I have a quite strongly held belief that trying to be mathematically precise (not to say pedantic) has never worked well in IETF protocols. (OK, there is one possible exception, namely NTP.) Our implementors are not going to understand subtleties such as 'low discrepancy'. IMHO, arguing about the correct mathematical formulation is not quite the point.
We do seem to agree that the important point is that the final output of the hash function used to select between alternate routes is uniformly distributed. Asking for flow label values that are reasonably well distributed as input to that hash function is enough. Any stateless method that provides this will do. Frankly I don't care whether we all it pseudo-random, quasi-random, or mashed potato. (Counting from 1 will not do, because it's stateful.) Regards Brian Carpenter On 2011-04-06 13:34, Hemant Singh (shemant) wrote: > -----Original Message----- > From: ipv6-boun...@ietf.org [mailto:ipv6-boun...@ietf.org] On Behalf Of > Thomas Narten > Sent: Tuesday, April 05, 2011 8:08 PM > To: james woodyatt > Cc: 6MAN Working Group > Subject: Re: Pseudorandom Flow Labels > > >> What is *required* is that the hash function (or whatever function >> that is used) on the router maps the tuples in a *uniform* way across >> the range of possible outputs. > >> If you have 10 links, and all your Flow Labels are clustered around >> low ten values, but in an approximately uniform way, a simple modulo >> hash will get you the kind of distribution you need. > >> The range of values of the flow label itself does not need to be >> uniformly distributed. > >> That is why incrementing a counter for each new flow provides adequate >> properties. I remain unconvinced that we need to require that they >> exhibit pseudo randomness, or even be uniform. > >> If that is a requirement, I'd like to see the justification, not just >> hand waiving. > > How about something like the text below: > > A property of the flow label is that when the flow label is used a key > to a hash function the function produces a uniform distribution of hash > values. Thus, incrementing a counter for each new flow is one means to > achieve the property described in the previous sentence. > > Hemant > -------------------------------------------------------------------- > IETF IPv6 working group mailing list > ipv6@ietf.org > Administrative Requests: https://www.ietf.org/mailman/listinfo/ipv6 > -------------------------------------------------------------------- > -------------------------------------------------------------------- IETF IPv6 working group mailing list ipv6@ietf.org Administrative Requests: https://www.ietf.org/mailman/listinfo/ipv6 --------------------------------------------------------------------
