<!DOCTYPE HTML PUBLIC "-//W3C//DTD HTML 3.2//EN">
<HTML>
<HEAD>
<META HTTP-EQUIV="Content-Type" CONTENT="text/html; charset=iso-8859-1">
<META NAME="Generator" CONTENT="MS Exchange Server version 5.5.2448.0">
<TITLE>RE: [Asterisk-Dev] Petition for IAX firmware</TITLE>
</HEAD>
<BODY>
<P><FONT SIZE=2>Hi,</FONT>
</P>
<P><FONT SIZE=2>Just find out this thread. My question is. If companies like Sipura already makes SIP (which can work with *), why do they want to add support for IAX. After all, that's extra resource they have to spend to get that done. I have no doubt what you guys talking will work; but does adding IAX support makes business sense for them?</FONT></P>
<P><FONT SIZE=2>-----Original Message-----</FONT>
<BR><FONT SIZE=2>From: Steve Kann [<A HREF="mailto:stevek@stevek.com">mailto:stevek@stevek.com</A>]</FONT>
<BR><FONT SIZE=2>Sent: Wednesday, April 06, 2005 9:49 AM</FONT>
<BR><FONT SIZE=2>To: Asterisk Developers Mailing List</FONT>
<BR><FONT SIZE=2>Subject: Re: [Asterisk-Dev] Petition for IAX firmware</FONT>
</P>
<BR>
<P><FONT SIZE=2>Paul wrote:</FONT>
</P>
<P><FONT SIZE=2>> Steve Kann wrote:</FONT>
<BR><FONT SIZE=2>></FONT>
<BR><FONT SIZE=2>>></FONT>
<BR><FONT SIZE=2>>> On Apr 5, 2005, at 10:49 PM, Paul wrote:</FONT>
<BR><FONT SIZE=2>>></FONT>
<BR><FONT SIZE=2>>>> denon wrote:</FONT>
<BR><FONT SIZE=2>>>></FONT>
<BR><FONT SIZE=2>>>>> Hi all,</FONT>
<BR><FONT SIZE=2>>>>></FONT>
<BR><FONT SIZE=2>>>>> I've put together a quick petition, in hopes that we can possibly </FONT>
<BR><FONT SIZE=2>>>>> persuade Sipura (or any other large-scale IP handset manufacturer) </FONT>
<BR><FONT SIZE=2>>>>> to include firmware support for IAX. The IAXy has proven that an </FONT>
<BR><FONT SIZE=2>>>>> IAX product is in demand, and very useful, and I think we'd all </FONT>
<BR><FONT SIZE=2>>>>> like to see a handset manufacturer follow Digium's lead. I'm not </FONT>
<BR><FONT SIZE=2>>>>> particularly endorsing Sipura, however I do know that they have </FONT>
<BR><FONT SIZE=2>>>>> seriously considered support for IAX, and have decided to hold off </FONT>
<BR><FONT SIZE=2>>>>> until "the demand is there". I'm hoping that with some numbers, we </FONT>
<BR><FONT SIZE=2>>>>> can prove to them that the demand is already here, and that IAX is </FONT>
<BR><FONT SIZE=2>>>>> already a viable technology.</FONT>
<BR><FONT SIZE=2>>>>></FONT>
<BR><FONT SIZE=2>>>>> I'd like to encourage everyone to show your support -- hopefully </FONT>
<BR><FONT SIZE=2>>>>> Sipura, and/or other manufacturers will see these hard names and </FONT>
<BR><FONT SIZE=2>>>>> numbers, and realize it's time to move something into production.</FONT>
<BR><FONT SIZE=2>>>>></FONT>
<BR><FONT SIZE=2>>>>> Petition:</FONT>
<BR><FONT SIZE=2>>>>> <A HREF="http://www.petitiononline.com/IAXPhone" TARGET="_blank">http://www.petitiononline.com/IAXPhone</A></FONT>
<BR><FONT SIZE=2>>>>></FONT>
<BR><FONT SIZE=2>>>>> Thanks,</FONT>
<BR><FONT SIZE=2>>>>></FONT>
<BR><FONT SIZE=2>>>>> -d</FONT>
<BR><FONT SIZE=2>>>></FONT>
<BR><FONT SIZE=2>>>></FONT>
<BR><FONT SIZE=2>>>></FONT>
<BR><FONT SIZE=2>>>> I like to see IAX support in any phone or ata. As for the benefits </FONT>
<BR><FONT SIZE=2>>>> of IAX trunking: Suppose that you have 4 SPA-2000 devices at a </FONT>
<BR><FONT SIZE=2>>>> branch office with IAX trunking back to headquarters? You won't be </FONT>
<BR><FONT SIZE=2>>>> getting the best bandwidth usage unless you also run a * server at </FONT>
<BR><FONT SIZE=2>>>> the branch office to aggregate those 8 extensions into one trunk. </FONT>
<BR><FONT SIZE=2>>>> Otherwise you have 4 trunks to the * server at headquarters. With 8 </FONT>
<BR><FONT SIZE=2>>>> IP phones it's even worse.</FONT>
<BR><FONT SIZE=2>>>></FONT>
<BR><FONT SIZE=2>>>> So it looks like I need to have a linux box running at each location </FONT>
<BR><FONT SIZE=2>>>> where I want to use IAX trunking. I'm sure I can get the cost of </FONT>
<BR><FONT SIZE=2>>>> that down to something reasonable. I guess the next question is: It </FONT>
<BR><FONT SIZE=2>>>> is much easier to do IAX trunking if all the ata's and IP phones use </FONT>
<BR><FONT SIZE=2>>>> IAX2 instead of SIP? A good example would be 8 IAXy devices as </FONT>
<BR><FONT SIZE=2>>>> compared to 4 2-port SIP ata's. Is the configuration going ot be </FONT>
<BR><FONT SIZE=2>>>> easier? How about performance and stability?</FONT>
<BR><FONT SIZE=2>>>></FONT>
<BR><FONT SIZE=2>>>> Rather than petition Sipura, I would prefer that we convince Digium </FONT>
<BR><FONT SIZE=2>>>> to go further with the IAXy concept. Give us a base unit that </FONT>
<BR><FONT SIZE=2>>>> accepts a mix of fxo or fxs modules. Offer it in 2, 4 and 8 port </FONT>
<BR><FONT SIZE=2>>>> versions. If those 3 sizes will do trunking and offer the best free </FONT>
<BR><FONT SIZE=2>>>> codecs, it will eliminate the need for a * server to do trunking at </FONT>
<BR><FONT SIZE=2>>>> a lot of locations. Digium should do it first and let the others </FONT>
<BR><FONT SIZE=2>>>> play catch-up when they finally see the light.</FONT>
<BR><FONT SIZE=2>>>></FONT>
<BR><FONT SIZE=2>>></FONT>
<BR><FONT SIZE=2>>> I have two points to make:</FONT>
<BR><FONT SIZE=2>>></FONT>
<BR><FONT SIZE=2>>> 1) If I were a vendor, and there was documentation on the IAX2 </FONT>
<BR><FONT SIZE=2>>> protocol (not even an RFC, but at least some kind of semi-official </FONT>
<BR><FONT SIZE=2>>> documentation), I'd be a lot more likely to implement it.</FONT>
<BR><FONT SIZE=2>>></FONT>
<BR><FONT SIZE=2>>> 2) The idea of an IAX2 trunk-aggregator is interesting. This is </FONT>
<BR><FONT SIZE=2>>> probably something that doesn't need a whole x86 linux box -- taking </FONT>
<BR><FONT SIZE=2>>> multiple IAX2 streams and putting them into trunks is trivial in </FONT>
<BR><FONT SIZE=2>>> terms of computational requirements.. A small microcontroller or an </FONT>
<BR><FONT SIZE=2>>> ARM chip is more than enough for this, and would be a neat idea of </FONT>
<BR><FONT SIZE=2>>> virtual PBX deployments..</FONT>
<BR><FONT SIZE=2>>></FONT>
<BR><FONT SIZE=2>>> -SteveK</FONT>
<BR><FONT SIZE=2>></FONT>
<BR><FONT SIZE=2>></FONT>
<BR><FONT SIZE=2>> There are a few linux-compatible boards being used right now for </FONT>
<BR><FONT SIZE=2>> applications such as wireless ISP's. A trunk-aggregator should not do </FONT>
<BR><FONT SIZE=2>> any transcoding. That would be handled by the * server at the other </FONT>
<BR><FONT SIZE=2>> end of the trunk(if needed). The dial plan is also handled by the </FONT>
<BR><FONT SIZE=2>> "master server". And so on and so on. That means maybe we can do it </FONT>
<BR><FONT SIZE=2>> with a fanless pentium-compatible board that boots from a compact </FONT>
<BR><FONT SIZE=2>> flash card. I have seen fanless products ranging from 133 to 800 mhz </FONT>
<BR><FONT SIZE=2>> cpu speed.</FONT>
</P>
<BR>
<P><FONT SIZE=2>You probably don't even need that. If you code it well, you can easily </FONT>
<BR><FONT SIZE=2>do a whole lot of channels of this kind of thing on a gumstix </FONT>
<BR><FONT SIZE=2>(www.gumstix.com), or even a less powerful chip.</FONT>
</P>
<P><FONT SIZE=2>></FONT>
<BR><FONT SIZE=2>> But I still don't know the answer to my question about SIP vs. IAX2. </FONT>
<BR><FONT SIZE=2>> Suppose the remote site has a mix of SIP and IAX2 devices. Does the </FONT>
<BR><FONT SIZE=2>> presence of SIP devices increase the computational requirements much? </FONT>
</P>
<P><FONT SIZE=2>Computational requirements won't change much, but the code you'll need </FONT>
<BR><FONT SIZE=2>will be a lot larger.</FONT>
</P>
<P><FONT SIZE=2>You can probably write (from scratch), a 500-1000 line C program which </FONT>
<BR><FONT SIZE=2>can aggregate single IAX2 calls from multiple sources into trunks to </FONT>
<BR><FONT SIZE=2>particular destinations. A version which handled SIP as well would need </FONT>
<BR><FONT SIZE=2>to be several times that much code.</FONT>
</P>
<P><FONT SIZE=2>> My thinking is that any SIP ata's or phones at the remote site are </FONT>
<BR><FONT SIZE=2>> going to be extensions of the master * pbx. Hopefully that makes it </FONT>
<BR><FONT SIZE=2>> easier on the trunk-aggregator cpu. Also I expect that in most </FONT>
<BR><FONT SIZE=2>> situations where this was deployed a codec other than g.711 would be </FONT>
<BR><FONT SIZE=2>> used since there is a motivation to conserve bandwidth. So we have SIP </FONT>
<BR><FONT SIZE=2>> traffic from a provider to the master server, IAX2 trunking to the </FONT>
<BR><FONT SIZE=2>> remote slave server and back to SIP over the LAN to somebody's </FONT>
<BR><FONT SIZE=2>> deskset. Will that conversion back to SIP to reach the deskset degrade </FONT>
<BR><FONT SIZE=2>> the call quality?</FONT>
</P>
<BR>
<P><FONT SIZE=2>It shouldn't, unless there's bugs :)</FONT>
</P>
<P><FONT SIZE=2>-SteveK</FONT>
</P>
<BR>
<BR>
<P><FONT SIZE=2>_______________________________________________</FONT>
<BR><FONT SIZE=2>Asterisk-Dev mailing list</FONT>
<BR><FONT SIZE=2>Asterisk-Dev@lists.digium.com</FONT>
<BR><FONT SIZE=2><A HREF="http://lists.digium.com/mailman/listinfo/asterisk-dev" TARGET="_blank">http://lists.digium.com/mailman/listinfo/asterisk-dev</A></FONT>
<BR><FONT SIZE=2>To UNSUBSCRIBE or update options visit:</FONT>
<BR><FONT SIZE=2> <A HREF="http://lists.digium.com/mailman/listinfo/asterisk-dev" TARGET="_blank">http://lists.digium.com/mailman/listinfo/asterisk-dev</A></FONT>
</P>
</BODY>
</HTML>