One keystone is not able to work as backend for another keystone currently, as 
far as I remembered.

So how about convert Multisite to one certification project: to certificate 
whether the installer support multi-site, and apply some tag or attribute on 
specific installer: it's "multi-site" support ( "mult-region" support, 
"keystone federation" support, "Tricircle" support) etc?

Or it should be part of Dovetail?

Currently most of installers only support single-site installation.

And I also hope that volunteer can step up and re-elect PTL.

Best Regards
Chaoyi Huang (joehuang)
________________________________
From: Julien [julien...@gmail.com]
Sent: 25 June 2017 20:44
To: joehuang; Fatih Degirmenci
Cc: opnfv-tech-discuss
Subject: Re: [opnfv-tech-discuss] multi-site next step ?

Maybe we'd better add a new scenario for multisite in OPNFV. In this scenario 
installer will support requirements as you listed.
Currently all installers install everything from the very beginning including 
keystone. If external keystone can be used as an internal keystone's backend, 
it is easier for installers. I am not sure about this configuration.


/Julien

joehuang <joehu...@huawei.com<mailto:joehu...@huawei.com>>于2017年6月24日周六 
下午4:56写道:
Hello, Faith,

Thank you for the information.

Except the Tricircle installation with OSA scripts, installer to support 
multi-site installation is also very important, either using shared KeyStone 
for multi-region deployment or KeyStone federation for multi-cloud deployment:

1. Each OpenStack instance can be installed with specified region name
2. the KeyStone endpoint used inside all services can be configured(KeyStone 
may be not in the same region), or federation configure could be configured.
3. Multiple OpenStack instances deployments.
4. compute nodes IP are routable if tunneling is expected to be established 
from compute node to compute node directly, or deploy L2 gateway in each 
OpenStack instance if L2GW will be used instead, L2GW can be ping-able with 
each other. Tricircle also support SDN controller as backend in local Neutron, 
and offload the networking automation to SDN controller if SDN controller 
itself support cross site tenant networking.

Best Regards
Chaoyi Huang (joehuang)
________________________________
From: Fatih Degirmenci 
[fatih.degirme...@ericsson.com<mailto:fatih.degirme...@ericsson.com>]
Sent: 23 June 2017 18:40
To: joehuang
Cc: valentin.bouc...@orange.com<mailto:valentin.bouc...@orange.com>; 
opnfv-tech-discuss; BLAISONNEAU David IMT/OLN
Subject: Re: [opnfv-tech-discuss] multi-site next step ?

Hi Joe,

As Valentin summarized, XCI uses OSA directly from upstream and deploys 
OpenStack from master so there is no dependency to OPNFV installers. All the 
features/scenarios are/will be implemented in upstream, getting part of the 
upstream.

XCI provides sandbox for the developers to use to  integrate their features to 
upstream OSA. It is available for quite some time and used for adding support 
for ODL, Tacker, OVS NSH and enabling SFC. Sandbox lets you deploy OpenStack 
using VMs with number of different flavors. Baremetal also works and we will 
enable those jobs in OPNFV CI soon.

The provisioning is done by bifrost (standalone ironic) which is an upstream 
project as well.

We are also in the process of setting up 3rd Party CI for OSA which will 
provide feedback on the patches you and/or others might be proposing to 
upstream from OPNFV directly on OpenStack Gerrit.

You can take a look at the documentation from below links and ask any questions 
you might have via mail or on #opnfv-pharos irc channel.

https://wiki.opnfv.org/display/INF/OpenStack
https://wiki.opnfv.org/display/INF/XCI+Developer+Sandbox

/Fatih

On 23 Jun 2017, at 09:28, joehuang 
<joehu...@huawei.com<mailto:joehu...@huawei.com>> wrote:

Hi, Valentin,

Good idea, would like to know how many installers support OSA(OpenStack 
Ansible), and contributors are needed :)

Best Regards
Chaoyi Huang (joehuang)
________________________________
From: valentin.bouc...@orange.com<mailto:valentin.bouc...@orange.com> 
[valentin.bouc...@orange.com<mailto:valentin.bouc...@orange.com>]
Sent: 22 June 2017 20:06
To: joehuang; opnfv-tech-discuss
Cc: BLAISONNEAU David IMT/OLN; Fatih Degirmenci
Subject: RE:[opnfv-tech-discuss] multi-site next step ?

Hello Chaoyi,

It's great summary of all work in OPNFV and also in OpenStack Tricircle !

For my part, I use Tricircle for the OPNFV demo, and I think the current 
fonctionnality is fine ! Of course, I found some bugs and now it's fixed.

I think for the next step, It's more the integration and deployment of an 
OpenStack with Tricircle. Yes, it can be supported with OPNFV installer but I 
think the best way is to integrate that in OpenStack upstream installer like 
openstack-ansible. Tricircle is an OpenStack project so the integration in OSA 
make sense. In addition, some people in the community works on the integration 
of OSA in OPNFV (Fatih and Yolanda and several Suse guys on XCI and David 
Blaisonneau with OSA on baremetal). It would be cool to have a 
os-nosdn-multisite-ha scenario based on OSA!

We can also deploy tricircle in real context, like Windriver guys aksing in the 
end of our session. 2 or more site with some latency between each site.

Best regards,
Valentin
________________________________
De : 
opnfv-tech-discuss-boun...@lists.opnfv.org<mailto:opnfv-tech-discuss-boun...@lists.opnfv.org>
 
[opnfv-tech-discuss-boun...@lists.opnfv.org<mailto:opnfv-tech-discuss-boun...@lists.opnfv.org>]
 de la part de joehuang [joehu...@huawei.com<mailto:joehu...@huawei.com>]
Envoyé : jeudi 22 juin 2017 08:59
À : opnfv-tech-discuss
Objet : [opnfv-tech-discuss] multi-site next step ?

Hello,

After several cycles running of the multi-site project, in last week OPNFV 
summit, we demonstrate mission critical application video-conference and vIMS 
running in multi-region (multi-site) environment, and leverage the L2/L3 
networking capability provided by Tricircle, the two applications can achieve 
high availability ( active-active, or multi-master ) across multi-region.

The presentation video is here: https://www.youtube.com/watch?v=tbcc7-eZnkY, 
the slides is here: 
https://docs.google.com/presentation/d/1WBdra-ZaiB-K8_m3Pv76o_jhylEqJXTTxzEZ-cu8u2A/
 And a video was recorded for video conference: 
https://www.youtube.com/watch?v=nK1nWnH45gI, and vIMS video is in recording.

If you compare what happened in Amazon AWS for mission critical applications: 
slides 
https://www.slideshare.net/AmazonWebServices/aws-reinvent-2016-moving-mission-critical-apps-from-one-region-to-multiregion-activeactive-arc309,
 video https://www.youtube.com/watch?v=tBM5MB22vik, you will found that the use 
case tried to deploy the PS store and  database back-end Casandra into two 
regions, with Casandra running in active-active mode, communicating with each 
other through EIP(same as FIP in OpenStack) in different region. In our demo, 
vIMS is also using Casandra, and also deployed with active-active in two 
OpenStack region, but Tricricle provide more flexible internal tenant level 
isolated L2/L3 networking across region, and support more versatile and 
flexible networking topology. To deal with mission critical application running 
inside cloud, we are not alone, Tricircle even provides some better capability 
to some extent.

We can have summary for what have done in the multi-site project, there are so 
many people contributing in the past for these 5 use cases which were discussed 
during the initiation of multi-site project:
Use case 1: multisite identity service management(Shared KeyStone in demo)
Use case 2: VNF high availability across VIM (vIMS & JITSI in demo, cross 
OpenStack L2/L3 networking provided by Tricircle)
Use case 3: Multisite VNF Geo site disaster recovery (not covered, covered by 
Karbor etc project)
Use case 4&5:

  *

Centralized Quota Management (Covered by Kingbird)

  *

Centralized management for distributed resources (Partly covered by Kingbird )

  *

A service to clone ssh keys, imgaes etc across regions (Partly covered by 
Kingbird )

  *

Global management for tenant level IP address / mac address space (Covered by 
Tricircle )

  *   Centralized monitoring service (not covered)

Most of them have been realized in Tricircle, Kingbird etc. Though Tricircle is 
developed in OpenStack, and has been accepted as OpenStack offical project 
since Nov.2016, and Kingbird was moved outside of OPNFV from E release, it's 
fine that we have solution for these use cases. Most of development activities 
are outside OPNFV now.

During the OPNFV summit, several guys expressed the interest of Tricircle 
integration, or contribution to Tricircle, it's great!

So what's the next step of multi-site project? Do we need a standalone project 
to do integration, or just treat the integration as one ticket of specific 
installer/sdn controller, different installer/sdn controller with different 
ticket, and Tricircle team just provide guide and support for the integration? 
Or follow OpenStack style, you can even do it yourself, and contribute in 
Tricircle for bugs and features if needed. Are there contributors interested in 
Functest/CI integration?

Tricircle can also work with CellsV2, which will be released in Pike, to scale 
single OpenStack instance. Tricircle has provided some documentation on how to 
run and integrate it https://docs.openstack.org/developer/tricircle/

Thoughts?

Best Regards
Chaoyi Huang (joehuang)

_________________________________________________________________________________________________________________________

Ce message et ses pieces jointes peuvent contenir des informations 
confidentielles ou privilegiees et ne doivent donc
pas etre diffuses, exploites ou copies sans autorisation. Si vous avez recu ce 
message par erreur, veuillez le signaler
a l'expediteur et le detruire ainsi que les pieces jointes. Les messages 
electroniques etant susceptibles d'alteration,
Orange decline toute responsabilite si ce message a ete altere, deforme ou 
falsifie. Merci.

This message and its attachments may contain confidential or privileged 
information that may be protected by law;
they should not be distributed, used or copied without authorisation.
If you have received this email in error, please notify the sender and delete 
this message and its attachments.
As emails may be altered, Orange is not liable for messages that have been 
modified, changed or falsified.
Thank you.


_______________________________________________
opnfv-tech-discuss mailing list
opnfv-tech-discuss@lists.opnfv.org<mailto:opnfv-tech-discuss@lists.opnfv.org>
https://lists.opnfv.org/mailman/listinfo/opnfv-tech-discuss
_______________________________________________
opnfv-tech-discuss mailing list
opnfv-tech-discuss@lists.opnfv.org
https://lists.opnfv.org/mailman/listinfo/opnfv-tech-discuss

Reply via email to