[Asterisk-Users] G.729 fallback
Chris A. Icide
chris at netgeeks.net
Wed Jun 2 00:35:49 MST 2004
Well here is my example.
I have a client, who has lots of work associates who call in from all over
the world to conference calls. For these calls, many of them use cell
phones because of local telco issues. This company then pays the cell
bills for these call ins. The bills are astronomical. They want to host
the conference calls via Asterisk, and want people who have any access to
IP to call in via VoIP. In many of these locations, these folks have
dial-up and need a codec with a very low bandwidth usage. Hello G729.
But whether or not you use G729 or any other codec doesn't negate the need
for a fail-over method for any codec that is limited by a license usage or
any other limit. The functionality needs to be there.
When I turned 16 and got my drivers license, I bought a car. That was back
before the japanese were importing cars in any numbers, and the american
built cars were not of the best quality, but hey, it didn't matter to me
because I really didn't know any better. I bought a car with an engine,
and 4 wheels which got me from place to place. It even had a radio with 4
speakers (self installed), A/C and heat. It also had alot of rattles,
cheap plastic that didn't fit perfectly, the doors had to be lifted a bit
when closing or they wouldn't latch, etc.
Today I have a car, of VERY nice quality, it's still an american mfg car,
but it's no better or no worse now that it's european and japanese
counterparts. If you came to me and offered me a car and said, but wait,
it's cheap and very inexpensive because it doesn't have a CD player (so we
don't have to pay fees for the license to that technology) and it doesn't
have anti-lock brakes (again no need to pay patent fees on that) and no
traction control, or automatic climate control, or cruise control, or
independent suspension..... BUT it's cheap and it gets you from point A to
point B.
Would I buy it? No, because my expectations for a vehicle have been set
beyond this.
Every once in a while someone on this list comes out and says something to
the fact of "just do this workaround". In many cases, they are correct,
and in a beta test environment, I fully understand the reasoning. However,
whether it's wise or not, people are coming up to the point where they need
to install, expand, or replace their current PBX systems and they can
either choose to go pay the nortel's of the PBX world lots of money, or
they can take the iPBX plunge. If Asterisk truly wants to play in the iPBX
world, then it MUST support the same features that are coming out in the
big players iPBX systems. Telephone is HUGE to almost every business. It
must work, and it must be able to perform in the manner they want it to
perform.
My client has said, "We are going to use G729 for our remote clients to
save bandwidth, and if we ever run out of licences, it need to complete the
call with another available codec". At this point, my choice is to either
make asterisk work as the spec requires, or install a different iPBX
system. I can't go back and say don't use G729, use GSM instead, because
some of their clients won't have the BW to use anything but G729.
So, to put Asterisk to work for this client, I really need this
functionality, and I suppose if it doesn't come out of the community, I can
hire a programmer (I'm too dumb to be a programmer) to do it for me,
because it will still be cheaper than taking the client down the nortel,
etc. road. But this particular instance set aside, I come from a formally
trained engineering background with quite a few years in a very stringent
engineering field (in other words, if something is poorly engineered people
died), and one of the basic tenants is that a well engineered system was
able to able to operate and function in any situation you could expect to
see under nominal conditions. This G729 codec failover is something that
seems to me to be a possible occurrence, and under nominal operating
conditions.
I recently spent some time chatting with John Todd about another feature
that fits in with this, and that is a bandwidth manager. In other words,
you set a maximum bandwidth allowed, and then asterisk will limit incoming
and outgoing calls that would overrun that limit. This failover system
would apply then in a situation where perhaps normally you might set up a
G711 connection, but that would overrun the limit, so instead, you fail
over to GSM...
I guess what I'm trying to say is that the function is needed and in the
end will probably be used for MANY things, but right now, the G729 license
limit is a strong candidate. Over-engineering is generally significantly
better than under-engineering.
Just my thoughts on the matter.
-Chris
On 05:19 PM 6/1/2004, Kevin Walsh wrote:
>Chris A. Icide [chris at netgeeks.net] wrote:
>> On 08:53 AM 6/1/2004, Kevin Walsh wrote:
>> > Mike Heininger [mike at heininger.at] wrote:
>> > > It's a pity ... it would be great to fallback to another (free)
>> > > codec.
>> > >
>> > Just use a relatively-free codec (iLBC or GSM etc.) in the first
>> > place, and avoid G.729. That strategy works for me. :-)
>> >
>> While in some cases, just not using G729 may be a valid option, there
>> will ALWAYS be exceptions to that.
>>
>I must be the lucky one then, as I have found no need for G.729. :-)
>
>If I find a supplier that only supports G.729 then I'll simply not
>do business with them. Giving in only encourages them.
>
>Perhaps, if enough people stopped using these closed "standards",
>suppliers would start to support the alternatives, and take up of
>VoIP would no longer be held to ransom by greedy monopolists.
>
>Just a thought.
>
>--
> _/ _/ _/_/_/_/ _/ _/ _/_/_/ _/ _/
> _/_/_/ _/_/ _/ _/ _/ _/_/ _/ K e v i n W a l s h
> _/ _/ _/ _/ _/ _/ _/ _/_/ kevin at cursor.biz
>_/ _/ _/_/_/_/ _/ _/_/_/ _/ _/
>
>_______________________________________________
>Asterisk-Users mailing list
>Asterisk-Users at lists.digium.com
>http://lists.digium.com/mailman/listinfo/asterisk-users
>To UNSUBSCRIBE or update options visit:
> http://lists.digium.com/mailman/listinfo/asterisk-users
More information about the asterisk-users
mailing list