Data centers today. But before we jump into Adam, as usual, could you share a little bit more about yourself?

今天的数据中心。但在我们讨论亚当之前，像往常一样，你能多分享一些关于你自己的事情吗？

Absolutely. I have experience in data center consolidation and infrastructures based on my experience for an active during the air force, also migrations and Accenture capacity planning at Google, unless recently as a consumer of aws and your baby.

绝对地。我在数据中心整合和基础架构方面有经验，这是基于我在空军期间的工作经验，以及在谷歌的迁移和埃森哲容量规划，除非最近作为AWS和您的孩子的消费者。

Awesome. Let's jump right into it. First question I just want to ask you after that is can you describe the type of failure scenarios that you might see in the data center?

极好的。让我们直接开始吧。之后我想问您的第一个问题是，您能描述一下您在数据中心可能看到的故障场景类型吗？

Great question. There's a lot of different ways. We see failures all the time and data centers, and I think it seems a pretty far away concept, but these days are hosting data for all the applications that we use and love every day.

问得好。有很多不同的方法。我们总是看到数据中心出现故障，我认为这似乎是一个相当遥远的概念，但如今我们每天都在为我们使用和喜爱的所有应用程序托管数据。

The first I like to mention is just component and serve level failures. So just think about what actually server is. The machine with the CPU rammed disc and actually running whatever application we're trying to access. There are failure rates of those Rams and CPU and Google scale. When I was working YouTube in the hundreds of thousands of millions of servers. We see failures every day. In fact, there's a dedicated team part hardware operators provide around scooters and swapping out components on daily basis. Another somewhat common failure scenario was cite power outages. This can happen. It's not so much on a daily basis. But if there's a failure in power, grid, local, fire, or disaster going on, or even sometimes cases with diesel generators, people who are familiar with data centers also know that they have back of power in the form of batteries and five wheels.

我首先要提到的是组件和服务级别的故障。所以想想服务器到底是什么。这台机器的CPU充满了磁盘，实际上运行着我们试图访问的任何应用程序。这些RAM、CPU和Google Scale都有故障率。当我在成千上万的服务器上为YouTube工作时。我们每天都看到失败。事实上，有一个专门的团队部分硬件运营商提供周围的滑板车，并每天更换组件。另一种较为常见故障情形是电力中断。这可能发生。在日常生活中没有那么多。但是，如果发生电力、电网、本地、火灾或灾难故障，甚至有时发生柴油发电机故障，熟悉数据中心的人也知道，他们有电池和五个轮子作为后备电源。

So definitely a lot of different things that can go wrong are those type of a power sources are primarily used to move from when the power goes out and then trying to get the diesel generators on. I have to mention with psychology is being very real on. Think about the season hurricane season in the us. Specifically the south is every year searching exercise of moving our jobs and loads from one site to another sort of trying to be ready to speak for this.

所以肯定有很多不同的事情可能会出错，这些类型的电源主要用于在停电时移动，然后试图启动柴油发电机。我不得不提到心理学是非常真实的。想想美国的飓风季节。具体来说，南方每年都在寻找将我们的工作和负载从一个地点转移到另一个地点的练习，试图准备好为此发言。

And then there's the least popular, but probably number one reason why we see a downtime of failures which is changes in software or configurations. I know, unfortunately, this can happen if source apple, someone puts out something there was a proper key way or testing down on it. Or if you have the situation where you're the third party cloud provider employer, there can make a policy change, revoking all of the access on accident. So these type of things definitely。

然后是最不受欢迎的，但可能是我们看到故障停机的首要原因，这是软件或配置的变化。我知道，不幸的是，这可能会发生，如果来源苹果，有人拿出一些东西，有一个适当的关键方法或测试下来。或者，如果您是第三方云提供商的雇主，可以更改政策，在发生事故时撤销所有访问权限。所以这类事情肯定。

sorry, thanks for familiar scenarios. I can't want to change gears in the interview this little bit, ask you follow up question, a little bit more like an estimation question, and sometimes interviews what sort of push you around in different ways.

抱歉，感谢熟悉的场景。我不想在面试中改变这一点，问你后续的问题，更像是一个评估问题，有时面试会以不同的方式推动你。

But the question I have you today is, could you estimate something like how much power or maybe how many winter bands that might take to power the global data center footprint for the company like Google?

但我今天要问你的问题是，你能不能估计一下，像谷歌这样的公司，需要多少电力或者多少冬季乐队才能为全球数据中心提供电力？

So just understand correctly and give you an idea of how I'm gonna approach this first, try and estimate how much power the data centers consume, and then how much they generate. You're sort of looking for a number of so instantaneously today how much power we use. That's right.

因此，请正确理解并让您了解我将如何处理这一问题。首先，尝试估计数据中心消耗多少电力，然后估计它们产生多少电力。你在某种程度上是在寻找一个数字，所以今天我们使用了多少电力。对的。

Okay, great. So I would say just in the number of data centers alone, there's 20, and it's actually publicly available Information for Google 12 or in the us five in Europe, in Asia. And one is in South America, on the Top of my head. I remember that data centers, consumer 15 megawatts or more of power. But I want to spend a little bit of time and validate that. Is that okay with you sounds good. A server consumes about 200 watts of power. And when you have multiple servers in Iraq is just like a cabinet in the data center about 50 servers, I can go into this rack.

好的，很好。所以我想说，仅在数据中心的数量上，就有20个，这实际上是谷歌的公开信息，在美国有12个，在欧洲和亚洲有5个。一个在南美洲，在我的头顶上。我记得数据中心消耗15兆瓦或更多的电力。但我想花一点时间来验证一下。你同意吗？听起来不错。一台服务器的耗电量约为200瓦。当你在伊拉克有多台服务器时，就像数据中心的一个机柜，大约50台服务器，我可以进入这个机架。

So if you multiply those two numbers out and be about 10,000 watts or 10 kilowatts per rack, now I think many racks are in a data center can depend on the side of the data center, but we use around 2,000 racks, just based on my experience and seen this number.

因此，如果将这两个数字相乘，则每个机架的功率约为10，000瓦或10千瓦，现在我认为数据中心中的许多机架可以依赖于数据中心的侧面，但根据我的经验和看到的这个数字，我们使用大约2，000个机架。

If I multiply this throughout, that ends up in about 20,000 kilowatts, or I when translate over the units 20 megawatts, so how much a single data center with take? But it is Google. And so we have to think that things are a little bit bigger. I'd like to have a scale factor here about 2 . 5 and just because there's, multiple views and facilities on each of these given campuses ever trying estimate. So use 2 . 5 multiplier on Top of my 20 megawatts. I will tell me each site is not 50 megawatts. I go back, I told you that there were 20 different sites, each consuming 50 megawatts, I want being but a thousand megawatts, translate over. That's one giga, one of power on. Yeah, so now that I have the power consumption of the data centers and move into. The second part, you mentioned the winter vines in our generation, like also Top. And I actually don't know how much a power winter wine can create. But I remember from my graduate research that solar powers, that one by one meter can generate 1 kilowatt power.

如果我把这个乘起来，结果是大约20，000千瓦，或者我在20兆瓦的单位上转换，那么一个数据中心需要多少钱？但它是谷歌。所以我们必须把事情想得更大一点。我想在这里有一个大约2的比例因子。5，只是因为在每个校园里都有多个视图和设施，所以一直在尝试估计。所以用2。在我的20兆瓦的基础上增加了5倍。我会告诉我每个站点不是50兆瓦。我回去，我告诉你有20个不同的站点，每个站点消耗50兆瓦，我只想要1000兆瓦，翻译过来。这是一个千兆，一个电源打开。是的，所以现在我有了数据中心的功耗，并进入。第二部分，你提到了我们这一代的冬季藤蔓，像也是顶级的。实际上，我不知道冬季葡萄酒能产生多大的能量。但我记得在我的研究生研究中，太阳能发电，一米乘一米可以产生1千瓦的电力。

So I'm just gonna assume at this point that one winter bind can do about 1 kilowatt, because I estimated one people want and being 1 million wind turbines to power Google status and difference.

所以我只是假设在这一点上，一个冬天的束缚可以做大约1千瓦，因为我估计一个人想要100万个风力涡轮机来驱动谷歌的地位和差异。
说话人1
Got it a lot of winter rents? Do we feel like that was an overestimate or underestimate? Or what would you change if you had more time to kind of analyze this answer?

有很多冬天的租金吗？我们觉得这是高估还是低估了？或者如果你有更多的时间来分析这个答案，你会改变什么？
说话人2
I think that's that's a big number, but also using it and operating Google scale, I think this is pretty appropriate, I think, out of the framework I laid out, and we can sort of go and change some variables.

我认为这是一个很大的数字，但使用它和操作谷歌规模，我认为这是非常合适的，我认为，在我制定的框架之外，我们可以去改变一些变量。

If a winter binding, use, open, generate 1 kilowatt, or if there's a different power calculation for servers or rocks, we can go back and adjust these things. But I feel pretty confident and good with my number.

如果冬季绑定，使用，打开，产生1千瓦，或者如果服务器或岩石有不同的功率计算，我们可以回去调整这些东西。但我对我的号码很有信心。
说话人1
Cool, that's awesome. We can kind of like a deeper for pause, the interview of just something here, anything else you want to share with our audience or any other high level thoughts on how the interview was for you.

酷，太棒了。我们可以更深入地暂停一下，采访这里的一些东西，你想与我们的观众分享的任何其他东西，或者任何其他关于采访对你来说如何的高层次想法。
说话人2
Not having worked in this world before I probably wouldn't have been able to answer it.

以前没有在这个世界上工作过，我可能无法回答这个问题。

So II didn't learn how to do this overnight. I definitely took practice and working with these different domain areas. I really take away for folks, is this question involved a lot of power and consumption also just understand the space in size of the data center but that might not be your area of expertise.

所以我不是一夜之间就学会的。我确实在这些不同的领域进行了实践和工作。我真的要告诉大家，这个问题涉及大量的电力和消耗，也只是了解数据中心的空间大小，但这可能不是你的专业领域。

And that's okay. When you receive these type of questions, it's important to focus on the process and sort of be able to regenerate and recreate the scenario that you did to create your number again. I II I wanna harp on really introducing those different factors and being able to defend them. And using this process, you can probably go forward. Tackle any question because the type of interview questions you get, they'll be looking for you and how assessing how you tackle and beauty.

这没关系。当你收到这些类型的问题时，重要的是要关注这个过程，并能够重新生成和重新创建你再次创建数字的场景。我想反复强调这些不同的因素，并能够为它们辩护。使用这个过程，你可能可以继续前进。解决任何问题，因为你得到的面试问题的类型，他们会寻找你，以及如何评估你如何解决和美丽。
说话人1
I thought you did a great job answering, and I'm very thoughtful and you went through it in very good detail, obviously, and more time being deeper into some of those things around how you might change your answer or tradeoffs or scenarios where you might change it. I'm always really valuable, but overall, I thought you didn't excellent job. And it's always a pleasure having you here. Adam, thank you again for being the show to provide your answer. Hi.

我认为你在回答问题上做得很好，我考虑得很周到，很明显，你非常详细地讨论了这个问题，并花了更多的时间来深入了解你可能会如何改变你的答案或权衡或你可能会改变它的场景。我总是很有价值，但总的来说，我认为你做得不是很好。很高兴你能来这里。亚当，再次感谢你在节目中提供你的答案