Server Consolidation with SPARC Enterprise M-series High-End Servers

  • 900 views
Uploaded on

Oracle’s SPARC Enterprise M-Series high-end servers, the SPARC Enterprise M8000 and SPARC Enterprise M9000, are ideal consolidation platforms. Mission-critical consolidation server requirements …

Oracle’s SPARC Enterprise M-Series high-end servers, the SPARC Enterprise M8000 and SPARC Enterprise M9000, are ideal consolidation platforms. Mission-critical consolidation server requirements include the scalability to handle large workloads; high reliability, availability, and serviceability (RAS) features to maintain application availability; virtualization to enable deploying many applications; and tools for managing the physical and virtual resources. The SPARC Enterprise M8000 and SPARC Enterprise M9000 are large SMP servers that provide as many as 64 processors, 288 I/O slots, and up to 4 TB of shared memory in a single OS instance and can run either the Oracle Solaris 10 or Oracle Solaris 11 operating system.

More in: Technology
  • Full Name Full Name Comment goes here.
    Are you sure you want to
    Your message goes here
    Be the first to comment
    Be the first to like this
No Downloads

Views

Total Views
900
On Slideshare
0
From Embeds
0
Number of Embeds
0

Actions

Shares
Downloads
23
Comments
0
Likes
0

Embeds 0

No embeds

Report content

Flagged as inappropriate Flag as inappropriate
Flag as inappropriate

Select your reason for flagging this presentation as inappropriate.

Cancel
    No notes for slide
  • Growing Administration and Support Costs Looming Consolidation and Security Costs Aging Servers Unable to Meet Service Levels Growing User and Application Demands Data Centers Reaching Power and Space Capacity Demand for Faster and More Flexible Deployment Platforms
  • There are numerous benefits from Server consolidation. The obvious benefits are those gained by improving system utilization….reduced power, space and cost and better ROI. However, there are other benefits as well. When consolidating older applications from older servers the new platforms normally provide increased performance, better virtualization, better manageability and better RAS. An environment where servers are treated a resources pools that can be partitioned to provide resources to various applications also means faster deployment of newer applications. It is easier to reassign resources inside a server to a new application than it is to deploy a new server. Note that in many cases “consolidation” means replacing older servers with newer servers. However, new applications can be deployed in a consolidated environment from the start thereby gaining most of the benefits articulated on this slide.
  • When combining applications it is best to combine “like” applications. E.g. while it is possible to combine the web tier, app tier and DB tier into one server the differences in those tiers make such a combination not optimal. DB apps work better with other DB apps. The servers used for a DB is normally much different in cost and capabilities and deployment architecture than the servers used for the web tier. So, as this slides shows when stacking applications try and combine similar applications.
  • The highend M-Series servers are the M8000 and M9000. They are the two “chassis” servers. The M3000, M4000 and M5000 are rackmount servers. Key specs of the M-Series High End are the number of processors, memory and I/O slots. The M-series are balanced systems that have enough cpus, RAM and I/O slots (and internal bandwidth) to handle the largest and most complex workloads. They are true SMP servers were all cpus, RAM and I/O can be assigned to a single OS instance. All memory, I/O and cpus are shared making it easier to deploy applications as data placement is not an issue. All cpus have equal access to memory and I/O slots.
  • Oracle runs its business on SPARC/Solaris – this example highlights the recent GSI conversion that took place, moving from 4 E25Ks to 2 M9000s, chosen because of the performance, reliability and large memory and single system image capabilities of the M-Series family. More details below. --------------------- Oracle's internal datacenters refreshed its mission critical database servers supporting core ERP applications with Oracle's optimized solution of M9000 and Solaris 10. See how Oracle achieved, in a Global Single Instance (GSI) architecture deployment, the following: a 2x service level performance increase, a targeted near 99.999% availability, a simplified deployment which reduces management complexity. non-stop database operations as the goal From 10 Hours to 30 minutes (oriented towards the requirements of the EBS and ERP applications in a Disaster Recovery Configuration) reduced RTO from 10 hours to target of 30 minutes for major disaster and transfer of operations from Austin Tx to the RMDC. timeframe savings such as 1/2 the amount of time it took to deploy previously Solaris Volume Manager being used to save the $347K verses Veritas In addition, that we have Disaster Recovery being taken care of via DataGuard, Dynamic Domains, and Dynamic Reconfiguration sponsored by an offsite location through OSC as another layer to provide access 7x24 for the Oracle Database. March 2011 Update – 200% Additional Increase Operational Workload   Oracle’s business success and the results of ongoing large investments engineering the evolution of enterprise computing via complete software and hardware solutions required Oracle’s PDIT department to quickly and seamlessly increase and expand operational capacities of the mission critical systems responsible for ERP and BPM. With the live system expansion capabilities of the Oracle SPARC and Solaris infrastructure, a third M9000 server was added to the datacenter’s core GSI architecture.   The Ease of Live SPARC/Solaris Non-Stop System Capacity Upgrades   This real world deployment demonstrates the relative ease of a mission critical enhancement of mission critical Oracle database application environments, facilitated by SPARC/Solaris technology’s feature functionality.   The additional M9000 SPARC system installation began at midnight March 3 rd , 2011, and was completed, in full, at 11:31am the next day, March 4 th , 2011. There was no down time of the live GSI database/ERP systems during installation by the Oracle PDIT staff.   The entire project from assessment, to planning to deployment took a total of about two and a half weeks.   This is largely due to the assistance and documentation from Oracle Systems Engineering, including sizing, assessment scoping tools and best practice reference architecture guides. All of these resources are available for all Oracle customer engagements via the materials created and provided by the Oracle Optimized Solutions group, responsible for engineering hardware infrastructure solutions for Oracle Database, Applications, and Middleware. For more information on Oracle Optimized Solutions please go to: www.oracle.com/goto/optimizedsolutions Actual TCO and ROI (as of May 2011): 2.5X increase in OLTP and ERP performance 50% less operational costs Over 95% acceleration in disaster recovery performance Near 75% savings in physical disk storage Over $400K savings in 3 rd party expenses and power costs 50% savings in licensing (with standard core multipliers) Over 5X faster App deployment and patch implementation Over 45+ IT management hours saved Before Details: System 4 x Sun E25k (88 total x UltraSPARC IV 1.2GHz dual core processors, 176GB RAM) 20 x drives (OS/boot, swap), 2 racks 6 x 1GB FC HBA cards (SAN connectivity) Oracle 11g - RDBMS 11.1.0.7.2, Total size: 16TB (14TB used) SGA: 40GB (20GB buffer, 20GB Shared Pool) PGA: 34GB IOPS: 15k average, 20k peak For additional detailed calculations, contact cathryn.grant@oracle.com
  • Mission-critical deployments REQUIRE high availability. Servers just be up and running and the apps available whenever a users needs access to applications. All of the features on this slide are designed to improve the availability of the system. The goal is to reduce or eliminate planned and unplanned downtime. The “Reliability” features are about preventing a failure in the first place. The “Serviceability” features are designed to reduced planned downtime for maintenance or reconfiguration and to reduce downtime due to fixing hardware problems. “Software” is a critical part of the high availability story as is “Monitor and Manage” Reliability: ECC (error corrections) means single bit errors on any data or instruction paths or on RAM are detected and corrected Memory extended ECC…better known in the industry as “chip kill”. If a a DIMM fails the system can still keep running. Instruction-level retry…failed processor instructions are retried. This eliminates system failures due to transient errors Component redundancy…”more than one” of each component means that a component failure does not keep the system down or in many cases does not even cause a failure. Serviceability: Fault-isolation (domains): A fault in one dynamic domain does not affect any other domain Dynamic reconfiguration: enables hot-swap of failed components so that they can be replaced or serviced while the server is running. Also used to resize dynamic domains Hot-swap of fans and power: No downtime required to replace power supplies and fans Hot-swap of CPU’s and RAM: No downtime required to add, remove or replace memory and processors. (M8000 and M9000 only) Hot-swap of I/O: No downtime required to remove or add I/O boards or PCIe cards Software: Solaris has 20 years of dependability. A very robust OS that scales up to hundreds of cores and threads Oracle Solaris cluster can provide cluster failover for many applications and all tiers (web, app, DB). Is also a good cluster manager for Oracle RAC Oracle RAC: cluster database whereby a server can fail but the clustered DB keeps running Oracle Solaris containers provide software fault-isolation so that software errors are confined to a single app/process Monitor and Manage: Fault-management architecture (FMA) is a fault reporting architecture that greatly simplifies fault management and diagnostics. Reduces mean time to repair. Predictive self-healing uses heuristics to predict components that might fail and takes them offline before they fail. With hot-swap many of the components can be removed and replaced with no downtime. OEM Ops center: Oracle Enterprise Manager OpsCenter is the OEM subset that is used to manage groups of SPARC servers. OpsCenter simplifies server management and speeds up system deployment which are key benefits for mission-critical deployments M-series XSCF is the software that monitors and manages individual M-series servers. It is the same software for all M-series servers thereby simplifying server management. It is used to manage FMA, predictive self-healing and domain manage it is critical to mission-critical deployments.
  • Any consolidation platform must be able to partition its resource so that many applications can run appropriately. All applications need IO, cpu, RAM resources but in a consolidated environment all applications much share server resources. To prevent applications from “hogging” resources or to prevent an error in one application affecting the other applications or to prevent a user on one application gaining unauthorized access to a different application in the same server it is necessary to have virtualization technology that provides the necessary separation. Not all workloads are the same so Sun offers three primary types of virtualization technologies. A very important “feature” of Sun’s SPARC virtualization technologies is that they are no charge to the customer. No per core licenses are required as is the case with HP and IBM and their virtualization technologies. IBM and HP charge in the 10’s of thousands of $$ for virtualization for their Itanium and Power servers. Sun customers do need to pay for support if they want patches or upgrades to the SPARC virtualization technologies but no license fees. This means that customers can deploy the appropriate virtualization technology when they need to without having to worry about license fees. These notes will not go into detail on each virtualization solution. There are other sources of much more detailed information. Some things to note: Solaris containers is available on all systems that run Solaris…even on non-SPARC systems. Dynamic Domains (formerly known as Dynamic System Domains on Sun Fire systems) is only on Mseries systems. Dynamic domains is a hardware partitioning scheme. Solaris containers has more than just zones. Solaris resource manager with fair-share scheduler, resource pools and processor sets are some key virtualization technologies besides zones.
  • Oracle offers a full portfolio of virtualization solutions to address your needs. SPARC is the leading platform to have the hard-partitioning capability that provides the physical isolation needed to run independent operating systems. Many customers have already used Oracle Solaris Containers for application isolation. Oracle VM Server for SPARC provides another important feature with OS isolation. This gives you the flexibility to deploy multiple operating systems simultaneously on a single T-Series server with finer granularity for computing resources. For SPARC T-Series processors, the natural level of granularity is an execution thread, not a time-sliced microsecond of execution resources. Each CPU thread can be treated as an independent virtual processor. The scheduler is built into the CPU, without the extra overhead for scheduling in the hypervisor. What you get is a lower- overhead and higher-performance virtualization solution. Your organizations can couple Oracle Solaris Containers and Oracle VM Server for SPARC with the breakthrough space and energy savings afforded by Oracle’s SPARC T-Series servers to deliver a more agile, responsive, and low-cost environment.
  • 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 Domains SPARC 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.
  • 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.
  • [Original slide author: Nicolas Droux] The Crossbow project provides was developed over the last few years and is being delivered as part of Solaris 11. Crossbow has re-architected the data link layer of the Solaris network stack, and focused on the functional areas listed here: Network virtualization : full data link layer virtualization which is closely integrated with Solaris zones and provides (1) virtualization of a hardware NIC into multiple virtual NICs, but also (2) the building blocks to enable full network virtualization and building networks in a box. [Note also that Crossbow is closely integrated with the LDOM network virtualization. In Solaris 11 the LDOM network virtualization is layered on top of Crossbow, and some features such as SR-IOV support are closely integrated between the two architectures.] Bandwidth partitioning : as soon as physical resources are virtualized, it is critical to be able to control how much of the physical resources each one of the virtual resources will be able to use. Crossbow provides built-in QOS which is closely integrated with the virtualized network stack. This new QOS implementation is not based on the traditional IPQoS, but rather a new design which provides easier management, and more efficient bandwidth control for virtualization. The new QOS implementation can also be used to control the bandwidth used by traffic flows. Resource control : another aspect of use of resource is being able to control where network traffic will be taking place on the machine for the various virtual stacks. We now provide the ability to specify on which CPUs or CPU pools network traffic processing will be taking place for better isolation and overall performance. Another key piece of the virtualization puzzle is to be able to measure the actual bandwidth used by the various VNICs. We have introduced several ways to do this in Solaris 11. We are introducing new tools observe data link (dlstat(1M)) and traffic flows (flowstat(1M)) in real time or through a history using extended accounting. Data link observability with dlstat(1M) can report statistics on a per data link basis (VNIC, etc), or on a per-hardware ring basis. The new stack is completely parallel for best scalability and performance . We used our early experience on large Sun Fire machines and SPARC CMT systems to design scalability into the architecture. This allows us to scale on systems with lots of CPUs and increase the aggregate throughput on 10 Gigabit Ethernet and beyond.
  • Ops Center provides real-time information on hardware assets Enables new application-to-disk management capability within Enterprise Manager Infrastructure awareness for application managers Application awareness for systems administrators Integrated application-to-disk compliance Ops Center is now available for free on servers that have Oracle premier support contracts.
  • Capacity planning and budgeting are different in a cloud world. Most IT departments procure hardware only after a particular project is approved, and that hardware is often dedicated to that project. In a cloud world, hardware is shared, and so a certain amount of capacity has to be on hand for users to obtain those resources on-demand. Most customers convert existing infras to cloud. When you consolidate for this purpose, how do you decide what needs to be done? You will need to apply business rules (e.g. avoid two departments on the same box) and technical rules (e.g. avoid mixing workloads on the same box). With EM, IT will get advice on how to consolidate. Then through in-context provisioning, EM will make it happen for all consolidation techniques like p2v and v2v. Consolidation Planner provides capabilities for the company's IT department to consolidate under-utilized servers in the company data centers, or to consolidate servers for migrating to company private cloud. Over the years, company data centers likely have more and more servers being added due to business requirements. While the excess servers occupy rack space, consume a lot of power for cooling, and require system maintenance such as security and patching, many of them are under-utilized. With Consolidation Planner, company IT department could identify the under-utilized servers and find a way to consolidate these servers to free up as many servers as possible, at the same time maintaining the service levels. When deploying a private cloud within enterprise, with Consolidation Planner, company IT department could identify the managed servers that are capable of migrating to the cloud, estimate the total demand of server resources for a purchase requisition of additional hardware and storage if necessary, and come up a way to group these servers to ensure some optimal consolidation with satisfactory performance and at the same time, complying with various business, compliance, and technical constraints. Customers can leverage the server hardware/software information collected by EM to optimize consolidation plans by maximizing the server density and minimizing resource contention while maintaining performance commitment and satisfying business constraints.
  • Oracle ’s Own Mission Critical Database Systems: Why SPARC M9000 Servers and Solaris 10? Over 20 years on SPARC and Solaris Oracle ’s Global Single Instance architecture for ERP Implementation requirements: Need high OLTP transaction rates for ERP workload 600 transactions per second requires additional threads/cores Need RAS (reliability, availability, scalability) for 100% uptime RDBMS for: E-Business Suite, Field Service, 24x7 Support Need large TB memory footprint 120GB SGA , 200GB PGA, +10,000 database processes Dynamic Domains (Virtualization) to maximize ROI 2nd environment supports test and DR requirements Maximizing legacy SAN investment for DR 100% Certified & Tested (Hardware and Software) Oracle's internal datacenters refreshed its mission critical database servers supporting core ERP applications with Oracle's optimized solution of M9000 and Solaris 10. See how Oracle achieved, in a Global Single Instance (GSI) architecture deployment, the following: a 2x service level performance increase, a targeted near 99.999% availability, a simplified deployment which reduces management complexity, and discovered almost $400,000 in operational savings. (Using SVM instead of VxVM saved us $347k) non-stop database operations as the goal From 10 Hours to 30 minutes (oriented towards the requirements of the EBS and ERP applications in a Disaster Recovery Configuration) reduced RTO from 10 hours to target of 30 minutes for major disaster and transfer of operations from Austin Tx to the RMDC. timeframe savings such as 1/2 the amount of time it took to deploy previously Solaris Volume Manager being used to save the $347K verses other volume manager. In addition, that we have Disaster Recovery being taken care of via DataGuard, Dynamic Domains, and Dynamic Reconfiguration sponsored by an offsite location through OSC as another layer to provide access 7x24 for the Oracle Database.   Savings: Before, we have 4 E25K running with 11 board on each of the E25K. Now, we are using 2 M9000 with 8 CMU each. Cost save due to reduce the space usage from 8 tile space to 4 tile space. Current ADC cost per server tile (calculated based on the fully loaded data center space including white space, e.g. hall way and space between rows in the DC) is $45 per sq ft * 35 sq ft (per tile) = $1575 per year. We saved $6300 per data center per year. Power: currently, we are based on book value. E25K is using 12KW each and total of 48 KW. M9000 is using 16 KW each and total of using 32 KW. So the power deduction is 16KW. Based on the information I got from our data center, it cost us $.14c per kw/hour. So, the deduction of 16KW saved us $53.76 daily and $19622.4 per year per data center. Actual TCO savings of $398,844.80 (so far) $347k saved using SVM instead of VxVM $51,884.80 in space, power & cooling savings Before Details: System 4 x Sun E25k (88 total x UltraSPARC IV 1.2GHz dual core processors, 176GB RAM) 20 x drives (OS/boot, swap), 2 racks 6 x 1GB FC HBA cards (SAN connectivity) Oracle 11g - RDBMS 11.1.0.7.2, Total size: 16TB (14TB used) SGA: 40GB (20GB buffer, 20GB Shared Pool) PGA: 34GB IOPS: 15k average, 20k peak

Transcript

  • 1. 1Copyright © 2012, Oracle and/or its affiliates. All rights reserved. Insert Information Protection Policy Classification from Slide 13
  • 2. Server Consolidation withSPARC Enterprise M-SeriesHigh-End ServersTom AtwoodSenior Principal Product DirectorGary CombsPrincipal Product Manager 2Copyright © 2012, Oracle and/or its affiliates. All rights reserved. Insert Information Protection Policy Classification from Slide 13
  • 3. Agenda  Benefits of Consolidation and Virtualization  Consolidation and Virtualization Strategies  Oracle SPARC Virtualization Solutions  SPARC Consolidation Examples  Summary3Copyright © 2012, Oracle and/or its affiliates. All rights reserved. Insert Information Protection Policy Classification from Slide 13
  • 4. Today’s Server Infrastructure Challenges Growing Costs and Complexity  Spiraling Operating and Administration Costs  Accelerating Demands for Performance and Capacity  Growing Complexity and Risk Driven by Server Proliferation  Need to Protect Investments in existing Applications and Skill Sets4Copyright © 2012, Oracle and/or its affiliates. All rights reserved. Insert Information Protection Policy Classification from Slide 13
  • 5. Benefits of Consolidation and Virtualization  Increased server utilization – Lower acquisition costs – Reduced service costs – Reduced power and cooling costs  Improved capacity and response time – Newer and faster processors, interconnects, and I/O  Smaller foot print – Reduced data center infrastructure costs  Increased flexibility – Faster time to deployment of new applications5Copyright © 2012, Oracle and/or its affiliates. All rights reserved. Insert Information Protection Policy Classification from Slide 13
  • 6. Consolidating & Sharing Headroom• One app per server leads to overload or extra headroom• Consolidate many apps to share and reduce headroom Many Smaller Servers Single Larger Server Overload Consolidated and 15% shared headroom Overload 70% Overload 40% 10% 5% 5% Overload 5% Distributed headroom6Copyright © 2012, Oracle and/or its affiliates. All rights reserved. Insert Information Protection Policy Classification from Slide 13
  • 7. Server Consolidation Change Deployment Topology • Keep different workloads separate • Mix like workloads App 1 App 2 App 3 Database Database Database Data-centricApplications Applications Applications Application-centric Edge/Web Edge/Web Edge/Web Web-centric App 1 App 2 App 3 7Copyright © 2012, Oracle and/or its affiliates. All rights reserved. Insert Information Protection Policy Classification from Slide 13
  • 8. High End M-Series Servers Over 20 Years Of Mission Critical Computing M8000 M9000/32 M9000/64• 16 CPUs • 32 SPARC64 VII+ CPUs • 64 SPARC64 VII+ CPUs• 1TB RAM • 2TB RAM • 4TB RAM• 16 SAS disks • 32 SAS disks • 64 SAS disks• 32 I/O slots/**112 I/O slots • 64 I/O slots/**224 I/O slots • 128 I/O slots/**288 I/O slots• 16 dynamic domains • 24 dynamic domains • 24 dynamic domains8Copyright © 2012, Oracle and/or its affiliates. All rights reserved. Insert Information Protection Policy Classification from Slide 13
  • 9. Meeting Mission-critical Requirements  Mission-critical Workloads: – Database deployments – High capacity and demanding applications – Enterprise application consolidation  SPARC M8000/M9000 Deliver: – Linear scalability from 16 to 64 quad-core processors – Upgradeability and future capacity – Highest Availability – Highly efficient “no-cost virtualization” – Security9Copyright © 2012, Oracle and/or its affiliates. All rights reserved. Insert Information Protection Policy Classification from Slide 13
  • 10. Benefits of SMP “Shared Memory Processing”  Single, shared memory image  Easier application deployment – Data partitioning not needed – Least complex software environment – Least complex storage and networking environment  Extreme system bandwidth  Extreme I/O bandwidth  Can be partitioned using virtualization tools  Easy to dynamically re-assigned resources  No limitations on workloads10Copyright © 2012, Oracle and/or its affiliates. All rights reserved. Insert Information Protection Policy Classification from Slide 13
  • 11. Meeting High Availability Requirements  Reliability  Serviceability – System-level ECC - Fault-isolation (dynamic domains) – Memory Extended ECC - Dynamic reconfiguration – Instruction-level retry - Hot-swap of fans and power – Memory Mirroring - Hot-plug of CPU’s and RAM – Component Redundancy - Hot-plug of I/O  Software  Monitor & Manage – Robust Oracle Solaris OS - Fault-management architecture – Oracle Solaris Cluster - Predictive-self healing – Oracle RAC - OEM Ops Center – Oracle Solaris Zones - M-Series XSCF11Copyright © 2012, Oracle and/or its affiliates. All rights reserved. Insert Information Protection Policy Classification from Slide 13
  • 12. Oracle M-Series Virtualization All are “Hard Partitions” for Oracle Software Licensing Purposes Dynamic Domains Solaris Zones M-Series All SPARC Systems Hardware partitions OS Virtualization Dynamic Dynamic No overhead Hundreds Per OS Instance Single-socket Granularity Software Isolation Resource Isolation Separate File Systems Security Isolation Upgrade OS and Upgrade All Fault and Service Isolation Zones Multiple OSes Single OS Included at No Cost! Included at No Cost!12Copyright © 2012, Oracle and/or its affiliates. All rights reserved. Insert Information Protection Policy Classification from Slide 13
  • 13. Solaris and M-Series Virtualization Dynamic Domains Oracle Solaris Zones Domain A Web DB App Web OLTP DB OLTP DB App App Domain B DW DB M-series Oracle Solaris13Copyright © 2012, Oracle and/or its affiliates. All rights reserved. Insert Information Protection Policy Classification from Slide 13
  • 14. SPARC Dynamic Domains• Complete isolation Domain 1 • Resource, security, service, fault Domain 1 Domain 2 Domain 2 Domain x Domain x• Single-CPU granularity• Dynamic• No overhead• Separate OS per domain Solaris Solaris Solaris• No cost to end user• M-Series SPARC servers CPU CPU CPU CPU CPU CPU CPU CPU CPU Mem Mem Mem Mem Mem 14Copyright © 2012, Oracle and/or its affiliates. All rights reserved. Insert Information Protection Policy Classification from Slide 13 14
  • 15. Oracle Solaris Zones• One OS instance for all zones• Separate file system Zone 1 Zone 1 Zone 2 Zone 2 Zone x Zone x• Complete software isolation• Sub-thread granularity• Dynamic and mobile• Low overhead• No cost to end user Solaris Solaris• All Solaris instances CPU CPU CPU CPU CPU CPU CPU CPU CPU Mem Mem Mem Mem Mem 15Copyright © 2012, Oracle and/or its affiliates. All rights reserved. Insert Information Protection Policy Classification from Slide 13
  • 16. Solaris 11 Zones Combine Privileges, Roles, Immutable Zones16Copyright © 2012, Oracle and/or its affiliates. All rights reserved. Insert Information Protection Policy Classification from Slide 13
  • 17. Solaris Zones Security Benefits  Restricted In-Zone Operations – Individual operating system hardening, RBAC, auditing, etc. – Prohibited from directly accessing kernel (modules), raw memory  External Enforcement of Zone Configuration – Configurable privileges, immutability, devices, file systems, resource controls, virtual network security controls, etc.  Observability with Integrity – Protected audit trails, file integrity verification, global zone has complete introspection capabilities17Copyright © 2012, Oracle and/or its affiliates. All rights reserved. Insert Information Protection Policy Classification from Slide 13
  • 18. Solaris 11 Network Virtualization Network Virtual NICs, Virtual Switching, Network in a Box Virtualization Bandwidth Built-in QOS: bandwidth limits for data links and on Partitioning a per-flows basis Resource Constraint traffic processing to CPUs or CPU pools Control dedicated to zones Real-time usage and history for VNICs, hardware Observability resources, and traffic flow Parallel traffic from hardware to applications, Scalability Dynamic Polling, NUMA I/O18Copyright © 2012, Oracle and/or its affiliates. All rights reserved. Insert Information Protection Policy Classification from Slide 13
  • 19. Data Architecture Strategies19Copyright © 2012, Oracle and/or its affiliates. All rights reserved. Insert Information Protection Policy Classification from Slide 13
  • 20. Managing Consolidate SPARC Servers Oracle Enterprise Manager Ops Center  Manage physical & virtual resources  Infrastructure Management  Application-to-Disk Management  Lifecycle Management  Systems Management & Support  Free with premier support20Copyright © 2012, Oracle and/or its affiliates. All rights reserved. Insert Information Protection Policy Classification from Slide 13
  • 21. Enterprise Manager Ops Center The management solution for all Oracle virtualization• Centralized interface for VM lifecycle management• Manage M-Series hard-partitions, OVM, and Solaris Zones from one place• Complete management of Virtual Storage and Networks21Copyright © 2012, Oracle and/or its affiliates. All rights reserved. Insert Information Protection Policy Classification from Slide 13
  • 22. Software Management• New Support for Image Packaging System• Enhanced Interaction with Alternate Boot Environments• New Support for Automated Installer• Improved Software Profiler• Better Dependency Handling• Increased Vendor Provider Patch Bundles 22Copyright © 2012, Oracle and/or its affiliates. All rights reserved. Insert Information Protection Policy Classification from Slide 13
  • 23. Enterprise Manager Cloud Control 12c Capacity and Consolidation Planning  Facilitates moving existing applications to the Cloud  Based on resource distribution and utilization analysis  Intelligence placement policies, technical and business constraints23Copyright © 2012, Oracle and/or its affiliates. All rights reserved. Insert Information Protection Policy Classification from Slide 13
  • 24. Take Control of Virtualization Ops Center Spends Deployment and Lifecycle Power+Cooling Space Daily Compliance Checking Analysis. Rack and Cable Fault Response/Replace Replication and Re-use Discovery and SOE deploy Perf Analysis and Recycle Resource Assignment Elasticity 1-2 weeks 2-6 weeks Lifetime Lifetime 1 week 2-6 weeks Down to Days Down to Hours Automate Orchestrate Down to hours Down to DaysOps Center PDU & Ops Center Intelligent Ops Center V12N CloneServer Energy Analysis Software Change Control and Cataloging and ASR Ops Center Server Pool Policy Ops Center Server Knowledge and Complex Prov Plans Ops Center OS Analytics, V12N, and Bare Metal Provisioning 24Copyright © 2012, Oracle and/or its affiliates. All rights reserved. Insert Information Protection Policy Classification from Slide 13
  • 25. Oracle Global Single Instance ERP System Database Infrastructure Modernization Project 2X Performance Gain 95% Reduction in Disaster Recovery time 75% Lower Storage Growth 50% Lower Licensing Costs 50% Lower utility and floor space cost $400K Saved on 3rd Party Licenses Two months from concept to production BEFORE AFTER• 4 x SPARC E25K Servers • 3 x Oracle SPARC M9000 Servers• Solaris 9 • Oracle Solaris 10 • DR - Oracle Solaris Volume Manager, Oracle Solaris Cluster,• DR – VxVM, Sun Cluster, Oracle Data Guard, Dynamic Oracle Data Guard, Dynamic Domains, and Dynamic Domains, and Dynamic Reconfiguration Reconfiguration 25Copyright © 2012, Oracle and/or its affiliates. All rights reserved. Insert Information Protection Policy Classification from Slide 13
  • 26. Summary  Server Consolidation enables the Efficient Datacenter  Consolidation Customer Benefits: – Increased system utilization – Lower TCO – Increased compute capacity – Faster time to deployment  Oracle M-Series servers are ideal consolidation platforms – High RAS – High capacity – Virtualization26Copyright © 2012, Oracle and/or its affiliates. All rights reserved. Insert Information Protection Policy Classification from Slide 13
  • 27. Oracle Solaris 10 System Virtualization Essentials • ISBN: 013708188X • ISBN-13: 9780137081882 • Chapter 2 covers Dynamic Domains on M-Series27Copyright © 2012, Oracle and/or its affiliates. All rights reserved. Insert Information Protection Policy Classification from Slide 13
  • 28. Server Virtualization for Dummies: Oracle Edition http://www.oracle.com/go/? &Src=7618691&Act=126&pcode=WWMK12044691MPP01228Copyright © 2012, Oracle and/or its affiliates. All rights reserved. Insert Information Protection Policy Classification from Slide 13
  • 29. For More Information29Copyright © 2012, Oracle and/or its affiliates. All rights reserved. Insert Information Protection Policy Classification from Slide 13
  • 30. 30Copyright © 2012, Oracle and/or its affiliates. All rights reserved. Insert Information Protection Policy Classification from Slide 13
  • 31. 31Copyright © 2012, Oracle and/or its affiliates. All rights reserved. Insert Information Protection Policy Classification from Slide 13