Re: IXP

[Matthew Moyle-Croft <mmc () internode com au>]

Arnold Nipper wrote:
> On 17.04.2009 20:52 Paul Vixie wrote
>   
> Large IXP have >300 customers. You would need up to 45k vlan tags,
> wouldn't you?
>   
Not agreeing or disagreeing with this as a concept, but I'd imagine that 
since a number of vendors support arbitrary vlan rewrite on ports that 
in simple environment you could do some evil things with that.  (ie.  
you could use QinQ "like" ATM Virtual Paths between core switches and 
then reuse the VLAN tag as a VC).  Then, as long as no peer has more 
than 4096 peers you're sweet. It'd hurt your head and probably never 
work, but heck, there's a concept to argue about.  (Please note: I don't 
endorse this as an idea).

I guess the other option is to use MPLS xconnect style or, heck, most 
vendors have gear that'll allow you to do Layer 3 at the same speed as 
Layer 2, so you could go for routing everyone to a common routing core 
with either EBGP multihop or MLPA with communities to control route 
entry and exit.  Then broadcast isn't an issue and multicast would kind 
of be okay.  (Please note: I don't endorse this as an idea).

None of these options, to be honest, are nice and all more complex than 
just a Layer2 network with some protocol filtering and rate limiting at 
the edge.  So, it's not clear what more complex arrangements would fix.

My feeling is that IXes are just a substitute for PNIs anyway, so 
peering does naturally migrate that way as the flow get larger.  If 
you're an IX as a business then this may afront you, but more 
IXes-as-a-business are Colo people (eg. S&D, Equinix) who make good 
money on xconnects anyway.  Or you have a business model that means you 
accept this happens.   Clearly, given the number of 10Gbps ports on some 
exchanges it's not that much of an issue.

MMC