Sparc SuperCluster
Upcoming SlideShare
Loading in...5
×
 

Like this? Share it with your network

Share

Sparc SuperCluster

on

  • 991 views

Discusses some of the developments concerning Solaris SPARC and a description of Oracle engineering system Supercluster

Discusses some of the developments concerning Solaris SPARC and a description of Oracle engineering system Supercluster

Statistics

Views

Total Views
991
Views on SlideShare
988
Embed Views
3

Actions

Likes
0
Downloads
44
Comments
0

1 Embed 3

http://www.slideee.com 3

Accessibility

Upload Details

Uploaded via as Microsoft PowerPoint

Usage Rights

© All Rights Reserved

Report content

Flagged as inappropriate Flag as inappropriate
Flag as inappropriate

Select your reason for flagging this presentation as inappropriate.

Cancel
  • Full Name Full Name Comment goes here.
    Are you sure you want to
    Your message goes here
    Processing…
Post Comment
Edit your comment
  • 2 Clicks
  • Oracle’s vision is to deliver an “integrated stack” that provides enormous benefit to customers in terms of reduced cost, simplified deployment and management, and greater investment protection while reducing overall risk.With the acquisition of Sun Microsystems, Oracle now has the most complete portfolio of best of breed technology across all layers of the IT stack. With this complete portfolio, Oracle is able to tightly integrate the technology to provide highest levels of performance, application availability, security, and investment protection. Through integrated development and test, Oracle can deliver innovative new capabilities to customers faster, with much lower risk to the business. For example, daily solaris builds are tested on DB and MW, weekly Application stack testing is conducted to assure no bottlenecks or bugs are introduced, and synchronized OS and DB patch releases ensure that customer environments have the latest and matching patch requirementsOur engineered and tested systems, ie. Exadata, Exalogic, Exalytics and Superclusters can be configured and deployed in a fraction of the time that a customer could do it in, and risk of problems are significantly reduced. Integrated systems management tools enable simplified management across the entire stack, from application and database layer all the way down to server, OS and disks. This allows for pro-active and preventive monitoring of systems, quick identification and of issues, as well as a single point of contact for vendor related escalations.But we also realize that customers need choice due to specific business and IT requirements. Through leadership in and rigorous adherence to open standards, Oracle products and technology work well in heterogeneous environments, and allow customers the flexibility to integrate with other systems, tools and applications.Oracle Advantage: Solutions, not ServicesOracle innovates in solutions and does not make the majority of its revenue through professional servicesOracle’s best-in-class suites and technologies in all layers of the stack aim to lessen the burden of heavy consulting servicesUniquely integrated self-healing, automated management, and control help lessen the need for outsourced servicesDeclarative development, policy-based configuration all speed time to implementation and deploymentOracle provides targeted services engagements to help customers realize goals: not vague, open-ended projects with dubious ROIOracle Advantage: Open Oracle can help modernize away from proprietary mainframe toward more open systemsBenefits to customers:Exploits new hardware and software architectures Flexibility: Can more easily interoperate with third-party software or custom applications into an open environmentFuture proof and minimized vendor lock-in: Adherence to open environment and standards shields customers from changes in product direction from vendorsLarger ecosystem of developers. Adherence to standards means more developers availableMore open, transparent collaboration with Oracle to influence future directions of products Deep IntegrationOracle lowers customer costs through unified, integrated solutions with easy-to-use development & management tools to bring down complexity and TCOIBM Software by contrast has 5 brand divisions, plus legacy software, all with redundant technology and disparate architecturesIBM does not have suite-based approach and relies on Global Services to fit solutions together – at considerable cost to the customersConfusing arrays of appliance not optimized for extreme performanceComplexity within each product and across the non-integrated stack leads to higher TCO and longer time to valueBenefits to customers:Easier to deploy and manageSimplification of IT environmentOptimized for solution performance and ease of scale-outQuicker time to valueLower TCOCompleteSingle vendor accountability for the complete stack of servers, storage, database, middleware, infrastructure management, and enterprise business applicationsChoice of traditional, optimized or fully engineered solutionsStandards-Based and Open SystemsOracle offers open alternatives to mainframes; provides interoperability and strict adherence to standards like Java and WS-* Solutions and Technology FocusOracle invests in building best-in-class solutions and technology. IBM invests in professional services, which constitutes about 57% of IBM’s total revenue
  • Slide Transition: When we say that Oracle hardware and software is engineered to work together, this refers to our unique ability to offer customers a complete hardware and software stack -- from applications through middleware and databases, and all the way down into servers and storage – that is integrated throughout.Oracle tests everything within a stack layer together–between different applications, between different middleware suites, between various database products, and so on. Oracle also tests everything across stack layers, from applications all the way down to the servers and storage. Oracle certifies the complete stack so that customers know which particular versions of software are designed to work together. For example, every major application that Oracle delivers, including Oracle E-Business Suite, PeopleSoft Enterprise, JD Edwards and Siebel CRM is certified with Oracle Fusion Middleware. Oracle packages the different technologies into standardized solutions, which customers can then deploy together.These complete solutions are designed to be upgraded effectively and efficiently together. And, the entire stack can be managed together, supported together and so on.
  • Develop best of breed ServersBetter IBM in Performance and Price/PerformanceBetter Intel in PerformanceCo-Engineer with the Oracle Stack to deliver a superior experienceFully TestedMore EfficientBetter SecurityBetter AvailabilityEase of UseBetter Manageability – Integrated full stack management Disk to Application viewEngineer unique proprietary capabilities – Application AcceleratorsBasis of future engineered systemsIncrease Application performance 2x every two years
  • Add slide for IBM comp? IP. Ecosystem.
  • Need x86 Solaris Numbers
  • The Oracle Real Application Clusters (RAC) distributed database product includes the Lock Management System (LMS), a user-level distributed lock protocol which mediates requests for database blocks between processes on the nodes of a database cluster. Fulfilling a request requires traversing and copying data across the user/kernel boundary on the requesting and serving nodes, even for the significant number of requests for blocks with uncontended locks. We have created a kernel accelerator (KA), which filters database block requests destined for LMS processes and directly grants requests for blocks with uncontended locks, thereby eliminating user-kernel context switches, the associated data copying, and LMS application-level processing for those requests.The KA exports shared memory in which the LMS locking daemon places its lock table. The KA intercepts DBMS block requests over the RDSv3 communications protocol used between cluster nodes and calls into a DBMS-provided kernel accelerator run-time (KA RT) module, which consults the shared-memory lock table. If the lock is available, the KA replies from the kernel, granting the request directly to the requesting node; if the lock is not available, the KA passes the request up to the LMS user process, which handles the request in the same fashion as when no KA is present.This not only speeds up the process of granting locks, but it also frees up CPU cycles, thus allowing for better throughput in the order of 30-40% depending on the workload.Note: The next generation Oracle Database technology uses the Oracle RAC Kernel Mode Acceleration. The kernel accelerator is available for Oracle Solaris 11 and Oracle Linux only.
  • Only available with Solaris. More DTrace integration to come in later DB releases.A simple example of an outlier I/O: We can check the v$kernel_io_outlier table to extract information about time spent in the kernel for I/O's whose end to end latency exceeds a given threshold (500ms be default but tunable via the '_io_outlier_threshold' tunable - the example below was on an instance with this set to 200ms): SQL> desc v$kernel_io_outlier  Name                                      Null?    Type  ----------------------------------------- -------- ----------------------------  TIMESTAMP                                          NUMBER  IO_SIZE                                            NUMBER  IO_OFFSET                                          NUMBER  DEVICE_NAME VARCHAR2(513)  PROCESS_NAME VARCHAR2(64)  TOTAL_LATENCY                                      NUMBER  SETUP_LATENCY                                      NUMBER  QUEUE_TO_HBA_LATENCY                               NUMBER  TRANSFER_LATENCY                                   NUMBER  CLEANUP_LATENCY                                    NUMBER  PID                                                NUMBER  CON_ID                                             NUMBER SQL> select IO_SIZE,PID,TOTAL_LATENCY,SETUP_LATENCY,QUEUE_TO_HBA_LATENCY,TRANSFER_LATENCY,CLEANUP_LATENCY from v$kernel_io_outlier; DEVICE_NAME --------------------------------------------------------------------------------    IO_SIZE        PID TOTAL_LATENCY SETUP_LATENCY QUEUE_TO_HBA_LATENCY ---------- ---------- ------------- ------------- -------------------- TRANSFER_LATENCY CLEANUP_LATENCY ---------------- --------------- sd@3,0:a,raw         64          0        402554 2020               107           400361              64 This example shows that this single 64k write to a scsi target had an end to end latency of just over 400 millisec (the timing numbers above are in microsec) and the breakdown is: SETUP_LATENCY: 2020 microsec - Time in microseconds spent during initial I/O setup before sending to scsi target device driver QUEUE_TO_HBA_LATENCY: 107 microsec - Time in microseconds spent in the scsi target device driver before being sent to the Host Bus Adaptor QUEUE_TO_HBA_LATENCY: ~400 millisec of this was spent being transferred to the physical device (in the Host Bus Adaptor and physically DMA'ing to the device). CLEANUP_LATENCY: 64 microsec- Time in microseconds spent freeing resources used by the completed I/O
  • 2013, rumor of x86 server sale to Lenovo: http://www.crn.com/news/data-center/240153148/ibm-in-talks-to-sell-x86-server-business-to-lenovo.htm--stalled in April/May 2013 over value of the deal: http://www.theregister.co.uk/2013/05/01/ibm_lenovo_x86_server_deal_skids/--Lenovo raising debt in US dollars may indicate it’s still in flight (late May 2013): http://blogs.barrons.com/techtraderdaily/2013/05/31/ibm-lenovo-bond-raise-may-signal-server-sale-says-rbc/2012, POS biz: http://www.theregister.co.uk/2012/04/17/toshiba_buys_ibm_retail_biz_followup/2007, printing: http://www.pcworld.com/article/128870/article.html2004, PCs: http://www.lenovo.com/news/us/en/2005/04/ibm_lenovo.html2002, disk: http://www.internetnews.com/bus-news/article.php/1183321/IBM+Sells+Hard+Disk+Drive+Biz+Cuts+Staff.htm1999, networking to Cisco: http://news.cnet.com/8301-1001_3-10228455-92.html
  • Execution strategy slideCore S3 is the common building block used in all 3 chipsets across the Oracle SPARC line now – using common building block accelerates time to market and enables the ability to make tradeoffs to meet application workload profiles – ie. T5 with 16 cores and 8M L3 cache vs. M5 with 6 cores and larger 48M cache for backend applicationsSystems now are modernized and have common features like IO, memory and virtualization (LDOMs)Complementary products deliver value and provide customers choices depending on data center requirements – regardless, SPARC systems are reliable and allow for secure operation with increased levels of performance vs. before
  • If customers are familiar with the SPARC T4 line of products, then the actual SPARC T5 systems should seem very familiar. The main change between the 2 generations is the processor and motherboard.
  • Details on the comparisonOracle:SPARC T5-2,2 x 3.6 GHz 16-core SPARC T5 (32 cores), 512 GB memoryOracle Solaris,Oracle VM Server for SPARCSPARC T5-4,4 x 3.6 GHz 16-core SPARC T5 (64 cores), 1 TB memoryOracle Solaris,Oracle VM Server for SPARCSPARC T5-8,8 x 3.6 GHz 16-core SPARC T5 (128cores), 2 TB memoryOracle Solaris,Oracle VM Server for SPARCSPARC M6-32,32 x 3.6 GHz 12-core SPARC M6 (384cores), 8TB memoryOracle Solaris,Oracle VM Server for SPARCIBM:Power 740, 2 x 4.2 GHz 8-core POWER7+ (16 cores), 512 GB memoryIBM AIX Standard Ed.,PowerVM Express Ed.Power 750, 4 x 4.0GHz 8-core POWER7+ (32 cores), 1 TB memoryIBM AIX Standard Ed.,PowerVM Express Ed.Power 780, 16 x 3.72GHz 8-core POWER7+ (128 cores), 2 TB memoryIBM AIX Standard Ed.,PowerVM Standard Ed.Power 795, 32 x 4.0GHz 8-core POWER7 (256 cores), 8 TB memoryIBM AIX Standard Ed.,PowerVM Standard Ed.
  • The SPARC M6-32 Virtualization Infrastructure provides No-cost virtualization which helps to improve system utilization.This slide shows the various levels of virtualization technologies available, at no cost, on the SPARC M5-32 and M6-32 servers At the bottom of the virtualization stack is the SPARC M5/M6 server patform. The first level of virtualization, Pdoms (formerly known as Dynamic Domains), are hardware partitions. Each Dynamic Domain can be further virtualized with the hypervisor-based Oracle VM Server for SPARC partitions, but is can also natively run the Oracle Solaris 11 Operating Environment. The next level of virtualization is Oracle Solaris Zones, a feature of Oracle Solaris 11. We can deploy Oracle Solaris 11 zones, but we are capable deploying legacy zones as wellToday, legacy zones can be configured to support applications running on Oracle Solaris 8 and 9, and Oracle Solaris Trusted Extensions. The beauty of these legacy zones is this platform being capable absorbing workloads who’s deployment started as early as Feb 2000, turning the M6-32 intro a consolidation as well as migration platformEach instance of Oracle Solaris has various resource management tools that are very useful for managing many applications on a single server
  • Dynamic domains have been the pre-eminent partitioning/virtualization technology on Sun’s Enterprise class SPARC systems since the mid-90’s. The Cray CS6400, the predecessor to the Sun E10K had three dynamic system domains that were well-received by Cray’s customers. When the Cray SPARC group was purchased by Sun in 1996 improved dynamic domain technology was introduced on Sun’s E10K 64 processor servers. Domains became a key enabling technology for consolidation as they provided the partitioning needed. With the introduction of the M-series servers “dynamic system domains” were renamed “dynamic domains”. With each new iteration of domain technology there have been improvements and they are now a well-established partitioning technology. About 75% - 80% of all highend SPARC Enterprise customers use dynamic domains to consolidate applications.Dynamic domains are very different from software or hypervisor-based virtualization technologies but have some significant advantages such as complete fault isolation, service isolation, resource isolation and security isolation. A domain can be brought down completely and not affect any other domains. It is not possible to log into one domain and access resources in any other domain (nor it is possible to access the service processor from a domain). Because the domains are hardware-based they have no overhead no matter how many domains or how large the domains are. The biggest improvement in the new M-series domains is granularity down to 1 processor.Dynamic domains have proven their value and worth for the last 15 years and make the M-series systems great consolidation platforms.Dynamic DomainsSPARC Enterprise M-Series Servers (except the M3000) have a unique hardware based virtualization implementation called Dynamic Domains. Dynamic Domains are included, a no cost in every M-Series server from the M4000 to the M9000. Domains were first introduced into Sun servers in the mid-1990’s and have been used by the majority of Sun’s midrange and high end customers to virtualize server resources achieving high levels of system utilization. Domains divide a system’s total resources into separate partitions that are created by linking hardware components together that are electrically isolated from all other hardware components in the server. Domains are a collection of hardware components controlled by the System Controller ASIC (SC). The SC, not software, maintains the links between components that form each Domain. There is no Hypervisor layer so Domains have no performance overhead. Each Domain has its own processors, memory, boot disks and its own instance of the Solaris OS. The I/O interfaces to network and disk resources are also unique to a given domain. The size of a domain can vary from one processor, one memory bank with eight DIMMs and one I/O switch with two PCIe slots up to and including all available resources in an M-Series server. The number of domains varies by server from two domains in an M4000 up to 24 Domains in a M9000. Domains are managed entirely by the M-Series service processor (XSCF) which is also used to monitor the M-Series platform.  Sun SPARC Enterprise M-Series servers have a feature known as Dynamic Reconfiguration (DR) that enables the movement of processors, memory, and I/O resources from one Domain to another without the need for downtime. DR can be used to resize domains to meet changing workloads. For example, a server with 3 domains might have one domain that is underutilized while another domain is running out of processor power. Using DR, a processor/memory/IO group (called an XSB) from the underutilized domain can be removed (virtually) from that domain and added (virtually) to the busy domain. DR enables flexibility in meeting workload requirements and improves system utilization. DR is also used to remove and add hardware components from a Domain for replacement, re-configuration or service without having to bring down that domain. If for some reason a Domain needs to be brought down all other domains in the server continue to operate normally because of the complete electrical isolation of Domains. A Domain that has a errant application that is using all the resources of that Domain will not affect any other Domain. Furthermore, Domains have complete security isolation meaning it is not possible for a user or an application in one domain to access any other domain. This allows many security sensitive applications to exist on the same server. In summary, Domains are hardware partitions that have complete resource, security, fault and service isolation and have no performance overhead. Domains are managed using the same XSCF service processor that manages the platform. Domains enable users to virtualize their SPARC Enterprise M-Series resources to improve system utilization, improve server availability and increase application flexibility.
  • Oracle VM Server for SPARC, previously called Sun Logical Domains, provides highly efficient, enterprise-class virtualization capabilities for Oracle’s SPARC T-Series servers. Oracle VM Server for SPARC leverages the built-in hypervisor to subdivide system resources (CPUs, memory, network, and storage) by creating partitions called logical (or virtual) domains. Each logical domain can run an independent operating system. Oracle VM Server for SPARC provides the flexibility to deploy multiple Oracle Solaris operating systems simultaneously on a single platform. This is the virtualization solution that fully optimizes Oracle Solaris and SPARC for your enterprise server workloads.
  • A no-cost option combining Oracle Solaris Zones and Oracle Solaris Resource Manager, Oracle Solaris Containers extend simple resource management by integrating it with OS partitioning. As a result, each application is able to have its own security and fault domain, name-space, locale, file-system, storage, and network resources. This in turn allows multiple applications to have what appear to be their own operating system instances, with resources such as CPU and memory allocated to them according to policies set with Oracle Solaris Resource Manager.  Different application administrators can be given root access to the containers they “own” without the risk of those administrators intentionally or accidentally escalating their privileges and affecting applications running in other containers. Fault isolation restricts the propagation of software faults to a single container. If an error causes a container to fail, it can reboot in just a few seconds because the underlying single OS instance runs intact.  A branded zone is a special type of container that can provide the illusion that it’s running a different operating system than the one controlling the hardware. Today, branded zones can be configured to support applications running on Oracle Solaris 8 and 9, and Oracle Solaris Trusted Extensions.The key benefit of Oracle Solaris Containers is the efficiency they afford. The overhead of running applications is nearly zero, as compared to virtual machine monitors that must trap every privileged instruction to create the illusion of each OS instance running on dedicated hardware.  Unlike virtual machines, the use of a single OS instance to support multiple containers means that less memory is required: Using Oracle Solaris Containers and sparse-root container configurations, all OS modules and commands share the same pool of memory. This means that even if multiple instances of a command are running, they all share the same executable code.  Oracle Solaris Containers are a preferred solution when multiple Oracle Solaris applications can share the same OS instance. If hardware fault isolation is required, along with OS isolation, consider using Dynamic Domains or Oracle VM SPARC (LDOMS).Oracle Solaris Containers can run on any platform running Oracle Solaris 10 (whether SPARC, UltraSPARC, or x64 servers), and they can be used in either a stand-alone instance of Oracle Solaris or in conjunction with any of the other virtualization technologies discussed in this paper. For example:Oracle Solaris Containers can be used in an Oracle Solaris OS instance running in a virtual machine or a Dynamic Domain. Oracle Solaris Resource Manager can be used to enforce fair-sharing of resources across containers as well as to manage the resources within a container.
  • ConclusionOracle continues to demonstrate its willingness to invest heavily in Solaris and SPARC, and customers that remain on these platforms have increasingly attractive solutions available to them. The general contraction of the Unix market continues to set a macro-level story that appears negative, but Oracle's investments in Solaris clearly bucks that trend.The company's willingness to embrace the best of open source for use in Solaris, through the integration of OpenStack, is a good indication of how the company plans to compete going forward. To the extent that Oracle Solaris can integrate with OpenStack private and public clouds, it expands the opportunities for Oracle. Further, given the dramatic performance improvements coming from the SPARC T5 and M6 processors, the company has a strong baseline platform for all of its products.
  • Add Benchmark info here (SAE 021814 slide 13)
  • One of the key values of an engineered system is that it’s pre-built and optimized, with all the components tested together long before they arrive at a customer data center.
  • Make this a family slide
  • As a simple and high-level overview, SuperCluster can be thought of as a combination of Exadata and Exalogic.
  • Data Sheet – http://www.oracle.com/us/products/servers-storage/servers/sparc/supercluster/supercluster-t5-8/oracle-supercluster-t5-8-ds-1964480.pdf
  • Key points to highlight:Initial products offered will have full configs – later in 2013 will allow lower mem densitiesNew hot plug IO carriers – different than EMs in T5-4 – allow use of standard LP cards and F40 flash acceleration
  • Comparable to 10,000 disks on 100 array frames20X more writes than previous Exadata version
  • Oracle SuperCluster T5-8 delivers up to 10X data compression reducing or eliminating the need to add more expensive SAN-based storageHybrid columnar compression support on ZFS
  • Apps too
  • While not high in IOPS due to a single tray, it is redundant with dual controllers.
  • Manageability through Ops CenterTransforming Complexity Into SimplicityOracle handles it; doesn’t become your problemIncluded with Oracle Premier supportNo financial barrier to effective managementAlways there, allows Oracle to leverage it to deliver simplicitySimple and Flexible ManagementManages all Oracle SuperCluster componentsSupports multiple levels of virtualizationDomains based on Oracle VM Server for SPARCOracle Solaris Zones and Oracle Solaris ContainersOps Center is included with every SuperCluster. Ops Center simplifies monitoring of SuperCluster and provides Oracle with a foundation to build upon for future manageability enhancements.
  • Oracle hardware and software are not only engineered to work together, they are engineered to be maintained, updated, and supported together. We are uniquely qualified to provide optimized performance at every level of the integrated stack,delivering the essential services and resources your business needs to maintain high availability, increase operational efficiency, and gain competitive advantage.Oracle Premier Support provides fully integrated system support with a single point of accountability… 24/7 support with access to Engineered Systems experts 2 hour onsite response Updates, upgrades and support for the Oracle operating system, database and integrated server and storage hardware. Access to My Oracle Support portal which contains a 1 million article database and many, many proactive tools to help you keep systems running at peak performance Oracle Automated Service Request – where your system phones home to Oracle to let us know if there is a problem with the hardware…and now, qualifying customers can also receive the enhanced coverage of Oracle Platinum Services for no additional cost.Oracle Platinum Services is a special entitlement under Oracle Premier Support, delivered at no additional cost.It’s exclusively available on Oracle Exadata, Exalogic and SuperCluster based on certified configurations.It provides 24/7 Oracle remote fault monitoring Backed by extremely fast response times:5 Minute Fault Notification15 Minute Restoration or Escalation to Development30 Minute Joint Debugging with Development And, quarterly patching deployed by OracleOracle Platinum Services takes standard support to a whole new level with additional, no cost services targeted to delivering high availability.To learn more about Oracle Platinum Services go to: www.oracle.com/goto/platinumservices
  • Oracle hardware and software are not only engineered to work together, they are engineered to be maintained, updated, and supported together. We are uniquely qualified to provide optimized performance at every level of the integrated stack, delivering the essential services and resources your business needs to maintain high availability, increase operational efficiency, and gain competitive advantage.Oracle Premier Support provides fully integrated system support with a single point of accountability… 24/7 support with access to Engineered Systems experts 2 hour onsite response Updates, upgrades and support for the Oracle operating system, database and integrated server and storage hardware. Access to My Oracle Support portal which contains a 1 million article database and many, many proactive tools to help you keep systems running at peak performance Oracle Automated Service Request – where your system phones home to Oracle to let us know if there is a problem with the hardware…and now, qualifying customers can also receive the enhanced coverage of Oracle Platinum Services for no additional cost.Oracle Platinum Services is a special entitlement under Oracle Premier Support, delivered at no additional cost.It’s exclusively available on Oracle Exadata, Exalogic and SPARC SuperCluster based on certified configurations.It provides 24/7 Oracle remote fault monitoring Backed by extremely fast response times:5 Minute Fault Notification15 Minute Restoration or Escalation to Development30 Minute Joint Debugging with Development And, quarterly patching deployed by OracleOracle Platinum Services takes standard support to a whole new level with additional, no cost services targeted to delivering high availability.To learn more about Oracle Platinum Services go to: www.oracle.com/goto/platinumservices
  • This shows the scope of remote patch deployment at a high level. For example, for Oracle Exadata, Platinum Services includes deploying the bundle patch or quarterly full stack down to system.While there is a limit to the number of databases that are patched under Platinum Services, ALL databases on the covered system are monitored. The gateway is patched quarterly and as needed to address critical security updates.Be sure to take a look at the Remote Patch Deployment Scope document which provides more details on the components that will be patched under Platinum Services.
  • This figure shows the ultimate solution for providing the ability to restore IT operations while spendingless for backup and recovery. Traditional solutions require expensive backup and recovery equipmentto restore the production system as quickly as possible. An alternative solution is to have a secondremote site with replicated applications and data. Using primary and secondary remote sites eliminatesthe need for high performance systems to recover production operations. Instead of deploying aOracle SuperCluster and a high-performance disk array to back it up, a secondary Oracle SuperClusterwith Oracle's lower cost Sun ZFS Storage Appliance or tape can be deployed. In the event of a failure,production can be switched over immediately to the remote site. The primary site's recovery time isirrelevant because production continues on the secondary systems. This approach provides betterbusiness continuance and results in significant savings because backup and recovery system costs arelower.When not running production operations as part of a business continuance effort, the remotereplicated location can be used for testing, development, and other functions. Oracle is developing theability to generate reports from a copy of the secondary site's production database, enabling IT staff tooffload work from the production system and keep the secondary database current. This solutionprovides savings for backup and recovery and maximizes system availability. Many organizationsdistribute their recovery systems in this manner, with applications that can fail over from city to city. Infact, this design can be used almost universally, with the exception of firms that are required to havelarge and elaborate backup and recovery systems for compliance reasons.
  • This is the FIRST slide must be shown with each presentation that show benchmark results!
  • This is the SECOND slide must be shown with each presentation that show benchmark results!

Sparc SuperCluster Presentation Transcript

  • 1. Copyright © 2014, Oracle and/or its affiliates. All rights reserved.1 SPARC/Solaris & SuperCluster Francisco Navarro Mayo 2014
  • 2. Copyright © 2014, Oracle and/or its affiliates. All rights reserved.2 Evolution of Information Technology Heterogeneous Application Silos Standardization, Consolidation, In frastructure as a Service Platform & Software as a Service 1990’s 2000’s 2010’s
  • 3. Copyright © 2014, Oracle and/or its affiliates. All rights reserved.3 Investment in Core Technology PROCESSORS SOFTWARE BANDWIDTH SYSTEMS  SPARC microprocessors are the fastest in the world, and continue to lead the industry into the multi- core era  Oracle Solaris is the leading enterprise cloud OS, many years ahead in scalability and reliability  SPARC servers reach incredible scale inside the server, or outside, through leading- edge InfiniBand technology  Oracle’s portfolio today has the performance and price / performance advantage, and offers breakthrough virtualization for business critical applications
  • 4. Copyright © 2014, Oracle and/or its affiliates. All rights reserved.4 VirtualizationNetwork Server ServerOS The Traditional Way of Engineering the OS Changing the Design Focus
  • 5. Copyright © 2014, Oracle and/or its affiliates. All rights reserved.5 Virtualization Drives Management Costs 0 10 20 30 40 50 60 70 80 90 $0 $50 $100 $150 $200 $250 $300 '96 '97 '98 '99 '00 '01 '02 '03 '04 '05 '06 '07 '08 '09 '10 '11 '12 '13 Servers (M)Spending ($B) Physical Servers (M) Virtual Servers (M) Management & Administration
  • 6. Copyright © 2014, Oracle and/or its affiliates. All rights reserved.6 Oracle Investment in Systems Oracle investment in:  Software-in-Silicon  Zero-Overhead Virtualization  Agile, Secure, Compl iant Life Cycle Management  Engineered Systems IaaS PaaS SaaS 10X Faster, 16X Higher Productivity, 10X Less Expensive
  • 7. Copyright © 2014, Oracle and/or its affiliates. All rights reserved.7 Middleware Database Application Virtualization & Cloud Orchestration SDN Server Platform A Better Way of Engineering the System Hardware and Software Engineered to Work Together High Performance, High Efficiency, High Value
  • 8. Copyright © 2014, Oracle and/or its affiliates. All rights reserved.8 1 128 2187 16384 78125 279936 823543 1 2 3 4 5 6 7 The Oracle Stack – drives down Complexity A workload optimised stack replaces homegrown Multivendor Stacks # of different Products p. Layer # of different Stack Variants . . . . . .
  • 9. Copyright © 2014, Oracle and/or its affiliates. All rights reserved.9 Engineered Together Best performance, availability, security, efficiency Tested Together 150,000 machine hours every week 10M+ machine hours annually Certified Together We take the risk for you Deployed Together Your application up in hours, not months Upgraded Together Risk free patching/updates Managed Together 3x lower administration and maintenance costs Supported Together Coordinated response for your deployment One Engineering Team
  • 10. Copyright © 2014, Oracle and/or its affiliates. All rights reserved.10 Engineered Systems: Breakthrough System Approach Extreme Application Performance System Software Your Application
  • 11. Copyright © 2014, Oracle and/or its affiliates. All rights reserved.11 Oracle SPARC Systems Direction Increase Application Performance 2x Every Two Years  Design the best servers in the world. Period.  Embed Oracle-specific enhancements  Lead Mission Critical computing  Not just about servers: Integrate into Engineered Systems
  • 12. Copyright © 2014, Oracle and/or its affiliates. All rights reserved.12 SPARC Establishes Dramatic Performance Leadership 2010 2011 2012 2013 2014 SPARC T-Series IBM Power HP/Cisco x86 SPARC T5 SPARC T4 SPARC T3 IBM Power7 Sandy Bridge Xeon E5 Enterprise performance SPARC: over 2x performance each generation x86: only ~20 to 50% each generation Power7+: only ~10% in 3 years Westmere Xeon E7 IBM Power7+ Nehalem Per chip based on internal testingSPARC redefines economics! Ivy Bridge Xeon E5 v2
  • 13. Copyright © 2014, Oracle and/or its affiliates. All rights reserved.13 M9000 2.88GHz M9000 3.0GHZ T4 T5 Power 6 Power 7 0 20000 40000 60000 80000 100000 120000 SPARC Power Record Breaking Database Performance Per-Processor Performance Trends (TPC-H@3000GB) Oracle Acquisition
  • 14. Copyright © 2014, Oracle and/or its affiliates. All rights reserved.14 Best Platform for Java SPECjbb2013 MultiJVM/socket  Large Page Support  Lambda Expressions  Java Locking Re- Architecture  SLA Integration  Zero Overhead Virtualization  Java GC Support in our Next Generation SPARC 2011 2013 2014 16,000 21,000 26,000 31,000 36,000 41,000 46,000 51,000 56,000 61,000 T4/Solaris 11/Java 7 T5/Solaris 11/Java 7 T5/Solaris 11/Java 8 X86 SNB/RHEL6/Java 7 X86 IVB/RHEL6/Java 7 X86 IVB/RHEL6/Java 8
  • 15. Copyright © 2014, Oracle and/or its affiliates. All rights reserved.15  New Oracle Solaris/RAC Kernel Mode Acceleration  Oracle Solaris responds directly to lock requests  In memory lock state shared by database and kernel Performance Optimizations for Oracle RAC Consistent RAC performance 30-40% lower latency lock grants Up to 20% higher throughput Oracle Solaris Oracle RAC Oracle Solaris Oracle RAC
  • 16. Copyright © 2014, Oracle and/or its affiliates. All rights reserved.16  Added DTrace I/O Event Tracing to Oracle Database 12c  V$ views for simple queries  Avoid complex and time- consuming report snapshot and diagnosis Simple Database I/O Diagnosis and Tuning Find I/O Outliers v$kernel_io_outlier TIMESTAMP IO_SIZE IO_OFFSET DEVICE_NAME VARCHAR2(513) PROCESS_NAME VARCHAR2(64) TOTAL_LATENCY SETUP_LATENCY … See demo: https://blogs.oracle.com/solaris/entry/data base_12c_and_solaris_dtrace
  • 17. Copyright © 2014, Oracle and/or its affiliates. All rights reserved.17 Java Mission Control Visualizing DTrace data
  • 18. Copyright © 2014, Oracle and/or its affiliates. All rights reserved.18 Price/Performance of SPARC Redefined $0 $2 $4 $6 $8 $10 $12 $14 $16 $18 $20 M9000 T5 $/Perf 78% lower Dramatic Improvement with T5
  • 19. Copyright © 2014, Oracle and/or its affiliates. All rights reserved.19 The Ultimate Software Optimization: Hardware Moving Oracle Database & Java Software Functions into Hardware Database Query Acceleration Application Data Protection Java Acceleration Data Compression / Decompression
  • 20. Copyright © 2014, Oracle and/or its affiliates. All rights reserved.20 2000 2001 2002 2003 2004 2005 2006 2007 2008 2009 2010 2011 2012 2013 20141999 Hardware Investment Profile Investment vs. Divestment *later acquired networking specifically for BladeCenter Sold networking business to Cisco* Sold HDD business to Hitachi Sold PC business to Lenovo Sold printing business to Ricoh Sold POS business to Toshiba Sold x86 Server business to Lenovo Divestment Investment IBM Microprocessors for Sale?
  • 21. Copyright © 2014, Oracle and/or its affiliates. All rights reserved.21 Oracle SPARC Processor Roadmap Solaris 11.xSolaris 11.1 Solaris 11.x Solaris 11.x Solaris 12 Solaris 11 T4 +1x Throughput +5x Thread Strength M5 +6x Throughput +1.5x Thread Strength M6 +2x Throughput >1x Thread Strength In Test M & T Series +2x Throughput +1.5x Thread Strength T5 +2.5x Throughput +1.2x Thread Strength 2011 2012 2013 2014 2015 2016 Delivered Software in Silicon •Database Query •Decompression •Encryption •Application Data Integrity •Java Optimization •Low Latency Clustering 2018 Oracle Numbers Optimization M & T Series +1.3x Throughput +1.5x Thread Strength Solaris 12.x 2017 2019 Software in Silicon Enhancements •Database Query+ •Java Optimization+ •Encryption+ •Low Latency Clustering+ Core Enhancements Increased Cache Increased Bandwidth Solaris 12.x Next Gen Core Planned
  • 22. Copyright © 2014, Oracle and/or its affiliates. All rights reserved.22 Best Enterprise Portfolio T4-1 T5-1B T5-2 T5-4 SuperCluster T5-8 and M6-32 Exalytics T5-8 Entry-Level Mid-Range High-EndEngineered Systems M6-32, M5-32 T5-8
  • 23. Copyright © 2014, Oracle and/or its affiliates. All rights reserved.24 SPARC T5-1B SPARC T5-2 SPARC T5-4 SPARC T5-8 Processor SPARC T5 3.6GHz SPARC T5 3.6GHz SPARC T5 3.6GHz SPARC T5 3.6GHz Max Processor Chips 1 2 4 8 Max Cores/Threads 16, 128 32, 256 64, 512 128, 1024 DIMM Slots 16 32 64 128 Max Memory 128GB, 256GB, 512GB 256GB, 512GB, 1TB 1TB or 2TB 2TB or 4TB Drive Bays 2 6 8 8 I/O Slots 2 x PCIe 2.0 EM, 2 NEM,1 FEM slots 8 LP x8 PCIe 3.0, 4 x 10GbE ports 16 LP x8 PCIe 3.0, 4 x 10GbE ports 16 LP x8 PCIe 3.0, 4 x 10GbE ports Form Factor/RU Blade Rack 3RU Rack 5RU Rack 8RU Max Power Consumption 689W 1927W 2410W 4850W SPARC T5 Servers Product Line Overview
  • 24. Copyright © 2014, Oracle and/or its affiliates. All rights reserved.25 Powered By Oracle Processor Technology SPARC M6: Higher Density and Lower Cost / Core  New: 12 S3 cores @ 3.6GHz  Large 48MB shared L3 Cache  Scalable to 32 processors  Integrated 2x8 PCIe 3.0  Integrated ISA-based crypto acceleration 2X THROUGHPUT PERFORMANCE OF M5
  • 25. Copyright © 2014, Oracle and/or its affiliates. All rights reserved.26 Oracle: Re-engineering Systems Economics WorseBetter Near Linear Pricing $/UnitofPerformance T5-4 $147,992 T5-8 $268,314 M6-32 $1,209,943 T5-2 $67,042 IBM Power 750 POWER7+ $204,982 IBM Power 780 POWER7+ $2,101,370 IBM Power 795 POWER7 $5,761,583 IBM Power 740 POWER7+ $101,571 2 Socket 4 Socket 8 Socket 32 Socket WorseBetter Note: Data as of 04/01/14
  • 26. Copyright © 2014, Oracle and/or its affiliates. All rights reserved.27 Removing the High End Pricing Premium Similar Price/Similar Performance 16 x SPARC T5-2: $1,073,000 1 x SPARC M6-32: $1,210,0004 x SPARC T5-8: $1,082,000
  • 27. Copyright © 2014, Oracle and/or its affiliates. All rights reserved.28 New l SPARC M6 Servers  Compute – Up to 32x SPARC M6 12 core 3.6GHz CPUs – Allows mixing with SPARC M5 6 core processors – Up to 1024x DDR3 DIMMs for max memory of up to 32TB  I/O and storage – 32x 2.5” SAS-2 internal drives – 64x PCIe Gen3 low profile internal slots  Scalability and investment protection – Upgradable with M6 processor  Availability and management – Advanced RAS with redundant and hot swap of key components – Extensive virtualization with Oracle VM Server for SPARC + hard partitions – Integrated Oracle ILOM system management 32-Socket High-End Datacenter Server
  • 28. Copyright © 2014, Oracle and/or its affiliates. All rights reserved.29 SPARC Server Availability M6-32 / M5-32: Highest Availability in the SPARC Server Portfolio RAS T4-4 T5-8 T5-4T5-2 T4-2 T4-1 M6-32 M5-32 VM Secure Live Migration Redundant / Hot-swap Fans, Power, Disks ECC, Predictive Self Healing Physical Partitions / Redundant Clock / Redundant & Hot-Plug SP Hot Plug PCIe Lane Sparing (L3$/Memory/Interconnect) + Redundant Interconnect
  • 29. Copyright © 2014, Oracle and/or its affiliates. All rights reserved.30 Built In, Zero Cost The Oracle Virtualization Advantage Enterprise Class Virtualization for SPARC and x86 Systems Oracle App Support Rapid App Deployment
  • 30. Copyright © 2014, Oracle and/or its affiliates. All rights reserved.31 SPARC & Solaris Virtualization Options Physical Partitions SPARC Hypervisor Oracle Solaris Zones Multiple OSes Single OS
  • 31. Copyright © 2014, Oracle and/or its affiliates. All rights reserved.33 SPARC Physical Domains 33 • Complete isolation • Resource, security, service, fault • No overhead • Separate OS per domain • No cost to end user Domain 1 Domain 2 Domain x CPU CPU CPU CPU CPU CPU CPU CPU CPU Mem Mem Mem Mem Mem Solaris Solaris Solaris
  • 32. Copyright © 2014, Oracle and/or its affiliates. All rights reserved.34 Oracle VM Server for SPARC (LDoms) • One OS instance for each LDom • Different patch levels for each LDom • Complete software isolation • Single-thread granularity • Dynamic • Secure Live Migration • Low overhead • No cost to end user Hypervisor LDom 1 LDom 2 LDom x CPU CPU CPU CPU CPU CPU CPU CPU CPU Mem Mem Mem Mem Mem Solaris Solaris Solaris
  • 33. Copyright © 2014, Oracle and/or its affiliates. All rights reserved.35 Oracle Solaris Zones • One OS instance for all zones • Separate file system • Complete software isolation • Sub-thread granularity • Dynamic and mobile • Low overhead • No cost to end user • All Solaris instances Solaris Zone 1 Zone 2 Zone x CPU CPU CPU CPU CPU CPU CPU CPU CPU Mem Mem Mem Mem Mem
  • 34. Copyright © 2014, Oracle and/or its affiliates. All rights reserved.36 SPARC M6-32 Virtualization Dynamically Resize LDoms and Zones for Best Utilization M6-32 Domain LDom A LDom B OLTP DB OLTP DB App App DW DB LDom A LDom B OLTP DB OLTP DB DW DB Zones inside LDoms App App OLTP DB M6-32 Domain M6-32 DomainM6-32 Domain OLPT DB Zones inside LDoms
  • 35. Copyright © 2014, Oracle and/or its affiliates. All rights reserved.37 Oracle Accelerates Momentum of Solaris Investment Al Gillen, Gary Chen, Peter Rutten May 06, 2014  Oracle continues to demonstrate its willingness to invest heavily in Solaris and SPARC, and customers that remain on these platforms have increasingly attractive solutions available to them. The general contraction of the Unix market continues to set a macro- level story that appears negative, but Oracle's investments in Solaris clearly bucks that trend.  The company's willingness to embrace the best of open source for use in Solaris, through the integration of OpenStack, is a good indication of how the company plans to compete going forward. To the extent that Oracle Solaris can integrate with OpenStack private and public clouds, it expands the opportunities for Oracle. Further, given the dramatic performance improvements coming from the SPARC T5 and M6 processors, the company has a strong baseline platform for all of its products.
  • 36. Copyright © 2014, Oracle and/or its affiliates. All rights reserved.38  Simple Administration  Leverages Solaris resource management and network virtualization  Seamless P2V and V2P  Locked-down root file system for both guest and host  Both SPARC and x86 Next Generation Virtualization New in S11.2: Solaris Kernel Zones OS and Virtualization – Engineered Together InfiniBand Fabric 10GbE Network S11.2 Virtual Router S11.2 S11.3 S11.4
  • 37. Copyright © 2014, Oracle and/or its affiliates. All rights reserved.39  PCIe Standard  Requires support on the card  Native I/O performance  Available today for – Ethernet – Infiniband – SAN  Dependency on Root Complex owner  With LDoms 3.1, this is dynamic! PCIe Virtualization OVM: SR-IOV
  • 38. Copyright © 2014, Oracle and/or its affiliates. All rights reserved.40 Solaris Roadmap ACCELERATING THE PACE Solaris 11 Solaris 11.1 Solaris11.2 Solaris 11.3 Solaris 11 … Solaris 12 Solaris 10 Update 9 Solaris 10 Update 10 Solaris 10 Update 11 2010 2012 2013 2014 … … T-3 T-4 T-5 T-7 M-7M-6M-5M-3 2011 X-4X-3X-2
  • 39. Copyright © 2014, Oracle and/or its affiliates. All rights reserved.41 Oracle Enterprise OpenStack Compute Virtualization Cloud Networking Cloud Storage Image Deployment Cloud Management Cloud APIs Elastic Virtual Switch ZFS File System Unified Archives Nova Neutron Cinder & Swift Glance Horizon Zones & Kernel Zones
  • 40. Copyright © 2014, Oracle and/or its affiliates. All rights reserved.42 Oracle Engineered Systems Performance AffordabilitySimplicity Exalogic Elastic Cloud Big Data Appliance Exalytics SuperCluster Exadata Backup, Log ging and Recovery Appliance Database Appliance Virtual Compute Appliance
  • 41. Copyright © 2014, Oracle and/or its affiliates. All rights reserved.43 Yesterday There is a Better Way I Engineered Systems Collection of vendors First integration is in your data center Complex matrix of products Staff busy keeping the lights on Engineered to work together One point of accountability Tested & proven architecture     Staff focused on strategic projects Tomorrow
  • 42. Copyright © 2014, Oracle and/or its affiliates. All rights reserved.44 1,000s of Customers In Production Existing Applications, Spectacular Results
  • 43. Copyright © 2014, Oracle and/or its affiliates. All rights reserved.45 Months PerformanceAchievement Assemble dozens of components Measure, diagn ose, tune and reconfigure Measure, diagn ose, tune and reconfigure Multi-vendor testing Test and debug production failure modes Live to production Build-It-Yourself Approach Oracle SuperCluster Days PerformanceAchievement Consolidation Accelerated Eliminate the Delays of Custom Assembly
  • 44. Copyright © 2014, Oracle and/or its affiliates. All rights reserved.46 Optimize All Workloads on a Private Cloud Oracle SuperCluster with Exadata Hardware and Exalogic Software COMPLETE Compute Storage Software Networking + + The best system for running databases and applications in memory Most scalable system for consolidation and cloud services Secure multi-tenant virtualization SuperCluster T5-8 SuperCluster M6-32
  • 45. Copyright © 2014, Oracle and/or its affiliates. All rights reserved.47 Oracle SuperCluster Oracle's Most Powerful Engineered Systems  World Record Performance  Breakthrough Efficiency and Scalability  Superior Affordability and ROI Oracle’s Hottest Technology  Exadata Storage Servers  Oracle M6 and T5 Microprocessors  24-Terabit Silicon Network  Oracle VM Server and Solaris Virtualization M6-32T5-8
  • 46. Copyright © 2014, Oracle and/or its affiliates. All rights reserved.48 SuperCluster Customer Overview Deployments by Originating/Competing Platform • IBM Replacement • Financial trading application and Oracle Database 11gR2 • 3x higher throughput • 4x higher trades • 30x consolidation ratio • X86 / Linux Replacement • Oracle E-Business Suite and Oracle Database 11gR2 • 22x system consolidation • 9x faster data operations • Deployment in 63 days Over 50% of customers are competitive take-outs or new Deployments 26% 18%47% 9% HP IBM SPARC New Oracle
  • 47. Copyright © 2014, Oracle and/or its affiliates. All rights reserved.49 What is Oracle SuperCluster?  SPARC servers – Better performance than IBM Power solutions  Exadata and ZFS Storage – Flash Cache – Executes Oracle Database and I/O operations  Low Latency Network Fabric – Massive bandwidth enabling the fastest network  Management Software – Cloud ready, zero overhead virtualization Best of Breed Technology Engineered to Work Together
  • 48. Copyright © 2014, Oracle and/or its affiliates. All rights reserved.50 Conceptual View of SPARC SuperCluster Exadata + Exalogic Heritage Exadata SPARC SuperCluster Exalogic Exadata Storage Database Servers ZFS NAS Storage Application Servers
  • 49. Copyright © 2014, Oracle and/or its affiliates. All rights reserved.51 Oracle SuperCluster Extending the Product Family  Best Price / performance  High consolidation ratios  Run database and applications on a single rack  Scalable configurations  Exadata database optimizations  Exalogic software and other application optimizations  Layered Optimized Virtualization  Virtual Tuning Assistant SuperCluster M6-32SuperCluster T5-8 All SuperCluster T5-8 benefits, PLUS…  Vertical scaling  In-memory applications using big memory capacity  Oracle Database12c and applications optimized  Mainframe-class RAS
  • 50. Copyright © 2014, Oracle and/or its affiliates. All rights reserved.52 SuperCluster SuperCluster T5-8 SuperCluster M6-32 Half Rack Full Rack Minimum Maximum Processors 8 16 16 32 Memory 2TB 4TB 8 TB 32 TB Exadata Storage Servers 4 8 9 Oracle SuperCluster At A Glance
  • 51. Copyright © 2014, Oracle and/or its affiliates. All rights reserved.53 Oracle SuperCluster T5-8  2 x SPARC T5-8 (128 cores, 2TB memory) – Each T5-8 configured with:  4 x SPARC T5 processors, 64 cores  1 TB memory  4 x InfiniBand HCAs (dual port)  4 x 10GbE NICs (dual port)  8 x 900GB SAS disks  Optional Fiber channel cards  1-4 Domains Half Rack Configuration  Storage – 4 Exadata Storage Servers (192TB HC or 57.6 TB HP disks) – ZS3-ES (80 TB disk and 292 GB write-optimized SSD)  Switches – 3 x InfiniBand 36 port switches – GbE Management switch
  • 52. Copyright © 2014, Oracle and/or its affiliates. All rights reserved.54 SPARC T5-8 Server  Compute – 4x or 8x SPARC T5 16-core CPU (1024 threads max) – 64x or 128x 16 GB DDR3 DIMMs  I/O and Storage – 16x PCIe 3.0 x8 slots, with PCIe carriers – 4x 10GbE ports built-in for management – 8x or 16x 10GbE ports for data access – 8x 2.5” SAS 3.0 drives  Availability and Management – RAID 0/1, Hot-swap fans & PSUs – Oracle Integrated Lights Out Manager Service Processor – Oracle Enterprise Manager Ops Center The Industry’s First Multithreaded Enterprise Datacenter Server for Mission Critical Workloads
  • 53. Copyright © 2014, Oracle and/or its affiliates. All rights reserved.55 SPARC T5 Processor  Features • 16 S3 cores, 16-128 Strands @ 3.6Ghz • Single or multi-threaded operation per core • System scalability to 8 sockets  SPARC Core S3 • 1-8 Strand Dynamically Threaded Pipeline • ISA-based Crypto-acceleration  8MB Shared L3$  Integrated I/O • Double I/O bandwidth over T5 • 2 x8 Lane PCIe 3.0 @ 8GT/s  System Scalability • 7 Coherence Ports for scalability to 8S  Power Management • Dynamic Voltage Frequency Scaling • Downclock, Overclock
  • 54. Copyright © 2014, Oracle and/or its affiliates. All rights reserved.56 Database Machines I Exadata Storage Server • Intelligent Scale-Out Storage • Optimized Oracle Database query processing and OLTP • Optimized Oracle Database CompressionExadata SuperCluster Exadata Storage Server
  • 55. Copyright © 2014, Oracle and/or its affiliates. All rights reserved.57 Exadata Storage Servers  Exadata Storage Servers run Oracle Enterprise Linux and Exadata Storage software. They currently consist of the following hardware: – X4-2L with:  2 sockets Xeon E5-2630 6 cores @ 2.6GHz 15MB L3  96GB Memory (4 * 8GB + 4 * 16GB DDR3 1600MHz LV DIMMs)  SAS-2 RAID HBA  12 disks, either high performance or high capacity • High Performance: 1.2TB@10Krpm SAS2 SSF 2.5" Drives (total 14.4TB) • High Capacity: 4TB@7200rpm SAS2 3.5" Drives (total 48TB) • 4 x 800 GB Sun Flash Accelerator F80 PCIe cards (total 3.2TB) • 2 InfiniBand 4X QDR (40Gb/s) Ports (PCIe 3.0)  They can only be used by Database Domains  Four are installed in the half rack config; eight are installed in the full rack config They are the EXACT same storage servers as the Exadata Storage Server X4-2L
  • 56. Copyright © 2014, Oracle and/or its affiliates. All rights reserved.58 Key Exadata Innovations + ++  Hybrid Columnar Compression – 10x compression for warehouses – 15x compression for archives Data remains compressed for scans and in Flash Space Savings Cascade to Copies compress primary DB standby test dev backup uncompressed  Smart Scale-Out Storage – InfiniBand connected servers – Smart Scan query offload  Smart PCI Flash Cache – Transparent cache in front of disk – Accelerates random I/O up to 30x – Quadruples data scan rate
  • 57. Copyright © 2014, Oracle and/or its affiliates. All rights reserved.59 Unique Software Optimizes Database Processing  Database offload in storage –Data intensive query operations offloaded to storage CPUs  Database storage compression –Hybrid Columnar speeds analytics –Reduces database size 10x  Database optimized PCI flash –Smart caching of database data –Millions of Database IOs/sec –Smart log writes to flash  Database optimized QOS – End-to-end prioritization from application to DB, and storage  Database optimized availability – Fast detection and recovery of failed database, server, storage or switch – Exachk validation of full hardware, software, settings & best practices  Database optimized messaging – SQL optimized InfiniBand protocol for high throughput low latency SQL
  • 58. Copyright © 2014, Oracle and/or its affiliates. All rights reserved.60 Improved Exadata Storage Server X4-2L Enhanced Smart Flash Cache • New software caches write I/Os in flash memory −20x write improvement −10x write improvement from previous storage servers • Cache is transparent, persistent, and fully redundant
  • 59. Copyright © 2014, Oracle and/or its affiliates. All rights reserved.61 Exadata Smart Flash Log Accelerate Transaction Response Times using Flash  Uses Flash for Database Logs in a clever way – Flash is fast but has slow outliers – Erase cycles, wear leveling, etc.  Smart Flash Log feature transparently uses Flash as a parallel write cache to disk controller cache – Whichever write completes first wins (disk or flash)  Better response time and more throughput  Uses almost no flash capacity (0.1% of capacity) Default (on left) - Choppy response - High Outliers Smart Flash Log - 3x faster response - Much lower outliers Transaction Response Times Smart Flash Log Enabled Automatic and Transparent
  • 60. Copyright © 2014, Oracle and/or its affiliates. All rights reserved.62 The Highest Levels of Data Compression  Results in needing 1/10 as much storage as competitive solutions  Increases effective disk bandwidth by 10X  Compresses even database content – unlike vanilla de-duplication  Drives unparalleled database and application performance Hybrid Columnar Compression (HCC) Drives Data Center Efficiency DATA COMPRESSION 10X DATA COMPRESSION
  • 61. Copyright © 2014, Oracle and/or its affiliates. All rights reserved.63 X4-2 Database Cell IO Performance from SQL X4-2 14 Celdas X4-2 8 Celdas X4-2 7 Celdas X4-2 4 Celdas X4-2 3 Celdas Flash Cache SQL Bandwidth1,3 High Cap Disk 100 GB/s 60GB/s 50 GB/s 29 GB/s 21.5 GB/s High Perf Disk 100 GB/s 60GB/s 50 GB/s 29 GB/s 21.5 GB/s Flash SQL IOPS2,3 8K Reads 2,660,000 1,520,000 1,330,000 760,000 570,000 8K Writes 1,960,000 1,120,000 980,000 560,000 420,000 Disk SQL Bandwidth1,3 High Cap Disk 20 GB/s 12GB/s 10 GB/s 6 GB/s 4.5 G/s High Perf Disk 24 GB/s 14GB/s 12 GB/s 7 GB/s 5.2 GB/s Disk SQL IOPS High Cap Disk 32,000 19,000 16,000 9,500 7,000 High Perf Disk 50,000 28,800 25,000 14,400 10,800 Data Load Rate4 20 TB/hr 10 TB/hr 5 TB/hr 1 - Bandwidth is peak physical scan bandwidth achieved running SQL, assuming no compression. Effective data bandwidth will be much higher when compression is factored in. 2 - IOPS – Based on read IO requests of size 8K running SQL, typically with sub-millisecond latencies. Note that the IO size greatly effects flash IOPS. Others quote IOPS based on smaller IOs that are not relevant for databases and measure IOs using low level tools instead of SQL. 3- Actual Performance varies by application. 4 –Load rates are typically limited by database server CPU, not IO. Rates vary based on load method, indexes, data types, compression, and partitioning
  • 62. Copyright © 2014, Oracle and/or its affiliates. All rights reserved.64 Exadata Storage Summary 4+4 Total Half + Exapansion Half SuperCluster Half SuperCluster (3 Celdas Activas) X4-2 Quarter Expansion Single Cell Exadata Storage Servers 8 4 3 4 1 Storage Grid CPU Cores 96 48 36 48 12 Raw Flash Capacity 25,6 12.8 TB 9,6 TB 12.8 TB 3.2 TB Raw Storage Capacity High Perf 114TB 57 TB 43.2TB 57 TB 14.4 TB High Cap 384TB 192 TB 144TB 192 TB 48 TB Usable mirrored capacity High Perf 50TB 25 TB 19TB 25 TB 6 TB High Cap 170TB 85 TB 63TB 85 TB 20 TB Usable Triple mirrored capacity High Perf 34TB 17 TB 13TB 17 TB 4 .25 TB High Cap 116TB 58 TB 43TB 58 TB 14.5 TB
  • 63. Copyright © 2014, Oracle and/or its affiliates. All rights reserved.65 Exalogic Elastic Cloud Software 2.0  WebLogic JDBC and Data Source  WebLogic Cluster State Replication  WebLogic Socket Direct Protocol  Coherence API support  Oracle Traffic Director EXALOGIC SOFTWARE OPTIMIZATIONS ON SUPERCLUSTER
  • 64. Copyright © 2014, Oracle and/or its affiliates. All rights reserved.66 Extreme Java and Web Performance Run Oracle Fusion Middleware Fastest JAVA MESSAGE SERVICE PERFORMANCE 10X 14XHTTPS PERFORMANCE
  • 65. Copyright © 2014, Oracle and/or its affiliates. All rights reserved.67 Oracle ZFS Storage ZS3-ES  Provides shared 80TB of storage for all Logical Domains – Not mounted by default in the Database Domain. Active mounts are established to all other Domains.  Data can only be accessed in the IB backbone. External access is via built-in GbE ports used exclusively for DR and data replication.  Two controllers for high-availability and Infiniband  292 GB of integrated Flash  Stores quorum device for Solaris Cluster  An additional IB-enabled ZFS SA device can be attached with IB ports connected to both IB Leaf switches  Accessible from Database Domain after manual mount – Can be used for RMAN backups or flat file staging Shared Storage
  • 66. Copyright © 2014, Oracle and/or its affiliates. All rights reserved.68  Layer 1 Virtualization – Domains Domains with zero performance overhead  Layer 2 Virtualization – Domains with zero performance overhead – Any mix of Database Domains and Application Domains  Oracle Solaris 11 or 10 for Application Domains  Layer 3 Virtualization – Both Database Domains and Application Domains support Oracle Solaris Zones M6-32 Dynamic Domain M6-32 Dynamic Domain Oracle SuperCluster M6-32 Virtualization Layered Optimized Virtualization Database Domain 1 Database Domain 2 Application Domain 1 Database Domain 1 Database Domain 2 Application Domain 1 Application Domain 2 DB Zone DB Zone DB Zone Zone Zone DB Zone DB Zone Zone Zone Zone Solaris Cluster Clusterware
  • 67. Copyright © 2014, Oracle and/or its affiliates. All rights reserved.69  Layer 1 Virtualization – Domains with zero performance overhead – Any mix of Database Domains and Application Domains  Oracle Solaris 11 or Oracle Solaris 10 for Application Domains  Layer 2 Virtualization – Both Database Domains and Application Domains support Oracle Solaris Zones T5-8 Compute Node T5-8 Compute Node Oracle SuperCluster T5-8 Virtualization Layered Optimized Virtualization Database Domain 1 Database Domain 2 Application Domain 1 Database Domain 1 Database Domain 2 Application Domain 1 Application Domain 2 DB Zone DB Zone DB Zone Zone Zone DB Zone DB Zone Zone Zone Zone Solaris Cluster Clusterware
  • 68. Copyright © 2014, Oracle and/or its affiliates. All rights reserved.71 Scale How You Want, When You Want Modular building-block design allows rapid expansion, on-demand ADD COMPUTE ADD STORAGE
  • 69. Copyright © 2014, Oracle and/or its affiliates. All rights reserved.72 Invest Today and Protect Tomorrow SuperCluster T5-8 Half to Full Rack Upgrades Simple in-rack upgrades
  • 70. Copyright © 2014, Oracle and/or its affiliates. All rights reserved.73 Deploy Seamlessly Into Your Datacenter Leverage Existing SAN Investments Existing SAN Data
  • 71. Copyright © 2014, Oracle and/or its affiliates. All rights reserved.74 Deploy Seamlessly Into Your Datacenter + ++ Oracle SuperCluster Half or Full Rack SuperCluster M6-32,T5- 8, and T4-4 Exadata Storage Expansion Rack Sun ZFS Storage 7420  Mixed multi-generation scaling Start Small, No Limits  Easily Expand Database Storage  NAS and Fibre Channel Storage
  • 72. Copyright © 2014, Oracle and/or its affiliates. All rights reserved.75  Discover Oracle SuperCluster as a single system  Hardware event interface view of Oracle SuperCluster as a single system  Monitoring and active management tasks separated by role Total Systems Management Oracle Enterprise Manager Ops Center 12c
  • 73. Copyright © 2014, Oracle and/or its affiliates. All rights reserved.76 Oracle SuperCluster Ops Center  Unified management of Servers, Storage, and Network Fabric  Simplified management of virtual infrastructure for easy application consolidation  Instant Network and Storage provisioning  Automated update of all firmware and software components  Direct connection to Oracle knowledge-base speeds problem resolution
  • 74. Copyright © 2014, Oracle and/or its affiliates. All rights reserved.77 Oracle Platinum Services  Higher support level for complete Oracle stack – Server, Storage, Network, Database software  24/7 Oracle remote fault monitoring  Industry-leading response times: – 5 Minute Fault Notification – 15 Minute Restoration or Escalation to Development – 30 Minute Joint Debugging with Development  Patch deployment by Oracle engineers High Availability Services At No Additional Costs
  • 75. Copyright © 2014, Oracle and/or its affiliates. All rights reserved.78 Oracle Support for SPARC SuperCluster Complete. Integrated. Proactive. High Availability Services. No Additional Cost. ORACLE PLATINUM SERVICES 24/7 support coverage with access to Oracle Engineered Systems Enterprise Support Team 2-hour onsite response to hardware issues1 Updates, upgrades and support for: – Oracle Database 11gR2, Oracle Exadata Storage Server software, Oracle Exalogic Elastic Cloud software, and the following certified and tested configurations for Oracle software: PeopleSoft HCM, WebCenter Content, Siebel CRM and Oracle E-Business Suite – Server and storage system hardware, integrated software (such as firmware), Oracle Solaris operating system, Oracle Solaris Cluster, and Oracle VM Server for SPARC 24/7 access to My Oracle Support proactive support portal "Phone home" automated service requests (ASR)  24/7 Oracle remote fault monitoring  Industry-leading response times: – 5 Minute Fault Notification – 15 Minute Restoration or Escalation to Development – 30 Minute Joint Debugging with Development  Patch deployment services  No additional cost under Oracle Premier Support Available now for certified configurations on Oracle SPARC SuperCluster T4-4
  • 76. Copyright © 2014, Oracle and/or its affiliates. All rights reserved.79 Scope of Remote Patch Deployment Oracle Exadata Oracle Exalogic Oracle SuperCluster Bundle Patch or Quarterly Full Stack Download Quarterly Patch Set Update Quarterly Full Stack Download Up to 8 DB and 2 DB homes (Exadata/SuperCluster full rack) Up to 4 DB and 2 DB homes (Exadata/SuperCluster 1/2 rack or smaller) NOTE: All DBs are monitored Oracle Advanced Support Gateway is patched quarterly and as needed to address critical security patches View more details on Remote Patch Deployment here: http://www.oracle.com/us/support/library/platinum-remote-patch-checklist-1958298.pdf Database Patching:
  • 77. Copyright © 2014, Oracle and/or its affiliates. All rights reserved.80 Backup and Recovery Best Practices Maximum Availability Architecture with Backup and Recovery Savings
  • 78. Copyright © 2014, Oracle and/or its affiliates. All rights reserved.81 Oracle Infrastructure as a Service (IaaS) On Premise with Capacity on Demand  Engineered Systems hardware & hardware support for a monthly fee – No upfront capital expenditure, often OPEX not CAPEX – Usually lower total cost than purchasing system  On Premise – Deploys behind customer’s firewall – Control and visibility over the system  Elastic Compute Capacity on Demand1 – For peak workloads and high availability – Pay only when needed  Oracle Platinum and PlatinumPlus Services2 – Higher level of support including patching services – Quarterly advisory services for system security, performance, and availability 1. Eligible systems include Oracle Exadata, Exalogic, Exalytics and SuperCluster. 2. Eligible systems include Oracle Exadata, Exalogic, and SuperCluster.
  • 79. Copyright © 2014, Oracle and/or its affiliates. All rights reserved.82 Required Benchmark Disclosure Statement  Copyright 2013, Oracle &/or its affiliates. All rights reserved. Oracle & Java are registered trademarks of Oracle &/or its affiliates.Other names may be trademarks of their respective owners..  TPC Benchmark C, tpmC, and TPC-C are trademarks of the Transaction Processing Performance Council (TPC). SPARC T5-8 (8/128/1024) with Oracle Database 11g Release 2 Enterprise Edition with Partitioning, 8,552,523 tpmC, $0.55 USD/tpmC, available 9/25/2013, New Order 90th% Response Time 0.410sec. IBM Power 780 Cluster (24/192/768) with DB2 ESE 9.7, 10,366,254 tpmC, $1.38 USD/tpmC, available 10/13/2010, New Order 90th% Response Time 2.10 sec. IBM x3850 X5 (4/40/80) with DB2 ESE 9.7, 3,014,684 tpmC, $0.59 USD/tpmC, available 7/11/2011. IBM x3850 X5 (4/32/64) with DB2 ESE 9.7, 2,308,099 tpmC, $0.60 USD/tpmC, available 5/20/2011. IBM Flex x240 (2/16/32) with DB2 ESE 9.7, 1,503,544 tpmC, $0.53 USD/tpmC, available 8/16/2012. IBM Power 780 (2/8/32) with IBM DB2 9.5, 1,200,011 tpmC, $0.69 USD/tpmC, available 10/13/2010. Source: http://www.tpc.org/tpcc, results as of 3/26/2013.  SPEC and the benchmark name SPECjEnterprise are registered trademarks of the Standard Performance Evaluation Corporation. Results from www.spec.org as of 3/26/2013. SPARC T5-8, 57,422.17 SPECjEnterprise2010 EjOPS; SPARC T4-4, 40,104.86 SPECjEnterprise2010 EjOPS; Sun Server X2-8, 27,150.05 SPECjEnterprise2010 EjOPS; Cisco UCS B440 M2, 26,118.67 SPECjEnterprise2010 EjOPS; IBM Power 780, 16,646.34 SPECjEnterprise2010 EjOPS. IBM PowerLinux 7R2, 13,161.07 SPECjEnterprise2010 EjOPS. SPARC T3-4 9456.28 SPECjEnterprise2010 EjOPS. SPARC T5-8 (SPARC T5-8 Server base package, 8xSPARC T5 16- core processors, 128x16GB-1066 DIMMS, 2x600GB 10K RPM 2.5. SAS-2 HDD, 4x Power Cables) List Price $268,742. IBM Power 780 (IBM Power 780:9179 Model MHB, 8x3.86GHz 16-core, 64x one processor activation, 4xCEC Enclosure with IBM Bezel, I/O Backplane and System Midplane,16x 0/32GB DDR3 Memory (4x8GB) DIMMS-1066MHz Power7 CoD Memory, 12x Activation of 1 GB DDR3 Power7 Memory, 5x Activation of 100GB DDR3 Power7 Memory, 1x Disk/Media Backplane. 2x 146.8GB SAS 15K RPM 2.5. HDD (AIX/Linux only), 4x AC Power Supply 1725W) List Price $992,023. Source: Oracle.com and IBM.com, collected 03/18/2013. SPEC and the benchmark name SPECjEnterprise are registered trademarks of the Standard Performance Evaluation Corporation. Results from www.spec.org as of 5/1/2013. SPARC T5-8, 27,843.57 SPECjEnterprise2010 EjOPS; IBM Power 780, 10,902.30 SPECjEnterprise2010 EjOPS. Oracle server only hardware list price is $298,494 and total hardware plus software list price is $1,565,092 http://www.oracle.com as of 4/24/2013. IBM server only HW list price is $835,555 and HW+SW cost of $2,174,152.00 and BM PowerLinux 7R2 server total hardware plus software cost of $819,451.00 based on public pricing from http://www.ibm.com as of 4/24/2013.  SPEC & the benchmark name SPECjbb are registered trademarks of Standard Performance Evaluation Corporation (SPEC). Results as of 3/26/2013, see http://www.spec.org for more information. SPARC T5-2 75,658 SPECjbb2013-MultiJVM max-jOPS, 23,334 SPECjbb2013-MultiJVM critical-jOPS. Sun Server X2-4 65,211 SPECjbb2013-MultiJVM max-jOPS, 22,057 SPECjbb2013-MultiJVM critical-jOPS. Sun Server X3-2 41,954 SPECjbb2013-MultiJVM max-jOPS, 13,305 SPECjbb2013- MultiJVM critical-jOPS. SPARC T4-2 34,804 SPECjbb2013-MultiJVM max-jOPS, 10,101 SPECjbb2013-MultiJVM critical-jOPS. HP ProLiant DL560p Gen8 66,007 SPECjbb2013-MultiJVM max-jOPS, 16,577 SPECjbb2013-MultiJVM critical-jOPS. HP ProLiant ML350p Gen8 40,047 SPECjbb2013-MultiJVM max-jOPS, 12,308 SPECjbb2013-MultiJVM critical-jOPS. Supermicro X8DTN+ 20,977 SPECjbb2013-MultiJVM max-jOPS, 6,188 SPECjbb2013-MultiJVM critical-jOPS. HP ProLiant ML310e Gen8 12,315 SPECjbb2013-MultiJVM max-jOPS, 2,908 SPECjbb2013-MultiJVM critical-jOPS. Intel R1304BT 6,198 SPECjbb2013-MultiJVM max-jOPS, 1,722 SPECjbb2013-MultiJVM critical-jOPS, HP DL980 G7 106,141 SPECjbb2013-MultiJVM max-jOPS, 23268 SPECjbb2013-MultiJVM critical-jOPS. Must be in SPARC T5 & M5 Presos with Benchmark Results (1 of 2)
  • 80. Copyright © 2014, Oracle and/or its affiliates. All rights reserved.83 Required Benchmark Disclosure Statement  Copyright 2013, Oracle &/or its affiliates. All rights reserved. Oracle & Java are registered trademarks of Oracle &/or its affiliates.Other names may be trademarks of their respective owners..  Two-tier SAP Sales and Distribution (SD) Standard Application benchmarks SAP Enhancement package 5 for SAP ERP 6.0 as of 3/26/13:SPARC M5-32 (32 processors, 192 cores, 1536 threads) 85,050 SAP SD users, 32 x 3.6 GHz SPARC M5, 4 TB memory, Oracle Database 11g, Oracle Solaris 11, Cert# 2013009. SPARC T5-8 (8 processors, 128 cores, 1024 threads) 40,000 SAP SD users, 8 x 3.6 GHz SPARC T5, 2 TB memory, Oracle Database 11g, Oracle Solaris 11, Cert# 2013008. IBM Power 760 (8 processors, 48 cores, 192 threads) 25,488 SAP SD users, 8 x 3.41 GHz IBM POWER7+, 1024 GB memory, DB2 10, AIX 7.1, Cert#2013004. Two-tier SAP Sales and Distribution (SD) Standard Application benchmarks SAP Enhancement package 4 for SAP ERP 6.0 as of 4/30/12:IBM Power 795 (32 processors, 256 cores, 1024 threads) 126,063 SAP SD users, 32 x 4 GHz IBM POWER7, 4 TB memory, DB2 9.7, AIX7.1, Cert#2010046. SPARC Enterprise Server M9000 (64 processors, 256 cores, 512 threads) 32,000 SAP SD users, 64 x 2.88 GHz SPARC64 VII, 1152 GB memory, Oracle Database 10g, Oracle Solaris 10, Cert# 2009046. SAP, R/3, reg TM of SAP AG in Germany and other countries. More info www.sap.com/benchmark  SPEC & benchmark names SPECfp, SPECint are registered trademarks of the Standard Performance Evaluation Corporation. Results as of March 26, 2013 from www.spec.org and this report. SPARC T5-8: 3750 SPECint_rate2006, 3490 SPECint_rate_base2006, 3020 SPECfp_rate2006, 2770 SPECfp_rate_base2006; SPARC T5- 1B: 467 SPECint_rate2006, 436 SPECint_rate_base2006, 369 SPECfp_rate2006, 350 SPECfp_rate_base2006. IBM Power 780 8-chip 3.92GHz: 2640 SPECint_rate2006. IBM Power 710 Express 1-chip 3.556GHz: 289 SPECint_rate2006.  TPC-H, QphH, $/QphH are trademarks of Transaction Processing Performance Council (TPC). For more information, see www.tpc.org, results as of 6/7/13. SPARC T5-4 409,721.8 QphH@3000GB, $3.94/QphH@3000GB, available 9/24/13, 4 processors, 64 cores, 512 threads; SPARC T4-4 205,792.0 QphH@3000GB, $4.10/QphH@3000GB, available 5/31/12, 4 processors, 32 cores, 256 threads; SPARC Enterprise M9000 386,478.3 QphH@3000GB, $18.19/QphH@3000GB, available 9/22/11, 64 processors, 256 cores, 256 threads; SPARC Enterprise M9000 198,907.5 QphH@3000GB, $15.27/QphH@3000GB, available 12/9/10, 32 processors, 128 cores, 128 threads; IBM Power 780 QphH@3000GB, 192,001.1 QphH@3000GB, $6.37/QphH@3000GB, available 11/30/11, 8 processors, 32 cores, 128 threads; HP ProLiant DL980 G7 162,601.7 QphH@3000GB, $2.68/QphH@3000GB available 10/13/10, 8 processors, 64 cores, 128 threads. ; HP ProLiant DL980 G7 158,108 QphH@10000GB, $6.49/QphH@10000GB available 4/15/13, 8 processors, 80 cores, 160 threads. Must be in SPARC T5 & M5 Presos with Benchmark Results (2 of 2)
  • 81. Copyright © 2014, Oracle and/or its affiliates. All rights reserved.84 Required Benchmark Disclosure Statement T-series Copyright 2011, Oracle and/or its affiliates. All rights reserved. Oracle and Java are registered trademarks of Oracle and/or its affiliates. Other names may be trademarks of their respective owners. Results as of 9/26/2011. SPEC and the benchmark name SPECjEnterprise are registered trademarks of the Standard Performance Evaluation Corporation. Results from www.spec.org as of 9/26/2011. SPARC T4-4, 40,104.86 SPECjEnterprise2010 EjOPS; Cisco UCS B440 M1, 17,301.86 SPECjEnterprise2010 EjOPS; IBM Power 780, 16,646.34 SPECjEnterprise2010 EjOPS.Focusing on the critical JEE server hardware & OS, the IBM result includes a JEE server with a list price of $1.30 million. The Oracle JEE servers have a list price of $0.47 million. The JEE server price versus delivered EjOPs is $77.97/EjOP for IBM versus $11.67/EjOP for Oracle. Oracle's $/perf advantage is 6.7x better than IBM ($77.97/$11.67). Pricing details for IBM, IBM p780 512GB based on public pricing at http://tpc.org/results/FDR/TPCH/TPC- H_1TB_IBM780_Sybase-FDR.pdf. Adjusted hardware costs to license all 64 cores. AIX pricing at: http://www- 304.ibm.com/easyaccess3/fileserve?contentid=214347 and AIX Standard Edition V7.1 per processor (5765-G98-0017 64*2,600=$166,400). This gives application tier hardware & OS Price/perf: $77.97/EjOPS (1297956/16646.34) Pricing details for Oracle, four SPARC T4-4 512 GB, HW acquisition price from Oracle's price list: $467,856 http://www.oracle.com. This gives application tier hardware & OS Price/perf: $11.67/EjOPS (467856/40104.86) The Oracle application tier servers occupy 20U of space, 40,140.86/20=2007 EjOPS/U. The IBM application tier server occupies 16U of space, 16,646.34/16=1040 EjOPS/RU. 2007/1040=1.9x TPC-H, QphH, $/QphH are trademarks of Transaction Processing Performance Council (TPC). For more information, see www.tpc.org. SPARC T4-4 201,487 QphH@1000GB, $4.60/QphH@1000GB, avail 10/30/2011, 4 processors, 32 cores, 256 threads; SPARC Enterprise M8000 209,533.6 QphH@1000GB, $9.53/QphH@1000GB, avail 09/22/11, 16 processors, 64 cores, 128 threads; IBM Power 780 QphH@1000GB, 164,747.2 QphH@1000GB, $6.85/QphH@1000GB, avail 03/31/11, 8 processors, 32 cores, 128 threads; HP Integrity Superdome 2 140,181.1 QphH@1000GB, $12.15/QphH@1000GB avail 10/20/10, 16 processors, 64, cores, 64 threads.
  • 82. Copyright © 2014, Oracle and/or its affiliates. All rights reserved.85