diff options
author | Manuel Buil <mbuil@suse.com> | 2017-08-21 14:10:22 +0200 |
---|---|---|
committer | Manuel Buil <mbuil@suse.com> | 2017-08-21 14:14:29 +0200 |
commit | 797f9156bfe3fbbf6fb68b62b906799484d4bbc6 (patch) | |
tree | 7be6a1b2d9a2903fc67c6b300b86b472bc1d902c /docs/development/design/architecture.rst | |
parent | 221fd4cfd7f86e110b3536066ede2910191dfc1e (diff) |
Update the docs/development directory
MS6 requires to have the docs directory structure ready and a preliminary
versio of the documents inside it.
- Moved requirements file to the requirements folder
- Removed the NSH workaround info as we don't need it anymore
- Added some new requirements
Change-Id: Icca89c738b1eb1865adff510ca2850c592c7a14a
Signed-off-by: Manuel Buil <mbuil@suse.com>
Diffstat (limited to 'docs/development/design/architecture.rst')
-rw-r--r-- | docs/development/design/architecture.rst | 27 |
1 files changed, 0 insertions, 27 deletions
diff --git a/docs/development/design/architecture.rst b/docs/development/design/architecture.rst index b190a18f..60183702 100644 --- a/docs/development/design/architecture.rst +++ b/docs/development/design/architecture.rst @@ -100,30 +100,3 @@ OPNFV SFC Network Topology The following image details the Network Topology used in OPNFV Danube SFC: .. image:: ./images/OPNFV_SFC_Brahmaputra_NW_Topology.jpg - - -OVS NSH patch workaround -++++++++++++++++++++++++ - -When using NSH with VXLAN tunnels, it is important that the VXLAN tunnel is -terminated in the SF VM. This allows the SF to see the NSH header, allowing -it to decrement the NSI and also to use the NSH metadata. When using VXLAN with -OpenStack, the tunnels are not terminated in the VM, but in the "br-int" OVS -bridge. There is work ongoing in the upstream OVS community to implemement NSH -encapsulation. To get around the way OpenStack handles VXLAN tunnels, the OVS -work will also include the ability to encapsulate/decapsulate VXLAN tunnels from -OpenFlow rules, instead of relying on the Vtep ports. The ongoing upstream OVS -work will not be finished by the time OPNFV Danube is released, so -a work-around has been created. This work-around will use a private branch of -OVS that has a preliminary version of NSH implemented. - -The following diagram illustrates how packets will be sent to an SF, when the -SFF has processed the packet and wants to send it to the SF: - -.. image:: ./images/OPNFV_SFC_BrahmaputraOvsNshWorkaround_toSf.jpg - -The following diagram illustrates how packets will sent from an SF to an SFF, -once the SF has processed a packet: - -.. image:: ./images/OPNFV_SFC_BrahmaputraOvsNshWorkaround_fromSf.jpg - |