Re: Ohta-san's draft

[Greg Maxwell <gmaxwell@martin.fl.us>]

Here is my perspective,

On 11 Apr 2001, Sean Doran wrote:

>     How does a host know its address?

I would purpose the same way it lears it's address normally.
 
>     That is, how does a host learn what set of path-indicators
> _can_ (not should, can) be advertised during the initial
> handshaking of a new TCP connection (or thereafter)?

The router/dhcp server tells it, or it could be manually configured. Link
local addresses should be ineligible (?), site locals should only be used
if the other host makes the connection to a site local on the system or
the other hosts share a common full prefix. 

(The same autoconfiguration method should also communitate traffic
engineering information).

>       An example set of issues that your answer hopefully will
>       make obvious based on this small graph follows.
>                                  
>                           MultihomedSender
>                            |          |
>                            v          v
>            LocalISPNumberThree       LocalISPNumberFour
>               |      |                 |       |
>               V      V                 V       V
>       RegionalISPZ RegionalISPY RegionalISPX RegionalISPW
>                         (Really big ISPs)
>       RegionalISPA RegionalISPB RegionalISPC RegionalISPD
>            |             |           |           |
>            |             |           |           |
>            |             |           |           |
>            v             v           v           v
>            LocalISPNumberOne         LocalISPNumberTwo
>                            |         |
>                            |         |
>                            v         v
>                           MultihomedSite
>                                 |
>                              Receiver
> 
>         How many path-indicators should MultihomedSender
>         announce to the Receiver?  Two?  More than two?
>         How many path-indicators will Receiver announce
>         back in reply to the Sender's SYN (which uses a
>         path-indicator from the DNS)?  

Assuming that all ISPs choose to pass their prefixes on (use of a
providers multihoming, or certan transits, might be a value added
service), and the sites on each end choose to pass on all the prefixes
they recieve (for traffic engineering or other local reasons, you might
choose to assign differnt prefix subsets to hosts):

Each end would send 4 potential prefixes. 

>         (We assume that the entity that informs the
>         Receiver what path-indicators it has can also
>         inform the DNS if desired, directly or indirectly)

The same mechanism can be used as is used for communicating the IP address
in any other dynamically configured enviroment.

>         Next: change.  What happens on the host if
>         MultihomedSite multihomes to another network?
>         What happens if some time (days?) subsequently
>         MultihomedSite permanently disconnects from
>         LocalISPNumberOne?  If the answer is: the host
>         acquires via some protocol (or manual
>         configuration) a new path-indicator, who assigned
>         the path-indicator in the first place?

The paths are acquired via manual or via DHCP via a link-local via host
autoconfiguration (or add it to autoconfiguration and remove DHCP
entirely).

>         Incidentally, any solution to this which does not
>         require changes to the current routing system is
>         also a general solution to site renumbering, which
>         is much-missed feature in IPv4 and IPv6, and
>         likely would be worthy of an RFC of its own.

Right. Renumber without losing you active SSH connection if there is a
way to push the new path-prefixes to the hosts. 

Also, this proposal will greatly increase consumer choice: Multihome via
cablemodem, DSL, (both often fast and cheap but not reliable) and a T1
(more expensive per bandwidth buy typically more reliable), no problem.
The providers don't even need to know about it *and* can still impliment
source address filtering if you use source based routing at egress.