Mark -<div><br></div><div> Thanks for sharing your thoughts, they are definitely helping to put the pieces of this puzzle together. Today I spent most of the day mapping out each office via the net and found the common backbone interconnects. At these x-connects I found 2 data centers. All offices are 30-40ms from one or the other and both DC's are 15-20ms from each other. I couldn't figure out what the distance would be from the DC to the provider, though I know the provider is in a major DC and one Province over so it can't be more than 15-20ms across the backbone.</div>
<div><br></div><div> Both DC's offer dedicated servers, so we are going to look into putting one server at each DC and ditch the original regional/national plan for a more conservative and easy to manage plan. I'm confident now that there will be better overall quality going this way.</div>
<div><br></div><div>Now its time to unscramble the mess that is my install notes and document a clean OpenSIPS+Asterisk install before moving further. After that I'm a bit lost though as I know that we need NAT but not sure which solution is best / easiest to work with (RTPproxy, NAThelper, MediaProxy). From what I've read up on each, Nathelper seems to be built into Osips whereas RTPproxy and MediaProxy require a possibly troublesome install vs loading module/adding code. Searching the mailing archives hasn't been enough for me to decide on a winner.</div>
<div><br></div><div> From what it sounds like, you have a lot of experience in the setup that I'm working on building. Out of curiosity, which method do you prefer for resolving far-end NAT issues?</div><div><br></div>
<div><br></div><div>Skyler</div><div><br></div><div> <br><br><div class="gmail_quote">On Mon, Jan 17, 2011 at 1:26 AM, Mark Sayer <span dir="ltr"><<a href="mailto:datapipes@avtb.co.nz">datapipes@avtb.co.nz</a>></span> wrote:<br>
<blockquote class="gmail_quote" style="margin:0 0 0 .8ex;border-left:1px #ccc solid;padding-left:1ex;">Skyler -<br>
<br>
We are a South Pacific regional provider of hosted PBX services so I<br>
may be prejudiced toward a like infrastructure. Some of our customers<br>
are 3000kms from our servers but the ping times are still less than<br>
50ms so I'm curious why yours are so long. That said, 200ms is sort of<br>
the magic number you don't want to exceed. (Having said that, we do<br>
get some pretty decent call quality connecting to some terminators who<br>
are over 250ms away. 50+250 and its still OK.) Call quality is 99%<br>
Internet connection. OpenSIPS + Asterisk works perfectly with every<br>
call but if the Internet (which you can't control) plays up you get<br>
flack for providing a bad service.<br>
<br>
I'd recommend spending some time looking at your Internet connections.<br>
Can you get them all from the same provider? (I don't even know what<br>
sort of connections you are talking about. We actually get business<br>
grade voice quality from ADSL over copper.) Can you locate your server<br>
in a data center that has good connections to both your ISP and your<br>
terminator? My dream has always been to have a large rack of equipment<br>
in the back office but to make our service work I've had to locate in<br>
a major data centre hundreds of kms away. Our office isn't nearly as<br>
impressive as our service is but that's what the customers pay for.<br>
<br>
I'd only put servers in the offices if there was some reason that<br>
functionality was needed there. Even if you need a receptionist at<br>
each office that can all be handled from a single Asterisk box.<br>
<br>
Just more thoughts.<br>
<font color="#888888">Mark<br>
</font><div><div></div><div class="h5"><br>
On Mon, Jan 17, 2011 at 6:11 PM, Skyler <<a href="mailto:skchopperguy@gmail.com">skchopperguy@gmail.com</a>> wrote:<br>
> Hi Mark,<br>
> Thanks for the reply. So if I understand correctly, I am thinking too big.<br>
> K.I.S.S as some say.<br>
> The existing PBX's are extremely old, so breakdowns & phones are a problem<br>
> and we don't want to repair anymore. In the suggested scenario would you<br>
> recommend replacing the existing hardware (as they breakdown) with IP phones<br>
> and Asterisk at each office then or just ditch the Asterisk and have all the<br>
> phones register to OpenSIPS directly at HQ? My concern is call quality with<br>
> 110ms to HQ then 75ms to provider = 185ms from furthest office, is this<br>
> still not an issue?<br>
> Thanks,<br>
> Skyler<br>
><br>
> On Sat, Jan 15, 2011 at 4:55 PM, Mark Sayer <<a href="mailto:datapipes@avtb.co.nz">datapipes@avtb.co.nz</a>> wrote:<br>
>><br>
>> Here is one suggestion:<br>
>> - single OpenSIPS & Asterisk at central office<br>
>> - use Asterisk as gateway to PSTN (for all offices)<br>
>> - connect remote office PBXs to central office using using multi-port<br>
>> FXS gateways<br>
>> - 110ms is no problem<br>
>> - single system admin point, single cpu, 200 or more concurrent calls<br>
>> - no admin, low cost at remote offices<br>
>><br>
>> Mark<br>
>><br>
><br>
</div></div><div><div></div><div class="h5">> _______________________________________________<br>
> Users mailing list<br>
> <a href="mailto:Users@lists.opensips.org">Users@lists.opensips.org</a><br>
> <a href="http://lists.opensips.org/cgi-bin/mailman/listinfo/users" target="_blank">http://lists.opensips.org/cgi-bin/mailman/listinfo/users</a><br>
><br>
><br>
<br>
_______________________________________________<br>
Users mailing list<br>
<a href="mailto:Users@lists.opensips.org">Users@lists.opensips.org</a><br>
<a href="http://lists.opensips.org/cgi-bin/mailman/listinfo/users" target="_blank">http://lists.opensips.org/cgi-bin/mailman/listinfo/users</a><br>
</div></div></blockquote></div><br></div>