On Fri, Sep 28, 2018, at 8:32 AM, Luigi Iannone wrote:
> Hi Alexey,
> 
> > On 27 Sep 2018, at 19:31, Dino Farinacci <[email protected]> wrote:
> > 
> >> Ok, maybe this is just me, but you don't actually define how to hash these
> >> things, you are only talking about what needs to be covered by the hash. I
> >> appreciate that picking a specific hashing algorithm is probably not 
> >> relevant
> >> for interoperability, but I feel adding a specific algorithm (as an 
> >> example or
> >> via a pointer) would improve the document.
> > 
> > We decided to leave this to the implementation and is a local matter ot the 
> > encapsulator. Most implementations use a “folded XOR”. Which means XOR the 
> > set of 32-bit fields from the 5-tuple hash, then XOR the 2 16-bit 
> > quantities, then XOR the 2 bytes. Mod the number of best priority RLOCs, to 
> > give you an index to choose one.
> 
> While you are right that what is in the document is just what can be 
> covered by the hash,  
> I agree with Dino on this point.
> 
> I do not think that we need a specific algorithm even as an example. 
> The load-sharing is local to the ITR, it just need to use any algorithm 
> that does that.
> 
> Would you prefer a clear statement?

Yes.

> Something like:
> 
> “The specific algorithm the ITR uses for load-sharing is out of the 
> scope of this document and does not prevent interoperability" 

Sounds good to me. Thank you.

> Ciao
> 
> L.
> 
> 
> > 
> > Dino
> > 
> 

_______________________________________________
lisp mailing list
[email protected]
https://www.ietf.org/mailman/listinfo/lisp

Reply via email to