甲骨文携手伙伴步入数据库云时代.ppt

下载

下载

16积分

20积分

下载

20积分

VIP价：16积分

甲骨文携手伙伴步入 数据库云时代 Oracle Confidential – Internal Copyright © 2014 Oracle and/or its affiliates. All rights reserved. | 1 应用程序 平台 定制 应用程序 定制 定制 客户 客户 服务 提供商 服务提供商 服务类型： IaaS、PaaS、SaaS IaaS 云 SaaS 云 PaaS 云 IT 专业人士 开发人员 业务最终用户 不同用户 整合 节省成本 应用程序开发 专注 新功能 速度 关键 驱动因素 Copyright © 2014 Oracle and/or its affiliates. All rights reserved. | 2 部署模型： 私有、公有、混合 私有云 公有云 混合云 运营支出 外包服务 由多个组织共享 资本支出和运营支出 控制和可见性 独自占有 云爆发 － 透支以应对峰值负载 开发/测试和生产 B2B 集成 专用 由多个组织共享 Copyright © 2014 Oracle and/or its affiliates. All rights reserved. | Today, customers have a choice of cloud models. They can choose among private clouds, public clouds and hybrid clouds. A private cloud is for the exclusive use of a single organization and is typically controlled and managed by in-house IT A public cloud is shared by multiple tenants on a shared basis, and it hosted and managed by someone else (a service provider). It’s a form of outsourcing. There are some basic trade-offs when deciding between private and public clouds. Public clouds are purchased out of operating expense, avoiding capital expenditures. Public clouds are faster and cheaper to get started, providing a “low barrier to entry.” However, private clouds can offer lower costs over some breakeven period. For IT projects, this tends to range from 2 to 4 years, averaging about 3. This is the similar to the decision some people make between owning vs. renting. Another trade-off is between public cloud outsourcing everything to a service provider, which is great if you don’t have the datacenter, hardware or employees, versus how in a private cloud, you are able to maintain control and visibility over security, regulatory compliance and quality of service. Both public and private clouds involved sharing or pooled resources, but in a private cloud, the sharing is different apps using shared platform or infrastructure all within a single organization. In a public cloud, the sharing may be the same or different apps among different tenants. Hybrid: fuzzy term with different meanings. Cloudbursting – steady state in private cloud, but when a spike or peak happens, use a public cloud, then return capacity to the pool when it’s no longer needed Hybrid across app lifecycle – dev/test in one cloud, production in another Hybrid can also mean different functions are performed in different clouds, such as for a business process that spans clouds, requiring B2B integration between them 3 PaaS 赶超 IaaS 平台即服务 (PaaS) 基础架构即服务 (IaaS) 信息来源： IOUG ResearchWire 成员于 2010 年 8-9 月份和 2012 年 8-9 月份的云计算调查 私有 2010 2012 2010 2012 公共 Copyright © 2014 Oracle and/or its affiliates. All rights reserved. | One last important finding: Our customers are adopting Platform-as-a-Service significantly more than Infrastructure-as-a-Service. As you can see from this graphic, the platform level services, such as app server, database, identity, are more prevalent than the infrastructure level services, such as raw compute, storage and dev/test (which is IaaS because it’s primarily the use of virtualization technology to share dev/test hardware). The rise of PaaS is something that others in the industry have also observed, but this survey really gives us quantitative evidence of this. When you think about the reasons for this, it makes sense. PaaS provides higher level of standardization, component re-use and sharing, with lower levels of heterogeneity, complexity and cost. PaaS also enables faster development, because developers can leverage the platform instead of re-inventing the wheel themselves. In contrast, IaaS enables sharing of hardware and some cost savings in hardware, power, cooling and data center space, but it does nothing to minimize high cost of managing the heterogeneous, complex and brittle environments above the hardware layer. 4 PaaS的快速增长 Oracle Confidential – Internal/Restricted/Highly Restricted 5 72% “未来5年的预计的接受率” 27% “当前 PaaS 的接受率 ” Source: ComputerWorld & Triangle research study, 2014 Copyright © 2014 Oracle and/or its affiliates. All rights reserved. | 5 在 PaaS 和 IaaS 层整合 应用程序 IaaS PaaS 整合到共享的 IaaS 而不进行标准化 整合到标准、共享和可灵活伸缩的 PaaS 对比 应用程序 应用程序 应用程序 应用程序 应用程序 应用程序 应用程序 应用程序 通过减少硬件、能耗和数据中心空间来节省成本 软件体系异构性、成本和复杂性没有改观 不能节省管理 (O&M) 成本 适用于所有应用程序的标准化 PaaS 可减少异构现象、降低成本以及减少复杂性 加快新应用程序开发速度 通过减少硬件、能耗和数据中心空间来节省成本 Copyright © 2014 Oracle and/or its affiliates. All rights reserved. | Consolidation means moving multiple applications onto pooled or shared computing resources. You have a very strategic decision here. You can move onto shared, standardized platforms or shared infrastructure. <CLICK>Consolidating at the infrastructure level is typically done using virtualization technologies to share compute servers, and it can be done without standardization of the software stack. Sharing hardware enables you to run at higher utilization rates, which delivers cost savings from using less total hardware, power, cooling and data center space. However, the software stacks are unchanged, which means you still have the cost and complexity of these heterogeneous software stacks. <CLICK>Consolidating at the platform level means standardizing on a common Middleware and/or Database architecture for multiple or all applications. This standardization reduces the heterogeneity of applications, which is the real driver of complexity and cost. At the same time, a standardized platform speeds development of new applications. As the IOUG survey and customer examples show, more of our customers are consolidation at the PaaS level. Although it’s more difficult to standardize, but customers are doing it because there’s more value to do it, in terms of cost, efficiency and speed. 6 “数据库是我们最大的软件生意，数据库也将是我们最大的云服务的生意。” Larry Ellison, CTO, Oracle Q1 2015 Earnings Call on 9/18/2014 Copyright © 2014 Oracle and/or its affiliates. All rights reserved. | This is a Quote with Picture slide ideal for including a picture with a brief quotation and attribution. To Replace the Picture on this sample slide (this applies to all slides in this template that contain replaceable pictures) Select the sample picture and press Delete. Click the icon inside the shape to open the Insert Picture dialog box. Navigate to the location where the picture is stored, select desired picture and click on the Insert button to fit the image proportionally within the shape. Note: Do not right-click the image to change the picture inside the picture placeholder. This will change the frame size of the picture placeholder. Instead, follow the steps outlined above. 7 DATABASE ESTATE 当前数据库部署情形 日益增长的数据库数量 多种版本和补丁级别 合规性要求 较低的系统资源利用率 高居不下的部署和运维成本 28% 的客户的数据库实例的数量每年增长超过 20% 不超出 50% 的数据库实例已经被整合 *IOUG Survey, 2013 烟囱的、分散的、多变的、复杂的 Copyright © 2014 Oracle and/or its affiliates. All rights reserved. | 8 The first challenge lies around the problem of silo’ed, dispersed and varied database deployments. With grid computing, we have witnessed a huge proliferation of databases across dev, test and production lifecycles. Large enterprises often have hundreds and thousands of databases of various versions, configurations and patch levels. At the bottom of the slide we see data from a recent survey which shows lot of customers having annual instance growth of more than 20% and 50% of those deployments are not consolidated. This leads to deployment, administrative and compliance challenges. 当前数据库部署的流程 耗时 低效 开发人员 申请 DB 获取数据库连接 经理 同意 数据库管理员 申请硬件 配置群集/创建数据库 系统管理员 安装OS/网络 存储管理员 分配存储 … Days to Weeks … Copyright © 2014 Oracle and/or its affiliates. All rights reserved. | The second challenge is around time to provision new databases. When an end user, be it a developer or a QA engineer, needs a database he or she typically has to go through an approval process, which then translates into a series of tasks for the DBA, the sysadmin and storage admin. The sysadmin has to provide the compute while the storage admin has to provide the space on a filer. Finally, the DBA would install the bits, create the database, optionally on Real Application clusters, and deliver that to the user. This is a cumbersome and time-consuming process that in case of some of our customers, may span days. 9 高风险 企业数据库平台面临的问题 糟糕的敏捷性 部署时间6+月 高昂的成本 服务器利用率= 5-10% 非计划停机时成本为每分钟 $5000 数据库许可证的费用 服务器，存储的费用 机房的开销 人员的成本 从项目立项，到项目实施完成 往往在一年以上，从而错过了 商机。 非计划停机往往不仅仅是经济成 本可以计算的。 Copyright © 2014 Oracle and/or its affiliates. All rights reserved. | 10 DBaaS ：使命 全面满足需求，全力达成所愿 最终用户的期许 方便的自助DBaaS自助服务 可计量的使用 依据服务水准的性能支撑 IT 的想法 标准平台，简化部署 更少的运维，更好的支持 更多地参与创新 DBaaS Copyright © 2014 Oracle and/or its affiliates. All rights reserved. | The above challenges provide the background behind the goals of database as a service. Users want database services on demand delivered via self-service. They do not want to be gated by IT for every single request. They prefer utilization based metering and a performance guarantee in the form of service levels. IT, on the other hand wants simplified deployments, standardization, less administrative headache and eventually, have more time for innovation as opposed to just keeping lights on. 11 Private Cloud Oracle Cloud DEPLOY ANYWHERE Same 技术 Same 标准 Same 产品 自由的在私有云和共有云之间移动 全面的选择 随需部署 Copyright © 2014 Oracle and/or its affiliates. All rights reserved. | Main Point: So no matter whether you want deploy your application in an On Premise Private Cloud or in Oracle Cloud you have a choice. With Oracle Cloud Technologies, you can transparently move workloads between on premise and oracle cloud. If you decide to develop in Oracle Cloud and later decide to deploy to a Private Cloud, know that you are sharing the same technology with the same standards using the same Oracle Products that you are already familiar with. This will save you time, effort and money… Private Cloud: Self-service provisioning of Oracle Fusion Middleware & Database on Virtual, Clustered or Engineered Systems Programmatic access with APIs & CLI On-demand delivery of resources with policy-based elasticity Load Balancer configuration for HA & throughput Automated Backup, Recovery, Patching & Upgrades Comprehensive dashboards & reports with rich self-service Oracle Cloud Instant access to full functionality of Oracle Fusion Middleware, Database, Compute, Storage & Network Only Public Cloud offering Fully Managed PaaS with Full VM Access & Control Automated Backup, Recovery, Patching & Upgrades Comprehensive dashboards & reports with Rich self-service DBaaS: 关键特性 DBaaS提供 共享、整合的平台用于数据库服务的供应和部署 服务目录 自助服务 服务的随需扩展，生命周期管理 计费/计量 DBaaS满足各种使用场景 新启动的项目需要新的数据库服务 QA需要完整的数据库用于性能测试 QA需要数据库用于部分的功能测试等 Copyright © 2014 Oracle and/or its affiliates. All rights reserved. | So, what’s the promise of an ideal database as a service aka “database cloud” solution? It should provide: First a consolidated platform to provision databases on in a repeatable manner It should enable self-service provisioning It should also provide facilities to scale up/scale down, scale out, scale back or relocate database services Finally, there should be a way to meter the usage Database as a service should also cater to various use cases and personas. For example, the requirements of a developer are very different from those of a DBA. And the requirements of a DBA are different from that of a QA Engineer. 13 Oracle客户通过建设DBaaS实现效益 10 倍 部署 速度提升 敏捷性  50% 总拥有 成本降低 成本  100% 服务级别 目标达成 风险  HDFC 银行把系统部署时间从56天降到3小时! Copyright © 2014 Oracle and/or its affiliates. All rights reserved. | HDFC银行：印度的一家私营银行 印度第四大银行 最应以自豪的是零售业务，最大的抵押贷款银行 挑战 建立在 168 个机架单元上的零售资产基础架构占用昂贵的场地 (DC) 空间 使用 5 年以上的现有老旧存储难以进行维护 容易出错的手动任务 有限的应用程序分区和 IO 重新平衡 实施 Exadata 和非 Exadata 上的云区域 RAC 数据库的自助式供应 补充管理功能的配置管理特性 优势 将空间占用从 168 个机架单元减少到 60 个 可扩展的系统能够最多支持三倍的当前业务量 批处理时间缩短 5 倍（从 10 小时降至 2 小时） 缩短有关贷款申请的“新产品”上市时间 IT 资源的 C 级可见性 阿尔法自然能源公司 Alpha Natural Resources, Inc. (Alpha) 是一家煤炭生产商﹐公司生产的煤主要是发热量高含硫量低的锅炉用煤和炼焦用煤。公司自己拥有44个地下煤矿﹑20个露天煤矿和11个预处理工厂。 美国500强公司 1, information of Alpha Natural Resources case EXTERNAL: Alpha is a leading global coal company and the world’s third largest metallurgical coal supplier. ANR has the production capacity of nearly 126 million tons of steam and metallurgical coal. Steam coal is primarily purchased by large utilities and industrial customers as fuel for electricity generation and manufacturing, and metallurgical coal is used primarily to make coke, a key component in the steel making process. ANR produces, processes and sells steam and metallurgical coal from approximately 150 active mines and 40 coal preparation plants located throughout Virginia, West Virginia, Kentucky, Pennsylvania and Wyoming. We are also actively involved in the purchase and resale of coal mined by others, the majority of which we blend with coal produced from our mines to create customized products for our customers. Alpha Natural Resources Gains 10X Performance Improvement Company:  Alpha Natural Resources  Video   |   Natural Resources   |   Published on January 03, 2013 ANR used to outsource the running of their Oracle E-Business Suite (for AP, AR payroll, etc.) to IBM at an IBM facility. They ran these workloads on hosted IBM x86 Systems. ANR purchased 2 Exadata Database Machines and deployed them in-house for their EBS work. The migration to Exadata was done quickly and without issues in a short period of time (how long). Performance improved 5x for many critical functions. Etc. Costs were dramatically reduced – how much, around 50% Other benefits  - our ability to manage and more transparent monitoring, more nimble 14 全面管理 指定的概要 容量大小 5GB, 20GB, 50GB 数据库访问: Java service, Oracle Application Express, Oracle RESTful Web Services 数据库概要 共享数据库 虚拟镜像 虚机 + 数据库 + 磁盘镜像 与 Azure image or AMI类似 磁盘镜像分发 与on premise相同 数据库云服务 虚机 + 数据库 + 全面供应 自动备份恢复 一键式补丁/升级 数据安全卫士配置 监控/管理门户 DBaaS：Oracle cloud 公有云解决方案 操作系统：Oracle Linux 硬件：On-demand storage & compute 版本：Choice of editions SE1, EE and Database 12c , and 11g 选件： EE High Performance (most options), EE Extreme Perf (all options) 网络：Full network, VM and OS isolation, Full SQL*Net access 接入：Self-managed with SSH access into VM with root privilege 基于数据库实例的服务 Database as a Service 2014年9月 2012年10月 基于Schema的服务 Database Schema Service Copyright © 2014 Oracle and/or its affiliates. All rights reserved. | DBaaS：Oracle cloud 丰富的管理工具和界面 Copyright © 2014 Oracle and/or its affiliates. All rights reserved. | Oracle Confidential - For Internal Use Only 16 DBaaS：数据库私有云的实现演进路线图 物理 专用和异构 静态、孤立的分析 标准化的硬件和软件体系 标准部署配置 数据库服务和服务级别目录 共享和安全的中央数据基础架构 动态优化和资源管理 自动化系统管理 按需、弹性、分层的自助服务 快速服务伸缩和自动化 计量、自动成本分配和收费 完全动态、统一的资源池 IT 作为云代理：仲裁和代理 安全的混合云集成（供应商、合作伙伴等） 传统孤岛 标准化的平台 整合的平台 服务交付平台 企业云平台 降低风险 降低运营支出 降低资本支出 提高敏捷性 全面优化 整合 私有 DBaaS 混合 DBaaS 孤岛 标准化 简化 高效 灵活 统一 当前 客户状况 1970年代 2000年代 2010年代+ Copyright © 2014 Oracle and/or its affiliates. All rights reserved. | 每个技术的最早出现具体年代 17 Oracle提供完整的从规划到实施的DBaaS服务 资源与规划 设计与构建 开发测试与部署 运行与维护 高级解决方案研讨会 上线技术支持 上线就绪准备情况检查 系统配置检查 系统性能检查 补丁检查 安装与配置 上线前 上线&投入生产 生产诊断检查 技术管理 高级技术支持 硬件清单, 软件清单 高级的监控与处理 云管理 Oracle数据库云 云安全 Copyright © 2014 Oracle and/or its affiliates. All rights reserved. | 规划，实施等 18 Oracle DBaaS规划与设计四个步骤 青铜级 白银级 黄金级 小 大 中 超大 RAC Data Guard 备份 PDB 数据库 模式 私有 公有 提供商 混合 Copyright © 2014 Oracle and/or its affiliates. All rights reserved. | 19 DBaaS 服务定义要素 服务类别 从DBaaS客户的角度来讲，服务定义必须包含服务能力、策略以及一些列流程 Copyright © 2014 Oracle and/or its affiliates. All rights reserved. | 攻击向量（attack vector）指的是黑客用来攻击计算机或者网络服务器的一种手段。攻击向量能够帮助黑客寻找系统可能存在的任何漏洞，包括认为因素。 关键词：attack vector 攻击向量 攻击向量（attack vector）指的是黑客用来攻击计算机或者网络服务器的一种手段。攻击向量能够帮助黑客寻找系统可能存在的任何漏洞，包括认为因素。 　　攻击向量的手段包括病毒、电子邮件附件、网页、弹出窗口、即时消息、聊天室以及其他的欺骗方法。所有的这些手段里面，除了人工欺骗，其余的都用程序完成，而人工欺骗这种手段是指利用常规的欺骗方法蒙蔽计算机管理人员，进而暴露出系统的漏洞。 　　防火墙和防病毒软件可以在一定程度上防止攻击向量的入侵。但是实际上，任何防护措施都不可能十全十美。黑客们已经变得越来越精明了，他们将传统的攻击向量不断进行升级，已经可以克服传统的防护手段，现在黑客们已经开始倾向于寻找新的攻击手段，这种攻击手段能够不经过授权就可以进入计算机以及服务器。 　　最常见的恶意攻击就是病毒了，比如特洛伊木马、蠕虫以及间谍软件。事实上，不要小看这些攻击向量的能量，它们可以造成的损失可能会比一个传统的导弹带来的攻击还要大 20 DBaaS的服务等级与SLA定义 服务等级 服务等级描述 DB 环境 RPO RTO 业务关键型应用：零应用停机，零数据丢失 生产环境 0~5分钟 < 1 小时 非业务关键型应用：分钟级的应用停机窗口，接近零的数据丢失 生产环境 5分钟 < 2 小时（至少） 所有的非生产运行环境：天级的应用停机窗口， 数据从最新的备份中恢复 所有的非生产运行环境 1 天 72 小时 正常运行时间不包括计划性停机时间 Recovery Time Objective (RTO) – 服务能够正常使用前的所有时间（不包括数据库恢复时间） Recovery Point Objective (RPO) – 业务能够容忍丢失的数据总量 金 银 铜 -------- -------- -------- -------- -------- -------- -------- -------- -------- Copyright © 2014 Oracle and/or its affiliates. All rights reserved. | 此处服务等级定义仅为参考，具体客户定义可以根据客户实际情况有所不同，比如后面瑞银的案例。 21 DBaaS 铜牌高可用服务 面向非关键应用的基本高可用 本地备份 RAC One Node 异地磁带存储备份 云 服务交付解决方案的入口 负载均衡 RAC One Node 数据防腐 DB parameters, ASM, RMAN, DBVerify 防止介质或站点失效进行备份 异地磁带存储备份的安全 Copyright © 2014 Oracle and/or its affiliates. All rights reserved. | RAC One Node是数据库企业版的一个选件， 11gR2 中引入的特性， 可以实现冷切换的高可用性，实例只在一个节点上运行，当出现问题时将实例转移到集群中中的其他节点上运行。 Oracle DB VERIFY是一个可以用来查看数据库物理文件正确性的命令工具，可以在线/离线情况下对数据库进行检查，甚至可以对数据库的备份文件进行检查。 22 DBaaS 银牌高可用服务 在铜牌高可用基础上，增加: 服务器和实例 HA Oracle RAC 数据库和站点 HA/DR Data Guard Active Data Guard 更高级的防数据腐败和数据丢失的保护措施 Data Guard Active Data Guard 面向关键业务应用的高可用 Backups Site A RAC Site B Active Data Guard RAC Backups Active Data Guard Active Data Guard Data Guard Copyright © 2014 Oracle and/or its affiliates. All rights reserved. | Site Guard 是一个灾难恢复 (DR) 解决方案，它让管理员能够自动完成站点故障切换。Site Guard 可编排 Oracle 融合中间件、Oracle 数据库的协调故障切换，并可通过扩展来支持其他数据中心组件。 23 DBaaS 金牌高可用服务 在银牌高可用 基础上，增加: 零应用停机 应用的业务连续性 通过WAN实现零数据丢失 Active Data Guard Far Sync 零维护停机 Oracle GoldenGate 自动化数据复制 Global Data Services 确保关键任务的业务连续性 Site A RAC Site B Active Data Guard RAC Backups Backups 应用的业务连续性 GoldenGate 全局数据服务 Active Data Guard Active Data Guard Copyright © 2014 Oracle and/or its affiliates. All rights reserved. | Active Data Guard Far Sync 和 Global Data Services 是Oracle DB 12c 的新特性。 24 DBaaS：整合模型 虚拟机 共享服务器 专用的schema 共享服务器、操作系统和数据库 整合度更好 专用数据库 共享服务器和操作系统 可插拔的数据库 共享服务器、操作系统和数据库 Copyright © 2014 Oracle and/or its affiliates. All rights reserved. | First, cloud architects and database administrators have to choose the right consolidation strategy. Let us look at some prevailing consolidation strategies. One can use server virtualization to provision databases packaged as VMs. This model is simple, provides operating system level isolation, but results in VM sprawl and other administrative challenges. The second model offers better consolidation where a number of databases can share the operating system. These databases can be deployed on single instance or RAC. One can also think of consolidating further by sharing databases for various applications. This can be achieved by deploying multiple schemas within the same database. As we see, all these models have their pros and cons…….. 25 整合模型的比较 Virtual Machines Dedicated DB Dedicated Schemas Pluggable DB 整合密度 较低 高 非常高 非常高 运维 非常复杂（虚拟机 无序扩张） 容易 相对容易 非常容易 隔离度 很好 好 一般 好 实施 简单 简单 困难 简单 应用适用性 部分（取决于负载） 所有 自行开发或者APP认证 所有 Copyright © 2014 Oracle and/or its affiliates. All rights reserved. | Let us now look at a summarized comparison of the various consolidation models. As we see here, pluggable database combines the best of all the other models and offers excellent consolidation, isolation, manageability and is suitable for any application that is certified to run on database 12c. With the other models, we see certain shortcomings. For example, server virtualization offers good isolation but creates compliance and administrative headaches. Schema based consolidation offers ease of management and patching, but limited isolation. Bottomline is: while Oracle provides best practice recommendations and blueprints for database as a service, enterprises need to choose among these models based on their individual needs. DBaaS：资源模型 数据库资源池 A group of homogeneous clustered or non-clustered resources exhibiting common characteristics. Example: Pool of Database Oracle Homes (for dedicated databases) Pool of 12c Container Databases (for PDBs) 区域Zone A logical grouping of cloud infrastructure resources based on QOS, functional, departmental or geographic boundaries. Example, Finance Zone, East Coast Zone Self-Service users provision into a Zone Zones can also be used to enforce access control and chargeback DB OH Pool DB Pool 12c CDB Pool DB Pool West Coast Zone East Coast Zone UK Zone RAC DB Pool RAC DB Pool Copyright © 2014 Oracle and/or its affiliates. All rights reserved. | After the administrator has decided on a consolidation model, he/she needs to pool software resources together to create a reusable deployment platform. A DBaaS resource pool is therefore, a homogeneous collection of software on which a database service can be provisioned. For dedicated database as a service it could be a bunch of Oracle Homes belonging to the same version. For schema as a service, it is a bunch of databases, again of the same version. The software pools are installed on a bunch of infrastructure comprising servers, storage and network. IT organizations may choose to segregate infrastructure based on Lines of Business or the lifecycle of the application-dev, test and production. Enterprise Manager exposes these resources to the end user as Zones. The end user of the database does not have to deal with the physicality of the database-hostnames, paths, etc 27 Oracle DBaaS解决方案 Oracle DBaaS 解决方案概览 应用 安全 事后审计 事前防范 事中拦截 Oracle数据库/多租户 Exadata Database Machine 业务连续性保障 集群及分区技术 云的测试管理 云生命周期管理 云的优化管理 数据库产品 管理 Copyright © 2014 Oracle and/or its affiliates. All rights reserved. | Oracle DBaaS 解决方案架构 Oracle DBaaS 所涉及的软件 •EM12c 管理包Packs of Oracle EM 12c for DBaaS –云管理包 –数据库生命周期管理包 •Oracle Database 数据库 •Oracle Database Options 数据库选件 –Depends on DBaaS configuration – RAC, Partition… –Database Security – AVDF, ASO, Database Vault… Copyright © 2014 Oracle and/or its affiliates. All rights reserved. | 29 Oracle Database 12c 用于云计算的 Oracle 多承租方 Copyright © 2014 Oracle and/or its affiliates. All rights reserved. | 传统数据库体系结构 将每个数据库的内存和进程分隔开来 CRM 数据库 文件 后台 进程 已使用的 内存 RAM DW 数据库 文件 后台进程 已使用的 内存 RAM ERP 数据库 文件 后台进程 已使用的 内存 RAM Copyright © 2014 Oracle and/or its affiliates. All rights reserved. | 31 全新的多承租方结构 仅在多承租方容器层面使用内存和进程 ERP 数据库 文件 已使用的 内存 后台进程 CRM 数据库 文件 已使用的 内存 后台进程 RAM DW 数据库 文件 已使用的 内存 后台进程 RAM RAM Copyright © 2014 Oracle and/or its affiliates. All rights reserved. | Add New in title 32 全新的多承租方结构 仅在多承租方容器层面使用内存和进程 RAM ERP CRM DW 后台进程 已使用的 内存 多承租方容器数据库 高效： 硬件减少为以往的 1/6，而可伸缩性提高 5 倍 降低运营成本： 将多个作为一个管理 透明： 无需更改应用程序 Copyright © 2014 Oracle and/or its affiliates. All rights reserved. | 33 OLTP 基准测试对比 50 个可插拔数据库仅需 3GB 内存，而 50 个独立数据库则需 20GB 可插拔数据库最多可扩展到 250 个，而独立数据库实例最多为 50 Multitenant 提高整合密度 相比于专用数据库，所需硬件资源是其 1/6，可扩展性是其 5 倍 Copyright © 2014 Oracle and/or its affiliates. All rights reserved. | 34 Multitenant 赋予新项目以敏捷性 快速供应、快照克隆 Copyright © 2014 Oracle and/or its affiliates. All rights reserved. | 35 多租户架构的优势 易于采用、降低成本、提高敏捷性 容器数据库 快速供应和迁移 提高数据库整合密度 将多个数据库作为一个管理 Copyright © 2014 Oracle and/or its affiliates. All rights reserved. | 36 用于云计算的 Oracle 集成式系统 Copyright © 2014 Oracle and/or its affiliates. All rights reserved. | 集成化系统 超强性能、集成化系统 性能无与伦比、部署简单、总成本较低 整合和云计算的构建块 Copyright © 2014 Oracle and/or its affiliates. All rights reserved. | “Engineered Systems” are a key part of Oracle’s overall strategy and a key part of Oracle’s cloud computing strategy. Exadata, Exalogic and the T4 SPARC SuperCluster are examples of Oracle’s Engineered systems.” They are made up of hardware and software designed to work together for extreme performance, security and manageability, and they are delivered to the customer in a pre-integrated and pre-configured manner. Oracle Exadata is a database machine, and Oracle Exalogic Elastic Cloud is a machine optimized for Java execution in the middle tier. Both are based on Intel x86 processors. SuperCluster is a new engineered system based on the T4 SPARC processor and Solaris, and can optionally run the Exadata storage software and Exalogic software. All 3 of these offer an order of magnitude improvement in performance and efficiency, making them ideal target platforms for mixed workload database and middle tier consolidation. They offer rapid deployment since they are pre-integrated and pre-configured by Oracle, not by you or your SI. And they are offer lower TCO because they can reduce the total hardware and complexity of your environment. While you can of course use Oracle Database and Oracle Fusion Middleware software running on other hardware, we consider Engineered Systems to be the ideal building blocks for private and public PaaS. They have the best performance and TCO in the market. 38 Exadata 提升整合性能 智能闪存缓存 使用闪存处理物理 I/O 智能闪存日志 使用闪存减少日志 I/O 延迟 智能扫描 在存储中运行部分数据库查询 混合列压缩 (HCC) 减少块和 I/O 数量 独一无二的技术可加快所有数据库负载的处理速度 Copyright © 2014 Oracle and/or its affiliates. All rights reserved. | 39 设置和管理云 Copyright © 2014 Oracle and/or its affiliates. All rights reserved. | Oracle Enterprise Manager 12c 完整云体系的完整云生命周期管理 1. 计划和设置云 容量和整合规划 资产发现 裸机供应 策略设置 2. 在云上构建、测试和部署 应用程序 将应用程序打包为组合件 测试应用程序 自助式供应 3. 管理和监视云 自动伸缩 完整体系管理 最终用户、业务级、应用程序监视 4. 计量、付费、优化 计量资源利用率 收费/计费 优化性能、容量、QoS Copyright © 2014 Oracle and/or its affiliates. All rights reserved. | 41 通过 EM12c 实现自助供应 自助供应专用数据库、可插拔数据库和模式 基于负载和配置的自动智能放置 开箱即用的DBaaS 高可用HA选件 使用配额、退役策略和反馈实现带治理的自助服务目录 灵活的克隆架构 利用备份实现完整数据克隆 Oracle ZFS 支持通过存储占用最低的“快照克隆”精简供应模式实现即时数据库供应 针对数据库生命周期管理的集成解决方案：监视、备份和修补 Copyright © 2014 Oracle and/or its affiliates. All rights reserved. | 42 自助服务：DBaaS 服务目录 Apps QA Project Mgr Apps IT 我需要一个数据库schema，去分析最新的市场活动 我需要一个数据库，用于新的人力资源应用 我需要生产数据库的复制，用于测试 我们需要最好的性能，去支持实时分析 我希望这个数据库是高可用的 仅仅用于测试，越便宜的系统越好 服务目录= ∑ ( 预先规划 and 预先配置的服务模板) 服务模板= 数据库配置 + 数据库定义 + 相关业务数据 Copyright © 2014 Oracle and/or its affiliates. All rights reserved. | 43 实现云管理计费/测量 付费管理员 自助门户用户 业务线用户 基于固定费率、使用量或配置的反馈 可并入其他量度，例如：业务事务 根据 LDAP 进行累计 针对不同用户类型定制的报告 Copyright © 2014 Oracle and/or its affiliates. All rights reserved. | The whole idea of chargeback is to get some visibility into determining how resources are being consumed. By getting this visibility you can then drive better decision making and planning. Since we have all the monitoring capabilities in the product we are able to track the resource usage at different levels (host, storage, network and so on). Based on that, we can create chargeback reports and since we can create charge plans we can do cost allocations based on all of this. As we showed earlier, this interfaces with the private cloud self service applications as well. 44 策略驱动的资源管理 在平台池上创建数据库池 定义位置策略 定义池的访问控制 基于CPU、存储、数据库数目等创建限额并分配给用户/角色 定义退出策略，例如一个月后将数据库设置为“不活动”状态 Copyright © 2014 Oracle and/or its affiliates. All rights reserved. | Another aspect of capacity planning is the consolidation planner. Nowadays new hardware is being made available more frequently than ever before, and this gives us a lot of opportunity to consolidate resources for more efficient use. A big part of this has been triggered by virtualization, but the release of products like Exadata and Exalogic mean there are opportunities for consolidation even in the physical world. There are clearly many benefits in undertaking this consolidation, from reduced server maintenance costs, to reduced administrations, energy and floor space costs to improved physical and network security, among other areas. Consolidation Planner analyzes all the historical utilization data we have collected over time, along with any constraints that may be defined (such as which workloads cannot run on the same machine as another workload or any business constraints), then based on all this information it produces a consolidation plan which might include consolidation of underutilized servers or migration to virtual machines. 45 容量规划 通过自动发现功能发现已有数据库 利用数据库整合规划程序 整合多个源数据库 基于技术指标和业务限制 单实例、RAC、非 CDB 和多租户数据库 通过整合重放测试应用 同时重放在不同数据库上捕获的负载 销售 HR ERP CRM 负载 整合 负载重放 Copyright © 2014 Oracle and/or its affiliates. All rights reserved. | The first thing we need to understand with capacity planning is the type of challenges administrators currently face. These include capacity prediction (administrators need visibility into when they will run out of capacity for key resources), efficient resource utilization and quality of service assessments. This whole area of capacity planning is an issue we have heard many times from customers and we are making concerted efforts to provide a solution. The benefits of such a solution include knowing when to make purchases so you can satisfy demand and at the same time ensure you are not underutilizing your systems, the ability to consolidate or repurpose any underutilized systems you have and improved quality of service. 46 管理与监控 管理云区域和底层资源 数据库池、服务器、存储 跟踪资源流动、租户、策略违规等 下钻到各个资源实现更深层次的监视 监视请求和故障率，识别要修复的潜在瓶颈 为 Exadata 提供深度软件和硬件管理 带仪表盘的合规性管理维护和显示最新的合规性评分 自动报告是否符合 Oracle 最佳实践和内部标准，涵盖 Exadata、RAC、单实例、可插拔数据库和底层基础架构 基于黄金标准的自动偏差控制 全面的补丁管理 建议、前期分析、执行和报告 通过同时将多个补丁应用于多个数据库目标，最大限度缩短停机时间 Copyright © 2014 Oracle and/or its affiliates. All rights reserved. | 47 产品方案 1/2配置Exadata X2-2 1/4配置Exadata X2-2 （两台）, 其中一台已扩容为1/2 配置 实施效果 整体提升8倍+业务执行效率 数据压缩比提升8~23倍 复杂统计性能提升18~352倍 案例：国内大型电信行业客户网管集中数据库云平台 基于Exadata实现云平台，整合OSS域17套Oracle数据库，提升8倍业务效率 客户挑战 OSS 17套系统分散运行在多个独立Oracle数据库主机上 数据资源分散，同步和共享难，统一监控和质量分析难，安全控制难 数据库利用率不均衡，计算和存储资源无法共享 各系统用独立数据库主机，维护和扩容成本极高 应用方案 OSS域17套Oracle数据库整合。混合负载，集中计算 第1期：OSS域17套系统迁移，含业务系统和分析系统 第2期：可查询的远程灾备 第3期：采用高容量盘，实现大量业务系统数据整合 第4期：扩容高容量盘Exadata ¼ Rack到1/2 Rack 客户概览 主营移动话音、数据、多媒体业务，IP电话以及互联网接入服务；用户数超过1000万 2011年3月上线 云平台-基于Exadata整合OSS域17套Oracle数据库，实现混合负载集中计算平台 Copyright © 2014 Oracle and/or its affiliates. All rights reserved. | August 16 BRONZE DB云 案例:瑞银集团Oracle DBaaS 特级保障级别：关键业务系统 系统举例：黄金交易系统 一级保障级别：重要业务系统 系统举例：托管中台系统 二级保障级别：普通业务系统 系统举例：ITIL系统 三级保障级别：开发测试系统 DIAMOND GOLD SILVER BRONZE SILVER DB云 GOLD DB云 DIAMOND DB云 CDB+PDB +RAC+BACKUP +DataGuard +GDS Copyright © 2014 Oracle and/or its affiliates. All rights reserved. | 欧洲最大的金融控股集团：瑞士联合银行集团（United Bank of Switzerland,英文简称UBS,中文简称瑞银集团） 1997年12月8日，总部位于巴塞尔的瑞士银行公司（Swiss Bank Corporation）与总部位于苏黎世的瑞士联合银行（Union Bank of Switzerland）宣布合并，成立瑞银集团。瑞士银行成立于1872年。因此，新成立的瑞银集团有着悠久的历史传承。 　　瑞银集团的业务遍及全球各大金融中心，在超过50个国家设有办事处，全球员工超过万名。 　　其中，39％的员工在美洲、36％在瑞士、16％在欧洲其他国家、9％在亚太区。瑞银为瑞士上市公司，于瑞士证券交易所、纽约证券交易所和东京证券交易所挂牌上市。瑞银集团的业务主要包括财富管理、投资银行及证券和资产管理三大块。 　瑞士是欧洲最富有的国家，也是全球最富有的国家之一。在瑞士有两大东西其它国家难以匹敌，一是钟表，二是银行。瑞士也是全球最大的离岸(offshore)金融中心，持有全球30%的离岸货币，总额接近2万亿欧元。瑞士还是全球名副其实的银行王国，银行产业是瑞士第一大支柱产业。2004年，瑞士金融产业创造了瑞士GDP的14%，银行产业对国民经济的贡献率同样属于金融强国的德国、法国和美国的两倍。瑞士的银行管理着全球1/4到1/3的国际投资私人财富，所以在私人银行服务和财富管理两大金融领域，瑞士银行占有极大的市场份额。 2011-2013， DBaaS Transformation Challenges/Opportunities: • 9600 instances today • 1,500 physical servers • 2,600 logical servers • 4,100 total servers • 5% YOY growth in compute • 3 PB data • 30% YOY growth in storage Solution: • 57 Exadata; 113 ODA’s; 51 T4’s • RAC, RAT, Lifecycle Mgmt, Cloud Mgmt • Exadata Storage Server Software • Migration Factory (OCS) Sales Cycle: • Start: July 2011 (Transformer) • Close: Feb 2013 • Oracle Executive contribution and commitment Competition: • IBM, HP, Cisco Business Benefits: • Reduced costs for automation, management and database provisioning. • Cut deployment times of new environments • Reduce the number of servers by 94% • Reduce Operating costs by 60% • Effective risk management & remediation. • Provide optimized infrastructure to support UBS Current Situation： Service Profile • 14 weeks to provision a database server, 2 weeks for just a database • Unique database processes and tools by division • Over 22 steps/ handoffs in typical provisioning workflow Workload Profile • More than two-thirds of database estate runs non-critical database workloads • Over 60% of estate non-production 49 铜 银 金 钻石 可用性 RTO < 15 min (尽力而为) < 15 min 最小的业务影响 DR RTO > 24 hr < 6 hr < 30 min RPO <= 24 hr <= 24 hr < 30 min 0 从备份恢复 < 72 hr < 24 hr < 8 hr < 30 min 性能 计划性的扩展 快速扩展 -------- -------- -------- -------- -------- -------- -------- -------- -------- 量体裁衣的服务定制 案例：瑞银集团实施Oracle DBaaS 业务服务 铜 银 金 钻石 数据库 HA 单实例MAA RAC + DR 数据复制 本地 无 无 Data Guard (同步) 远程 Data Guard (异步) 备份/恢复 每周全备份 每天备份日志 每周全备份 每天增量备份 银 + 闪回日志 金 + 日内的日志+ RMAN合并 技术服务 Copyright © 2014 Oracle and/or its affiliates. All rights reserved. | 50 案例：瑞银集团实施Oracle DBaaS效果分析 物理服务器减少94% - 从3300台到187台 数据库CPU 核数减少 71% - 从36388到10640 总体成本减少42%到$232M 未来五年TCO降低26% 物理服务器 Oracle数据库许可证 Copyright © 2014 Oracle and/or its affiliates. All rights reserved. | 30 UBS Achievements: • Standardize and simplify 736 different database stack variants • Radically reduce database provisioning time • Significantly reduce operational inefficiencies • Reduce physical servers by 94% (from 3,300 to 187) • Reduce DB Cores by 71% • Run-rate cost reduction by 42% to $232m • 26% TCO reduction expected over first 5 years 总结：Oracle数据库云（DBaaS）优势 应用上线 时间大幅缩短 降低风险 降低成本 1000% 提升部署新的数据库速度 (典型部署时间4+月) 50% 降低TCO (典型的服务器利用率= 5-10%) 100% 可用性SLA (典型的非计划停机时成本为每分钟 $5000 ) Copyright © 2014 Oracle and/or its affiliates. All rights reserved. | 30 Great Conversation Starters 1. CFO: Do you want to radically improve your margins? Oracle DBaaS solutions helped clients reduce IT OpEx costs by 50%. Can we accelerate your strategy in this area? (CBA reduced OpEx by 50-60%) 2. CIO/CFO: Oracle have helped customers to reduce IT CapEx by 80% through maximizing the consolidation density of their DB infrastructures. Would you like to achieve this within the next 2 years with Oracle? (Turkcell required 80% less power, Sabre required 90% less servers) 3. CIO/CTO: Established Oracle IT-as-a-Service customers are already provisioning DB’s extremely rapidly (in minutes rather than hours or days) and removing overall inefficiencies in service delivery. Can Oracle assist you to achieve this? (BT went from weeks to minutes, CBA from months to minutes, Allied Irish Bank reduced leads time from 10 days to 2 days and man effort from 5 days to ½ a day) 4. CIO/CTO: If you have virtualized your servers to take advantage of server consolidation density, are you aware of the significant OpEx in maintaining such environments ? (Allied Irish Bank reduced database patching time from 240 days/year to 80 days/year for their entire estate) 5. CTO: Oracle solutions can increase the agility of both receiving or delivering IT services. Can we discuss ways to improve time-to-production performance within your organization? Oracle数据库被广泛应用到各行各业，市场占有率第一 Oracle提供业界最优的集成平台和端到端云管理工具 Oracle具备经验丰富的、业务与技术兼备的咨询团队 Oracle提供的DBaaS技术与产品是开放的、标准的 Oracle DBaaS解决方案已验证为各行业客户带来效益 总结：Oracle DBaaS提供最好的端到端方案 Copyright © 2014 Oracle and/or its affiliates. All rights reserved. | 53 Q & A Copyright © 2014 Oracle and/or its affiliates. All rights reserved. | 54 Chart1 0 2010 2012delta Sheet1 2010 2012delta 2012 Dev/Test as a Service 7% 15% 22% Public 24% 0% 24% Private Compute as a Service 10% 12% 22% Public 17% 8% 25% Private Storage as a Service 9% 23% 32% Public 30% 7% 37% Private Identity as a Service 1% 9% 10% Public 8% 9% 17% Private Database as a Service 12% 25% 37% Public 35% 15% 50% Private Application Server as a Service 18% 12% 30% Public 36% 15% 51% Private To resize chart data range, drag lower right corner of range.