More on Vonage service disruptions...
bonomi at mail.r-bonomi.com
Fri Mar 4 19:04:41 UTC 2005
> From owner-nanog at merit.edu Fri Mar 4 11:44:17 2005
> From: Christopher Woodfield <rekoil at semihuman.com>
> Subject: Re: More on Vonage service disruptions...
> Date: Fri, 4 Mar 2005 12:45:54 -0500
> On Mar 3, 2005, at 10:25 PM, Scott Morris wrote:
> > Perhaps it varies by state, but I thought part of the E-911 service
> > regulations was that if you were offering (charging) for it, you had to
> > offer it as "lifeline" service which meant it had to survive power
> > outage.
> > *shrug*
> > I guess the original regs weren't written with these things in mind!
> This does bring up a hardware design question...I'm wondering how
> difficult of an engineering/marketing problem it would be to design
> VoIP adapters with built-in backup batteries.
Relatively trivial. Or you could just plug it into a UPS.
> How does the power
> consumption profile of a VoIP adapter compare to, say, a cellphone?
Look at the typical "wall wart" that powers it. Draw is "a few watts".
Compare this to the battery rating for a cell-phone. (a few amp-hours, at
a few volts -- run-times in the *multiple* "tens of hours")
Cell-phone is "order of magnitude", at least, lower in power consumption.
VOIP adapters _could_ be designed (at higher cost) to consume considerably
less power, but present-day cost/benefit analysis dictates against doing so.
> What would this add to the cost of the device,
Meaningless question, as stated. Something on the order of "10 seconds" can
be handled by simply adding a SuperCap -- cost measured in pennies. OTOH,
to ride out a 72-hour power-failure probably adds $100 or more.
> and how long could the
> battery last?
How much battery are you willing to pay for? <grin>
However, it isn't _only_ the VOIP adapter that has to have back-up power,
*EVERY* piece of gear 'upstream' of it has to be powered, too.
This includes, at a minimum the following CPE:
the VOIP adapter,
any hub/switch/routers in the circuit,
any dedicated "firewall" hardware,
and the DSL/cable "modem" CPE, itself.
"Cable" connections may have in-line amplifiers/repeaters between the
customer location and the "head-end" -- gotta have back-up for the power
feed (going down the cable, itself) for those devices.
*PLUS* back-up power for the 'head end' gear -- DSLAM at the C.O., or cable
the concentrator. -- as well as the routers that switch the traffic to the
Now, the CPE can be simply plugged into a stand-by UPS with reasonable
battery capacity, and you'll get as much run-time as you're willing to
pay for. <grin> Assuming you've only got a VOIP adapter, and a cable/DSL
modem to deal with -- call it 20 watts total -- then a UPS rated for
1000 watts, with a run-time of 10 minutes at full load, will probably run
at least 10-15 _hours_ powering only the VOIP adapter and cable/DSL modem.
Of course, having the CPE powered is _worthless_ if the immediate upstream
gear has lost power.
I can't see how Vonage, or *any* other "3rd party" (i.e. anybody other than
the provider of the gear for the infamous 'last mile' connectivitiy) can
_hope_ to claim service 'survivability' in the event of a power failure.
Considering that the "Packet8"s, etc. have *no* contract whatsoever with that
last-mile provider, they can't even 'specify requirements'.
They can "encourage" the customer to put the CPE on UPS, but that's all.
Getting somewhat "far afield" from 'network operations', but this issue
also arises with regard to "real telephone service" offered by the CABLE TV
companies. All too frequently _this_ service relies on CPE that draws
'utility' power. It is frequently the case that they fail to mention to
the customer the "desirability" of putting _that_ equipment on UPS. With
the resultant loss of dial-tone in the event of _power_ failure.
More information about the NANOG