Re: soft state (was Re: shim6 and bit errors in data packet headers

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

On 12-mei-2005, at 15:30, marcelo bagnulo braun wrote:

> I mean, is it possible to create two contexts between the same two  
> nodes with the same two ULIDs (and perhaps a different context  
> tag)? would this make any sense?

Having two contexts between the same hosts can lead to trouble. On  
the other hand, making sure this can't happen is also hard. But we  
should definately not make this something that happens routinely, IMO.

> The above assumption is based in that i think this is needed to  
> properly protect from flooding attacks.
> I mean, the goal of reachability tests in a shim protocol can be  
> two-folded:
> - to explore if a given address pair is working
> - to determine if a given host is willing to receive packets at a  
> given address (i.e. to prevent flooding attacks)

> the first goal can be achieved with a kind of ping (enhanced ping  
> in order to determine unidirectional reachability) but it may not  
> require that the reachability test is associated with a given  
> context. I mean, the goal here is just to obtain reachability  
> information

> The second goal is somehow different, and since what is being  
> queried is the willingness of the node to receive traffic through a  
> certain address, it is needed to inform the node which traffic are  
> we talking about.

The thing is, that you need some kind of semi-reliable communication  
to accomplish all of this, and to do the negotiation too. (In the  
case of the reasability testing, it would be useful if the receiver  
of a probe would have a semi-reliable way of telling the sender that  
the probe was received.) There is also some circularity: you need  
connectivity to negotiate locators, but you need to know the locators  
already to discover connectivity.

I'm thinking we could come up with a kind of "UDP on steroids" that  
similarly to SCTP works over multiple addresses. Such a semi-reliable  
multi-address datagram protocol would be very useful for the  
negotiations and similar exchanges, and it would supply reachability  
information as sort of a by-product.

Separating this layer also has the advantage that a simple  
unprotected multi-address datagram protocol can be used for other  
purposes as well, and it allows us to layer the solution so that we  
can work on different aspects independently.

> :-)
> Context state AND path failure AND unidirectional connectivity....  
> this seems amusing enough

Since the situations where context state doesn't exist (= before the  
start of a session) and unidirectional reachability (caused by  
ingress filtering) will be very common, we really need to cover this  
case.