multithreading - Why does Google App Engine support a single thread of execution only? -


Do anyone have any idea why Google App Engine only allows a single formula of execution for deployed applications gives?

I personally believe that he has something with an application's approximation so that Google can shape its performance more accurately. No argument has been written about Google Threaded execution on Google's site, so my question is.

An application which is already multi-threaded and currently posted on VM, means that it is difficult for me

edit : I have marked the answer given below because it looks quite well that thread is not allowed due to horizontal scaling requirements. Naturally all are executed within the same process space and, GAE can run many processes for your application, so the threads will be difficult to share. He said, I still think a small thread pool per process will be useful and will help to migrate the application in the cloud. I will request it as a feature thanks for the discussion!

There is a limited option for creating threads in Google App Engine by the name of the work rows:

Edit

From:

Allow app engines to distribute the request for applications on many web servers, and An application can be prevented from interfering with another, this application runs in a restricted "sandbox" environment. In this environment, the app engine can execute code, store and query data in the datastore, app engine mail, URL fetch and user services, and can examine the user's web request and generate feedback. .

Like other people, it has been said that threads are not supported for securities due to sandbox applications.

There are several other restrictions of Google App Engine that force developers to create scalable apps. I believe that the work queue is just one of these restrictions, as opposed to opposing the creation of a thread on a machine that handles the current HTTP request, is placed in a queue which can then be scheduled and the other Can be executed by machines. Work queue allows the machines to share and distribute in a scalable manner.


Comments