Re: shim - transport/app communication

[Iljitsch van Beijnum <iljitsch@muada.com>]

On 16-mrt-05, at 12:44, Pasi Sarolahti wrote:

> I was wondering what would
> happen, for example, when these addresses are being advertised out in
> the network as a part of some application protocol.

This is one of the reasons why it's a good idea to have the addresses 
that are exposed to the application be addresses that are actually 
reachable (most of the time). This means that as long as there aren't 
any failures, referrals work just as good or bad as they work now. 
Obviously when the initially used address becomes unreachable, the shim 
will move the session to another set of addresses and now trying to 
contact the original address will probably not be problem free. But 
it's still better than what you get without the shim.

> Another example, I guess that bind() with IP address should still 
> select
> the interface used for the socket, right? In principle, if assuming
> ULIDs, the parameter address would represent the host, not the
> interface.

That's a very interesting change in viewpoint, which may or may not 
have important consequences. Note that we often give routers loopback 
interfaces with real addresses (not ::1) exactly because they need an 
address that remains reachable despite the status of any one interface.

Iljitsch