毁掉云计算的3个好办法区域

毁掉云计算的3个“好办法”

中国IDC圈11月8道：目前云计算市场仍处于跳跃性发展阶段，许多公司在这方面也已颇有建树。但是随着新技术的不断涌现，有些项目只是昙花一现。这些项目失败的原因渐渐明晰，并且极易避免。

下面有三个让云计算必死无疑的方法。你应该从中吸取教训，避免重蹈覆辙。

首先，人事安排不当。发展，转移，项目实施失败这就是云计算项目运营的怪圈。云计算被炒得很火，所以那些IT公司里最有影响力的人便很快涉足其中。但是，他们和首席信息官关系密切并不意味着他们同样有能力运营好云计算项目。

同样，他们在选择技术类型和供应商方面也错误频出。许多机构在管理方面生搬硬套。当你在迎合潮流和最优结构之间徘徊时，你就是在自找麻烦。最近我正在处理一堆这样的问题。

第二，安全问题被抛之脑后。那些项目负责人往往在项目部署之后才考虑到安全和法规问题。但是云计算项目的安全措施，一旦部署便很难改造。所以在使用技术（譬如加密技术）时需要系统地考虑运营环境。忽略安全问题一个低级错误。

第三，用云计算来解决不相关的业务问题。

对于企业来说，新应用程序的开发或者旧的程序的转型都是不错的选择。但这并非重中之重。

下列两种运营方式必然会导致项目失败。第一种

，将一次运行中断便会毁掉业务的程序植入云中，然后向那些互联中的神灵祈祷千万不要出任何差错。这太冒险了。另外一种便是将没人感兴趣的程序植入云中，希望能够以此引起用户兴趣。这种方法太不显眼了。我们应该寻找一种更为折中的办法。

英文原文：

As best practices continue to emerge， so are the things that will kill your cloud project guarantee

Many companies are having great success with cloud computing， and clear that the market continues to grow by leaps and bounds. However， with any new technology plays， there are those projects that do quick face-plants. These paths to failure are also emerging and highly avoidable.

Here are three surefire ways to fail with cloud computing， and what you can learn from them to avoid suffering that same fate.

First， put the wrong people on the project. This is the most common way that cloud computing development， migration， and implementation projects fail. Cloud computing is a hyped "cool" space. So those who have the most political clout in an IT organization quickly position themselves on cloud computing projects. However， just because they are buddy-buddy with the CIO does not mean they have the architectural and technical skills to make the cloud work for the enterprise.

Bad decisions are also made in terms of deciding how to select technology types and technology providers. It's a manage-by-magazine world at many organizations. When you select what's popular versus what's a true architectural fit， you shoot yourself in the foot. I'm fixing a ton of these mistakes these days.

Second， security is an afterthought. This means that those driving the project do not consider security and compliance requirements until after deployment. It's almost impossible to retrofit security into a cloud computing deployment， so the approach and use of technology （such as encryption） should be systemic to the environment. This is a rookie mistake.

Third， select the wrong business problem to solve with cloud computing. The right approach is to pick new application development or existing application migration that is meaningful to the business， but that is not mission-critical.

There are two paths to failure here. The first is to pick the "kill the business with a single outage" type of application， put it in the cloud， and then pray to the Internet gods that nothing goes wrong. Too risky. The second is to pick a meaningless application that nobody cares about， move it to the cloud， and hope that somebody notices. Too underwhelming. Find something that falls in the middle.