hotel networks, was One of our own in the Guardian.

Sun Jul 14 17:32:33 UTC 2013

On 7/14/13 7:22 AM, John Levine wrote:
>> I suspect the problem is the (offsite) hotel that Mark and I are at was not
>> really prepared for a full house of folks interested in viewing streams,
>> downloading documents, etc. (despite attempts to inform the hotel of the
>> impending tsunami). I imagine folks involved in setting up NANOG-related
>> networks might be familiar with this sort of situation...
> I've talked to people who do conference arrangements, and no matter
> what you tell the hotel, the hotel talks to their outsourced Internet
> provider, who tells them it will be fine, which of course it will not
> be.  The hotel outsourcers also tend to have poorly trained staff who
> think that the way to increase wifi capacity is to turn the power on
> all of the APs up to 11.
Simply put they were'nt designed and built to be operated with 100% 
concurrency. Short of some kind of exceptional contractual arrangement 
you shouldn't expect them to be different when you arrive then when the 
facility was contracted.
> The IETF deals with this problem by writing into the conference
> agreement that their netops people will take over the hotel's network
> for the duration of the meeting, and bring in their own adequate
> backhaul.  Dunno what ICANN does.
Building a network for a week is expensive.  it's gotten a lot simpler 
and cheaper but it's still relatively extrodinary. Taking over existing 
infrastructure operating it and putting it back is a new challenge 
everytime.
>