Print

Print


Note that this about who can be a member, not who is a member. There's
the separate step of electing or appointing.

On Fri, 2015-03-27 at 20:40 +0000, Justin Krejci wrote: 

> I would suggest an amendment to your 1A and 1B proposals
> section (ii) changed from
> an operator of equipment providing approved indirect connections.
> to
> an operator of equipment providing approved indirect connections with
> at least one active member connection


I considered this. I originally went so far as to write "two or more".
But in the end, I took it out as I don't think this matters. The Board
shouldn't approve a switch that's not going to have any members. If they
do and it turns out nobody connects, they can just de-approve it after a
while. Also, they can hold up approving the switch operator's membership
(separate from "approving the switch") until such time as there are
connections behind the switch.

If we were to adopt your proposal, I'd like to see it amended by
striking the word "member". We should not assume that all connected
networks are necessarily members.


> Depending on how your legalese translates, this could cover a fiber
> provider not otherwise associated with the exchange, eg a provider who
> leases wave service or dark fiber to connect the 511 fabric to a
> remote colo switch. With proposal 1A you could have both the switch
> operator and the fiber owner as members. With proposal 1B you could
> have the switch operator (unless it's an official MICE switch), the
> fiber owner, and the colo facility all of which could be separate
> companies.


I considered this possibility. But, even if the fiber owner tries to
claim they're eligible for membership, the Board/members need not
appoint/elect them.


> Also, are we afraid of using the term "switch" instead of the generic
> term "equipment?" Are there situations where there is member
> connectivity that does not involve the use of a switch specifically?
> Are just preparing in case ATM makes a move to over? 53 bytes FTW!


I believe CNS still connects remote participants. (Halstad Telphone
Company is the example now. Wikstrom Telephone Company was connected
this way in the past.) This is done by extending the MICE VLAN through
the CNS network out to the remote participants. It is not the case (at
least it wasn't for us) that the remote participants have purchased a
layer 2 circuit to 511 that they can cross-connect to arbitrary
destinations. So it's not that they have a presence in 511 and are
connected to a single, identifiable "remote switch".

Granted, CNS is directly connected, and also runs an identifiable remote
switch, so this is moot for them. And it may not be likely to occur with
another network. But, for another example that might be more likely....
Imagine a carrier neutral data center operator carries the MICE VLAN
through their switching fabric, whether that's in one building or more
than one around the metro area. They might very well not run an IP
network that connects to MICE.

-- 
Richard