PaaS 和应用服务器托管有什么区别?
就在互联网上部署 Java Web 应用程序而言,提供应用程序服务器托管(例如 Tomcat、Jboss 和 Glassfish)的网站与提供平台即服务(例如 Google 应用程序引擎、Heroku 和 Amazon EC2)的网站有什么区别?
In terms of deploying a java web app on the internet, what is the difference between sites that offer app server hosting such as Tomcat, Jboss and Glassfish, and sites that offer platform as a service such as Google app engine, Heroku and Amazon EC2?
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(1)
这是一个很难回答的问题,因为 PaaS 一词经常被误用。
但我会尝试:PaaS 通常涉及(另一个)抽象层(您通常并不真正知道您的应用程序到底在哪台服务器上运行)。另外,您的应用程序可能会也可能不会与一个或多个其他租户共享硬件。
传统上,应用程序服务器托管提供商仅在每台服务器上使用 N(例如 1000)个客户并以此方式进行计算。 GAE 等尝试对此进行优化(更巧妙地分配负载),并且通常根据您实际使用的内容收费。
恕我直言:理论上,您可以在应用程序服务器上使用比实际支付的更多的资源。但由于他们通常过度销售,你会发现很难得到你所支付的价格。理论上,通过按使用量付费的 PaaS 产品,您所付出的代价就应该得到。 /恕我直言
Difficult question, since PaaS is often misused term.
But I'll try: PaaS usually involves (another) layer of abstraction (you usually don't really know on which server exactly your app is running). Plus your application may or may not share the hardware with one or many other tenants.
Traditionally providers of app server hosting just used N (like 1000) customers per server and calculated that way. GAE etc. try to optimize (distribute the load more cleverly) this and usually charge you on what you actually use.
IMHO: In theory you could use more resources on a app server than you actually pay. But since they usually over-sell, you'll find it hard to get what you pay for. In theory, with pay-what-you-use PaaS offerings you get should what you pay for. /IMHO