vlan translation in Quantum

asked 2013-10-23 18:17:31 -0500

Giuliano
3 ●2 ●4 ●5

updated 2013-10-23 18:20:57 -0500

I don't understand tha VLAN translation mechanism. In the http://docs.openstack.org/network-admin/admin/content/under_the_hood_openvswitch.html (OS official guide) it's said the net01 and net02 have VLAN ID 1,2 while VLAN 101,102 are EXPLICITY specified in the quantum CLI commands.
Why the need to change the VLAN ID internally, that is inside the compute node and inside the network node?

Can anyone please explain me the mechanism and the reason behind this?

I'm using Openstack Grizzly.

answered 2013-10-25 07:07:08 -0500

RomilGupta

844 ●7 ●17 ●29

Hi ,

Earlier I was having the same doubt , the segmentation ID we explicitly specified are Global VLAN ID( 101,102...) and the VLAN ID we see on the Compute node or NW node are local VLAN ID (1 ,2 ...). The OVS agent internally maintains a mapping table which contains the mapping entry for every unique global vlan id to local vlan id for ingress and egress traffic.

edit flag offensive delete link

Comments

Yes, but I don't understand why doing such a translation/mapping... Maybe is it necessary to allow the independence from the actual implementation of the network segregation on the Data Network (VLAN, GRE, VXLAN...) ?

Giuliano ( 2013-10-28 08:57:29 -0500 )edit

add a comment

1 answer

Comments

Get to know Ask OpenStack

Question Tools

Stats

Related questions

Feedback About This Page

OpenStack

Community

Documentation

Branding & Legal

vlan translation in Quantum edit

1 answer

Comments

Get to know Ask OpenStack

Question Tools

Stats

Related questions

Feedback About This Page

OpenStack

Community

Documentation

Branding & Legal

vlan translation in Quantum