Open Stack

On 04/23/2013 10:41 AM, Mark McLoughlin wrote:
> On Tue, 2013-04-23 at 10:11 -0400, Jay Pipes wrote:
>> This project (Falcon) is only about 4 months old. While it looks to be
>> good solid code, it would be great if OpenStack projects could settle on
>> a WSGI framework and everyone work towards the same end goal. By my
>> count, to date OpenStack projects have used Bottle, Flask, Pecan, WSME,
>> and custom-built frameworks.
>>
>> Why don't we get this into Oslo and at least make an attempt at driving
>> the car in the same direction? If the performance benefits of Falcon are
>> as advertised, I would think the OpenStack dev community would be eager
>> to create a simple Oslo wrapper around the Falcon library and start to
>> standardize on a single framework for WSGI handling.
>
> I wholeheartedly agree with the premise - "this is a frickin mess and it
> would be awesome to see it get fixed"
>
> We've probably discussed it at every design summit to date e.g.
>
>    https://etherpad.openstack.org/havana-common-wsgi
>    https://etherpad.openstack.org/grizzly-common-wsgi-frameworks
>
> However, the solution isn't for someone to add a Falcon based API to
> Oslo and leave it rot there. The hard part is moving projects to a
> common framework.
>
> Doug's proposal is to use Pecan/WSME for new APIs based on his
> experience using it in Ceilometer - it would be a step forward if we
> could have consensus on at least this much.

+1000. I spent some time looking over Doug's shoulder on the Pecan/WSME 
front. I imagine porting to this in Nova would drop 1000s or 10000s of 
lines of code and complexity, and provide us much better validation on 
the front side.

> However, the existing WSGI code in our projects isn't going away any
> time soon and continues to evolve. I think there'd be huge value in
> someone incrementally consolidating those WSGI implementations, bit by
> bit.
>
> The issue here isn't a lack of WSGI frameworks to choose from, or a lack
> of understanding that this is a huge issue, or a lack of frustrated
> people who think it should be easy to fix ... :-)
>
> The issue here is a lack of anyone willing to grab the bull by the horns
> and start taking baby steps to improve the situation.

Agreed. Hopefully in the Nova v3 time frame we can pull some of this in 
there, help doing that with something that's been proved out in another 
OpenStack project (like Pecan/WSME) would be far more appreciated than 
another web services stack.

	-Sean

-- 
Sean Dague
http://dague.net