Transcript (MPLS-1).sub-ip

A New IETF Work Area?
Fred Baker
IETF Chair
A long time coming
Sorry for the last minute confusion
 The IESG has been talking about this for a long time
 Its hard because changing the IETF structure is hard

• And should be hard

Not quite done yet
• Waiting feedback from this week
49th IETF CCAMP BOF
Above and below

Traditionally the IETF has been:
• “Above the wire and below the application”
• Not defining user interfaces
• Not defining physical wire types

But doing “IP over foo”
• “Foo” has been network types
• Ethernet, Token Ring, ATM, Sonet/SDH, ...
• But foo is changing
49th IETF CCAMP BOF
IP over “trails”, “circuits”, “paths”, ...

What looks like wires to IP may not be physical
wires
• May instead be something where paths can be configured
• Where a path looks like a wire to IP
– E.g. ATM VCs
• Might also be routed datagrams another layer down
– E.g. Ipsec tunnels

And then there is MPLS
• A progressively more important “foo”
49th IETF CCAMP BOF
Layer violations

Another complexity if the sub-IP technology is
configurable
• E.g. MPLS, ATM, Frame Relay, ...

Might want to have IP-level routing be able to know
about sub-IP paths
• Question may be “could a new path exist with certain
characteristics”
• Not just “does a path exist”
49th IETF CCAMP BOF
Architectural Issues

End to end?
• Means host to host
• Internet layer is end to end
• Service provider network is edge to edge
– Precludes application optimizations

Internet: heterogeneous network of networks
• Optimizations can preclude future developments
• Optimizations might not work out
49th IETF CCAMP BOF
UNIVERSITY
Not limited to MPLS

Other current and proposed IETF efforts deal with
sub-IP
•
•
•
•
Traffic Engineering,
General Switch Management Protocol
IP over Optics
Provider Provisioned VPNs...
49th IETF CCAMP BOF
A new area?

A systematic approach to sub-IP issues would be
nice
• But exact scope is not yet clear

IESG created a temporary pseudo-area for sub-IP
• Temporarily, in General area, managed by IETF Chair
• IESG will evaluate effort before next IETF meeting
49th IETF CCAMP BOF
Options with respect to area
New area not needed: distribute Working Groups to
other areas
 Only short term requirement: continue pseudo-area

• Disband when done
Might ask current Area Directors to manage (like
with IPng)
 New IETF area needed: ask nomcom to search for
AD(s)

49th IETF CCAMP BOF
Non-Objectives

The IETF is not expanding into standards for
physical or virtual circuit technologies
• No new circuit switch architecture from IETF
• Leave them to others

But may form Working Groups to help advise other
standards organizations on how to make things IPfriendly
• E.g. Ipoptr
• Need to communicate with other standards organizations
on what we are actually doing
49th IETF CCAMP BOF
Objectives

Layer 2.5 protocols:
• E.g. MPLS

Protocols that monitor, manage or effect logical
circuit technology
• E.g. IP Over Optical, Traffic Engineering, Common
Control and Management Protocols

Protocols that create logical circuits over IP:
• E.g. Provider Provisioned VPNs

Protocols that interface to forwarding hardware
• General Switch Management Protocol
49th IETF CCAMP BOF
Area management
Housed in General area under IETF chair
 New area directorate

• Area Directors from O&M, RTG, INT & TSV
• Working Group chairs for Working Groups in area
• People that the directorate thinks would be helpful

Directorate to advise IETF chair about management
of area
49th IETF CCAMP BOF
Entities in new area

Working Groups:
•
•
•
•

Internet Traffic Engineering (tewg)
Multiprotocol Label Swapping (mpls)
Generic Switch Management Protocol (gsmp)
Provider Provisioned Virtual Private Networks (ppvpn)
BOFs this week
• Common Control and Management Protocols (ccamp)
• IP over Optics (ipo)
• IP over Packet Transport Rings (ipoptr)
49th IETF CCAMP BOF
Status

Working Group list not final
• Looking for advice

Working Group charters (and re-charters) not final
• Looking for advice

Working Group chairs not final
• Looking for advice

Internet Draft partitioning not yet done
• Looking for advice
– Where should IDs be considered (if at all) ?
– Which IDs should be consolidated?
– For this meeting IDs stay where they were
49th IETF CCAMP BOF
A New IETF Work Area?
Discuss:
Request:
Discussion...
[email protected]
[email protected]