[openstack-dev] [all][massively distributed][architecture] Coordination between actions/WGs

lebre.adrien at free.fr lebre.adrien at free.fr
Tue Aug 23 10:17:24 UTC 2016

Hi Folks, 

During the last summit, we suggested to create a new working group that deals with the massively distributed use case:
How can OpenStack be "slightly" revised to operate Fog/Edge Computing infrastructures, i.e. infrastructures composed of several sites. 
The first meeting we did in Austin showed us that additional materials were mandatory to better understand the scope as well as the actions we can perform in this working group. 

After exchanging with different persons and institutions, we have identified several actions that we would like to achieve and that make the creation of such a working group relevant from our point of view. 

Among the list of possible actions, we would like to identify major scalability issues and clarify intra-site vs inter-site exchanges between the different services of OpenStack in a multi-site context (i.e. with the vanilla OpenStack code). 
Such information will enable us to better understand how and where each service should be deployed and whether it should be revised.  

We have started an action with the Performance WG  with the ultimate goal to analyse how OpenStack behaves from the performance aspect as well as the interactions between the various services in such a context. 

Meanwhile, we saw during this summer the Clynt's proposal about the Architecture WG.

Although we are very exciting about this WG (we are convinced it will be valuable for the whole community), we are wondering
whether the  actions we envision in the Massively distributed WG will not overlap the ones (scalability, multi-site operations ...) that could be performed in the Archictecture WG.

The goal of this email is to : 

    (i) understand whether the fog/edge computing use case is in the scope of the Architecture WG. 

    (ii) if not, whether it makes sense to create a working group that focus on scalability and multi-site challenges (Folks from Orange Labs and British Telecom for instance already told us that they are interesting by such a use-case).

    (iii) what is the best way to coordinate our efforts with the actions performed in other WGs such as the Performance and Architecture ones (e.g., actions performed/decisions taken in the Architecture WG can have impacts on the massively distributed WG and thus  drive the way we should perform actions to progress to the Fog/Edge Computing target)

According to the feedback, we will create dedicated wiki pages for the massively distributed WG. 
Remarks/comments welcome. 

Further information regarding the Fog/Edge Computing use-case we target is available at http://beyondtheclouds.github.io

More information about the OpenStack-dev mailing list