下一代数据中心解决方案范例.ppt

分布式智能处理 分布式处理，解决计算能力问题：每个服务器包括存储、CPU及相应的软件, 处理SQL 及压缩，恒定的计算能力与存储的比值。 横向并发计算，解决带宽和可靠性问题：分布式处理单元横向部署，ASM提供镜像保护，处理能力与数据量同比扩展。 廉价标准化，解决价格及更新问题：完全标准的PC服务器硬件，智能存储单元数量不受限，极大降低成本，并随着PC技术的发展而发展。 Exadata Cell InfiniBand Switch/Network 集群数据库或单实例数据库 Exadata Cell Exadata Cell 分布式处理层 集中处理层 超高速、并发网络层 Exadata Cell Exadata Cell Exadata Cell Exadata Cell 云中存储的高可用性通过自动存储技术(ASM)避免磁盘或存储节点的单点故障 自动存储管理保证数据的冗余和条带化 保证数据的分布是均匀的 保证存储的动态添加和删除 保证热点数据获得更好的性能 Exadata Cell Exadata Cell Hot Hot Hot Hot Hot Hot Cold Cold Cold Cold Cold Cold ASM Disk Group ASM Failure Group ASM Failure Group Insert Picture Here 案例介绍 Oracle奥斯丁数据中心 奥斯丁数据中心: 超过 20,000 服务器 全世界最大的 Dell/Linux 数据中心 每星期增加100服务器 超过 5,000 TB 数据 全世界最大的 NetApp 数据中心 每月增加 60TBs 数据 支持超过 400 其他客户的关键应用 为超过65000甲骨文员工服务 超过7000平方米 Oracle的应用架构整合 1998 Today Business as Usual Global Single Instance (GSI) 52 application instances 1 application instance of Oracle E-Business Suite 40 data centers 3 data centers Inconsistent business processes Global standard processes Fragmented data Consistent data and global business intelligence Decentralized decision making Centralized decision making 结果: 持续发展和收益增长 Note: All figures after FY2004 are provided on a non-GAAP basis.?GAAP to non-GAAP reconciliations are included in our earnings releases, which are available on the Oracle Investor Relations website at /investor. TotalRevenue($Billions) De-centralized IT Consolidated IT Corporate Expansion $10.0 $30.0 1998 LTM OperatingMargin(%) 20.0% 30.0% 40.0% 50.0% $20.0 Operating Margin (%) 1999 2000 2001 2002 2003 2004 2005 2006 2007 Total Revenue ($Billions) * * * BRU_ * Having said that though, metrics on resource utilisation levels within data centers globally is tragically low. This slide references a study done by the Uptime Institute (on behalf of McKinseys) to determine how well resources are being utilised and the results are, frankly, eye-opening and a source of major concern. Some key metrics of the study - around 29% of servers within the polled DC’