Softwire wg
Alain Durand, ComcastDavid Ward, Cisco
Note WellAny submission to the IETF intended by the Contributor for publication as all or part of an IETF Internet-Draft or RFC and any statement made within the context of an IETF activity is considered an "IETF Contribution". Such statements include oral statements in IETF sessions, as well as written and electronic communications made at any time or place, which are addressed to: the IETF plenary session, any IETF working group or portion thereof, the IESG, or any member thereof on behalf of the IESG, the IAB or any member thereof on behalf of the IAB, any IETF mailing list, including the IETF list itself, any working group or design team list, or any other list functioning under IETF auspices, the RFC Editor or the Internet-Drafts function
All IETF Contributions are subject to the rules of RFC 3978 and RFC 3979. Statements made outside of an IETF session, mailing list or other function, that are clearly not intended to be input to an IETF activity, group or function, are not IETF Contributions in the context of this notice. Please consult RFC 3978 for details.
Wg status• Charter went to internal & external review• Comments received• Chairs, AD & IESG members started addressing the
comments…• …AD had a baby! (Congratulation!)
• As of this morning, we are approved as a wg by the IESG!– Secretariat still needs to make it formal
Agenda• Overview of meeting in Paris (Chairs)
• Hub and Spoke Problem Overview (Durand)– Hub and Spoke Illustration (Miyakawa, Palet, Williams)
• Mesh Problem Overview (Ward)– Mesh Illustration (Li)
• Status of draft problem statement (Chairs)
• Next steps (Chairs, all)
Paris Interim Meeting• We held an interim meeting in Paris on October 11th-12th
– 18 participants, intense discussions, very productive meeting
• Focus on problem statement– draft-durand-softwire-problem-statement-00.txt
edited in rush just before the cut-off date (excuse typos…)
• 2 problems identified, topology driven:– Access network, customer initiated, one exit path [Hubs & Spokes]– Core network, ISP initiated, complex routing topology [Mesh]
• We will look at both problems independently– Hopefully, they will share enough common technology
Hub & Spoke Description
Hubs & Spokes Problem• Description:
– Access network problem, customer initiated, one exit path
• Applicability:– ISPs with Dual Stack core and a number of dual stack Points of
Presence (“Hubs”) where they connect their customers.
• 3 usage cases have been identified:– the networks between the CPE router and the hub supports only one
address family.– the CPE router cannot be easily upgraded to support both address
families, a softwire is created from a node behind the CPE router– Same, but initiated from another router behind the CPE router
Usage Case 1Dual AF
Single AF
CPE Router Dual AF
Softwire Initiator
SoftwireConcentrato
r
Usage Case 2Dual AF
CPE Router Single AF
Dual AF HostSoftwire Initiator
SoftwireConcentrato
r
Usage Case 3Dual AF
CPE Router Single AF
Dual AF RouterSoftwire Initiator
SoftwireConcentrato
r
Hubs & Spokes Assumptions• NAT/PAT (in IPv4) is present
• Not always upgradeable CPE router
• “Stable” IPv6 prefix desired
• Softwires initiated by customer– Customer side: softwire initiator
• May be a host or a router– ISP side: softwire concentrator
• Routing:– default route from softwire initiator to concentrator
• (CPE routers do not generally run a routing protocol, but the softwire solution will work even if it does.)
Hubs & Spokes Properties (1)• Scaling:
– to the millions of softwire customers
• Set-up time (a.k.a. “latency”)– A fraction of the total set-up time of the CPE router
• Multicast– Classic multicast solution run over the softwire
Hubs & Spokes Properties (2)
• Security– Must support secure user authentication
• May be turned off.– Must be able to support payload security when desired outside
of the softwire mechanism
• Operation And Management– Keep alive– Usage accounting– End point failure detection (inner address of the softwire)– Path failure detection (outer address of the softwire)
Hubs & Spokes Encapsulations• Critical path
– IPv6/IPv4– IPv6/UDP/IPv4– IPv4/IPv6
• Other encapsulations to be supported later(e.g. IPv6/IPv6)
Hub & Spoke Illustrations
Slides from Shin, Carl & Jordi
Mesh Description
Mesh Problem• Description:
– Core network problem, ISP initiated, complex routing topology
• Applicability:– ISPs (or large enterprise networks acting as ISP for their internal
resources) establish connectivity to 'islands' of networks of one address family type across a transit core of a differing address family type.
Mesh Diagram
IPv6-only Transit Core
BGP Dual-Stack AFBR
BGP Dual-Stack AFBR
BGP Dual-Stack AFBR
BGP Dual-Stack AFBR
IPv4Access Island
IPv4Access Island
IPv4Access Island
IPv4Access Island
IPv6Access
Network
IPv6Access
Network
AFBR• To provide reachability across the transit core dual-stack
devices are installed that act as "Address Family Boundary Routers”.– Creates a limited dual-stack edge network– Core can be solely one AF and islands don’t require upgrade
• AFBR provide peering across AS or within an AS• Can be used inconjunction w/ route reflectors
Full Mesh Overlay for Many2Many connnectivity
AFBR
AFBR
AFBR
V6 transit
V4 island
V4 island
V4 island
May have different encaps available
Must have solution to allow for negotiation and preference of encap
AFBRMGREIPsec
V6 transitcore AFBR
MGRE,L2TPv3
AFBRL2TPv3MPLS
V4 island
V4 island
V4 island
L2TPv3 MGRE
IPsec
Must support Applications…. L3-VPN using 2547bis
Route Reflector
AFBR
AFBR
V6 transit
V4 island
V4 island
V4 island
AFBR
VPN VPN
VPN
Mesh properties (1)• Scaling
– Number of AFBR related to the number of islands and exit points from islands (x0-x00 islands)
• We know of no cases of x0000++ islands– Full routing table needs to be supported
• Islands can carry x00000 of routes
• Services / Encapsulation– v4/v6 or v6/v4– L2VPN– L3VPN (overlapping address spaces)– Multicast a must in all cases
• Security– No “user” authentication– Authentication for control plane
• may be turned off– Support for IPsec in data plane (outside of softwires)
Mesh properties (2)• Operation And Management
– No need for keepalive– Usage accounting– End point failure detection– Path failure detection– Flexible encapsulation possibilities– Interconnection at L2 or L3– Cannot require full mesh of all AFBRs under all circumstances
Mesh Illustrations
Slides from Pr Li
Problem Statement Draft Status• Problem statement described in
draft-durand-softwire-problem-statement-00.txt• Comments received on the ML
– Typos– Some minor stuff– n engineer that comes up with n+1 design syndrome– 3 issues raised about the Mesh problem:
• Scale– Presented today
• Should this be solved at layer 2 or layer 3– Crystal ball says both (This belongs to the solution space)
• Should the softwires be initiated from the PE or CPE or both?– Crystal ball says most commonly PE (for mesh)
Next Steps• Mark finish the creation of the wg!
– Done, minor nits on charter + secretariat action
• Rev problem statement draft– draft-ietf-softwire-problem-statement-00.txt Nov. 14th– draft-ietf-softwire-problem-statement-01.txt Dec. 1st
• WG Last Call on problem statement draft– Target: Dec. 8th
• Interim meeting on solution space (Jan/Feb 06)– Last was in Europe, Hong Kong?