Re: [Bgp-autoconf] Minutes and action points of the conference call

["Dongjie (Jimmy)" <[e-mail]>]

Hi Xiaohu,

Thanks for providing the list of requirements in your draft.

Best regards,
Jie

From: Bgp-autoconf [mailto:[e-mail]] On Behalf Of 徐小虎(义先)
Sent: Friday, February 28, 2020 9:11 AM
To: Bgp-autoconf <[e-mail]>; [e-mail]
Subject: Re: [Bgp-autoconf] Minutes and action points of the conference call


Hi all,



The following requirements are quoted from https://tools.ietf.org/html/draft-xu-idr-neighbor-autodiscovery-12#page-4



4<https://tools.ietf.org/html/draft-xu-idr-neighbor-autodiscovery-12#section-4>.  Requirements





   This section describe the requirements for the BGP hop-by-hop routing

   deployments that were considered for the definition of the BGP

   Neighbor Discovery extensions proposed in this document.



   Following are the key requirements related for the BGP neighbor

   discovery process:



   1.  It should perform discovery of directly connected BGP routers.

       Mechanism should support either IPv4 or IPv6 or a dual stack

       design and it should be generic for any link-layer.



   2.  It should include exchange of BGP peering addresses (IPv4 or IPv6

       or both) that routers can use to automatically setup BGP TCP

       peering between themselves.  The mechanism should leverage the

       existing capability negotiation process performed as part of the

       BGP TCP session establishment.



   3.  When BGP peering is desired to be performed over loopback

       addresses of the routers, then the mechanism should automatically

       setup reachability to the loopback over one or more underlying

       directly connected links between them.  In this scenario, the

       mechanism should also provide resolution for the BGP next-hop

       address (i.e. the loopback address) for the BGP routes exchanged

       over these sessions between the loopback addresses.



   4.  Mechanism should enable exchange of link-level information such

       as IP addresses and link attributes between the directly

       connected BGP routers.  It should be extensible to include other

       information in the future.



   5.  Mechanism should be limited to link scope for security and use

       link-local addressing only.  Cryptographic mechanisms should be

       also provided for additional security.



   6.  Mechanism should support capabilities for performing optional

       validation of parameters to detect misconfiguration (e.g. link

       address subnet mismatch, peering between incorrect AS, etc.) in

       an extensible manner before going on to use the link and the

       setup of the BGP TCP peering session over it.



   7.  The mechanism should not affect or change the BGP TCP session

       establishment procedures and the BGP routing exchange over the

       TCP session other than the interactions for triggering the setup/

       removal of peer session that is based on discovery mechanism.



   8.  The mechanism should leverage existing fast-detection techniques

       for failures that are used currently for EBGP sessions over

       directly connected links like fast-external-failover and BFD.



   9.  The mechanism should focus on the discovery process and exchange

       of status as a control plane procedure and be sufficiently

       loosely coupled with the base BGP operations to enable

       implementations to ensure scalability of BGP operations when

       using the discovery procedures.

Best regards,
Xiaohu
------------------------------------------------------------------
From:Dongjie (Jimmy) <[e-mail]<mailto:[e-mail]>>
Send Time:2020年2月25日(星期二) 23:28
To:[e-mail] <[e-mail]<mailto:[e-mail]>>
Subject:[Bgp-autoconf] Minutes and action points of the conference call

Hi all,

First of all, thank you for attending the conference call, we had good discussion on several items, and many thanks to Sue and others who helped with the minutes.

The minutes is on https://etherpad.ietf.org:9009/p/bgp-autoconf-feb-25, please check if you want to make any change to it.

I also summarized our discussion and action points as below:

- We agreed to work on DC first, while keep an eye on what makes the other cases different.

- For DC case, will focus on intra-data center which uses BGP as underlay, either interface addresses or loopbacks can be used for peering. Inter-DC and VPNs are out of scope.

- We had some discussion about whether we should cover the misconnectivity or misconfiguration detection, there is still no clear result.

- The team agrees that discovering the role of a device could be useful, something similar has been used in RIFT.

- We discussed the security requirements, the authentication has two properties, transport or object security, and authenticating the peer.

- In order to reduce the number of parallel peers between two nodes, need to use loopbacks to setup only one session, then need to consider how to automatically generate the configuration of routes to peer’s loopback.

- Enablement of BFD based on auto-discovery is considered as one optional function, while the complexity needs to be considered.

- Some considerations on Zero touch provisioning (ZTP). It is assumed some level of provisioning/pre-configuration is still required.

The team agrees to look at the list of solution drafts, summarize the requirements from each draft, and find a minimal common set.

Action points:
1. The author of each draft to summarize the list of requirement from his draft, and share it on design team list. (by end of this Friday)
2. Other people can provide additional comments to the summarized requirements. (in next week)
3. The team to provide a minimal common set.

Best regards,
Jie