Version 2 of working group notes

[[e-mail]]

I've included the comments I've received.  Please review
these notes.  If I don't hear anything by 3/21/96, the notes
will be considered correct.   Please note ">" indicates a
line that was changed.

			Sue Hares

=======

BGP Working Group   

Date: 3/5/96 time 3:30pm -5:30pm PST 
(note taker: Susan Hares)
(version 2: changes have > in front of line)

1) BGP-4 MIB (draft-ietf-bgp-mibv4-06.txt)

	Jeff Johnson (Cisco), the latest editor,
	of the BGP-4 MIB document will make the
	few minor editorial changes given by the list. 

	The working group needs to forward two
	reports of those using the BGP-4 MIB.
	Cisco and GateD have BGP-4 mibs.

	Someone from Cisco, and Sue Hares from
	Merit will forward experience reports to
	Yakov to send to the IESG.  At this point
	the BGP-4 MIB will be forwarded to
	Draft Standard.

	MIB entries for newer attributes (Destination Preference Attribute,
	BGP communities, and Conferations) will go in a separate
	document. 

2) IDRP for IPv[46] (draft-ietf-idr-idrp-v4v6-02.txt)

	Yakov will produce a complete IDRP specification
	based on the the difference document and the 
	IDRP specification.  Anyone wishing to comment
	on features in BGP-4 or the features in the
	existing document should send mail to the
	mailing list or Yakov Rekhter.

3) Confederation  (Ramesh Govindan) (draft-ietf-idr-rdc-config-00.txt) 

	One problem with IDRP confederations is that
	every BGP speaker must know what confederations it
	belongs to.  This amount of configuration guarantees
	that the correct processing of confederations even if
	individual nodes are mis-configured.  If instead you,
	configure 
		- only the border routers AND
		- if there is ambiquity, assume nesting

	You may be able to limit configuration amounts. 
	
	This proposal had some discussion from the floor and
	will continued to be discussed on the mailing list.

4) Destination Preference Attribute (DPA)

>	(draft-ietf-idr-bgp-dpa-05.txt
>	 draft-ietf-idr-dpa-application-02.txt
	 draft-ietf-idr-symm-multi-prov-02.txt)

	The processing of the the DPA value is currently
	included in the route policy calculation as:

	Multi-Exit-Discriminator (MED)  - 1st
	Destination Preference Attribute (DPA) - 2nd 

>	Consensus has been reached by all parties and will be put out as
>	draft-ietf-idr-bgp-dpa-06.txt.   After the document
>	to the working and Internet Drafts this document will be
>	moved to Proposed Standard.  

>	The associated documents:  
>	 draft-ietf-idr-dpa-application-02.txt
>	 draft-ietf-idr-symm-multi-prov-02.txt 
>
>	will be moved to informational RFCs.
>
>
>  New text for this decision in the DPA document is:
>	
>  A router may use DPA to influence LOCAL_PREF computation. DPA shall
>  not directly affect route selection.
> 
>  DPA influence of the LOCAL_PREF attribute computation is a local
>  matter. In general a route with a higher DPA indicates a higher
>  preference by the originator of the DPA attribute.	
		 

5) BGP communities
	
	The BGP community specification has been release for
	over a year. It has been implemented
	by Bay networks and Cisco.  It is suggested that
	Yakov make the last call, and move it to proposed standard. 
>	(document was at: draft-chandra-bgp-communities-00.txt
>	 will be at: draft-ietf-idr-bgp-communities-00.txt)	

	Tony Bates and Enke Chen has released a draft
	on the usage of communities.  Enke Chen made a presentation
	of this feature at the NANOG meeting in February 1996.
>	This document is currently listed as:
>		draft-ietf-idr-community-usage.txt

6) BGP Confederations

	draft-trainia-bgp-confed-00.txt

	It was moved that is be moved to experimental.
	In deployment Cisco discovered some problems with
	this approach.

7) Route Reflection

	draft-ietf-idr-route-reflect-00.txt
	
	Cisco has implemented the route reflection a year ago.
	MCI has used this in it's network for a year.
	(Has Bay implemented this?) 	
	 
	The working group recommends that this move forward to 
	experimental.  A work item for the group is to
	decide which solution to the IBGP mesh problem will
	be selected by the group: the Route Reflection 
	or the Route Server (Dimitry Haskins) or a cobmination
	of both. 

Fun quotes:

	"You forced this @#* MIB down my throat."
	"Standards don't matter if the code is running 
	 in all the routers"