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

joehuang joehuang at huawei.com
Wed Aug 31 06:48:01 UTC 2016


Hello, Joshua,

According to Peter's message, "However that still leaves us with the need to manage a stack of servers in thousands of telephone exchanges, central offices or even cell-sites, running multiple work loads in a distributed fault tolerant manner", the number of edge clouds may even at thousands level. 

These clouds may be disjoint, but some may need to provide inter-connection for the tenant's network, for example, to support database cluster distributed in several clouds, the inter-connection for data replication is needed.

There are different thoughts, proposals or projects to tackle the challenge, architecture level discussion is necessary to see if these design and proposals can fulfill the demands. If there are lots of proposals, it's good to compare the pros. and cons, and which scenarios the proposal work, which scenario the proposal can't work very well. 

So I suggest to have at least two successive dedicated design summit sessions to discuss about that f2f, all  thoughts, proposals or projects to tackle these kind of problem domain could be collected now,  the topics to be discussed could be as follows :

   0. Scenario
   1, Use cases
   2, Requirements  in detail
   3, Gaps in OpenStack
   4, Proposal to be discussed

  Architecture level proposal discussion
   1, Proposals
   2, Pros. and Cons. comparation 
   3, Challenges
   4, next step

Best Regards
Chaoyi Huang(joehuang)
________________________________________
From: Joshua Harlow [harlowja at fastmail.com]
Sent: 31 August 2016 13:13
To: OpenStack Development Mailing List (not for usage questions)
Subject: Re: [openstack-dev] [all][massively distributed][architecture]Coordination between actions/WGs

joehuang wrote:
> Cells is a good enhancement for Nova scalability, but there are some issues in deployment Cells for massively distributed edge clouds:
>
> 1) using RPC for inter-data center communication will bring the difficulty in inter-dc troubleshooting and maintenance, and some critical issue in operation. No CLI or restful API or other tools to manage a child cell directly. If the link between the API cell and child cells is broken, then the child cell in the remote edge cloud is unmanageable, no matter locally or remotely.
>
> 2). The challenge in security management for inter-site RPC communication. Please refer to the slides[1] for the challenge 3: Securing OpenStack over the Internet, Over 500 pin holes had to be opened in the firewall to allow this to work – Includes ports for VNC and SSH for CLIs. Using RPC in cells for edge cloud will face same security challenges.
>
> 3)only nova supports cells. But not only Nova needs to support edge clouds, Neutron, Cinder should be taken into account too. How about Neutron to support service function chaining in edge clouds? Using RPC? how to address challenges mentioned above? And Cinder?
>
> 4). Using RPC to do the production integration for hundreds of edge cloud is quite challenge idea, it's basic requirements that these edge clouds may be bought from multi-vendor, hardware/software or both.
>
> That means using cells in production for massively distributed edge clouds is quite bad idea. If Cells provide RESTful interface between API cell and child cell, it's much more acceptable, but it's still not enough, similar in Cinder, Neutron. Or just deploy lightweight OpenStack instance in each edge cloud, for example, one rack. The question is how to manage the large number of OpenStack instance and provision service.
>
> [1]https://www.openstack.org/assets/presentation-media/OpenStack-2016-Austin-D-NFV-vM.pdf
>
> Best Regards
> Chaoyi Huang(joehuang)
>

Very interesting questions,

I'm starting to think that the API you want isn't really nova, neutron,
or cinder at this point though. At some point it feels like the efforts
you are spending in things like service chaining (there is a south park
episode I almost linked here, but decided I probably shouldn't) would
almost be better served by a top-level API that knows how to communicate
with the more isolated silos (edge clouds I guess u are calling them).

It just starts to feel that the architecture you want and the one I see
being built are seemingly quite different and I haven't seen it shift to
something different so maybe it's time to switch the problem on the head
and accept that a solution may/will have to figure out how to unify a
bunch of disjoint clouds (as best you can)?

I know I can say that such a thing I'd like as well, because though
godaddy doesn't have hundreds of edge clouds, it is approaching more
than a handful of disjoint clouds (across the world) and a way to join
them behind something that can unify them (across just nova) as much as
it can would be welcome.

-Josh



__________________________________________________________________________
OpenStack Development Mailing List (not for usage questions)
Unsubscribe: OpenStack-dev-request at lists.openstack.org?subject:unsubscribe
http://lists.openstack.org/cgi-bin/mailman/listinfo/openstack-dev



More information about the OpenStack-dev mailing list