[151886] in North American Network Operators' Group
Re: airFiber
daemon@ATHENA.MIT.EDU (Josh Baird)
Mon Apr 2 09:45:17 2012
In-Reply-To: <218AB54691EB49439829EFD136F473CF27752437@exchange2k10.corp.power1.com>
Date: Mon, 2 Apr 2012 09:44:06 -0400
From: Josh Baird <joshbaird@gmail.com>
To: Dylan Bouterse <dylan@corp.power1.com>
Cc: NANOG list <nanog@nanog.org>
Errors-To: nanog-bounces+nanog.discuss=bloom-picayune.mit.edu@nanog.org
I was told to expect 0.1ms by UBNT. Haven't seen this published, though.
Josh
On Mon, Apr 2, 2012 at 9:38 AM, Dylan Bouterse <dylan@corp.power1.com> wrot=
e:
> What published specs have you seen on the airFiber latency? I asked one o=
f the UBNT guys and they said it's microsecond. On any network I've managed=
, anything sub 1ms is acceptable.
>
> Dylan
>
> -----Original Message-----
> From: John van Oppen [mailto:jvanoppen@spectrumnet.us]
> Sent: Saturday, March 31, 2012 2:22 PM
> To: 'Andrew McConachie'; Marshall Eubanks
> Cc: NANOG list
> Subject: RE: airFiber
>
> We actually have a lot of the old gigabeam radios in service, they are fa=
ster than the published specs of the airfiber links (1G full duplex vs 750 =
mbit/sec fd) and lower latency due to their very simplistic design. =A0 =A0=
To be honest, from a network engineering standpoint, the gigabeams were co=
nveninet as path issues would show up as ethernet errors that can be used t=
o trigger reroutes or other events. =A0 =A0That being said, we did not have=
a large variety of switches as the microwave side of our house is made up =
entirely of just a couple of cisco models. =A0 =A0The gigabeams also have a=
pure OOB management setup.
>
>
> John
>
>
