背景：我知道，谷歌的表现可以（而且经常是）很容易归因于高素质的员工谁爱做什么，他们的爱， 但严格从技术上来讲，下面的公式只是没有任何意义对我说：谷歌的 = 更多的数据（实际上是大量的数据）的 + 还快（??）。我认为这是合乎逻辑的和更有道理，我认为它是周围的其他方法：更多的数据=速度较慢。不过它看起来像 谷歌几乎令人重力和不按物理学的传统规律，当谈到自己的搜索服务的性能。 他们似乎遵循无理方程以上强调（其实它看起来像他们只是不停地越来越快，随着时间的推移：更多的数据进来，他们得到更快 并简单意义不大能玩弄更多的球再少球更好 - 我可以兼顾pretty的还有一个球，但 与玩弄两个球是一个很大的困难给我;但如果谷歌是我，它会更好地忙里忙外。国际海事组织这是一个有点古怪，至少可以说）。

Background: I know that Google's performance can (and often is) easily attributed to quality staff who love and do what they love, BUT strictly technically speaking, the following equation just makes no sense to me: Google = more data(actually huge amounts of data) + yet faster(??). I think it's logical and makes much more sense to me that it's the other way around: more data = slower .. however it looks like Google almost defies gravity and its not following the conventional laws of physics when it comes to the performance of their search service. They seem to follow the irrational equation highlighted above (actually it looks like they just keep getting faster and faster as time goes on: the more data comes in the faster they get and simply makes little sense to be able to juggle better with more balls then less balls - I can juggle pretty well with one ball but with juggling with two balls is a great difficulty to me; yet if Google was me it would juggle better. IMO this is a bit odd to say the least).

问：什么是具体实施范例（如特定的 编程算法实现的细节，设计模式和放大器;整体解决方案的体系结构），谷歌正在实施/雇用你和我是不是在我们的应用中这使谷歌为 使远程应用程序（即谷歌的搜索服务）经常执行的速度，然后我们的本地应用程序？ 就好像你和我1 + 1 = 2，而是谷歌1 + 1 = 3 - 这几乎是没有意义的。

Question: What are the specific implementation paradigms (such as specific programming algorithms implementation details, design patterns & overall solution architecture) that Google is implementing/employing that you and I aren't in our applications which enable Google to make a remote application (namely Google's search service) often perform faster then our local applications ? As if to you and me 1+1=2 but to Google 1+1=3 - it almost makes no sense at all.

是什么让他们（谷​​歌）分开？什么是背后的执行的酒吧没有业绩的秘密技术规划/设计的成分？请尽可能详细。

What sets them(Google) apart ? What are the secret technical programming/design ingredients behind the bar-none performance of their implementation ? Please be as specific as possible.

*免责声明* 常规，关于这个问题如的这个线程或的这个或其他流行和知名知道来源，如维基百科（新的答案/事实/结论，新的可能性hypothesis-我期待的软件实现细节（不宽：即使非常相关的），除非你提供论据来支持一些新的'东西'和见解，就像将不被视为一个有效的答复候补理论），这些小知识，你不能很容易找到，而在所关注的对象搜索在谷歌（或其他搜索引擎），或者根本无法发现或者是极不可能找到=不知道大多数开发;上可接受的答案的例子可以是：你个人的意见，写上该其提供通常是不广泛已知的和/或处于某种方式显着不同，从广泛地提供信息的信息。在广泛使用的信​​息只能被用作在你的答案的介绍和/或片尾曲，但并不像为核心的参数（preferably它不应该在核心参数的话）。请后鲜为人知和/或新技术就只有主体信息。谢谢你。

*Disclaimer* General, widely available information on the subject such as the info found on this thread or this one or other popular and well-known know sources such as Wikipedia (even though very relevant) will not be considered as a valid answer candidate unless you provide arguments to support some new 'things' and insights like: new answers/facts/conclusions, new possibilities hypothesis- I am looking for software implementation details(not broad theories) which are little-know and you can't easily find while searching on Google (or other search engines) on the subject in question or cannot be found at all or is highly unlikely to find=not know by most developers; examples of acceptable answers can be: your personal opinion, a write up which which provides information which is generally not widely-known and/or is noticeably different in someway from widely available information. The widely available information can only be used as in the intro and/or outro of your answer but not as the as the core argument (preferably it should not be in the core argument at all). Please post little-known and/or new technical info on the subject only. Thanks.

推荐答案

我觉得你提供的链接（和直接从这些页面引用的信息）提供了如何在后端处理负荷非常详细的视图。没有任何额外的'秘密武器'，超越已知成分，如缩放，高速缓存和无情的优化。

I think all of the links you provided (and the information referenced directly from those pages) provide a very detailed view of how the back end handles the load. There isn't any additional 'secret sauce', beyond the known ingredients such as scaling, caching, and ruthless optimization.

有一件事情你没有提到不过是战术快速加载网页。这些在史蒂夫Souders的谷歌技术讲座覆盖好。以下是YouTube上的这些会谈的链接。

One thing you didn't mention however was the tactics for fast loading pages. These are covered well in Steve Souders' Google Tech Talks. Below are the links for these talks on Youtube.

http://www.youtube.com/watch?v=BTHvs3V8DBA&功能=频道 http://www.youtube.com/watch?v=52gL93S3usU&功能=频道 http://www.youtube.com/watch?v=BTHvs3V8DBA&feature=channel http://www.youtube.com/watch?v=52gL93S3usU&feature=channel