The MPLS WG Archive

[Curtis Villamizar <curtis@fictitious.org>]

In message <4B6D09F3B826D411A67300D0B706EFDE0115C837@nt-exch-yow.pmc-sierra.bc.
ca>, Shahram Davari writes:
> >A PID is not non-standard and proprietary if it is standardized and
> >that is what we are discussing.
> 
> That is exactly what I said. You want to standardize a proprietary PID that
> uses the first 4 bits. If it is approved by IETF then off course it won't be
> proprietary any more.
> 
> >
> >> Do you think that all hashing key selection algorithms by 
> >all vendors should 
> >> be supported
> >> by all IETF standards? What if one decides to use CRC to 
> >distinguish IP packe
> >> ts? should IETF
> >> standards be backward compatible with that too?
> >> 
> >> -Shahram
> >
> >RFC2991 is fine as is.  Load balancing algorithms exist.  There is no
> >need to standardize them.  Some are based on hashing.
> >
> 
> You didn't answer my question. I asked what if a vendor wants to use CRC as I
> P PID. 
> Should IETF protocols support them too by requiring that all other protocols 
> that
> get transported over MPLS not to produce a valid CRC?
> 
> BTW I never said we need to standardize load balancing algorithms. Where did 
> you get
> that from?
> 
> -Shahram


I don't feel obligated to answer nonsense questions.  No one has
proposed putting a CRC in the PID.

Curtis

ps - "requiring [..] not to produce a valid CRC" doesn't even parse in
english so your question is complete nonsense.