Sometime back we had news about how VoLTE is battery killer and how it would suck our 4G phones dry. Well, I agree. I am no fan of VoLTE and think that CSFB solution can suffice in mid-term. Having said that, there is a solution which would be soon available to sort this battery issue during VoLTE call. I had a post on this topic earlier titled SPS and TTI Bundling. I am not sure about exactly how much saving would occur if either of the features are implemented.
ST Ericsson has recently released a whitepaper on this topic that is embedded below. If you have more idea on this, please add it in comments.
Zahid, this is great info.
ReplyDeleteIn my battery life modeling, VoLTE is a LOT more power hungry than CDMA or WCDMA voice calls. This basically comes down to the power efficiency of the hardened modem vs. a general purpose apps processor.
This situation may change over time but I do not see a reversal for several years at least.
Joe Madden
Mobile Experts
www.mobile-experts.net
In their summary ST Ericsson predict DRX will reduce VoLTE power consumption by 30%. In this more detailed paper Nokia-Siemens measure it at 14% (see page 6)
ReplyDeletehttp://www.nokiasiemensnetworks.com/system/files/document/more_battery_life_with_lte_connected_drx_0.pdf
This all is not really surprising, especially for anyone who actively perceived the advent of UMTS in GSM areas.
ReplyDeleteThe situation was very similar back then, so I don't see why people still act so surprised.
Anyway, thanks for sharing the great ST-Ericsson paper; it's nice to see some realistic numbers supported by technological facts, instead of some mindless ranting about how new technology sucks just because, in the very beginning, the user experience may be a little worse in parts.
@rupello: This is not necessarily a difference in propositions.
ReplyDeleteThe STE paper refers to the likely 30% reduction during a VoLTE call, while the NSN presentation takes a "app"-scenario, i.e. bidirectional burst package-traffic.
The latter should account for less power saving opportunities through DRX, because often high level bi-directional protocols are used which don't allow for long sleep periods of the transmitter.
I assume that's why NSN also suggests "20 - 40%" savings for "background" services (like notifications) which often don't ACK a received message and often do barely any sending and thus have greater opportunities for DRX.
@Joe Madden - I agree, with general purpose/software radio approach to LTE you will take a huge hit on battery life. However there are good LTE modems (like the Thor M7400) which are as hardened as any 2G or 3G modem. You will see them in handsets very soon.
ReplyDelete@rupello - The NSN paper refers to general internet/app type of traffic for which DRX provides much less benefit.
The interesting thing will be to see if the chipset suppliers don't just put VoLTE into the modem, but also implement various of the needed codecs & media processing for WebRTC as well (VP8, Opus etc)
ReplyDeleteIt is clear that VoLTE will not be the only basis for future telco-run telephony, so it makes sense to hedge bets. It certainly seems sensible for Google/Android/Motorola to push WebRTC down into hardware as fast as possible.
Don't want to divert the discussion to far but wanted to comment on the point in the original post that 'I am no fan of VoLTE and think that CSFB solution can suffice in mid-term'.
ReplyDeleteFor anyone that doesn't know, I will declare a bit of vested interest up front - I led the GSMA's work to define VoLTE, and am something of an IMS zealot. My point is more about the principle of having any kind of 'interim' solution however. I think the industry has pretty much accepted that VoLTE will happen, but I would actively encourage as many operators as possible to skip the CSFB step if they possibly can. The main reason is that if you put CSFB in your network as an interim solution, you have to invest in it and then maintain that investment for a considerable time to come, long past your VoLTE deployment taking off. Not only that, but the industry as a whole ends up needing to suppot interworking between existing CS, CSFB and VoLTE for roaming and some interconnect scenarios.
It was these kind of points that killed off VoLGA - not much wrong with the technology, but adding more options for voice service has an exponential effect on the number of interworking scenarios everyone has to support. VoLGA was one option too far. I agree that some operators need an 'interim' solution if they need a voice solution in their LTE network before they have IMS operational, but don't do it by rote - you do not have to do CSFB, and if you do do it, be prepared to support it for longer than you might expect. It is not just the network that has to churn it out again, but also all of the handsets (and while some of us change handset regularly, plenty of people don't). If you can step straight to VoLTE, do. You're going to go through the pain of VoLTE deployment at somepoint anyway - why compound the pain by also deploying and then extracting CSFB?
There is an interesting comment on the last page - Services like Skype would continue to run on power hungry application processor compared to VoLTE service that would run on more efficient modem. Could this be an issue for OTT voice applications including WebRTC. I guess the outcome will depend on typical voice versus non-voice usage pattern of LTE subscribers and as corollary % power consumed by voice application.
ReplyDelete