[Asterisk-Dev] A crazy idea... Skype channel in Asterisk

[Chris Lee]

Benjamin on Asterisk Mailing Lists wrote:
> On Tue, 19 Oct 2004 13:10:27 +0100, Chris Lee
> <cslee-list at cybericom.co.uk> wrote:
> 
> 
>>Where Skype wins customers is ease of use;
> 
> 
> I have to disagree there.
> 
> Where Skype wins customers is hype. SKYPE = HYPE.
> 
> Also, don't confuse the protocol with the client application. It would
> seem that you are talking about the client application when you say
> "ease of use". However, Firefly or Apple's iChat are just as much ease
> of use as the Skype client is.
> 
> 
>>IAX still needs a centrally administered server - Asterisk.
> 
> 
> You've got it upside down. You can call peer-to-peer with IAX. You
> cannot with Skype. Skype needs Skype servers (hardcoded "supernodes"
> in the Skype code). IAX needs no such thing.
> 
> 
>>If "IAX over peer-to-peer" and "Client plugins" for the popular IM
>>products out there were to be produced;
> 
> 
> Well, I agree on the plug-ins. It would be very helpful if there was a
> helper that could be called by a web browser so people could put an
> "IAX address" on their website, something like a SIP URL along the
> lines of
> 
> iax:fred at flintstones.com
> 
> then when you click on that, the helper would do either of two things:
> 
> - launch your IAX softphone and instruct it to dial Fred at
> Flintstones using IAX
> 
> - talk to your Asterisk server and initiate the call, then to be
> bridged to your phone.
> 
> depending on how the preferences of the helper have been defined.
> 
> 
> Someone else mentioned interoperability with Gnomemeeting. Since
> Gnomemeeting is open source, it would be possible to create a
> plugin/module for it that teaches it how to speak IAX.
> 
> As far as the other services are concerned, we only have to make
> ourselves heard. I said this before: we are not a small minority
> anymore. There are quite a few thousand folks on the Asterisk mailing
> lists. We do have a voice. All we have to do is use it.
> 
> Of course there is little point asking Microsoft to support IAX in MSN
> messenger, but how about Yahoo? How about AOL and Apple? If we all
> send those guys email asking for IAX support, they won't jump at it
> right away, but it won't be ignored either. Somebody will be tasked
> with finding out what "this IAX thing" is all about and then one thing
> could lead to another.
> 
> That's why I say, let's not waste our time with Skype; Instead let's
> put the effort into lobbying for IAX. When was the last time you told
> your phone supplier that what you *really* want is an IAX phone?! When
> was the last time you sent an email to Yahoo, AOL or Apple to tell
> them that what you *really* want is a chat client with IAX support?!
> So, why don't you invest five minutes of your time and do that right
> now?
> 
> rgds
> benjk
> 

Ok, all fair comments, though Hype may be a substantial part of why 
skype is doing well it is also not true to say it is not ease of use.

A little history here;
I first was told about Skype by a friend of mine in South Africa, the 
two of us gave up on other products because between S.A. and the UK most 
VoIP products (IMs and the like) just did not cut it, usually extreme 
delay and loss of voice data.
So I set up skype and it was as good as my normal telephone, I was amazed.
I then got my parents and a few others on skype and all was well. 
(Though sometimes, very seldom, the voice was all corrupted)

I then discovered Asterisk and IAX and saw a superior technology (being 
open and very useful) and got a soft phone and set my parents up as an 
extension of my soon to be created home telephone system.
IAX was great for my parents as an extension but to set up my friends 
etc. as individual extensions was getting to be a bit too much admin, 
changing IP addresses and the like. (had I not been the type to dabble 
my parents would never have been able to use an IAX client to call peer 
to peer to an IAX client on my system, there were just too many things 
they needed to know to get a link up.)

So I have a little experience on the ease of use and the features.
Skype except the occasional scrambling of voice was very good at call 
quality, IAX is also very good but it does drop the call far more often 
than skype, this I can only assume is because of the multiple concurrent 
links skype uses.
I have not worked out which is less latent as they both vary during 
use.( I have not tried to use skype locally on my network, skype is no 
longer installed on any of my machines)

So my point was that people find the "supernode" provided features very 
useful, they make life easy.
I admit I mixed application and protocol a bit, but was thinking of the 
following situations:

1. I set up skype and can call any named user, and any user can call me 
using my name. so no more concerns with configuration. Happy days!

2. I set up an IAX client, I now have to find out how to receive calls 
through the NAT that is hiding me away, I can not get the NAT 
re-configured (If I even know what NAT is), so I am stuck. (solution 
start a session with a peer outside the NAT who will answer calls for me 
and use my session to send them to me - how do I find said peer with 
valid IP address?, need some kind of fabric to ask)

3.In the scenario that I do have a fixed IP address or a re-directable 
port (not too many people behind the firewall/router or we would have a 
major admin job) I still have to find a way to let people map my "Name" 
to an IAX address, This probably requires DNS configuration and well I 
am a layman so have no idea how you get that working, what is DNS I 
thought it was a set of 4 numbers? etc.

so what I was suggesting was a peer-to-peer fabric that can help with 
the membership, naming and call receipt issues so that IAX could pick up 
some of the ease of use provided by the supernodes.

Another advantage of each IAX-peer client having a "peer server" is that 
I can receive multiple path options, direct or routed via "IAX nodes" 
which have a calculated "better path" than direct, and possibly find 
some advantage of using two slow paths to make a quicker one.

These are the thought behind my submission.

Chris.