[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]
Re: Draft AUP
good progress! some whining, of course :-)
>>> 19. Members may not directly connect customers who are not UIXP Members
>>> via circuits to their router housed in any UIXP rack.
>> i.e. they may not put in private peering inter-member interconnects? why
>> not?
> Not that actually. But imagine I am a customer of ISP A and I am probably
> nearer to the XP than to ISP A's premises. An ISP may be tempted to
> instead connect me to their router at the IXP instead of all the way to
> their premises. But otherwise inter-member connects should be okay.
i imagine that this is because of some scarce resource at the ixp. if so,
then maybe add a "because" clause so folk can understand the physics.
>>> 20. Members should not routinely use the UIXP for carrying traffic between
>>> their own routers.
>> if i have two routers on the mesh (for redundancy i choose to pay for two),
>> then 50% of the traffic from a peer (unless meds are used) will go to the
>> 'wrong' router. so 50% of my inbound traffic will be routinely shifted
>> between two of my own routers.
> That is a valid case. But imagine ISP A has two POPs in X and Y, and it
> has a direct link between these two POPs but each of them is also
> connected to the IXP. It would not be nice for say when their link between
> the POPs dies for the ISP to decide to route their own internal traffic
> through the IXP. i.e traffic from X destined for Y
yup. another seemingly valid use. so what scarce resource is being
preserved by proscribing it?
>> i am not a big fan of technology by democracy.
> True. But it is the only way of ensuring that everyone is happy with
> decisions made.
in the very short term, while life is meetings as opposed to moving packets.
when it comes to moving packets, people are most happy by well-working
technology. is the goal meetings or moving packets. i suggest optimizing
the design for the latter.
> We wouldn't like a member to come up and say that the decision making
> process is dominated by only a few and that their views are not heard.
in my culture, it is guarranteed that someone will do that. when that
happens, i would much rather be able to justify the decisions on sound
technical principles than a vote taken back in 1948.
randy
--
This is the techies mailing list. The list archives can be found at http://uixp.co.ug/archives .