Mixing python with a faster language for optimization in GAE

Question

I'm a newbie in the Python and GAE world and I have a question. With Python the normal approach is to only optimize the code when needed, fixing the more urgent bottlenecks. And one of the ways to achieve that is by rewriting the most critical parts of the program in C.

By using GAE are we losing this possibility forever? Since Google's Go language is now (or it will be as soon as it is compiled more efficiently) the fastest language in GAE, will there be a way to mix Python and Go in the same app? What other ways could be used to achieve a similar result?

Answer 1

See Can I write parts of the Google App Engine code in Java, other parts in Python? for how to use multiple languages.

Basically, each version of a given app can only use one runtime language.

But, you can have two different versions of your app, written in different languages, and they can pass information back and forth through the datastore.

Also, you can have two different apps, in two different languages, and you can have then pass information back and forth through requests.

Answer 2

I think you're falling for premature optimisation here. For nearly all webapps, the majority of time spent is in RPCs, waiting for the rest of the system to do something such as process datastore queries. Of the remainder, a significant fraction is often spent in C code anyway. There are relatively few webapps that need to do a lot of processor-intensive work in order to serve a typical query.

If your app is one of those, you may want to reconsider writing your entire app in Python, given the unavailability of C extensions on App Engine, and choose Java or Go. If your app is one of the 99% that don't need to do much processor intensive work for typical requests, don't worry about it.