1) Just to add on Overhead stuff as mentioned above, I've honestly
never seen a webapp that connects to a Database ( Mysql, Postgresql,
Couch/Mongo/etc ) or exists in a SOA talking to other daemons/apis,
where the webapp was the bottleneck.  In my experience, it's always
been the database or internal APIs holding stuff up, or the frontend
server not being able to handle concurrent connections well ( and not
getting them to the backend fast enough ).  I'm pretty much not
worried by the overhead of things like SqlAlchemy -- there are better
places to optimize... like Pyramid itself, which is incredibly
unlikely to be a bottleneck itself.

> I have been primarily a C/C++ system level programmer and architected and 
> developed large Telecom systems (carrier grade) requiring great performance 
> levels. Currently looking into Python and corresponding web frameworks for 
> our platform.

My 02¢ -- If i were doing anything Telecom related, unless you're
looking for a website, i'd be focused on Erlang.


> How is the performance of Paster Web server as compared to something like 
> lighttpd (which is supposedly a light weight server)? If lighttpd, then I 
> would need to embed Pylons/Pyramid there. (I have embedded Pylons with Apache 
> but haven’t checked the performance).

Few people use lighttpd, as it has a long history of random bugs and
never really being all that great.  Nginx is the current leader in
light http servers.

You wouldn't run python within the http server, but use nginx to
either proxy to a server or whatever handles wsgi.


>  Regarding DB, SQLAlchemy can be a choice for SQL based DB servers. 
> Considering something like MongoDB, not sure how would SQLAlchemy help there. 
> But certainly if I have my own DB access library using native APIs, probably 
> a better performance can be achieved.

The point of SqlAlchemy, Pyramid, Python, etc, is that you have a
tradeoff -- you give up a marginally faster runtime ( vs a native API,
a custom framework, programming in C ) for significantly faster times
in your release schedule, management cycle, and a much lower cost-to-
change and iterate.

So yes, you can go crazy trying to optimize the theoretical max
performance of an application - or you can choose something like
Pyramid and SqlAlchemy and optimize the actual performance of your
development team.  I prefer the latter.

-- 
You received this message because you are subscribed to the Google Groups 
"pylons-discuss" group.
To post to this group, send email to [email protected].
To unsubscribe from this group, send email to 
[email protected].
For more options, visit this group at 
http://groups.google.com/group/pylons-discuss?hl=en.

Reply via email to