Ericsson and Qualcomm just pressed that they performed successful voice handovers back on December 23, 2011 with VoLTE (Voice over LTE) to WCDMA handover using Single Radio Voice Continuity (SRVCC.) The significance of this is that VoLTE/SRVCC allows a good user experience while reducing the required radio components thus saving cost and battery life and of course answers the question, what happens when the mobile user, on a voice call, leaves the LTE coverage area. There is a back up plan called Circuit Switched Fall Back (CSFB) but it’s painful in terms of latency and reliability. Exactly what must have been behind the response Steve Jobs gave previously when asked when Apple devices would have LTE . The delay from the beginning of time until in your hands is availability of broadband processing that support these features for the mobile device. The networks are more or less ready with some exceptions that can be worked around. (Note, 3GPP Release 9 added emergency services including location, SIM less emergency calling and some other fixes like faster CSFB so it’s to the implementers to make it happen.) Oh yeah, Ericsson provided the radio edge network (EUTRAN) and core network with IMS and Qualcomm used their MSM8960 in a VoLTE capability demonstrating device. As Martha Stewart would say, “It’s a good thing.”
Missed the boat…after the break
A friend of mine, Jim, sent me a link to AnandTech’s article. They missed a key point so I created this write up instead of just letting links fly like the rest of the blogosphere. The butler did it in the library with the candlestick….I found it interesting that the MSM8960 performed the SRVCC testing as it is a dual radio device. DRVCC would be the obvious milestone so SRVCC was an interesting choice. Unless Qualcomm disables the second radio chain, or sleep it whatever, (actually sleeping the 2nd TX chain to be more precise) then there’s going to be a lot of avoidable power consumption based on the hardware. Was the testing to focus the world on the possible power savings of using this multicore multi radio chip in single radio mode? Not saying that’s a bad thing, just seems a little like a misdirection?
What we do know is: 8960 has multiple cores and multiple radio RX and TX chains on board. We also know it’s easier to transition a VoLTE call call to WCDMA since it supports packet and circuit switched (voice) on the same channel and just so happens that the voice coding, AMR, the default in VoLTE, is used for WCDMA calls. On the other hand, LTE to 1X gets challenging because there are 2 radio channels typically, a circuit switched voice 1xRTT and packet switched domain 1xEVDO. The UE (user device) would need to go through a selection process (EVDO would need some sort of VoIP capability with various options including IMS or legacy implementations) to do a VoIP handover or otherwise nontrivially, transition a call from the packet switched (PS) to the circuit switched (CS) domain. Another sticky point is if LTE to CS then the 1X default voice coding is EVRC, so now there is a transcoding situation. So, what we do know is the MSM8960 to WCDMA SRVCC handover with VoLTE was demonstrated, but we don’t know is if that setup supports 1XRTT/1XEVDO and DRVCC.