Download presentation
Presentation is loading. Please wait.
Published by孙 昝 Modified 7年之前
1
This presentation will discuss the performance challenges faced by both the business and DBAs and will demonstrate how EMC can address these challenges with EMC flash and storage tiering solutions for Oracle Database environments. Oracle 性能和优化 2013 年 11 月
2
课程安排 客户面临的难题 检查 Oracle 技术层 FAST 和闪存技术 性能参考体系结构 问题
3
第一 第一 第一 第一 第一 第一 第一 EMC 在应用程序存储方面占据第一宝座 2012 年在 EMC 存储上部署的应用程序
Oracle SAP Exchange VDI 主目录 SharePoint 分析 第一 第一 第一 第一 第一 第一 第一 The result today is the more EMC storage is deployed for Oracle environments as well as Microsoft, SAP, VDI, Analytics and Home Directory according to the 2012 IDC Storage Users Demand Study. 2012 年在 EMC 存储上部署的应用程序 比任何其他供应商都要多 — IDC 来源:IDC,2012 年存储用户需求研究 — 秋季版:客户仍然希望拥有选择余地,#234651
4
DBA 活动不断增加 管理容量、性能、数据保护和可用性 监视可用性 77% 诊断性能问题 75% 管理备份和恢复 71% 容量规划 61%
(Page 14) In the 2013 IOUG survey of data science skills, over 92% of respondents indicated at least one of the skills listed here is part of their day to day routine and the extent to which those activities are increasing vs. decreasing. 199 Respondents (37% DBAs) Respondents say the are being asked to get more involved in monitoring system and platform availability, including storage, and developing better automated reporting. Diagnosing database performance issues more quickly and capacity planning (our plug- ins hit this) Overseeing backup, recovery, clustering, mirroring and failover (our BRS and VPLEX demos hit these) And ~90% of those surveyed sees these activities staying the same or increasing 容量规划 61% 管理数据存储和检索 60% 修补系统 59% 大数据愿景:2013 年数据科学技能调查
5
主要的数据库管理难题 故障排除性能高居榜首 诊断数据库性能问题 47% 确定资源密集型 SQL 语句 44% 保持数据库的当前修补程序级别
31% In the IOUG study entitled, “From Database Clouds to Big Data” the number one challenge Oracle DBAs have is Diagnosing database performance problems. Here is a quote from the study, “Diagnosing performance and identifying SQL statements that may be taxing systems remains an ongoing challenge, the survey finds. As data volumes grow, and more parts of the business demand greater access to data for analytics, there is increased pressure on data professionals to maintain high availability and performance on a 24x7 basis.” Oracle DBAs are now part of the teams that review and recommend which servers and storage infrastructure will meet the business needs of their mission critical applications. Working with DBA teams to solution how EMC can provide a platform for performance and manageability must be part of the strategy to provide the customer the best storage and to win the the key voters: Oracle DBAs, IT Directors, Application Owners and Senior management. 管理更多数据库 31% 安全威胁数量增加 29% 资源和容量规划 28% 从数据库云到大数据:2013 年 IOUG 数据库可管理性调查
6
存储工作负载要求 性能 低服务 级别 高服务 级别 容量
7
EMC 存储策略:同类最佳的开放策略 引领 高端 引领 中端 领导者 横向扩展 领导者 备份 领导者 对象 存储 存储 存储 存储 存储
性能 XtremSW XtremSF XtremIO 服务器端 存储 即将推出: 全闪存 存储 VMAX 引领 高端 存储 引领 中端 存储 VNX 低服务 级别 高服务 级别 领导者 横向扩展 NAS 存储 Isilon Data Domain Avamar 和 NetWorker 领导者 备份 领导者 对象 存储 Atmos 容量
8
Oracle Enterprise Manager 插件
简化目前的常见任务 可见性、性能、数据保护、可用性 The Oracle DBA can use EMC’s Oracle Enterprise Manager (OEM) 12c plug-in to gain visibility into storage performance and configuration for the VNX and VMAX. Using the free OEM plug-in the database can see IOPS of disks supporting the mission critical database and review the configuration. This enables the DBA to do their own analysis and collaborate with the storage administrator for improving the performance of production databases and applications. In addition to the performance of all our systems it is important to showcase EMC solutions as offering the best in backup & recovery and continuous availability too. A customer choosing EMC solutions has the best performance, protection and continuous availability in the market today. 备份和恢复 DD Boost for RMAN 存储可见性 Oracle Enterprise Manager 插件 连续可用性 适用于 Oracle 的 EMC VPLEX
9
Symmetrix VMAX The Symmetrix VMAX sets the standard for performance. In the following section we will show way the VMAX storage can accelerate your mission critical database and applications.
10
0.5 毫秒延迟 Vblock for Oracle 3,700,000 次读取 IOPS
The new Vblock system for high performance databases uses the VMAX to drive storage performance. EMC’s VMAX gives customers choices within an open infrastructure, enabling them to easily integrate into existing data center operations and take advantage of new technologies such as XtremCache which also fully supports Oracle RAC now. Using the following architecture EMC was able to drive performance to 3.7 million read IOPS: VMAX 40K Xtrem SF flash cards XtremCache © 版权所有 2013 EMC Corporation。保留所有权利。
11
适用于 Oracle 的 VMAX 任务关键型存储平台
多引擎体系结构为混合的 Oracle 和其他工作负载(OLTP、DW、Exchange...)提供高性能 混合工作负载 六个九可用性 高度冗余和组件级别的故障切换设计提供了业界领先的六个九可用性 快照与克隆集成 Oracle DBA 可以在几分钟内创建克隆和/或快照,以满足测试/开发和其他目的。与 Oracle Clonedb 工具集成 Before exploring performance it is important to highlight the features that make the VMAX number one for enterprise storage: VMAX Platform Mixed Workload Performance Six 9’s Availability FAST (perf and cost savings) Virtual Provisioning (perf & cost savings) Data Protection SRDF VPLEX RP SRM Integration RAC integration, certification, affinity (extends RAC, quicker failover) 全自动存储分层 利用全自动存储分层技术提高性能,同时显著降低存储成本 虚拟资源调配 确保 Oracle 数据库在满足容量需求的同时,降低存储成本并简化管理
12
开放式性能解决方案的优点 VMAX FAST VP:
EMC Symmetrix VMAX 40K、EMC XtremeCache、NEC Express5800/A1080a-E 和 VMware vSphere 5 VMAX FAST VP: 加快 Oracle 数据库的速度 响应时间从 35–20 毫秒减少为不足 5 毫秒 FAST VP 与 XtremSF(位于服务器中)PCIe 闪存卡相结合,可提供非常短的响应时间 毫秒 In a recent case study the combination of VMAX FAST VP and XtremCache reduced Oracle latency for this virtualized database. The latency saving was outstanding: the original baseline was approximately 35 milliseconds and after FAST VP and XtremCache were applied the latency dropped to under 5 milliseconds a 31 ms savings. This represents a six times latency performance improvement for Oracle. When taking into account that the number of calls for storage I/O wait events can easily number in the 10k to 100k range in a 30 minute time period a savings of 31 ms per call can be substantial.
13
来源:Wikibon,2013 年 4 月, Oracle 的虚拟化演变成生产系统的最佳做法
虚拟化 Oracle TCO 分析 3 年硬件/软件/维护 530 万美元 物理且无闪存 390 万美元 439 万美元 总体拥有成本降低 26% VMware FAST VP 缓存/内存 275 万美元 In the Wikibon research study entitled, “ Virtualization of Oracle Evolves to Best Practices for Production Systems” on of the key finding was, “IT operations should ensure that the best quality servers and that flash is aggressively used to minimize IO response times and variance.” This finding of using flash combined with the fastest possible servers (processors) translated into having to use fewer CPUs for Oracle Databases. The first bar shows a physical environment (no virtualization or flash) costing 5.3 million dollars of which Oracle software and licensing and maintenance accounted for million. The next bar in the graph shows the effect of using VMware with VMAX FAST VP enabling the use of fewer server processors and having a big savings on Oracle software licensing and maintenance. In fast there was a 26% lower TCO with 6 times the performance using FAST VP and VMware together. 性能提高 6 倍 Oracle 软件和 维护 Oracle 软件和 维护 来源:Wikibon,2013 年 4 月, Oracle 的虚拟化演变成生产系统的最佳做法
14
Accenture 和 Navitaire VMAX 帮助技术解决方案提供商实现零宕机 难题 解决方案 应用程序 Andrew Wilson
满足任务关键型数据的高可用性要求 需要技术更新和数据迁移 需要提高可扩展性、稳定性和性能 解决方案 VMAX 20K(SRDF 和 FAST) VMware 应用程序 Oracle Customer Video: vmax.htm Company background: Navitaire is a wholly owned Accenture Company providing technology and business solutions to the airline and travel industry since Accenture is a global management consulting, technology services and outsourcing company with 259,000 employees serving clients in more than 120 countries. Accenture generated net revenues of $27.9 billion for the fiscal year that ended Aug. 31, 2012. Navitaire’s integrated solutions include reservations, distribution, travel commerce for ancillary revenue, loyalty, revenue management, data store for business intelligence and revenue accounting. More than 50 airlines worldwide rely on Navitaire systems, including many of the world’s most successful hybrid and low-cost carriers. Navitaire reservations systems have processed more than 2.2 billion reservations since the initial product launch. In 2011, Navitaire reservation services processed 355 million passenger reservations. More than 70% of these reservations, or 250 million, were taken via the Internet, making Navitaire a recognized leader in online sales. Quotes: “What VMAX brings to us is the ability to scale beyond that which we could do, to scale predictably and then to operate predictably with a very high performance characteristics.” Andrew Wilson “VMAX gives us real-time, permanent uptime, highly available and an ability to introduce both new customers, new clients and to apply patches and maintenance releases in a very seamless way that has changed the service experience and changed the uptime characteristics of the end to end service and that was vital.” - Andrew Wilson “The business value of VMAX is the ability to deliver mission critical services that enable an entire airline for multiple airlines with hundreds of thousands of customers flying daily. And VMAX is a vital part of that. VMAX is needed because I need to deliver highly available services. VMAX is needed because I need to be able to maintain them in an efficient manner and I need to be able to operate and upgrade things in a manner that deals with very fast moving market that has a high demand for new functionality on a broad customer base. For all of those reasons I need world class enterprise level storage.” - Andrew Wilson VMAX brought a new set of technical capabilities to us. It brought the ability to upgrade. It brought the ability to maintain. It brought the ability to operate in highly available manner without taking down time.” - Andrew Wilson “We selected VMAX because we needed enterprise level storage that would deliver high availability performance in a predictable manner. VMAX has given us scale, it has given us very, very high availability and it has given us an ability to deliver more more quickly to more clients. “ – Andrew Wilson Andrew Wilson 高级基础架构经理 “我们选择 VMAX 是因为我们需要可以按预期提供高可用性和性能的企业级存储。VMAX 为我们提供了可扩展性。它不仅为我们提供了极高的可用性,而且帮助我们更加快速地向更多客户提供服务。” 结果 全球基础架构实现零宕机 系统响应时间缩短了 25–50% CPU 利用率从 50% 增加到 70% 复杂的报告作业减少了 60-80%
15
Lafarge 转变 IT 以加快创新并降低成本 Brent Wolfram 技术和安全体系结构总监 难题 解决方案 应用程序 结果
全球企业面对爆炸性数据增长,需要可扩展性 在异构存储环境中保护任务关键型应用程序 解决方案 EMC VNX 统一存储 利用 FAST VP 的 VMAX VPLEX RecoverPoint 应用程序 VMware vCloud Oracle JD Edwards ERP、Oracle 数据库 Microsoft SharePoint 和 SQL Server Customer Profile: Company background: World leader in building materials especially in cement, aggregates and concrete. Operates in 64 countries and has over 64,000 employees Previous Environment & Existing Applications Oracle JD Edwards, Oracle Database, IBM Cognos, WebSphere, IBM Maximo, CommandAlkon, Software AG webMethods, Microsoft SQL Server, Microsoft SharePoint, Lotus Notes Heterogeneous storage environment with mostly HP. Had experienced hardware issues Environment is 80% virtualized with goal to reach 100% in the next two years Company Challenges: Explosive data growth across global business leading to scalability issues Difficulty in protecting business-critical applications on its heterogeneous storage environment Competition: IBM, HP, Dell EMC Solution & Important Benefits to Customer: VCE Vblock VMAX VNX and VNXe VPLEX Metro (placed 40 km apart) RecoverPoint VMware vCloud Suite Managed exponential growth with less staff Reduced administration Results: Reduced storage floor space by 72 percent and power usage by 54 percent Improved application performance by more than 40% with VMAX Now offering ITaaS to the rest of the organization worldwide. Operating IT in a similar fashion as a Service Provider Quotes: “By innovating with virtualization and IT as a Service, we're managing resources with more precision and aligning IT services to the individual business units' needs. So we're reducing our costs while helping the business grow and innovate.” – Brent Wolfram “The combination of Vblock, VPLEX, and RecoverPoint will allow us to federate application processing across locations and enable workload mobility. We're also fine-tuning our metro cluster to provide the exact recovery point objectives and recovery time objectives across our full range of critical application environments." - Brent Wolfram “With Vblock systems and EMC VMAX, we can manage resources with much greater precision and align IT services exactly to each business unit’s needs.” – Brent Wolfram “The lack of availability of key business processes such as distribution, sales, and operational logistics would have significant cost and resource impacts, costing many thousands of dollars per hour. Now that we've moved to Vblock and VMAX, availability and performance have improved dramatically such that our operations are far more reliable and predictable.” – Marino DeLuca Technolgy Architect “Application performance has improved by more than 40 percent for some application environments,” De Luca reports. “We no longer have those bottlenecks that used to drag down productivity and slow down the pace of business.” “VPLEX 与 RecoverPoint 相结合可在一系列关键应用程序中提供精确的 RPO 和 RTO 目标。” Brent Wolfram 技术和安全体系结构总监 结果 VMAX 将应用程序性能提高了 40% 以上 VPLEX 改进了面向任务关键型应用程序的可用性 存储占用空间减少 72%,电耗减少 54%
16
助力 Oracle:VNX Powering Oracle Performance with the VNX
17
为什么选择将 VNX 用于 Oracle 一流的虚拟化 使用“闪存优先”策略提高性能 利用 VPLEX 实现最强的恢复能力
全面的 VMware 和 Windows Server 2012 集成 VAAI、VASA、ODX、SMB 3.0、Oracle Virtualization Manager (OVM) 使用“闪存优先”策略提高性能 FAST VP 自动分层 + FAST Cache = 最高的性能 利用 XtremSW Cache 提高 Oracle 服务器性能 Oracle RAC/dNFS 闪存性能 + 可管理性 利用 VPLEX 实现最强的恢复能力 跨阵列的主动-主动模式 跨站点的无缝应用程序可用性 So why VNX? We have superior virtualization integration than any other storage vendor. Why is this important. Well, Oracle licensing is per CPU. By virtualizing Oracle, you are using less CPUs per Oracle instance and therefore Oracle costs you less. And you know what, Contrary to popular believe, Oracle supports virtualizing Oracle. And another good selling point is that fact that Exadata does not support virtualized Oracle. VNX is VAAI, VASA and SMB3.0 integrated. VNX also9 support Oracle Virtualization Manager (OVM) Lets address that manual tuning issue we discussed on the previous slide. Why are Oracle Administrators always tuning Oracle, because they want more transaction per second out of their Oracle deployment. They are constantly tuning their setup to get the maximum number of transaction. Well guess what. With the FAST suite we can ensure that some workloads that require more or faster transactions per second are able to do so by utilizing flash drives. And those other workloads that don’t require that performance, can be placed on cheaper storage tiers. This is all done automatically. What is also good is that this is an on going process, meaning every day the system is rebalancing the system and ensuring the workloads are maximizing the performance / cost tradeoffs. Now, we go even further. If your Oracle administrator requires even more performance, for example for ready intensive workloads, VNX is integrated with XtremSF Cache for that extra performance layer. You want Oracle over FC, no problem. You want Oracle dNFS support again, no problem, the VNX has you covered. Additionally, for those mission critical Oracle deployments, those deployments where a Oracle customer cannot afford any downtime, well, VNX is integrated with VPLEX providing you that extra layer of availability like stretching Oracle RAC over distance. You want a highly available DR solution? Simple add RecoverPoint to your Oracle on VNX deployment. Whether you want local or remote protection VNX with RP has you covered. 利用 RecoverPoint 实现高可用性 跨平台的 CDP 和 CRR 解决方案 自动执行 VMware 灾难恢复测试、故障切换和回切
18
面向 Oracle E-Business Suite 的行业领先的性能
Oracle 和 Cisco 执行 Oracle 基准测试* 不到 1 小时即可处理 100 多万人 的薪酬! Here is an interesting one, Cisco did test for Oracle E- Business Suite. The conducted a test where they were able to process over 1 million paychecks in < 1 hr …setting a new industry 1st. And note, this was conducted with the 1st generation VNX….just imagine what they would be able to do the the new VNX. Use Case HR Payroll Processing (Batch) Application Oracle E-Business Suite R12 Oracle 11g Cisco UCS B200 M3 24-Core 128GB Memory VNX5500 FC Interconnect 75 Drives (39 TB Raw) Oracle ASM f * 2013 年 7 月 — Oracle 和 Cisco 使用一台 Cisco UCS B200 M3 服务器和一台 EMC VNX5500 执行了 Oracle E-Business Standard 基准测试,以衡量采用 Oracle11g™ 数据库的 E-Business Suite R12 (12.1.3) 的批处理性能
19
利用 FAST 扩展用户和性能 dNFS 132% > 2 倍 ESG 经验证的解决方案 可扩展性平稳期
(不采用 FAST Cache) 可扩展性平稳期 (采用 FAST Cache) 550,000 500,000 450,000 400,000 350,000 300,000 250,000 200,000 每分钟事务量 132% > 2 倍 When enabling FAST VP, five Flash drives were added to the data pool as RAID 5, the TPM increased by about 20 percent and stabilized at around 290,000, and the read response time was reduced by 42 percent. When enabling FAST Cache, we used four Flash drives. The TPM surged to around 510,000 and stabilized at that level. The response time was dramatically decreased to less than 10 ms. See the Wait statistics from Oracle AWR reports section for detailed analysis from the database side. The other two test results in Figure 17 show the performance of combining FAST Cache and FAST VP, which are two complementary technologies. When configuring four Flash drives for FAST Cache and five Flash drives for FAST VP, the TPM is slightly higher than that of using four Flash drives for FAST Cache only, and the read response time was reduced by 14 percent accordingly. When configuring two Flash drives for FAST Cache and five Flash drives for FAST VP, the TPM is slightly lower than when using four Flash drives for FAST Cache only, and the read response time was tripled. VNX File for Oracle / dNFS Transactions increased by 132% Response time reduced to <10ms 基准 采用 FAST Cache 每个节点的并发用户 ESG 经验证的解决方案
20
利用 VNX 和 FAST 来提高 Oracle 性能
dNFS 利用 FAST Suite, 读延迟缩短 5.2 倍 利用 FAST Suite, TPS 增加 2.2 倍 42% 63% 每秒事务处理数 读取延迟 38% This is a great example of how good the FAST suite is for virtualized Oracle dNFS – i.e. Oracle direct NFS. As you can see from both graphics VNX with the FAST Suite supercharges your Oracle dNFS environment. On the left, you can see the benefits of FAST VP and then the added benefit of FAST Cache on read latency. FAST VP alone has a 42% benefit on read latency and add to that the 67% benefit from FAST Cache resulting in a cumulative benefit of over 5x improvement on read latency. Likewise, on the right hand graphic, you can see the benefits of FAST VP and then the added benefit of FAST Cache on Transactions Per Second (TPS). FAST VP alone has a 38% benefit on TPS and add to that the 63% benefit from FAST Cache resulting in a cumulative benefit of over 2.2x improvement on TPS. Now imagine what it can do for your block environment. Note to Presenter (competitive differentiation): The unified design of VNX enables users to simplify Oracle provisioning and management by deploying on NFS, not available from HP 3Par or HDS HUS. These waterfall charts also demonstrate how VNX offers a fully flash-enabled solution. NetApp as well as emerging competitors such as Nimble provide flash as cache, but lack both the intelligent control used by VNX to focus on performance-centric data and an available flash tier to selectively maintain performance of continuously hot data. IBM V7000, HP 3Par and HDS HUS utilize autotiering, but lack the response advantage of FAST Cache. Data from: EMC VNX7500 SCALING PERFORMANCE FOR ORACLE 11gR2 RAC ON VMWARE VSPHERE 5.1 White paper: vnx7500-scaling-performance-oracle_11g-vsphere-wp- final.pdf?mtcs=ZXZlbnRUeXBlPUttQ2xpY2tTZWFyY2hSZXN1bHRzRXZlbnQsZG9jdW1lbnRJZD0wOT AxNDA2NjgwNmUwMDkyLGRhdGFTb3VyY2U9RENUTV9lbl9VU18w 67% 基准 基准 (5 个 SSDs) (5 个 SSDs) (7 个 SSD) (7 个 SSD) ESG 经验证的解决方案
21
提高 Oracle 性能,同时降低成本 American Tire 利用 EMC 技术优化 Oracle 性能和保护
挑战 业务发展导致硬件成本超额 不一致的 Oracle 数据库利用率和性能 耗时的备份和恢复流程 解决方案 VNX Data Domain VMware 应用程序 Oracle RAC Oracle E-Business Suite MySQL 群集 American Tire Distributors is one of the nation’s largest independent suppliers to the replacement tire market with 84 distribution centers servicing 37 states. In recent years tremendous growth led to a surge in transactions of 20-40% creating storage infrastructure requirements that could not be met cost-effectively by the company’s infrastructure. To address this the company went on an aggressive path to virtualize 99% of its physical infrastructure and application tiers on Vmware together with EMC VNX Unified Storage. The company has virtualized it’s Oracle E-Business Suite, Fusion Middleware and Oracle RAC database as well as MySQL database and Linux. Since virtualizing Oracle and deploying VNX the company has seen dramatic performance impact. Back-end processing times were reduced from 32 seconds to 14 seconds and on the front end 80% of customers can complete their orders in 1 minute. This delivers a faster, more streamlined ordering process for customers. The impact on their distribution centers has been equally as positive enabling them to load and ship products more efficiently. During peak periods, utilization on the Oracle database would exceed 90% slowing the system to a crawl. Now database utilization peaks at about 3%. Since virtualizing Oracle on EMC with VMware, American Tire has also seen better ROI on their Oracle software licenses. Their Director of IT writes: “When we’re paying Oracle license fees per CPU, cutting down the number of physical servers decreases our overall Oracle licensing expense significantly. With VMware, new tiers for Oracle Applications can be deployed within 20 minutes while applications remain online. The company has also reduced server count from 79 servers to just 9 not only saving on server costs but Oracle licenses as well. By moving from tape to EMC Data Domain the company has reduced backup times from 12 hours to 5 hours for their 2.5 terabyte Oracle database and restores take 4 hours compared to 36 hours before. The amount of headcount required to manage those backups has been reduced as well from 2 to 1. Go to: 结果 Oracle 后端处理时间缩短 65% 利用基于 VNX 的 VMware 将数据库利用率提高了 97% 利用 Data Domain 将备份和恢复速度分别加快了 2 倍和 9 倍,同时 管理人手减少了 50% 服务器数量减少 9/10,从而降低了 Oracle 许可成本 “由于我们按 CPU 支付 Oracle 许可证费用,因此可以削减物理服务器数量,我们的总体 Oracle 许可证费用进而大幅减少。” ANGELIQUE GIBSON IT 运营总监
22
利用“闪存优先”将 Oracle 性能提高 5 倍
Carglass 在提升 Oracle 性能的同时,减少了 50% 的磁盘 挑战 需要提高任务关键型应用程序的性能 漫长的恢复时间,导致业务处于风险之中 解决方案 VNX FAST Suite RecoverPoint 应用程序 Oracle Carglass is an auto glass specialist based in Germany, operating over 260 service centers. In 2011 it became clear their storage systems were going to reach their limits. They needed to find a way to ensure long term high levels of performance and reduce bottlenecks for their Oracle databases and applications. Carglass could either increase capacity of their existing storage array or migrate to a new storage platform. Carglass wanted to future proof their infrastructure and decided upon deploying EMC VNX Unified Storage together with FAST Cache allowing them to store frequently accessed data blocks in cache and less frequently accessed data blocks on lower cost disk. With FAST Cache despite an increase in storage capacity, Carlgass needs 50% fewer drives to manage them Performance has increased as well. Their IT manager writes: “Especially during peak loads, for example, when preparing quarterly financial statements, EMC’s Unified Storage system proved more than five times faster than the previous storage system. It’s a completely new paradigm for us.” Three VNX systems are deployed across three locations with EMC RecoverPoint providing closely spaced recovery points and application consistent recovery in the event of a site failure or data loss. What used to take 3 days now takes 2 hours in terms of recovery and tests can be run during replications to test recovery plans. Go to: Other Quotes: “Management simplicity is another major benefit that we’ve seen from deploying VNX unified storage. Thanks to the simplicity of VNX we now spend far less time on storage administration, maintenance and configuration.” – Frank Miller, IT Manager, Carglass “Our EMC VNX has worked very smoothly, without issues. It has proven to be powerful enough to withstand the highest possible loads with its simple and efficient design, which delivers consistent and predictable service levels, even as out workloads fluctuate constantly.” – Bernd Kuxdorf, IT administrator UNIX, Carglass “尤其是像准备季度财务报表这样的高峰负载时段,EMC 的统一存储系统证明其运行速度比以前的存储系统快 5 倍以上。这为我们带来了全新的模式。” FRANK MULLER IT 经理 结果 Oracle 性能提高 5 倍,同时磁盘使用量减少 50% 连续远程数据保护在 2 小时内提供恢复, 而非 3 天时间,并提 供无中断测试
23
利用 FAST Cache 将 Oracle 性能提高 15 倍
Ensco 基于 VNX 加快 Oracle 性能并缩短恢复时间 挑战 业务规模在 12 个月内加倍 数据中心与虚拟化实现整合 解决方案 VNX RecoverPoint VMware Site Recovery Manager 应用程序 Oracle PeopleSoft Microsoft SQL Server、Exchange 和 SharePoint The ENSCO group of companies has been providing ingenious engineering, science and advanced technology solutions that guarantee mission success, safety and security to governments and private industries around the globe. Ensco’s business had doubled in size in just over a year leading to rising IT costs and difficulty ensuring performance and data protection. To address these challenges, Ensco moved to a virtualized infrastructure on EMC VNX Unified Storage with VMware, together with EMC RecoverPoint and VMware Site Recovery Manager. Since deploying EMC VNX Unified Storage with FAST Cache and VMware, their Oracle/PeopleSoft batch processing times have gone from 10 hours to 40 minutes and Oracle/PeopleSoft recovery times have gone from 10 hours to 15 minutes. In addition, with Vmware virtual machines their IT department can now deploy test environments in minutes vs. hours it took before. Setting up virtual machines on EMC was much faster than with their previous storage, NetApp due to EMC, Vmware integration. It used to take about 10 minutes to create one VM on NetApp. With VNX, we can deploy several VMs in that same amount of time. Their Manager of IT Operations states: “PeopleSoft batch processes used to take eight to ten hours, but now using FAST Cache, they complete in 30 to 40 minutes. This allows our business users to make faster decisions, affecting everything from productivity and safety out on the rigs to customer satisfaction and profitability.” “PeopleSoft 批处理过去需要八到十个小时,而现在利用 FAST Cache,30 到 40 分钟就能完成。如此一来,我们的业务用户便可更加快速地做出决策,这影响着从工作效率到井台安全再到客户满意度和盈利能力的方方面面。” MARSHALL BOSE IT 运营经理 结果 利用 FAST Cache 将批处理速度提高 15 倍 利用 RecoverPoint/SRM 将恢复速度加快 40 倍 使用虚拟机在几分钟内创建测试环境,而以往需要几个小时
24
VNX 静态数据加密 (D@RE)* 基于控制器的加密 (CBE) 方法,可提供最大限度的保护和灵活性
2014 年第 3 季度 VNX 静态数据加密 基于控制器的加密 (CBE) 方法,可提供最大限度的保护和灵活性 在驱动器级别对阵列上的所有用户数据进行加密 支持所有 VNX 数据服务 嵌入式加密密钥管理器 高级加密标准 (AES) 加密 将提交进行 FIPS 级验证 在 VNX2 上进行无中断原位升级 VNX2 arrays will support controller-based data at rest encryption or in Q VNX2 uses Advanced Encryption Standard (AES) 256 encryption has been submitted for FIPS Level 1 validation. Designed to protect against physical disk loss, Controller-based encrypts the entire array with minimal performance overhead and works with both file and block data. supports ALL drives, and ALL VNX2 features. The key management is embedded with a unique key for each drive. Data at rest encryption is a requirement for Healthcare, Government and Financial environments. The key use cases are for physical disk security and in some cases can eliminate the need for data erasure services. ALL VNX2 arrays have already shipped with the encryption hardware included. The chargeable software to manage the encryption and keys will be available in Q Also all existing VNX2 systems can be upgraded in the field for to non-disruptively encrypt the existing drives as a background task. Note to Presenter (competitive differentiation): VNX Series controller based encryption supports ALL drive types (unlike self-encrypting drives) and can be enabled after the system is deployed -- at any time. The only hybrid array to offer controller based encryption is HDS – NetApp, HP 3PAR, and IBM use self encrypting drives. 防止 驱动器丢失 联邦条例 行业法规遵从性 往往消除了驱动器粉碎需求 * VNX 将于 2014 年第 3 季度提供
25
XtremCache In the following section we will show why XtremCache offers fast response times and very low latency to accelerate your business applications.
26
利用 XtremCache 加速业务 XtremCache 31% 更多的 IOPS 65% 更低的 延迟
利用主机端缓存提高数据库服务器性能 XtremCache 31% 更多的 IOPS 65% 更低的 延迟 As you have seen throughout this performance presentation XtemCache can be used in combination with both the VMAX and VNX storage arrays. The XtremSF card is an in server solution that uses Single-level Cells (SLC) or Enterprise Multi-level Cells (eMLC) memory to cache reads in the server. Generally, SLC memory is faster and eMLC memory is slightly slower but higher capacity. Using these cards in database server(s) can increase the number of IOPS as nearly 100% reads could be addressed in the server on by the XtremCache. The memory on these cards are extremely fast offering microsecond (one millionth of a second) performance which is exponential faster than the millisecond (one thousandth of a second) offered by storage arrays. Using these cards for Oracle databases could reduce the latency by as much as 65% for the majority of read related I/O.
27
Oracle RAC 支持* 通过分布式缓存一致性提供主动-主动共享存储
在运行 Oracle Clusterware 11g 的 Windows、RHEL 或 OEL 上运行 Oracle Database 11g,并采用以太网互连 每个群集多达 8 个节点 使用 SCSI-3 Persistent Reservation 阻止加入节点访问存储,直至得到 XtremSW Cache 的批准 XtremCache support for Oracle RAC enables active-active shared storage in an Oracle environment using a distributed cache coherency algorithm. This enables peer-to-peer communication between nodes. When a certain block of information is overwritten in shared storage and on a cache device, the other cache devices in the cluster are instructed to delete that block in their caches to prevent the use of invalid data. At installation time, XtremCache automatically recognizes the presence of Oracle RAC, and switches operation to clustering mode. When a node leaves or joins the cluster, XtremCache must know about it in order to prevent modification to shared storage by that node. This is done through integrating with the Oracle cluster management, and by using SCSI-3 Persistent Reservations to notify the backend storage to block the joining node from accessing the storage until XtremCache “approves” it is recognized. When a node leaves the cluster, all caches will move to pass through mode and be purged, to ensure coherency. It is recommended to use this feature to cache LUNs that hold data files. It is not recommended to cache redo logs, archives, temporary and grid data. XtremCache supports Oracle Database 11g on Windows, RHEL or OEL running Oracle Clusterware 11g with Ethernet interconnect and up to eight nodes per cluster. Note that support for Oracle RAC will be available in October 2013. * 2013 年 10 月发布
28
针对 Oracle 的 XtremCache 210% 性能提高 3.1 1 每分钟相对事务数 时间
Similarly, XtremCache was able to provide over two-times the IOPS. Obviously, every environment will differ, but this just gives you a sense of the results that can be achieved using XtremSW Cache. 解决方案体系结构由 Oracle 数据库、Cisco UCS C-460 服务器和 EMC Symmetrix VMAX 构成。EMC 测试了 XtremSW Cache 的各项功能。测试采用了一个类似 TPC-C 的标准 OLTP 工作负载,带一个 1.2 TB 的数据库,读/写各占 70%/30%,热数据 250 GB。XtremSW Cache 在所有数据 LUN 上启用,但在日志 LUN 上不启用。没有对 Oracle 数据库进行任何调整。Symmetrix VMAX 存储系统:4 个引擎,512 GB 的镜像缓存;在 64 个 450 GB 15K rpm 光纤通道驱动器 (RAID 5) 上有 16 个 250 GB 数据 LUN,在 16 个 450 GB 15K rpm 光纤通道驱动器 (RAID 5) 上有 4 个 200 GB 日志 LUN;数据库缓冲区缓存 4 GB; 服务器使用 4 个 8 Gb/s 光纤通道连接到 Symmetrix VMAX。
29
缩短 60% 针对 Oracle 的 XtremCache 1 0.4 响应时间 相对的应用程序级延迟 时间
Let’s talk about performance. Simply put, XtremCache is about making your application wait less while simultaneously enabling more transactions. The next few slides show some of the results we have seen in the testing of XtremCache. XtremCache has been tested in physical and virtual environments, in Oracle and Microsoft SQL environments, on Symmetrix VMAX and VNX, with EMC FAST software and without. And there are a variety of white papers on EMC.com that document this testing and the results. In this example, EMC tested the capabilities of XtremCache to accelerate OLTP performance in an Oracle Database environment running on Cisco UCS and Symmetrix VMAX. Testing employed a standard TPC-C like OLTP workload with a 1.2 TB database and a 70–30% read/write mix with 250 GB of hot data. The results show before and after XtremCache was implemented. As you can see, XtremCache reduced response time by 60 percent. 解决方案体系结构由 Oracle 数据库、Cisco UCS C-460 服务器和 EMC Symmetrix VMAX 构成。EMC 测试了 XtremSW Cache 的各项功能。测试采用了一个类似 TPC-C 的标准 OLTP 工作负载,带一个 1.2 TB 的数据库,读/写各占 70%/30%,热数据 250 GB。XtremSW Cache 在所有数据 LUN 上启用,但在日志 LUN 上不启用。没有对 Oracle 数据库进行任何调整。Symmetrix VMAX 存储系统:4 个引擎,512 GB 的镜像缓存;在 64 个 450 GB 15K rpm 光纤通道驱动器 (RAID 5) 上有 16 个 250 GB 数据 LUN,在 16 个 450 GB 15K rpm 光纤通道驱动器 (RAID 5) 上有 4 个 200 GB 日志 LUN;数据库缓冲区缓存 4 GB; 服务器使用 4 个 8 Gb/s 光纤通道连接到 Symmetrix VMAX。
30
用于 Oracle 服务器整合的 XtremSF
IOPS 115K 27 倍 性能提高 相较于本地磁盘 In this next example, EMC tested the capabilities of XtremSF to accelerate OLTP performance in an Oracle database environment running on Cisco UCS. Testing employed a standard TPC-E like workload with a 1.2 TB database. Again, the results show the performance difference between placing the data on ten 600 GB SAS drives versus one EMC XtremSF 2.2 TB device. XtremSF delivered 27 times better performance in terms of IOPS and 115,000 transactions per minute. This shows that XtremSF greatly minimizes the number of back-end mechanical drives required to achieve similar optimal performance. And XtremSF frees up server resources, thereby allowing a single server to handle a greater workload serving multiple databases or applications. This reduced infrastructure translates into lower TCO with regard to reduced cooling and power requirements and lower maintenance costs. 115,000 个事务/ 每分钟 4.3K 解决方案体系结构由 Oracle 11g R2、RedHat Enterprise Linux 5.7、配备 512 GB 内存和十个 600 GB SAS 驱动器的 Cisco UCS C-460 M2 架装式服务器以及 EMC XtremSF 2200 构成。测试采用了一个类似 TPC-E 的标准 OLTP 工作负载,带一个 1.2 TB 的数据库。没有对 Oracle 数据库进行任何调整。
31
利用 XtremCache 提高 Oracle BI 性能
挑战 数据超过 1 PB,并且每年以 30% 的速度增长 复杂的数据库查询影响性能 需要采用可以提供更出色的可用性和数据保护的模式 解决方案 VNX 和闪存优先 XtremCache Data Domain/Networker Cancer Research U.K.’s mission is to save lives from cancer through funding world class research, providing information for patients, health professionals and the public. Over a petabyte of data running on two systems growing at 30 percent a year was straining their storage systems. Complex database queries used by marketing and sales to drive fundraising took longer than necessary. Opening a simple could take 30 seconds or more. In an effort to increase performance and better utilize capacity, their IT organization deployed a virtual private cloud including Vmware for Oracle and SQL Server databases as well as Siebel and CRM applications, a total of more than 780 windows and Linux server machines. CRUK deployed EMC VNX Unified Storage with the EMC Total Protection Pack and FAST Suite as well as flash drives. CRUK also replaces its legacy outsourced backup and recovery architecture with EMC Data Domain integrated with EMC Networker. Through EMC VNX and a Flash 1st policy, CRUK has been able to achieve 30% more performance. With the deployment of XtremSW Cache, CRUK expects a 4X increase in the number of complex queries they can run per day. Their head of infrastructure states: “IOPS from the flash drives results in a 15 percent performance increase. Add to that the high level of VNX performance and Flash 1st and the result is a 30 percent performance increase for many applications” Go to: “闪存驱动器提供的 IOPS 将性能提高了 15%。除此之外,还提供高级别的 VNX 性能和闪存优先功能,因而许多应用程序的性能提高了 30%。” MICHAEL BRIGGS 基础架构主管 结果 应用程序性能提高 30% 数据库查询时间从 4 个小时减少到 40 分钟 Total Protection Pack 简化了可用性
32
针对 Oracle 的 最佳做法 Leveraging EMC flash and storage tiering solutions enable customers a better performance standard for mission critical databases and applications.
33
面向 Oracle 数据库管理员的 EMC 社区
活动日历和员工培训 互动式博客 实况演示 数据库管理员 数据库 管理员 EMC 针对 Oracle 的虚拟基础架构 闪存以及提高 存储可见性和控制力 Oracle 性能 加快部署、提高可用性并降低 IT 成本 Oracle 融合 基础架构 DBA 控制备份、恢复和复制 Oracle 备份和 恢复 逻辑和物理恢复与主动-主动可用性 Oracle 连续 可用性 The Everything Oracle at EMC community is the place for all EMC solutions for Oracle. If you are interested in learning more about any of the topics discussed today you can find solutions, blogs, videos and other great content on the community. 优化 Oracle 基础架构并最大限度地提高 DBA 工作效率
35
XtremIO The XtremIO all flash sets the new performance standard for mission critical databases and applications. In the following section we will show why the XtremIO offers fast response times and very low latency to accelerate your business applications.
36
想必您已经听说过 XtremIO 令人难以置信的性能以及闪存优化的数据服务 精简资源调配 内联式数据减少 数据保护 可写快照
Inline Data Reduction: Truly inline, global, always on, Improves array performance, Scales-Out, Lowers cost of ownership, Extends flash endurance, Array services built on top Thin Provisioning: Always thin provisioned, 100% performance, Perfect 4KB allocation granularity, No fragmentation or garbage collection Data protection: Not pre-existing RAID, N+2 redundancy per X-Brick, 8% ultra-low capacity overhead, 2.4x superior flash endurance, 1.6x better than RAID 1 performance, Rapid, distributed, content-aware rebuilds, User fault tolerant Writeable Snapshots: Instant, completely space efficient (both data and metadata) writeable volume copies that can be in any topology and which can be used just like the parent volume – full flexibility and full performance. Allows huge consolidation of dev/test environments with a small capacity footprint. 绝不写入 重复数据 4 KB 分配 最佳利用率 闪存优化 零配置 即时拷贝 完整性能
37
XtremIO — OLTP 价值主张 令人难以置信的性能 提高灵活性 — 免费测试/开发数据拷贝 管理接近零 — 易用性
一致的低延迟性 线性可扩展性 从不调整! — 不需要规划 提高灵活性 — 免费测试/开发数据拷贝 管理接近零 — 易用性 最低的 TCO — 节省资本和运营成本 节省 Oracle 许可费用 DB 整合 — 节省大量数据中心成本 Flash drives are faster than traditional high capacity drives as they have no moving parts (spinning platter and spindle) and maintain data in memory cells. Because there is no physical dependency on a rotating platter access to data is very fast (low latency). Flash drives offer approximately 2000 IOPS and 1 ms latency times and the fastest traditional drives (FC, SATA, SAS) peak at 200 IOPS and 3 ms of latency time. Adding more flash drives to the storage array means the business can take advantage of near linear scalability and not have the management overhead of tuning related to traditional disk drives. XtemIO increases scale-out with a building block approach to adding more performance and capacity. To scale-out simply add more ‘X-Brick’ scaling units. XtremIO arrays allocate capacity to volumes on-demand in granular 4KB chunks. Volumes are always thin provisioned without any loss of performance, overprovisioning of capacity, or fragmentation. Oracle database initialize very block in the database meaning that each block is unique so in terms of thin provisioning XtremIO will very quickly provision a full copy of the Oracle database. The scale-out approach to improving performance has benefits in terms of management too. Traditionally, storage administrators had to architect placement of databases and applications however, with XtremIO everything (all data) is uniformly stripped across all disks and X-bricks. This completely eliminates complex setup and tuning steps, while inherently delivering maximum performance. Create and size volumes in a few clicks, map them to house with a few more, and the array is configured. A few other key features include: Deduplication-enhanced copy services Advanced hypervisor integration Ultimate efficiency Surprisingly affordable With all these features and performance XtremIO has one of the lowest TCO: capital and operational savings
38
CMA 客户概要信息 Oracle RAC 数据仓库 结果 BRIAN DOUGHERTY 挑战 解决方案 应用程序 五个九可用性和零宕机
极高性能(6 GB/秒大型数据块随机工作负载) 高用户并发性 需要极快的响应速度 复杂的设置/调整 解决方案 XtremIO 全闪存阵列 VMAX 40K SRDF TimeFinder 应用程序 美国最大的医疗补助分析数据库 (Oracle RAC) 结果 五个九可用性和零宕机 可预测且可靠的响应时间 灵活的开放式体系结构支持可扩展性 通过易管理性降低 TCO BRIAN DOUGHERTY 首席数据仓库架构师 “我们能够更加快速地设计 Oracle RAC 系统、提高其可扩展性并满足更多并发用户的需求,同时在数据中心内的占用空间仅为以前的 20%。”
Similar presentations