Comments on: Separation of Application and Transport http://www.phoneboy.com/1037/separation_of_application_and_transport VoIP, Mobile Phones, Telecom, and Technology Made Simple Sat, 05 Jul 2008 10:39:15 +0000 http://wordpress.org/?v=2.5.1 By: Anonymous http://www.phoneboy.com/1037/separation_of_application_and_transport#comment-60 Anonymous Wed, 31 Dec 1969 16:00:00 +0000 #comment-60 Not likely, Phoneboy<br />It's hard to control consumers in an all-IP environment, Phoneboyeeeee. Hence the time-division hijinks and UMA. The cell carriers are in no hurry to move to IP. Not likely, Phoneboy
It’s hard to control consumers in an all-IP environment, Phoneboyeeeee. Hence the time-division hijinks and UMA. The cell carriers are in no hurry to move to IP.

]]> By: Anonymous http://www.phoneboy.com/1037/separation_of_application_and_transport#comment-61 Anonymous Wed, 31 Dec 1969 16:00:00 +0000 #comment-61 The following may be out of<br />The following may be out of context in which you state that, "[in telecom networks] the application ("voice") and the transport mechanism seem inexorably tied together". For a long while now, this seems to be the conventional wisdom, thanks to the "Stupid Network" paper. I am of the contrary view. If this is indeed true then how come we are able to use the telecom network for a Fax application or a data (modem) application? The Stupid Network paper claimed that one can not change the codec in the old telecom network. Of course it is factually erroneous. The STU III telephone uses a different codec and that predates the paper. With the current advances in the technology, one can run a wideband codec end-to-end, all without making any changes to the "Middle". Just grant intelligent devices at the end. The following may be out of
The following may be out of context in which you state that, “[in telecom networks] the application (”voice”) and the transport mechanism seem inexorably tied together”. For a long while now, this seems to be the conventional wisdom, thanks to the “Stupid Network” paper. I am of the contrary view. If this is indeed true then how come we are able to use the telecom network for a Fax application or a data (modem) application? The Stupid Network paper claimed that one can not change the codec in the old telecom network. Of course it is factually erroneous. The STU III telephone uses a different codec and that predates the paper. With the current advances in the technology, one can run a wideband codec end-to-end, all without making any changes to the “Middle”. Just grant intelligent devices at the end.

]]> By: Anonymous http://www.phoneboy.com/1037/separation_of_application_and_transport#comment-64 Anonymous Wed, 31 Dec 1969 16:00:00 +0000 #comment-64 Another approach?<br />While I understand you are dealing with GSM in Europe, I thought this info regarding CDMA2000 1xEV-DO Rev. A was worth mentioning. Taken <a href="http://www.qualcomm.com/press/releases/2006/060607_successfully_demonstrates_fully.html">from this Qualcomm announcement </a> (June 07, 2006) - <blockquote>By basing all communications services on an IP network, operators are able to use their network capacity in a much more flexible manner, through the dynamic allocation of capacity to an ever-increasing array of 3G services, such as: <ul> <li>Push to talk, which is VoIP offered as walkie-talkie-like service</li> <li>Packet-based video telephony, which combines realtime video with VoIP</li> <li>Simultaneous VoIP calls and data sessions</li></ul></blockquote> Sprint planned to begin rolling CDMA2000 EV-DO Rev. A out in 4th QTR of 2006/early 2007, but I believe they have already started at this time. Another approach?
While I understand you are dealing with GSM in Europe,
I thought this info regarding CDMA2000 1xEV-DO Rev. A was worth mentioning.

Taken from this Qualcomm announcement (June 07, 2006) -

By basing all communications services on an IP network,
operators are able to use their network capacity in a much more flexible manner,
through the dynamic allocation of capacity to an ever-increasing array of 3G services, such as:

  • Push to talk, which is VoIP offered as walkie-talkie-like service
  • Packet-based video telephony, which combines realtime video with VoIP
  • Simultaneous VoIP calls and data sessions

Sprint planned to begin rolling CDMA2000 EV-DO Rev. A out in 4th QTR of 2006/early 2007,
but I believe they have already started at this time.

]]> By: Anonymous http://www.phoneboy.com/1037/separation_of_application_and_transport#comment-76 Anonymous Wed, 31 Dec 1969 16:00:00 +0000 #comment-76 Ahem. Economics...<br />"IP is built on the 7-layer OSI model, thus demanding separation between application and transport." SS7 has never followed the ISO model as most of the protocol predates it. ISUP for example isn't even nicely conformant with the rest of the SS7 protocols -- it covers several SS7 levels. Anyway, TCP/IP just isn't good enough for telephony signalling protocols that demand 911/111/000/112 levels of service -- SCTP/IP is but not well adopted. "So when will the telcos realize that it is in their best interest to move everything to IP as quickly as possible, including "basic" phone calls? " I'd love it if my network ripped out all their 1000's of basestations, routers, MUXs, and all the MSCs. But how is it going to pay for it? If the cost of wireless data continues to decay but users demand more bandwidth, coverage, and services I can't see how this is going to work. For these businesses to succeed on an ongoing basis the customer base has to end up paying at least the cost+margin, and "internet" model retail prices are never allowed to go up. Ripping out all this equipment and replacing it with more complicated equipment to provide basically the same service isn't economic. We can debate the "more complicated equipment" another time...certainly they are less mature and don't inter-op well. UMA is overly complicated. The CDMA equiv. will be too. I'd love to do this on CDMA/GSM/UMTS with fast MobileIP switching like WiMax is planning, but...I'm not even sure fast MobileIP handoff is a current reality. Ahem. Economics…
“IP is built on the 7-layer OSI model, thus demanding separation between application and transport.”

SS7 has never followed the ISO model as most of the protocol predates it. ISUP for example isn’t even nicely conformant with the rest of the SS7 protocols — it covers several SS7 levels. Anyway, TCP/IP just isn’t good enough for telephony signalling protocols that demand 911/111/000/112 levels of service — SCTP/IP is but not well adopted.

“So when will the telcos realize that it is in their best interest to move everything to IP as quickly as possible, including “basic” phone calls? ”

I’d love it if my network ripped out all their 1000’s of basestations, routers, MUXs, and all the MSCs. But how is it going to pay for it? If the cost of wireless data continues to decay but users demand more bandwidth, coverage, and services I can’t see how this is going to work. For these businesses to succeed on an ongoing basis the customer base has to end up paying at least the cost+margin, and “internet” model retail prices are never allowed to go up.

Ripping out all this equipment and replacing it with more complicated equipment to provide basically the same service isn’t economic.

We can debate the “more complicated equipment” another time…certainly they are less mature and don’t inter-op well.

UMA is overly complicated. The CDMA equiv. will be too. I’d love to do this on CDMA/GSM/UMTS with fast MobileIP switching like WiMax is planning, but…I’m not even sure fast MobileIP handoff is a current reality.

]]>