Re: [Hipsec] HIP DEX Aggresive Transmission Mechanism

hmmm. On further study of the HIP BEX state machine, one slight change...

On 06/23/2010 03:20 PM, Robert Moskowitz wrote:
> HIP BEX has been observed to behave poorly (or so I am told) on 
> networks with packet loss.
>
> HIP DEX will be used on networks with high packet loss, and MUST 
> perform well. So I have developed an Aggresive Transmission Mechanism 
> for DEX. Think if you want this added to 5201-bis.
>
> Here is the text from the DEX ID:
>
> HIP DEX may be used on networks with high packet loss. DEX deals
> with this by using an aggressive transmission practice for I1 and I2
> packets. The Initiator SHOULD continually send I1 and I2 packets at
> some short interval t msec, based on local policy. The transmission
> stops on receipt of the corresponding R1 or R2 packet, which acts as
> an acknowledgment receipt.
>
> Since the Responder is stateless until it receives an I2, it does not
> need any special behaviour on sending R1 other than to send one
> whenever it receives an I1. The Responder sends an R2 after receipt
> every I2. The Responder does need to know that R2 was received
> by the Initiator. It can learn this by NOT receiving any I2 packets
> after sending R2 after waiting the interval t msec that was in I2.

BEX learns this when R starts receiving data. That should also work for 
DEX. One less change.

>
> The Initator SHOULD NOT send any datagrams for the interval t msec
> after receipt of R2 to ensure that the Responder has transitioned to
> the ESTABLISHED state.
>
> =========================================
>
> Comments, of course, are welcome and invited.
>
> _______________________________________________
> Hipsec mailing list
> Hipsec@ietf.org
> https://www.ietf.org/mailman/listinfo/hipsec
>