Hi Donald,

 

Thank you! Please refer to response inline.  (<<<JAY).

 

Thanks,

Jay

 

 

On 6/28/18, 5:09 PM, "Donald Sharp" <sharpd@cumulusnetworks.com> wrote:

 

    A) OSPF v2 supports both MI and VRF's.  Use one or the other, but not

    at the same time.  No-one has done the work to make IS-IS vrf aware at

    this point in time(patches welcome! HA!)

(<<<JAY) Ok I will add IS-IS to the to-do-list which needs work to support MI and VRF’s.

   

    For B, I am not really quite sure what you are asking.  I guess that

    you are asking does FRR support l3vpn's and EVPN in a VRF context and

    yes it does.

 

(<<<JAY) Yes, (b) means VPN VRF

   

    For C, I do not understand your question.

(<<<JAY) Contrast to (b), non-VPN routing instance; Does FRR VRF support both cases, may be the context tell the difference?

    https://ipwithease.com/difference-between-vrf-and-virtual-router-in-juniper/

 

   

    As for logical router, just create linux namespaces and run separate

    instances of FRR.

(<<<JAY) OK.

   

    donald

   

    

    On Thu, Jun 28, 2018 at 4:59 PM, Jia Chen <jchen1@paloaltonetworks.com> wrote:

    > Hi All,

    >

    > A question about FRR way of supporting vrf, multi-instance, or logical router.

    >

    > A. OSPF/IS-IS multi-instance is a control plane concept, not forwarding

    > B. VRF is both routing and forwarding, for example, L3vpn and EVPN, which uses RT, RD in BGP updates

    > C. Virtual router is a non-VPN instance, which does not require RT, RD

    >

    > Both B and C, has a one-to-one interface to routing instance mapping.

    >

    > Are A, B, C all supported by FRR?

    >

    > Juniper has a logical system (or router), which divided physical router into logical sub-routers. Each logical router has dedicated interface, routing protocols, routing and forwarding table. Are there equivalent in FRR?

    >

    > Thanks,

    > Jay

    >

    >

    >

    > On 6/28/18, 9:31 AM, "dev on behalf of Donald Sharp" <dev-bounces@lists.frrouting.org on behalf of sharpd@cumulusnetworks.com> wrote:

    >

    >     Mark -

    >

    >     Great idea, We'll start the meeting off with your talk.

    >

    >     thanks!

    >

    >     donald

    >

    >

    >     On Thu, Jun 28, 2018 at 11:58 AM, Mark Stapp <mjs@voltanet.io> wrote:

    >     > Hi,

    >     > I've been doing some work on the dataplane side of zebra, and I'd like to

    >     > talk a bit about the approach and outline some of the possible next-steps.

    >     > Could I have a few minutes at next week's dev meeting to introduce the

    >     > topic? The broad goals are sort of:

    >     >

    >     > * to harden the interface that zebra has with the dataplane or -planes that

    >     > are in use,

    >     > * to improve support for remote dataplanes,

    >     > * to allow dataplane/kernel updates to be moved to a dedicated pthread,

    >     > * to improve integration of FPM, and to include more events

    >     >

    >     > I've been working at first just to support asynchronous route/fib updates,

    >     > just to start exploring the possible approaches and to raise some questions.

    >     >

    >     > Thanks,

    >     > Mark

    >     >

    >     > _______________________________________________

    >     > dev mailing list

    >     > dev@lists.frrouting.org

    >     > https://urldefense.proofpoint.com/v2/url?u=https-3A__lists.frrouting.org_listinfo_dev&d=DwIGaQ&c=V9IgWpI5PvzTw83UyHGVSoW3Uc1MFWe5J8PTfkrzVSo&r=yetdj-aXQpuqTCJGs-93hOpK3740MIRXowfUNLByeos&m=0aLyQFqXsyUE3eFVSmsBMSQzPcFi8wkUtaBrS_9RhKM&s=rttXUUgi9mh45PL0z88r5pPfh1dDBfOaglxXATnOa-s&e=

    >     >

    >

    >     _______________________________________________

    >     dev mailing list

    >     dev@lists.frrouting.org

    >     https://urldefense.proofpoint.com/v2/url?u=https-3A__lists.frrouting.org_listinfo_dev&d=DwIGaQ&c=V9IgWpI5PvzTw83UyHGVSoW3Uc1MFWe5J8PTfkrzVSo&r=yetdj-aXQpuqTCJGs-93hOpK3740MIRXowfUNLByeos&m=0aLyQFqXsyUE3eFVSmsBMSQzPcFi8wkUtaBrS_9RhKM&s=rttXUUgi9mh45PL0z88r5pPfh1dDBfOaglxXATnOa-s&e=

    >

    >