Babel BoF 
1550-1720 4 April 2016

Introduction: Tony Przygienda, 5 minutes

What are trying to do ? 

    Not a barfight

    Tough on problems/Soft on People

    Not at attempt to kill Babel with Burearcracy


Gauge interest


Summary of charter, for context: Joel Halpern, 5 minutes
(Note: discussion of charter is later.)

Move 6126/7557 combined to Proposed Standard.

    Include security and manageability


Work Item Summary

    revise 6126 toward Proposed Standard

    Security needs for Babel

    Appliability Statemed

    Multicast

    ECMP

Applicability statement: Juliusz Chroboczek 10 minutes
http://tools.ietf.org/html/draft-chroboczek-babel-applicability
https://www.ietf.org/proceedings/95/slides/slides-95-babel-5.pdf

    Babel is

    robust

    flexible and extensible

    implementable


    Deployed Successfully:

    hybrid networks - classic prefix based, and pure mesh

    Global Overlay Networks

    Source Specific Routing

    Small, Simple Networks


    Could be Useful


    Better Protocols Exist:

    Large Stable Networks that use OSPF/IS-IS/EIGRP



Required changes to Babel documents: Juliusz Chromoczek 10 minutes
This is to focus on things we know need review,
based on RFC 6126 and RFC 7557.

https://www.ietf.org/proceedings/95/slides/slides-95-babel-4.pdf

    Uma Chunduri: 1) Major Problems looking to solve with protocol, and
2) future extensions to be added

    3) Security Aspects


    JC: 1)Simple, implementable suitablable for hybrid and highly
dynamic networks. 

    2) TLS 

    3) Security talk coming


    Erik Normark: Place about where routing interopable where possible.


    JC: one version speaks IPv4/V6 and one version speaks IPv6 only.
They will speak IPv6. 


    Emmanuel Baccelli: OSPF had talked extensions about dynamic
extensions for mesh networks. 


    JC: three extensions for mesh, should work in principle.


    Enno Rey: Work like to see move forward


Overview of some technical items: Tony Przygienda 10 minutes
This is to highlight some issues that may or may not need
to be addressed, but which were noted in several recent reviews.
The purpose of this presentation is to give an idea of the kinds
of technical work the WG is likely to need to do.


https://www.ietf.org/proceedings/95/slides/slides-95-babel-4.pdf


    PS Document Quality

    Serious discussions on oscillations

    Reference Implementation vs Specification

    Extensiability



Implementation issues: Toke Høiland-Jørgensen 10 minutes
https://www.ietf.org/proceedings/95/slides/slides-95-babel-3.pdf

    Summary: We should all try to implement it!


Security considerations: Denis Ovsienko 10 minutes
This will review RFC 7298
https://www.ietf.org/proceedings/95/slides/slides-95-babel-2.pdf


Discussion of draft charter and interest in forming a working group:20
minutes, led by Chairs
https://www.ietf.org/proceedings/95/slides/slides-95-babel-1.pdf

Teco Boot: Most IETF protocols can be tuned. Babel also supports this.
But a zero-conf network should have a consistent configuration, e.g. for
timers, metrics. New task to define such, currently missing in proposed
charter. Task for Homenet WG? or Babel WG? Define a Babel-Homenet
profile?

David Lampeliter: Multicast in Babel in the general sense. 
JH: Not requiring WG to address it, but needs to decided

Kiretti Kompella: Not sure why charter is in the WG. 
JH: Applicability statement. Can't talk smack about another WG
KK: Explain why don't use OSPF, IS-IS, etc. 
JH: Let's not go there. 

JH: number of people worthy of writing text:
    about 9 or so
    
JH: people worth

JH: worth forming WG (hum)

Large Hum in interest
Smaller Hum in negative

Barbara Stark:  Managability - ISP provided router, netconf suggested
but many millions of TR69

    Informational model to reflect

JH: add to minutes

Mikhal Abhamranson: Do you see other things being added to the charter
JH: Charter will have

JC: Proposed charter speaks of metrics

David Schinazi: Playing wiht it at his employer (Apple) but says no
more. 
 
Bob Hinden: Supports this work as former Routing AD. 1) used in
unmanaged environments; 

    2) multicast more interesting in unmanaged


Dave Taht: OK with basic charter would like zeroconf and L2 issues
explored -
 babel over 802.11, 6LOWPAN, 802.14, Moca, ethernet over powerlineet al 
Adrian Farrell:  Have done some routing protocols.  Issues about charter
text: risk in documentating babel as 
a standard. Must not confuse charter with standardizing babel.

Sunati ?: Extended to low power networks