Tag Archive: Asterisk

QoS overview

The most helpful overviews I’ve found to date on QoS for VoIP:


http://www.networkworld.com/community/node/37507 (You can browse through the preceding several posts until you arrive at the initial discussion on the QoS byte. Interesting reading.)

The key is to tag your VoIP traffic as DSCP class EF (Expedited Forwarding; the numeric value for EF is 46, binary is 101110(00), ToS 184).

The DSCP classes were created by Cisco and are the default markings applied to VoIP traffic. They are the evolution of the IETF’s ToS byte, which allowed the first bits to be used to specify a rough QoS (called “IP Precendence”). The other bits were never used, so Cisco used several of them to create a more granular set of QoS markings. EF 46 is one of them. EF 46’s first bits indicate IP Precedence 5, which is the class for voice traffic on the older IP Precedence system. So, if you mark traffic as DSCP EF, you’re also marking it as IP Precedence 5. Convenient!

Here’s a helpful tool to assess line quality:

http://myspeed.visualware.com/servers/namerica/iad.php?testtype=-1&codebase=mcsiad.visualware.com&location=USA: Dulles, Virginia&ver=9&cm=&map=namerica&lines=&pps=50&bpp=80&codec=G.726 (32 Kbps)&provtext=Visualware&provtextextra=&provlink=http://www.visualware.com

Better Asterisk voice prompts

To switch out the default system recordings in Asterisk (I’m using Tribox), download the Voice Vector replacement files from http://www.voicevector.com/Downloads.php?PHPSESSID=ca76ce14c2a011256980418e7b1b7f05, extract them (e.g., with 7zip on Windows), connect to the server using WinSCP, rename the /var/lib/asterisk/sounds folder to sounds.old and upload the extracted sounds folder to the same place. Asterisk will pull the better recordings instead.

One especially nice feature is that the recorded attendant says “Asterisk Mail” instead of “Comedian Mail”.