AWS、Heroku 与其他可扩展平台？
考虑到您是一家初创公司，没有资金用于自己的服务器场。哪种现有解决方案可以让您高枕无忧，流量的突然增加不会导致一切崩溃。
我知道这不仅仅取决于硬件，因此我们计划至少有一个负载均衡器、内存缓存和一些数据库服务器。
是否可以在 AWS 上设置一个在流量增加时自动添加实例和带宽的设置？
您还可以向部署菜鸟提供哪些其他建议？谢谢。
ps：如果问题太广泛或反映出对上述主题缺乏经验，我提前道歉，但这就是我问的原因。
Considering you're a startup with no funds for own server farm. Which existing solution can give you a peace of mind that any sudden increase in traffic won't bring everything down.
I know it's not just up to hardware, so we plan to have at least a load balancer, memcache and few db servers.
Is it possible to have a setup on AWS that would automatically add instances and bandwidth if the traffic increases?
What other advice you could give to deployment noobs? Thanks.
ps: I apologize in advance if a question is too broad or reflects inexperience on mentioned topics, but that's why I ask.
如果你对这篇内容有疑问，欢迎到本站社区发帖提问 参与讨论，获取更多帮助，或者扫码二维码加入 Web 技术交流群。
绑定邮箱获取回复消息
由于您还没有绑定你的真实邮箱，如果其他用户或者作者回复了您的评论，将不能在第一时间通知您！
发布评论
评论（8）
赫罗库。因为你是一家初创公司，所以要保持精简，而且它不会比几乎免费（使用 1 个 dyno + 小型共享数据库）更精简。花时间构建你的产品，而不是基础设施。当您应该与客户交谈时，您不想安装补丁。 Heroku 也很灵活，允许您随着流量的增加而扩展“dynos”，因此不必担心流量的增长。不过，Heroku 不会自动为您扩展，因此您自己的服务器监控也是如此。 Heroku 附加组件也不错。
Heroku. Because you're a start-up, keep things lean and it doesn't get leaner than almost free (with 1 dyno + small shared DB). Spend time building your product, not on the infrastructure. You don't want to be installing patches when you should be talking to customers. Heroku is also flexible and allows you to scale up 'dynos' as your traffic increases so no worries about growing there. Heroku won't scale automatically for you, though, so do your own server monitoring. Heroku add-ons are also nice.
最近我们对 AWS 和 Heroku 进行了很好的比较，我们决定迁移到 Heroku，这里是详细信息 http://www.confiz.com/blog/tech-session/selecting-the-right-cloud-platform/
Recently we have done a very good comparison between AWS and Heroku and we decided to move to Heroku, here is the detail of this http://www.confiz.com/blog/tech-session/selecting-the-right-cloud-platform/
如果您使用的是 Python，可以尝试 Google App Engine。
将 Python 应用从一个平台迁移到另一个平台并不是太困难 一旦你克服了哪些功能可用（不可用）的学习曲线。 GAE 提供数据存储、memcache、blobstore 以及其他一些好东西，如 dJango 和 Jinja（模板）。值得查看 Python 起始页，我并没有花很长时间也将其集成到 Facebook 和 Dropbox 中。
If you're on Python, you can try Google App Engine.
Migrating the Python app from one platform to another isn't too difficult once you get the past the learning curve as to what features is (not) available. GAE offers datastore, memcache, blobstore plus a few other goodies like dJango and Jinja (templating). Worth checking the Python start page and it didn't take me long to integrate it into Facebook and Dropbox too.
远离赫罗库。您可以从 Amazon 免费获得一年的 EC2。扩展 Heroku 的成本极其高昂。他们的定价往往不明确，而且客户服务总体很糟糕。
Stay away from Heroku. You can get EC2 for free for a year from Amazon. Scaling up heroku is extremely costly. Their pricing tends to be unclear and their customer service in general sucks.
适用于 Amazon EC2 的 BitNami 包括可立即运行的 Apache、MySQL、PostgreSQL、SQLite、Python、Django、Git 及其所需依赖项等等。它可以通过一体化的免费本机安装程序、虚拟机和云映像进行部署。也许值得一看。
BitNami for Amazon EC2 includes ready-to-run versions of Apache, MySQL, PostgreSQL, SQLite, Python, Django, Git, their required dependencies, and much more. It can be deployed via all-in-one free native installers, virtual machines and Cloud Images. maybe worth checking out.
我个人的经验是，你通常应该从 Heroku 开始。让您的应用程序投入使用，并找到适合的产品/市场或某种类型的吸引力。你会知道你要去某个地方，因为客户会引起扩展问题。在这种情况下，Heroku 将允许您以很少的开销进行扩展。在一段时间内，这种扩容不会让你的钱包受到影响。
准备好后跳转到 AWS。你什么时候准备好？当你在钱包里有足够的痛苦时，你需要对堆栈有更多的控制权。您可以聘请 AWS 开发人员或自行了解。
Heroku 和 AWS 都有自动扩展解决方案，但 Heroku 的学习曲线相当平坦（这就是您所付出的代价），而 AWS 可以很快变得广泛且陡峭。 Udemy AWS 课程或其他数百个在线资源中的任何一个将帮助您开始构建强大的 AWS 架构。
最后，虽然性能不应该是您最关心的问题，但请确保您在代码中使用最佳实践。您的第一个用户不应该让您的系统变得缓慢。如果她这样做，AWS 也不会提供帮助。
希望这能在某种程度上有所帮助。
这是我的经历。出于这个原因，我的 saas 启动套件是为了开箱即用地部署到 Heroku 而构建的。然而，启动套件也是容器化的。我知道您明确提到了 AWS，但对于容器，您可以与基础设施无关。这是值得考虑的！
特德 [at] https://stacksimple.io
My personal experience is that you should generally start with Heroku. Get your app out in the wild and find that product/market-fit or some type of traction. You will know you are going somewhere because customers will cause scaling issues. In this case, Heroku will allow you to scale with very little overhead. And for some time, this scaling will not hit you in the wallet.
Jump to AWS when you are ready. When will you be ready? When you have enough pain, in the wallet, where you need more control over the stack. You can hire a AWS devops type or learn about it, yourself.
Both Heroku and AWS have auto-scaling solutions, but whereas Heroku has a fairly flat learning curve -- that is what you are paying for -- AWS can get broad and steep fairly quickly. A Udemy AWS course or any of the hundred other online resources will get your started down building a robust AWS architecture.
Lastly, while performance should not be your primary concern, make sure that you are using best practices in your code. Your first user should not bring your system to a crawl. And AWS will not help if she does.
Hope this helps in some way.
This has been my experience. My saas start kits are built to deploy to Heroku out of the box for this reason. However, the start kits are also containerized. I know that you spoke of AWS explicitly, but with containers you can be infrastructure agnostic. This is worth considering!
Ted [at] https://stacksimple.io
请查看我开始撰写的这个博客系列，因为我发现与 EC2 和 Digital Ocean 相比，从财务角度来看 Heroku 根本无法扩展。将展示如何使用 Docker 将 Ruby 应用程序放在 Digital Ocean 上，这使您能够具有相同的灵活性和快速扩展和缩减的能力 https://medium.com/@karimbutt/weaning-off-heroku-part-1-b7f123ae855f
Check out this blog series I'm starting because I found Heroku to not be scalable at all from a financial perspective compared to EC2 and Digital Ocean. Going to be showing how to put a Ruby application on Digital Ocean using Docker, which allows you the same flexibility and ability to scale up and down very quickly https://medium.com/@karimbutt/weaning-off-heroku-part-1-b7f123ae855f
这在很大程度上取决于您正在寻找 PaaS、IaaS 还是 SaaS，以及您使用的语言是什么。
其他平台根据您的需求发挥作用，您可能还想看看这个比较：https://dictativ.com/compare/paas
It greatly depeneds whether you're looking for a PaaS, IaaS or SaaS, and what is the language you using.
Other platforms come into play depending on your needs, you might want to take a look at this comparison as well: https://dictativ.com/compare/paas