recovery as well as impact this can have on applications is a concern for many Oracle customers. For disaster recovery, Oracle provides certain capabilities in the database, such as Oracle Data Guard, but these capabilities can be limiting in their ability to protect the entire applications environments Tailored Solutions: Finally, every customer we talk to today has in most cases a pre-existing investment in Oracle…meaning existing Oracle applications, middleware, databases as well as DBAs and DBA Managers with specific processes and best practices. Our belief at EMC and VMware is you should be able to leverage those existing investments and the right virtualized infrastructure can easily shape itself into your Oracle environment, and then attach to the Oracle software features and enhance them. We believe this is a much more effective model than moving to net, new verticalized appliance architectures. Virtual Infrastructure vs. Vertical Appliances EMC Virtualization Assessment for Oracle EMC, Oracle Technology Centers and EMC Global IT experts (House of Brick Consulting??) Solutions Tailored to Your Business 4

More Performance and Utilization Let’s start the discussion with performance and utilization 5

Oracle Performance and Utilization Challenges Targeting the performance you really need vs. “extreme performance” to avoid excessive cost Oracle workloads will change over time requiring adjustment and tuning Traditional “one workload, one box” approach leads to over-provisioning of servers Physical storage provisioning creates underutilization of capacity and impedes database performance These are many of the top challenges we hear from you our customers about managing performance and utilization of Oracle environments. Many customers are being told today they need to go for “extreme performance” to fix their problems. Our belief is that it’s more efficient to rather target the performance you need, when you need it is a more cost effective approach to take. Given the changing nature of Oracle OLTP workloads, tuning of storage infrastructure can become a constant and very time consuming task during which the database and application is not performing at maximum. Infrastructure should adapt itself automatically to remove the burden of tuning the Oracle database constantly Finally, traditional physical provisioning models have now been replaced with virtual provisioning capabilities on servers and storage, which provide significant efficiency and cost savings to increase performance while also increasing utilization.

Fully Automated Storage Tiering (FAST) VP NEW Q4 ’10 FAST VP (Sub-LUN) THIN PROVISIONING Traditional FAST (Volume/LUN level) Configure storage 80% faster without downtime Scale Oracle performance while increasing utilization Reclaim 50% un-used capacity with ASM Storage Reclaim Utility Flash–10% SATA–90% Flash–10% Fibre Channel–40% SATA–50% We’ll start with storage layer and how EMC has helped IT rethink Oracle storage architecture….. A few years ago, EMC was the first in the industry to introduce the concept of tiered storage for Oracle environments. At the time users had essentially two tiers of storage, online FC drives and offline TAPE. EMC introduced lower cost, higher density SATA drives into our platforms as a means for customers to keep less active data online and available, but without the high cost of FC drives. This was a great fit, for example with Oracle Flashback technology for quick online recovery of Oracle files. We then evolved this with the introduction of Flash Drives, non-mechanical, non-rotating drives which greatly enhance the speed with which data can be retrieved off of disk. Flash Drives are a perfect fit for those Oracle workloads where you applications are spending a lot time seeking for data on disk, small random I/O workloads. For the past several years we worked actively with Oracle customers essentially taking Oracle AWR reports and using those to find the most latency sensitive applications and manually place those on FLASH Drives. The challenge with this approach was as Oracle workloads change, the tiering process becomes very time consuming and in many cases the Oracle database and application needs to be taken offline while the tiering is performed. EMC introduced Fully Automated Storage Tiering (FAST) as a means to automate this provisioning process. Now let’s look at FAST VP, and how it is adding increased efficiencies for automated storage tiering. Note to Presenter: Click now in Slide Show mode for animation. FAST provides the automation to place and move data to the appropriate tier while significantly reducing storage administration time and complexity. FAST has been innovated by EMC to now operate at the sub-LUN level where we are able to actually tier Oracle database data down to the block level, identifying the specific parts of the LUN that are the most active and can most benefit the from Flash drives, as well as the least active parts of the LUN that can run on SATA. This allows the most active 5–10 percent of active data to be placed on Flash drives and get more performance, while the 90-95 percent of the inactive data can be placed on SATA drives at a much lower cost than traditional Fibre Channel. In EMC’s testing of this Solution with Oracle, we demonstrated that FAST VP can actually remove 80% of the time it takes storage administrators to configure storage for Oracle. This is a great benefit to Oracle DBAs to spend less time tuning and less time waiting for storage to adjust to the workloads required by Oracle. This complements Oracle RAC models as you scale out Oracle FAST also operates in a virtual provisioning model, which means we are able to present Oracle hosts more virtual capacity than is currently physically allocated. If you consider a simple example where you have a 10 TB storage array with 3TB allocated for the Oracle database, 3TB for email and 3TB for Oracle homes data, in actuality you may only be consuming 2TB of space for each of these environments, leaving 4TB of un-used capacity. With thin provisioning we are able to present this un-used 4TB to other applications for use as a shared pool. The key with this model is we need to very carefully monitor capacity allocation and utilization to ensure we don’t provision more storage than is actually available. EMC FAST VP automatically monitors and manages this, but also you can set alert policies to notify you if certain capacity limits are being reached. EMC worked with Oracle to integrate the use of the Oracle ASM Storage Reclamation Utility to optimize reallocation of virtual storage pools and reduce storage capacity required up to 50%

VMware vSphere and ESX Server Increase Oracle x86 server utilization from 10-15% up to 80% with vSphere Run multiple Oracle workloads on same hardware Convert physical machines to virtual machines without downtime Reduce hardware requirements by up to 90% with VMware ESX Each instance of VMware ESX can support more than 100 VMs Virtualizing your Oracle database servers provides similar benefits in terms of increasing server utilization and avoiding over-provisioning of server hardware resources. Increase Server Utilization to as Much as 80% The traditional “one workload, one box” approach to server provisioning inevitably leads to over-provisioning and underutilization of hardware assets. Most servers operate at only about 5-15% of their total load capacity. VMware virtual machines run independently from the underlying hardware and are supported on a broad array of physical servers, from 2-CPU to 32-CPU systems Each VMware virtual machine represents a complete system—with processors, memory, networking, storage and BIOS—enabling you to run Windows, Linux, Solaris and NetWare operating systems and software applications on the same piece of hardware at the same time. Running multiple workloads on highly configured x86 servers lets you increase the utilization of your server hardware from 10-15% to as much as 80%. Oracle environments can be converted to virtual machines, without service disruption or system downtime. VMware vCenter Converter can be run on a wide variety of hardware and supports most commonly used versions of the Microsoft Windows operating systems. Reduce Hardware Requirements by a 10:1 Ratio or Better Using VMware vSphere, your organization can significantly reduce the amount of IT hardware in the data center. Each instance of VMware ESX can support more than 100 virtual machines. In fact, the vast majority of our customers run as many as 10 or more applications on a single piece of hardware. Fewer servers means less money spent on energy consumption due to reduced power and cooling requirements. When totaled with the hardware cost savings and other capital reductions, VMware solutions can help your organization save more than $3,000 annually for every server you virtualize. VMware virtual machines cumulatively have saved over five billion Kilowatt-hours of electrical power. The impact of virtualization on energy consumption is so significant that utility providers in California such as PG&E, SoCal Edison and SDG&E are paying customers for server virtualization projects that result in the removal of server computing equipment.

< 15% overhead for 8 vCPU VM 8,900 total DB transactions per second Single VM Performance: Oracle Database OLTP Workload † < 15% overhead for 8 vCPU VM 8,900 total DB transactions per second Near-linear scalability from 1 to 8 vCPUs 60,000 I/O operations/second Intel® Xeon® processor 5500 series based 8-pCPU server RHEL 5.1 Oracle 11gR1 In-house ESX Server Transaction Rate (Ratio to 1-way VM) When we discuss VMware with Oracle customers, two questions are usually raised right away. The first is will VMware reduce the performance of my Oracle databases and applications. Over the past several years as VMware vSphere has evolved, today Oracle databases and applications can operate at near native performance to physical hardware. To demonstrate this, VMware ran a large DB OLTP benchmark to show the tremendous VM scalability for these types of apps. Running a resource-intensive OLTP benchmark, based on a fair-use implementation of the TPC-C* workload specification, VMware achieved 85 percent of native performance when running Oracle DB on VMware ESX hosted on an Intel Xeon 5500 based 2 socket server. This workload, which demonstrated 8,900 database transactions per second and 60,000 disk input/outputs per second (IOPS), is the most resource- intensive load ever shown in a virtual environment to date. This benchmark was on 3.5 – 4.0 is even better. † A fair-use implementation of the TPC-C workload; results are not TPC-C compliant 9 9

Oracle Support for VMware Oracle MyOracleSupport (Metalink) 249212.1 defines Oracle’s VMware support policy most broadly Support facts: Oracle RAC included for 11.2.0.2 and above (Updated Nov 8,2010) Oracle will accept SRs on VMware for bugs already known to Oracle Oracle may accept SRs on VMware for bugs that are not seen by Oracle as being caused by virtualization Oracle maintains (as do most ISVs) the right to require physical reproduction if they suspect VMware is “at fault” VMware's technology partner EMC and system integrator House of Brick have support statements for Oracle on VMware: http://www.emc.com/solutions/application-environment/oracle/oracle- virtualization-vmware.htm www.houseofbrick.com/oracle-on-vmware The second question we get on VMware with Oracle is “does Oracle support this?” The answer simply put is “yes” . There has been a lot of confusion in the marketplace on this issue, with many customers having heard Oracle does not in fact support VMware. The truth is Oracle’s support for VMware has evolved over the past several years, ,and the most important thing you can do as Oracle customers is to check with official Oracle Support Website, MetaLink where there is an official support statement on VMware. Oracle’s MyOracleSupport (Metalink) 249212.1 defines Oracle’s VMware support policy Most recently, Oracle also added support for Oracle on RAC. The key is Oracle RAC for 11.2.0.2 and above (Updated Nov 8,2010) There is a caveat to this support which states: Oracle will accept SRs on VMware for bugs already known to Oracle Oracle may accept SRs on VMware for bugs that are not seen by Oracle as being caused by virtualization Oracle maintains (as do most ISVs) the right to require physical reproduction if they suspect VMware is “at fault EMC and VMware work with Oracle support daily and to date have not seen any issue where Oracle support was not provided. 10

Aligning DBAs and Infrastructure Management Oracle Software Layer Oracle AWR Reports End-to-end visibility and control of virtual infrastructure Simplify monitoring, tuning and reporting Oracle AWR reports provide before-and-after performance analysis EMC, VMware simplify management for Oracle Managing these virtual storage and server environments is a top concern for many Oracle DBAs. From an overall perspective, managing Oracle in a virtual environment vs. physical does not require any changes to how a DBA manages his or her Oracle database or applications. Note to presenter: click for animation EMC and VMware have invested heavily to integrate our management technologies so that they are bi-directionally aware. Meaning, both storage administrators and VM administrators have access to the same topology views of the virtual server, network and storage architecture, and are therefore integrated in now we Deliver end-to-end visibility and control for a single management screen Discover, visualize, and manage storage-to-virtual machine relationships Simplify the ability to monitor, provision, and report on virtual servers From a DBA perspective, EMC and VMware work actively with DBAs in many of our accounts to leverage Oracle AWR reports as a means to help prioritize from an Oracle perspective this applications are best targeted for certain tiers of virtualized infrastructure, and after virtualization is applied, ensure we can check to monitor the impact this had had on the Oracle database and applications. Virtual Servers Virtual Infrastructure Layer EMC <-> VMware Plug-ins Thin Devices Thin Pool

Oracle Use Case: EMC Symmetrix FAST VP with VMware NEW Q1 ’11 Oracle Use Case: EMC Symmetrix FAST VP with VMware Virtualized Infrastructure EMC Symmetrix VMAX FAST VP for high performance at lowest cost Oracle RAC and ASM over FCP Tiered Storage: 128 x 200GB 15k FC Drives; 28 x 2 TB 7,200 rpm SATA; 8 x 200 GB EFD drives EMC Control Center VMware vSphere and vCenter Oracle Software Oracle E-Business Suite R12 Oracle Application Server 10.1 Oracle Database 11g R2 Enterprise Edition Real Application Clusters (RAC) and ASM RHEL 5.1 Walking away from today’s session we want to share with you that EMC and VMware invest heavily in Oracle Solutions. These are where we take these EMC and VMware technologies we describe and test them in real-world complete Oracle software environments. This testing is done by a team of dedicated Oracle DBAs and EMC and VMware technologists at our Global Solution Centers in Hopkinton, MA; Santa Clara, CA; Tokyo, Japan; and Cork, Ireland. The goal of this testing is to demonstrate to you beyond the marketing metrics, practical real-world achievable results from combining our technologies in an Oracle environment. More importantly, we document not just storage and virtualization best practices, but the Oracle software best practices required in setup, configuration, deployment and management. Here is one example just announced in January 2011 from EMC and VMware, leveraging EMC Symmetrix VMAX storage infrastructure with VMware vSphere to provide a high performance, FAST VP environment running a virtualized Oracle environment with Oracle RAC, supported by Oracle E-Business Suite of Applications. EMC Symmetrix VMAX

Documented Best Practices 44% More Oracle TPM and 30% lower cost Setup test of multiple 400 Oracle user sessions with OLTP test suite Deployed FAST VP and tested physical Oracle RAC performance at scale Performed non-disruptive physical-to-virtual migration of Oracle RAC Ran same tests on virtualized Oracle RAC scaling to 8 VMs Response times within 10% of physical, with more performance, less cost Our goal with this test was to essentially prove out the fact that we can deploy Oracle on FAST VP, and move Oracle from a physical two node Oracle RAC cluster to a virtual cluster with VMware. We then scaled out Oracle RAC across 8 virtual machines. The test suites run were using Swingbench to simulate a real-world Oracle OLTP workload. We setup multiple “Sessions” consisting of groups of 400 users with similar workload profiles and ran these against the RAC cluster. We were able to first, compare the performance of running Oracle on physical hardware vs. virtual hardware, and then measure the virtualized Oracle RAC performance as we scaled. The results were very compelling in that we were able to show that running Oracle RAC in virtual vs. physical provided response times within 10% of physical, but because of FAST VP and tiered storage we were able to scale up to 45% more TPM, while also reducing overall TCO by 30%.

Faster Provisioning Let’s continue the discussion with provisioning Oracle environments. 14

Oracle Provisioning Challenges Time to provision new Oracle environments (Apps, Middleware, Database) and infrastructure Minimizing Oracle production application downtime during dev/test/QA process Time required in recreating copies in the event of test failure Managing this process at scale in large, multi database Oracle environments These are many of the top challenges we hear from you our customers about provisioning in Oracle environments. Many customers are being told today they need to go for “extreme performance” to fix their problems. Our belief is that it’s more efficient to rather target the performance you need, when you need it is a more cost effective approach to take. Given the changing nature of Oracle OLTP workloads, tuning of storage infrastructure can become a constant and very time consuming task during which the database and application is not performing at maximum. Infrastructure should adapt itself automatically to remove the burden of tuning the Oracle database constantly Finally, traditional physical provisioning models have now been replaced with virtual provisioning capabilities on servers and storage, which provide significant efficiency and cost savings to increase performance while also increasing utilization.

EMC Replication Manager Replicate with 90% Less Downtime Speed and Automation Replicate Oracle in minutes vs. hours / days Auto-discovery of Oracle, replication technology, VMware datastores (VMFS [Virtual Machine File System]/NFS), virtual machines OAUG “Innovator of Year Award” 2009 Application consistency Understands how to place applications into proper state for application-consistent replicas Oracle RMAN Hot backup for Oracle VMware snapshots for VMFS and NFS datastores Instant restore Application-consistent replicas of Oracle running on VMware virtual machines VMFS or NFS datastores containing virtual machines ORACLE-AWARE SNAPS / CLONES Oracle production Oracle copy Oracle dev/test Oracle copy Replication Manager is designed for ease of use through the following: Auto-discovery of applications, their associated storage, and the replication technology available; VMFS/NFS datastores; and virtual machines. Intelligence to place applications in the proper state for application-consistent replicas, such as VMware snapshots for VMFS or NFS datastores, VSS for Exchange, VDI for SQL Server and SharePoint, and online hot backup mode for Oracle. Instantly recover back to production application data for Exchange, SQL Server, SharePoint, and Oracle running on VMware virtual machines or Microsoft Windows child partitions as well as VMFS or NFS datastores containing virtual machines. Perform a simple restore through VMware vCenter Server of a single virtual machine from a mounted replica created by Replication Manager. With the built-in automation, application intelligence, and flexibility of Replication Manager, protecting your VMware virtual infrastructure is a seamless alternative to scripting. Oracle copy Oracle Backup HIGHEST Replication Manager Administrator Power Database Administrator Database Administrator Power User Operator USER ROLE PRIVILEGES LOWEST

Deploying Virtual Machines with Templates vSphere Client Centralized VM management across heterogeneous x86 operating environments Linux, Windows Accelerate provisioning time 50-70% with VM Templates Eliminate repetitive configuration and installation steps across Application, Middleware and Database tiers Combine with EMC replication to provision complete environment Deploy Template wizard

Use Case: EMC Unified Storage with VMware NEW Q1 ’11 Virtualized Infrastructure Oracle RAC and ASM over FCP Oracle App Server and Appl_TOP over NFS 6GB 10k SAS Drives VMware vSphere and vCenter VM Templates on App Server and Database Tiers VMware vStorage APIs for Array Integration Oracle Software Oracle E-Business Suite R12 Oracle Application Server 10.1 Oracle Database 11g R2 Enterprise Edition Real Application Clusters (RAC) and ASM 200 GB Database RHEL 5.1 FC SAN NAS EMC VNX Unified Storage

Use Case: Rapid Provisioning of Complete Environment EMC Replication Manager Oracle EBS Copies Production Oracle Snapshot Oracle Snapshot vSphere Client Clone Be sure to talk to how fast it is to recreate dev/test with EMC Replication Manager as well Deploy Template wizard Virtual machine name Virtual machine location Pre-created Custom Specification (defines network configuration for machine) All 10 virtual machines deployed in less than 10 minutes

Documented Best Practices Deploy up to 10 Oracle VMs in 10 Minutes Configured Oracle EBS Application shared file system and ASM Configured VMware vSphere and create virtual machines Deploying templates across Oracle Application and Database tier nodes Deploy Oracle EBS across shared file system using RAC and ASM Perform functional test of application tier and add app server nodes

Enhanced High Availability 21

Oracle High Availability Challenges Backup, recovery can become time consuming and costly as Oracle environments scale Oracle software alone limited as complete disaster recovery solution for applications Traditional physical disaster recovery infrastructure remains idle and limits application mobility

Enhanced Oracle Backup and Recovery Control Tier Ex: 84-92% Less Storage in 5 weeks Backup and Archive Applications Target Tier CIFS, NFS, NDMP, OpenStorage Ethernet Virtual Tape Library (VTL) over Fibre Channel To continue next we’ll talk about backup and recovery. Here we want to highlight a new and exciting EMC acquisition has made with Data Domain. Data Domain is a storage system with shelves of disks and a controller. It’s optimized, first to back up and second to archive applications, and supports most of the industry-leading backup and archiving applications. Data Domain storage serves as a direct disk target for Oracle RMAN and Oracle Secure Backup, as well as for enterprise backup software, archiving applications and other data movers. This technology provides 3 primary benefits to Oracle customers: Retain backups longer: Through de-duplicating data before it is backed up, we can allow for more data to be stored online at lower cost. This allows us to retain more Oracle data for longer periods of time, and perform more frequent full backups vs. incremental backups. Replicate smarter: Data Domain systems only move de-duplicated data over existing networks with up to 99 percent bandwidth efficiency for fast, cost-effective disaster recovery. Recover reliably from disk: Through allowing for recovery from full backups vs. incremental, we are able to recover more reliably and with less complexity. As data is sent to a Data Domain system, a Checksum is taken of everything that’s about to be stored. The system checks that it has already stored the data and if not, the data is then stored through the file system to disk. Then the backup I/O is internally read back from disk, and re-Checksummed. Next, the system compares it to the original Checksum taken at write time. This allows the system to know for sure that the data is reachable, the blocks are readable, and the data within those blocks is correct. The ideal target Oracle customers for this solution would include those that are: Dependent on tape for data protection and have minimal or no retention on disk Unable to meet RPO/RTOs Experiencing frequent failures and delays with Oracle database recovery or cloning Intolerant to changes to their existing RMAN procedures Using expensive primary storage capacity to mitigate any of these challenges DD880 appliance Retain longer Store weeks or months of full backups with 10-30x less disk Perform Oracle full backups to disk, reducing tape dependence Replicate smarter Move only de-duplicated data over existing networks with up to 99% bandwidth efficiency for cost-effective disaster recovery Recover reliably Recover from full backups vs. incremental reducing complexity Checksum and re-checksum ensures data blocks readable and correct Direct target for Oracle RMAN and Oracle Secure Backup as well as enterprise backup software 23

Complete Oracle Disaster Recovery EMC RecoverPoint WAN SAN SAN RecoverPoint Appliance RecoverPoint Appliance Oracle Applications Oracle Middleware Oracle Database Over time as the data within the Oracle database becomes more critical to the business, disaster recovery will become a focus. EMC has worked with Oracle in this area for many years and today including best practices for the use of EMC disaster recovery solutions with or without the use of Oracle Data Guard. This slide describes a joint engineering initiative with EMC and Oracle to document using RecoverPoint to protect the entire environment including Oracle Applications, Fusion Middleware and Database. For those customers who use Oracle DataGuard, DataGuard can protect the database files, but not the applications or middleware. So Oracle engaged EMC to document this Solution. We have two best practices available, one which outlines using EMC RecoverPoint to protect the entire environment, or one where we use Oracle DataGuard to protect the database, but use RecoverPoint to protect the applications and middleware. RecoverPoint—EMC RecoverPoint is a comprehensive data protection solution providing concurrent local and remote data protection. RecoverPoint can protect data both locally and remotely. For local replication, RecoverPoint writes data to a local journal and then a local volume. Here RecoverPoint can synchronously track every write in the local journal and distribute the write to the local volume all without impacting application performance. For remote replication, RecoverPoint transfers significant writes, based on application recovery time or recovery point objectives, to a remote site where they are saved in a history journal. Once the RecoverPoint appliance receives the write it will bundle up the write with others into a package. Redundant blocks are eliminated from the package and remaining writes are sequences and stored with their corresponding timestamp and bookmark information. The package is them compressed and an MD-5 checksum is generated for the package. The package is then scheduled for delivery across the IP Network to the remote appliance. Once received the remote appliance validates the checksum to ensure the package was not corrupted in the transmission. The data is then uncompressed and written to the journal volume. Once the data has been written to the journal volume it is then distributed to the remote volumes to ensure write order sequencing is preserved. In the case of Oracle we support both synchronous, asynchronous and snapshot mode. All Oracle LUNs are grouped into a consistency group and replication sets are created to map the production LUNs to the local and/or remote LUNs. We then provide several options for recovery: Application consistent recovery from a shut down – also known as a cold backup – the user creates a consistency group that represents the Oracle application, containing all files for the application including data files, online redo log files, configuration files and optionally control files. This method allows for recovery with 100% reliability. Because operations must be halted while the cold backup is taken, this is not ideal for 24x7 operating environments. Crash consistent recovery during operations – this enables crash consistent images without shut down. As a write is issued to the production volumes, RecoverPoint splitters intercept it and send a copy of the write to the RecoverPoint appliance for further processing. These captured writes represent on-disk consistent data. When Oracle is restarted from a server crash, it must open the database then perform recovery operations. Oracle uses the redo logs to recover the committed data and undoes any in process transactions on the failed instance. Application consistent recovery during operations – this enables creation of application consistent images without application shutdown. It requires that all data files belonging to the relevant Oracle tablespaces, archive logs and control files are flushed from the servers in memory buffers to disk. This procedure involves placing the appropriate tabelspaces or databases into hot backup mode. Once backup mode is set, RecoverPoint creates a bookmark for the specific consistency group to identify the image as an application consistent image. History Journal Recovery Journal Local Remote Benefits Track every Oracle write without application performance impact Remotely replicate with compression reduces bandwidth up to 90% Multiple modes (hot/cold application consistent or crash consistent) Integrates with Oracle Data Guard for complete protection

Integration with VMware vCenter Site Recovery Manager (SRM) Heterogeneous, network-based replication EMC RecoverPoint Servers Heterogeneous Storage Production VMware Infrastructure Virtual Machines vCenter SRM Disaster Recovery Uses RecoverPoint replica image to test disaster recovery failover Provides on-demand failover of a selected subset or all virtual machines Automates failback with vCenter Server plug-in for RecoverPoint RecoverPoint is also integrated with VMware vCenter Site Recovery Manager. Virtual machines can be brought back online rapidly with no data loss when RecoverPoint is used with VMware vCenter Site Recovery Manager to orchestrate and streamline data protection and failover processes. RecoverPoint is the most flexible approach to protecting virtualized data—replicating VMware vStorage VMFS to protect and recover a single virtual machine or the entire VMware ESX server.

Active-Active High Availability with VPLEX EMC VPLEX Local EMC VPLEX Value to Oracle Environments Seamless Oracle application mobility - Supports the movement of VMs up to 17x faster without downtime Active-active data over distance - Increases utilization of passive DR assets up to 50% Migrations and tech refresh - Removes any application downtime There are two VPLEX products currently available today. VPLEX Local, which provides local federation and simplified storage management at a single site, and VPLEX Metro, which allows you to connect two clusters together within synchronous distances and move or share data between them. EMC VPLEX Metro

Solutions Tailored to Your Business 27

Summary: Oracle on EMC, VMware Gain Everything, Sacrifice Nothing Optimized performance and utilization 45% more TPM with reduce cost through FAST VP Scale 50% more users using same physical capacity with FAST VP Up to 90% less hardware consumption through virtualization Up to 90% more efficiency for Oracle environments 80% faster storage provisioning with FAST VP and VMware 90% less downtime during replication (dev/test/backup/D/R) 84%+ less storage for Oracle RMAN backups 90% less bandwidth for remote-site application recovery Active-active high availability increases use of D/R assets 50% Virtualize Oracle at best pace for your business Extensively tested joint best practices and Solutions Full support for Oracle production (including RAC)

Where is a good place to start Where is a good place to start? Virtualization Readiness Assessment for Oracle Create the Optimal Virtualization Approach for Your Oracle Applications Approach Assess the business and service level requirements Identify Oracle application, server, and database dependencies Capture and review business transactions for CPU, memory, disk, and network during peak and non-peak loads Review Oracle disk group layout and recommended appropriate virtual environment Review HA requirements and recommend solutions Analyze physical performance and estimate virtual resources required Results Detailed assessment of your Oracle environment and its ability to meet readiness criteria Recommendations for virtualization of your environment The EMC Application Virtualization Readiness Assessment for Oracle will help your organization develop a virtualization strategy. It will also provide a design and implementation plan for a state-of-the-art solution that will yield the agility, control, security, and efficiency you expect. This on-site assessment will evaluate constraints and dependencies within and across your Oracle applications and infrastructure providing the information necessary to scope and plan your virtual Oracle environment. We work directly with your IT and business teams and employ a combination of automated tools, best practices, and subject matter experts to: Assess the business requirements and service level requirements of the Oracle environments Identify Oracle application, server, and database dependencies Capture and review CPU utilization, memory utilization, disk IOPS, network I/O, network bandwidth utilization and throughput in business transactions per minute during peak and non-peak loads Review current Oracle disk group layout and recommended appropriate layout for virtual environment Review high availability business requirements and technologies currently in use and recommend solutions for the next generation Oracle environment Analyze physical performance characteristics and estimate resources required for the virtualized environment. The result is an executive-level presentation that contains a detailed assessment of your Oracle environment and its ability to meet readiness criteria with high-level recommendations for addressing the assessment findings.

Resources Documented reference architectures and configuration guides Note: will add some graphics here Resources Documented reference architectures and configuration guides Oracle Technology Centers (Atlanta, GA; Reston, VA) EMC Solution Centers (Hopkinton, MA; Santa Clara, CA; Cork, Ireland) EMC Global IT Experts available to talk to customers www.emc.com/oraclesolutions www.emc.com/virtualizingoracle

Q&A