Presentation is loading. Please wait.

Presentation is loading. Please wait.

Microsoft的商业智能概览 ZhaoHui Tang 唐朝晖

Similar presentations


Presentation on theme: "Microsoft的商业智能概览 ZhaoHui Tang 唐朝晖"— Presentation transcript:

1 Microsoft的商业智能概览 ZhaoHui Tang 唐朝晖 zhaotang@microsoft
Microsoft的商业智能概览 ZhaoHui Tang 唐朝晖 Program Manager SQL Server BI Microsoft公司

2 纲要 当前我们面临哪些商业挑战? SQL Server 2000的商业智能(BI) Microsoft的BI技术概述及相关战略 案例研究
演示 问与答 I’m going to walk you through the following top-line issues that many of our customers worldwide – and representing all of the major vertical industries – face as they determine how to evaluate, implement, and deploy business intelligence solutions. [Read bulleted list] Are there any other topics that, if we have time at the end of the presentation, you would like me to address today?

3

4 商业智能市场的发展趋势 BI Advantage 日益激烈的市场竞争和日渐衰退的经济环境 赋能技术的广泛采用 Internet的影响
Web 服务 XML / SOAP / HTTP 移动设备 ASP / Hosting 带宽 全球化竞争 增长缓慢的经济形势 不断提高的生产力/关注于利润 高速增长 / 商业灵活性 BI Advantage Internet的影响 企业应用程序的标准化 庞大的数据量(CRM、ERP、SCM) 实时信息 / 通过Internet访问信息 同商务伙伴和用户间的交互 新的数据来源 KPI以及针对 CRM、ERP、SCM和点击流应用的架构 开放和标准的API 由需求推动 赋能元素

5 分析麻痹症: 听着很熟悉,是吗? 我有太多的数据,但是没有足够的信息 为什么我所有的报告不能都订在一起? 管理层需要了解昨天的业务进展情况
分析麻痹症: 听着很熟悉,是吗? 我有太多的数据,但是没有足够的信息 为什么我所有的报告不能都订在一起? 管理层需要了解昨天的业务进展情况 谁有时间进行分析?

6 分析麻痹症: 你并不孤独 可用商业数据的数量呈指数级增长 在过去5年中,每周所需制定的重要决策的数量增加了3倍
分析麻痹症: 你并不孤独 可用商业数据的数量呈指数级增长 在过去5年中,每周所需制定的重要决策的数量增加了3倍 在过去5年中,专业分析人员的数量减少了50% 资料来源: Gartner, Inc.

7 认清对BI的需要 运营报告不匹配 对用户和产品获利情况缺少企业高度的整体印象 不堪报告请求重负的IT部门 业务单位雇用单独的IT人员
缺少用户细分和有针对性的市场推广 Sales figures don’t match Inventory figures are in disagreement Detail reports don’t add up to summary reports Who are our customers? What products do they have? How much inventory are we using?

8 需要BI的公司 在今后12个月中,用户计划购买: 对分析应用的购买计划: 商业智能/DW解决方案 57% 基于Web的产品和服务 40%
硬件/服务器 % 电子商务产品和服务 % 对分析应用的购买计划: 客户关系管理 % 知识管理 % 分析应用程序开发 % 财务分析 % 性能管理/平衡记分卡 % If that sounds too futuristic, consider where customers are spending their IT dollars. Purchase plans include both the platform and other building block pieces of data warehousing and BI, but it also includes analytic applications for all of those disparate data systems that I showed on the previous slide…so that the customer channels like CRM can really provide actionable information, not just store data about the customer. DM Review 读者调查

9 BI的用途 BI具有两方面的主要用途: 它可以对组织的财务和运营健康状况进行监视。 它可以规范 组织的运营 来自用户社区的观点
报告、警告、报警、分析工具、关键性能指标(KPI)和仪表板 它可以规范 组织的运营 同运营系统、信息反馈系统的双向集成 来自用户社区的观点

10 BI的使用状况(节选) 工具产品 目前 工具产品 现在 用户类型 用户需要 5%-10% 70%-80% 15%-20% 强大的分析工具
- 成熟的报告工具 - 先进的计算能力 - 评级、预测 MS合作伙伴的产品 --ProClarity --Cognos --Business Objects --更多… MS合作伙伴的产品 --ProClarity --Cognos --Business Objects --更多… 5%-10% 70%-80% 15%-20% 分析人士 Excel 2002 Data Analyzer --数据可视化 --简单分析 --发布到Excel、 PowerPoint和Web 信息探索者 简单的分析工具 --数据可视化 --简单过滤/排序 --轻松发布 Excel 2002 --数据透视表 --数据透视图 Now, if you look inside an enterprise we see three main worker types: The Analyst: needs robust analytical tools The Information Explorer: simple analysis tools and easy publishing Information Consumer: simple access to views Now, you could probably debate the % of information explorer – it’s actually probably less than 70 – 80%, but the reality is that there is probably about that many people in an enterprise that WOULD LIKE to use more data analysis to their benefit. Right now, MS has great partners that are delivering high-end analysis tools to the analyst community. And in the middle-tier – folks that do not need super-robust analysis – we have Excel. The feedback from the Excel community (again, the mass of folks in the middle that do not need high-end tools) is that they want more than what Excel can provide. And that’s our strategy with Data Analyzer: release a new member of the Office family that brings a new level of visualization and analysis to the knowledge worker. Introduce a product that essentially picks up where Excel leaves off: with Excel today we have pivot tables, etc. but ultimately it’s simply a 2 dimensional grid of numbers that is often difficult to see trends. With Data Analyzer, we help with that challenge by providing a great product that (1) quickly enables users to visualize data, (2) enhances their productivity, and (3) allows for quick and easy publishing. 信息的使用者 查看报告 共享 .xls 文件 简单“另存为Web页” 集成化的分析 和发布体验

11 SQL Server 2000 商业智能

12 什么是OLAP? 在线分析处理(Online Analytical Processing) - 由EF Codd在1994年首次提出
词义与一些早期术语基本相同,例如:决策支持、商业智能、执行信息系统 OLAP = 多维数据库 MOLAP:多维OLAP(Arbor Essbase、Oracle Express) ROLAP:关系型OLAP HOLAP:混合OLAP

13 多维的OLAP “嘿 … 我卖掉了价值1亿美元的货物” 维度:产品、地区、时间 地区 分层汇总路径 产品 地区 时间 行业 国家 年
产品 地区 时间 行业 国家 年 类别 地区 季度 产品 城市 月 周 部门 天 地区 西 产品 牙膏 果汁 可乐 牛奶 奶油 肥皂 月份 1 2 3 4 7 6 5

14 OLAP服务器的架构 XML/A Object Model PTS OLAP ADO MD Server OLAP Manager
应用程序 Object Model RDBMS - 数据源 & ROLAP Store OLE DB Rep. PTS OLAP Server 应用程序 ADO MD MOLAP Store 应用程序 Rep.

15 XML for Analysis 查询OLAP数据源的行业标准 充分利用OLE DB for OLAP/DM的优点 利用XML、SOAP
发现、执行 Microsoft XML/A SDK

16 ADOMD.Net

17 BI Accelerator 使用Excel的BI开发工具 面向零售和市场推广的数据模型示例 数据处理和数据移动的基础结构

18 SQL Server 2000的功能特性 数据挖掘预测
使用决策树理解和预测个体的属性 所有学生 决定高中学生进入大学的因素是… 进入大学: 55% 是 45% 否 进入大学: 79% 是 21% 否 IQ=高 45% 是 55% 否 IQ=低 IQ ? 进入大学: 94% 是 6% 否 家庭富裕 = True 79% 是 21% 否 家庭富裕 = False 家庭富裕? 进入大学: 70% 是 30% 否 31% 是 69% 否 父母鼓励 = No 父母鼓励 = Yes 父母鼓励?

19 通过数据挖掘进行用户细分 创建“最不同的”的用户群 每个用户群采用 不同的市场战略 促进有针对性的市场推广活动
理解不同用户的需求差异,对产品或服务进行细微调整,例如移动电话呼叫计划等。 每个用户群采用 不同的市场战略

20 OLE DB for DM API 通过以下手段使DM成为一种 大众化的市场技术: 利用现有的技术和知识 普通的行业概念和数据呈现方法
SQL和OLE DB 普通的行业概念和数据呈现方法 从“专有”到“开放”不断变化的DM市场认知

21 DM 查询语言 Create Mining Model CreditRisk (CustomerID long key,
Gender text discrete, Income long continuous, Profession text discrete, Risk text discrete predict) Using Microsoft_Decision_Trees Insert into CreditRisk (CustomerId, Gender, Income, Profession, Risk) Select CustomerID, Gender, Income, Profession,Risk From Customers Select NewCustomers.CustomerID, CreditRisk.Risk, PredictProbability(CreditRisk) From CreditRisk Prediction Join NewCustomers On CreditRisk.Gender=NewCustomer.Gender And CreditRisk.Income=NewCustomer.Income And CreditRisk.Profession=NewCustomer.Profession

22 Microsoft的BI技术概述 及相关战略

23 普遍深入的商业智能 Microsoft 的商业智能战略 全面的BI平台 最佳的互操作性 快速的投放 市场时间
集成化的RDBMS、ETL、OLAP、数据挖掘和Office工具 最佳的互操作性 Data Warehousing Alliance 和 Framework、开放的数据访问API、数据源连接 快速的投放 市场时间 易于构建和管理 最低的TCO 普遍深入的商业智能

24

25 Microsoft BI – “专为大众构建”

26 全面的BI产品 SQL Server 2000: Office XP 完整的平台 – 可伸缩的RDBMS,包括 数据转换服务
分析服务(OLAP和数据挖掘) 元数据服务 英语查询工具 Office XP Excel – 内置数据透视表服务 为定制开发提供的Web组件 数字仪表板 Microsoft Data Analyzer

27 Microsoft Office XP Excel数据透视表中的 OLE DB for OLAP
为基于浏览器的分析任务提供的Office Web组件

28 数字仪表板 个性化的BI数据汇集视图

29 Data Analyzer的重要功能 快速查看数据 改进了数据分析效率 轻松共享和发布信息 与SQL Server 2000全面集成
灵活的部署和可扩展性 Intuitive Visual Interface - Quickly Analyze Business Data New methods for data retrieval in Data Analyzer make it easier to connect to your critical information. Perform detailed data analyses using graphical analysis tools across multiple dimensions and get quick results with the Data Analyzer View. Find business data anomalies with ColorScale, a dynamic and customizable tool that automatically highlights Powerful Analysis Capabilities - Improve Data Analysis Productivity Options for displaying multiple measures, such as gross profit, unit sales or quantity, or displaying relationships between unlimited business dimensions, such as customer, region, or product, in a single, easy-to-use interface. - The Business Center offers a guided analysis tool that enables you to easily identify trends and answer common business questions. Use Filter by Criteria to quickly select and filter data to locate the most pertinent information. Built-in Template Measures enable you to easily find key performance indicators. Change from Last Year, Average Children's Length, Number of Children, Change in Year to Date, Percent of total length of members in filter, Change from Previous Period Easily Publish and Share Data Export to Excel, Export to PowerPoint, or Publish to HTML to quickly publish data and share it with others in your work environment. Flexible Deployment With MSDA, users can easily access data analysis tools over a network, over the Web, or in remote/offline scenarios. In addition to client-server access, MSDA supports remote connection to OLAP cubes, as well as analysis of local cubes. Developers can customize Data Analyzer by using the documented application programming interface (API) that comes with Data Analyzer. They can also automate Data Analyzer features with Microsoft Visual Basic®, Microsoft Visual Basic for Applications (VBA), Microsoft Visual Basic Scripting Edition, and Microsoft Visual C++®. Examples… -Embed the XML control on a Web page or digital dashboard. The ActiveX control interacts with the OLAP cube through an XML file. -Send periodic views as static HTML files; e.g., in messages to top managers or shareholders. -Embed MSDA views in any Office document. -Anywhere you have a solution that does data mining or reporting against an OLAP data source, you can add the MSDA controls to add a highly intuitive, visual element. Capitalize on SQL Server 2000 Analysis Services Take advantage of vastly expanded navigational and analytical capabilities with support for Specialized Dimension Types. Data Analyzer includes full support for all new dimension types available in SQL Server Analysis Services, including parent-child dimensional hierarchies. Easily access data analysis tools via the Web with Cube Access over the Web, a powerful feature that enables remote connection to online analytical processing (OLAP) cubes anywhere on the Internet, with full security. By utilizing Actions and launching a Web page, you can further analyze and share information based on a specific data element.

30 Microsoft BI 产品的成就 同其它竞争产品相比,Microsoft 产品最受用户欢迎,80%的用户选择了Microsoft的产品
Microsoft OLAP服务器的购买和安装数量位居市场第一 Microsoft拥有最大的席位部署数量。比与其最接近的竞争对手高出 40% Microsoft 产品的用户忠诚度最高 资料来源:2001 – The OLAP Survey

31 商业智能 MS 在OLAP领域排名第一 -OLAPsurvey.com MS的市场份额增长最快(86% CAGR, IDC 2001)
MS的DW/BI收入 = 5亿美元(根据IDC的数据) DW/BI代表了40%的SQL Server收入 $12b 31.4% 32.5% 22.0% 29.9% “在BI平台幻方图中最值得注意的变化就是 Microsoft不断提高的执行能力,以及市场对Analysis Server及其相关工具不断提高的接纳程度。”' - Howard Dresner, Gartner Group, 7/01 合计 = 26.9% CAGR (%)

32 业界领先的基准测试结果 指标 世界记录 获胜者 SQL Server Teradata IBM Onyx 32,000 个并发用户
PeopleSoft CRM 30,000 个并发用户 PeopleSoft eBill Pmt 191,694笔支付/小时 Pivotal eRelationship 20,000 个并发用户 SAP R/3 Retail 316.5万个销售数据行项目/小时 SAP R/3 Sales & Dist. 24,000 个并发用户 TPC-C 709,220 tpmC TPC-W 10,439 WIPS TPC-H 18,830 3TB Teradata J.D. Edwards OneWorld 6,000 个并发用户 IBM PeopleSoft Financials 15,000 个并发用户

33 关系型数据库的市场份额 Microsoft分别位于第一和第二
Windows Oracle 37% IBM 19% Informix 1% Sybase 2% 其它 3% Microsoft 38% 总计24亿美元 增长40% Unix+Windows Oracle 53% IBM 16% Informix 4% Sybase Other 6% Microsoft 17% 总计54亿美元 增长24% 2000 Revenues 数据来源: GartnerDataquest

34 几个案例研究

35 T3 – 企业级的商业智能 世界最大的多维数据集 - MOLAP Cube 中1.2 TB的源数据 DW
77亿行 RDW进入到 MOLAP中(440GB) 50 个并发用户 中值查询时间 = sec 反映了真实的公司架构、业务流程和数据 只有SQL Server能够支持这种规模;易管理性 1.2 TB source data – not including indexes 7.7 billion input rows 449 GB cube size (=37% of source) vs data explosion in competing products (10x to 20x) Performance Processing times = all the data in 53 hours; 145 million rows / hour; 40,000 rows per second Warm cache Median 0.02 seconds Mean 0.08 Percentile # Calculated member sample query Anecdotal 126 concurrent 1500 queries (95% < 1 second) cold cache 100 million row = 1 hour Quad = <$100k

36 AT&T Growth Markets 商业问题 解决方案 – 关注于关键指标 结果 6个月内获得了300%的ROI
改进对主要市场动态的理解 制定更有效的市场渗透战略 提高公司的业务增长速度 解决方案 – 关注于关键指标 建立在SQL Server、数字仪表板和地图信息插件程序之上的ProClarity 500位AT&T业务分析员、销售人员和产品经理 11,000位现场销售代表使用一个Web浏览器解决方案 结果 6个月内获得了300%的ROI AT&T Growth Markets is a major business unit of AT&T charged with big task: winning more business in a highly competitive, multibillion-dollar marketplace. To give its sales force faster access to better data, AT&T is rolling out its “Attack” database. It’s a powerful, Web-accessible solution that uses Microsoft SQL Server 7.0 to gather internal and external data that can be used by sales personnel to make better business decisions. Let’s work through the business problem, strategy, solution, and results achieved. $8 billion operator of world’s largest digital wireless network Business Problem – need to improve their understanding of key market dynamics, plan more effective market penetration strategies and increase their business growth rate Increase customer focus and enable growth Reduce Costs without Impacting Quality Support People for Success Strategy - Focus on Key Metrics Customer Churn, Product Affinity / Customer Demographics Uptime of network & services Response time to requests & problems Solution - Implemented ProClarity, SQL Server, Digital Dashboard and Map Info plug-in Users 500 AT&T business analysts, sales professionals and product managers 11,000 field sales representatives via a Web-browser solution Results 300% ROI in 6 months

37 案例研究:Verizon 位于美国的8个主要的数据中心 20,000多名呼叫中心员工 先前GTE Territories中的8TB数据
COFEE 部署在4个数据中心里 20,000多名呼叫中心员工 每天有7500名CSR使用 COFEE 先前GTE Territories中的8TB数据 单个数据库实例拥有2.4TB数据 1亿多份在线帐单

38 Disco 解决方案 Microsoft SQL Server 2000 超过1TB的原始数据 “因为我们的数据仓库建立在Microsoft SQL Server 7.0的基础之上,对于我们来说,使用Microsoft SQL Server 2000 Enterprise Server中所附带的数据挖掘技术是一种最自然和最适合我们需要的方法。” -- Horacio Díaz 数据仓库项目负责人 Disco S.A. 关键的成功因素 伸缩性 集成化的数据挖掘功能 效果 改善了市场营销和市场推广活动的效率 技术特性 SQL Server 2000 Windows® 2000 Advanced Server

39 资源 如需获得有关Microsoft商业智能解决方案的更多信息 请访问以下Web站点
Microsoft SQL Server Microsoft Data Warehousing Alliance 关于Microsoft Consulting Services,请访问:

40 如果您有任何问题,请加入 微软中文新闻组 继续讨论
加入微软中文新闻组

41


Download ppt "Microsoft的商业智能概览 ZhaoHui Tang 唐朝晖"

Similar presentations


Ads by Google