Host Centric Multi 6 Christian Huitema Architect Windows

Host Centric Multi 6 Christian Huitema Architect Windows Networking & Communications Microsoft Corporation

Principle Site is connected to multiple providers Internet n P 1 P 2 R R R H H 7/20/2002 Prefixes are propagated to all site routers n Router renumbering? Hosts get as many addresses R n H H Get as many prefixes n n Prefix = provider Subnet number Host ID Publish addresses in DNS Host centric multi 6 2

Issue: destination address selection Fairly common n Many hosts are multi-homed. Debate whether hosts have sufficient information n n Hard for small appliances, not enough information Easy for large servers It is not unrealistic to expect progress in this area, n n n 7/20/2002 communication between the hosts and the routers, sharing of experience between hosts, innovative application design At worst, a host can always try the proposed addresses one by one, and pick the first one that actually works -- not very elegant, but definitely workable. Host centric multi 6 3

Issue: source address selection Existing software ties source address selection to interface selection n Select outgoing interface Pick one address on interface as source Only consider address scope, and possibly “privacy” status Choosing the source address will affect the reverse path of the connection n Issue similar to “destination address selection” We need some improvement for multiaddressing 7/20/2002 Host centric multi 6 4

Issue: rapid reaction to topology change One of site X providers (A) becomes unreachable How do we avoid picking a source address A: X ? How do peers avoid picking a destination address A: X ? 7/20/2002 Host centric multi 6 5

Issue: site exit & ingress filtering /--( A )--( )–-( C )-- (RXA) ( ) (RYC) X (site X) ( IPv 6 ) (Site Y) Y (RXB) ( ) (RYD) --( B )--( D )--/ X picks source address A: X, dest D: Y Routing fabric sends packet to exit router RXB Provider B sees source = A: X, perform ingress filtering, rejects the packet 7/20/2002 Host centric multi 6 6

Classification of the issues Destination address selection Nice to have. Should at least “retry” Source address selection Nice to have. Reaction to topology changes Nice to have Ingress filtering MUST SOLVE 7/20/2002 Host centric multi 6 7

Comparison of ingress filtering solutions Relax address filtering n n Requires provider involvement Easy to deploy for large sites Source address dependent routing n Variant = tunnels between exit routers Packet rewriting at exit router n Inferior to “exit tunnel” solution Source address selection by the host n n 7/20/2002 Complement to source dependent routing Requires “exit router” discovery Host centric multi 6 8

The “dumb host” requirement Unmodified host n n Picks a single source address Must work at least as well as “not multihomed” Consequence n n 7/20/2002 Ingress filtering must work for all destinations if source provider available Imply either “relaxed filtering” or “per source routing” Host centric multi 6 9

Solution’s principle (dumb host) X tunnel rxa A Y B ate d p U rxb 7/20/2002 Host centric multi 6 10

Solution’s principle (exit discovery) rxa try X A e c r u o s d ba Y B rxb • Try can be as simple as sending a “ping”, maybe with source address = site local… 7/20/2002 Host centric multi 6 11

Solution’s principle (exit tunnel & redirect) t c e ir red X Upd ate (in tun nel) tunnel rxa A Y B ate d p U rxb • There alternatives, e. g. don’t use update, just a direct tunnel to the “right” exit. 7/20/2002 Host centric multi 6 12

Proposed solution Facilitate site exit n n n Site exit “logical” address (for tunnels) Site exit redirect ICMP Tunnel to appropriate exit Router advertisements for rapid reaction n n preferred lifetime > 0 source is OK Need router renumbering Host improvements n n n 7/20/2002 Source and destination address selection Exit router discovery (understand site exit ICMP) Binding update / Mobile IPv 6 for “reassignment” Host centric multi 6 13

Going forward Reconcile / merge with Bagnulo’s draft n Compare binding update versus advertisement of multiple addresses Study possible provider help n Some form of tunneling when provider link is broken… Get consensus for a narrow scope WG charter, or progress document without a WG 7/20/2002 Host centric multi 6 14