Re: Echo Function and Asymmetry - Timer negotiation

[Dave Katz <[e-mail]>]

On Aug 11, 2005, at 1:58 PM, Tanya Shastri ((tanyas)) wrote:

>
> For an Asynchronous BFD session that has Asymmetric Echo function  
> the following statements in the spec seem to be misleading, unless  
> I am missing something.
>
> Section 6.4 says: "When a system is using the Echo function, it is  
> advantageous to choose a sedate transmission rate for Control  
> packets since liveness detection is being handled by the Echo  
> packets."
> Shouldn't it be the receive rate - or the transmission rate of  
> Control packets of the remote system - that should be sedate? Also  
> the end that does not have echo mode active relies on Control  
> packets for liveness detection and may still want to receive  
> Control packets at a faster rate.

Yes, the language is a little slippery.  It dates from a time in BFD  
prehistory when echo mode had to be symmetric, and the use of  
"transmission rate" was meant to be generic, that is, the packet  
rate.  As you note, in the asymmetric case, the guy not running echo  
mode may like to be receiving control packets at a faster rate.

>
> Section 6.7.3 says: " When the Echo function is active, a system  
> SHOULD set bfd.DesiredMinTxInterval to a value of not less than one  
> second"
> From the description in 6.7.4 this will force the remote end to  
> have a Detect Timer no faster than 1 second, since the Detect Timer  
> is based on the transmit rate of the remote system. Also this will  
> not reduce the rate at which it sends control packets to the end  
> that has echo mode active.
>
> Wouldn't it be better to suggest that "RequiredMinRxInterval" in  
> BFD control packets should be set to a minimum of 1 second by the  
> end that has echo function active?

Same issue, but you're right.  I'll fix it in the next round.

--Dave