Re: RoHCv2 : Problems in decoding sequentially late packets

On Jan 13, 2009, at 09:31, Carl Knutsson wrote:

> If reordering is causing an lsb encoded field is decompressed
> incorrectly, then you have an incorrect the reording_ratio and/or a to
> small sliding window (see section 5.1.2).


Well, the whole thing is a probability game.
So, more precisely put: If that happens often enough to result in a  
performance problem, you should use different values and sacrifice  
some compression performance.
If it happens rarely enough that it doesn't result in performance  
degradation, you probably don't care about the occasional mis- 
decompression.
Most of these cases will be caught by the header checksum, and then  
there is often a UDP checksum as well.
(Link layer checksums are not perfect even without decompression,  
either.)

Gruesse, Carsten