Re: [asterisk-users] Re: Bandwidth requirements
Benny Amorsen wrote: rJ == raphael Jacquot [EMAIL PROTECTED] writes: rJ ATM cell tax is actually 10% as there's 5 header bytes for each 53 rJ bytes cell, For VoIP the cell tax is much larger. In the example, each RTP packet contains 20 useful bytes and 40 bytes IP overhead. 60 bytes doesn't fit in one cell, so you end up with 106 bytes at the ATM layer to transport 20 bytes of G.729. The ATM-caused overhead is thus 46 bytes per voice packet, thereby making the needed bandwidth 77% larger. CRTP solves this issue (40byte waste) -- J. Oquendo http://pgp.mit.edu:11371/pks/lookup?op=getsearch=0x1383A743 sil . infiltrated @ net http://www.infiltrated.net The happiness of society is the end of government. John Adams ___ --Bandwidth and Colocation provided by Easynews.com -- asterisk-users mailing list To UNSUBSCRIBE or update options visit: http://lists.digium.com/mailman/listinfo/asterisk-users
Re: [asterisk-users] Re: Bandwidth requirements
True but you need to look at the actual speed of a DSL line. a 1.5m/ 384k DSL line is actually 1.7m/467k so a chunk of that ATM cell tax is already factored into the speed you buy from your DSL provider. If you get an integrated DSL modem/router like a Zoom X5 you can see the actual speed of the link. On Oct 5, 2006, at 9:31 AM, Benny Amorsen wrote: rJ == raphael Jacquot [EMAIL PROTECTED] writes: rJ ATM cell tax is actually 10% as there's 5 header bytes for each 53 rJ bytes cell, For VoIP the cell tax is much larger. In the example, each RTP packet contains 20 useful bytes and 40 bytes IP overhead. 60 bytes doesn't fit in one cell, so you end up with 106 bytes at the ATM layer to transport 20 bytes of G.729. The ATM-caused overhead is thus 46 bytes per voice packet, thereby making the needed bandwidth 77% larger. All in all VoIP over ADSL adds 430% overhead, when using G.729 and 20ms packets. Lovely, isn't it? /Benny ___ --Bandwidth and Colocation provided by Easynews.com -- asterisk-users mailing list To UNSUBSCRIBE or update options visit: http://lists.digium.com/mailman/listinfo/asterisk-users -- Matthew S. Crocker Vice President Crocker Communications, Inc. Internet Division PO BOX 710 Greenfield, MA 01302-0710 http://www.crocker.com ___ --Bandwidth and Colocation provided by Easynews.com -- asterisk-users mailing list To UNSUBSCRIBE or update options visit: http://lists.digium.com/mailman/listinfo/asterisk-users
RE: [asterisk-users] Re: Bandwidth requirements
J. Oquendo wrote: Benny Amorsen wrote: rJ == raphael Jacquot [EMAIL PROTECTED] writes: rJ ATM cell tax is actually 10% as there's 5 header bytes for each 53 rJ bytes cell, For VoIP the cell tax is much larger. In the example, each RTP packet contains 20 useful bytes and 40 bytes IP overhead. 60 bytes doesn't fit in one cell, so you end up with 106 bytes at the ATM layer to transport 20 bytes of G.729. The ATM-caused overhead is thus 46 bytes per voice packet, thereby making the needed bandwidth 77% larger. CRTP solves this issue (40byte waste) cRTP is a great idea that is not widely implimented. Not very many endpoint support it, and Asterisk currently does not. In version 1.4, Asterisk will support configurable RTP packetization. So you could use 40ms for your G729 payload and reach a 50/50 split between payload and overhead (before ATM encapsulation for DSL). At 50ms for g729, you get a nice fit into two ATM cells, while a small bump to 60ms will force you into three ATM cells. Of course you need to be mindful of the latency between endpoints when you start increasing the payload, but a healthy network and tuning the payload can greatly reduce bandwidth requirements. AND most endpoints do support configurable RTP packetization. If Asterisk and more endpoints supported cRTP, then combining that feature with larger RTP payloads would result in very low overhead numbers. Dan ___ --Bandwidth and Colocation provided by Easynews.com -- asterisk-users mailing list To UNSUBSCRIBE or update options visit: http://lists.digium.com/mailman/listinfo/asterisk-users
Re: [asterisk-users] Re: Bandwidth requirements
Unfortunately cRTP is not supported by Asterisk, and normally it is implemented between routers that support it. On 10/5/06, Dan Austin [EMAIL PROTECTED] wrote: J. Oquendo wrote: Benny Amorsen wrote: rJ == raphael Jacquot [EMAIL PROTECTED] writes: rJ ATM cell tax is actually 10% as there's 5 header bytes for each 53 rJ bytes cell, For VoIP the cell tax is much larger. In the example, each RTP packet contains 20 useful bytes and 40 bytes IP overhead. 60 bytes doesn't fit in one cell, so you end up with 106 bytes at the ATM layer to transport 20 bytes of G.729. The ATM-caused overhead is thus 46 bytes per voice packet, thereby making the needed bandwidth 77% larger. CRTP solves this issue (40byte waste) cRTP is a great idea that is not widely implimented. Not very many endpoint support it, and Asterisk currently does not. In version 1.4, Asterisk will support configurable RTP packetization. So you could use 40ms for your G729 payload and reach a 50/50 split between payload and overhead (before ATM encapsulation for DSL). At 50ms for g729, you get a nice fit into two ATM cells, while a small bump to 60ms will force you into three ATM cells. Of course you need to be mindful of the latency between endpoints when you start increasing the payload, but a healthy network and tuning the payload can greatly reduce bandwidth requirements. AND most endpoints do support configurable RTP packetization. If Asterisk and more endpoints supported cRTP, then combining that feature with larger RTP payloads would result in very low overhead numbers. Dan ___ --Bandwidth and Colocation provided by Easynews.com -- asterisk-users mailing list To UNSUBSCRIBE or update options visit: http://lists.digium.com/mailman/listinfo/asterisk-users -- Omar E.P.T - Certified Networking Professionals make better Connections! http://omarept.blogspot.com/ Usysnet Corp Open Source Solutions www.usysnet.com.pe ___ --Bandwidth and Colocation provided by Easynews.com -- asterisk-users mailing list To UNSUBSCRIBE or update options visit: http://lists.digium.com/mailman/listinfo/asterisk-users