Re: REJECT and SN-NOT-VALID

Klaus Warnke <[email protected]> Fri, 14 Aug 2009 17:58:43 +0200
Newsgroups gmane.ietf.rohc
Message-ID <u8whmbd58.wl%[email protected]>
At Fri, 14 Aug 2009 21:15:45 +0800,
  wrote:
> RFC 3095 describe the compressor's action when receive REJECT or
> SN-NOT-VALID feedback options, However, It seems that the
> decompressor does have sufficient resources definitely to handle the
> packet stream, taking into account the consistency of MAX_CID.
> 
> Moreover,How the decompressor decide whether carry REJECT or
> SN-NOT-VALID?

CaiWei,

the SN-NOT-VALID is for the case, that the de-compressor could not
determine the SN of the packet causing that feedback. This occurs, if
the CRC check fails. Then the de-compressor does not know, where the
error in the packet is and it could be, that the SN was damaged while
transmission (or the CRC value). The de-compressor must put a SN in
the feedback packet, but this SN is (maybe) not valid.

The REJECT option is for the case, that the de-compressor could not
handle the packet (at this moment), but maybe later, I think. Low
memory could cause it. The de-compressor receives a IR packet with a
new CID, but has not enough memory to allocate a new context, for what
reason ever.

br
Klaus
> _______________________________________________
> Rohc mailing list
> [email protected]
> https://www.ietf.org/mailman/listinfo/rohc