超大数据中心运维管理方案.ppt

成功案例--国内（部分） ………. Thanks！ 如果我们更仔细地研究一下数据中心，就可以发现其运营成本占了 IT 预算的很大一部分，高达 70%（包括支持与维护），而用于更具战略性或创新性的预算所占比例则少得可怜。企业无疑需要降低针对运营和维护的资源总量，并释放出更多资源以用于更具战略意义和创新性的任务。 预算已陷入于高成本、供应过剩但利用率不高的怪圈 业务创新所占比例降至 30%： 需要快速响应业务及客户需求以提升客户满意度及客户忠诚度。客户期望他们能够提高服务级别并使 IT 对业务需求更具响应性。如今，我们看到业务价值有了很长的延期。 此外，我们发现业务敏捷性受到一定约束。为了解决公司拥有太多应用程序、太多定制内容和太高复杂程度等实际情况，必须进行改变。 最后，需要更好地控制风险并提高服务质量以取得业务成功。为了实现此目的，他们需要确保业务持续性并保护他们的资源和数据资产，以防范层出不穷的各种安全威胁。此举不仅会提升服务质量，还可以帮助公司满足企业要求和政府条例。 为了解决这些挑战，数据中心需要进行根本性的变革。他们需要打破孤岛、自动化运行那些极易出错的人工流程，同时加强管理控制。他们需要整合硬件、软件和服务以创建可以更有效运行的新一代数据中心。有一种更好的方法。 Let’s take a look at how Mercury LoadRunner Transaction Breakdown for J2EE really works. The Transaction Breakdown module is one of the holy grails of LoadTesting because it takes an end-to-end transaction response time and breaks it down into the various areas the transaction time was spent. For example we can tell you exactly how much time was spent in the network, in the database and the application server. Further, within the application server, we can drill down and split the response time across EJBs, JSPs, JNDI and JDBC and down to a method level. Most importantly, all correlated from a single business transaction. The reason this is so valuable is that previously a load tester could only tell a developer or architect that the application was slow. Now he can pinpoint the particular method that is causing the application slowness in the context of a particular business transaction. It eliminates guesswork and accelerates time to resolution. * J2EE Web Transactions Monitor helps to better visualize J2EE Web Transactions end to end, with performance metrics of all components starting from URLs to SQL queries. With this feature, application teams can quickly identify the Performance metrics of WEB components, EJB, Java and SQL statements in production. They can view trace that will chart the sequence of the internal invocations (methods) of the URL * 超大数据中心运维管理解决方案 ——卓豪（ZOHO）超大数据中心运维管理方案分享—— 主要内容 2 超大DCIM目前面临的挑战 32