Sprint to join Rural Operators Roaming Hub (CCA and RRPP thread)

[bretton88]

What's tmobile's long term plan for iwireless?

Really T-mobile doesn't have that control to make a plan. They're a minority partner with Iowa Telecom. They've even moved out of the Iowa market and handed it their remaining presence over to iwireless. But it's up to Iowa Telecom on what capex to do and strategic planning with Iwireless.

[gnoj]

According to the information we have about Sprint's VoLTE plan there will be VoLTE to CDMA handoff.

I would imagine this is the main reason why it's taking Sprint so long to go public with their plans for VoLTE- They are working around the clock to get eSRVCC (and perhaps the reverse as well) to function reliably enough that they don't throw away all the gains Sprint has made over the last couple years in voice coverage & reliability. Verizon's LTE network is big enough (and their leaders arrogant enough) that they can afford to give up on getting handovers to CDMA to work. Sprint, justifiably, feels that they need to work on the problem more before they should give up entirely.

 

Other reasons for the delay may be that they're waiting for B26 POP coverage to get closer to that of B25 (especially if handover to CDMA remains iffy), and also for the finalization of the EVS codec and other features of 3GPP Release 12, which is expected to be frozen this March. If HD Voice is an inevitable casualty of the handover process, then they may as well move to the newer EVS for VoLTE, rather than sticking with EVRC-NW. On top of that, they have to ensure interoperability with all of their CCA partners. USCC, for their part, has said they'll have some test markets up this year. I expect we'll know much more by this summer.

[Conan Kudo]

I would imagine this is the main reason why it's taking Sprint so long to go public with their plans for VoLTE- They are working around the clock to get eSRVCC (and perhaps the reverse as well) to function reliably enough that they don't throw away all the gains Sprint has made over the last couple years in voice coverage & reliability. Verizon's LTE network is big enough (and their leaders arrogant enough) that they can afford to give up on getting handovers to CDMA to work. Sprint, justifiably, feels that they need to work on the problem more before they should give up entirely.

 

Other reasons for the delay may be that they're waiting for B26 POP coverage to get closer to that of B25 (especially if handover to CDMA remains iffy), and also for the finalization of the EVS codec and other features of 3GPP Release 12, which is expected to be frozen this March. If HD Voice is an inevitable casualty of the handover process, then they may as well move to the newer EVS for VoLTE, rather than sticking with EVRC-NW. On top of that, they have to ensure compatibility with all of their CCA partners. USCC, for their part, has said they'll have some test markets up this year. I expect we'll know much more by this summer.

EVRC-NW is not going to be supported for VoLTE ever. The 3GPP has indicated it has no intent to add support for the EVRC family, and 3GPP2 has fallen apart over the last 18 months (useful activity has dropped to nearly nothing these days). The only codecs supported for VoLTE at this time are EVS (in draft process now), the AMR codec family, the GSM codec family, and PCM. Potentially, Opus will be added in order to support interworking with WebRTC, which declares Opus as a "mandatory-to-implement" codec for the technology. If Sprint was to do VoLTE now, it'd have to use the AMR codec family to provide a high-quality voice service.

[gnoj]

EVRC-NW is not going to be supported for VoLTE ever. The 3GPP has indicated it has no intent to add support for the EVRC family, and 3GPP2 has fallen apart over the last 18 months (useful activity has dropped to nearly nothing these days). The only codecs supported for VoLTE at this time are EVS (in draft process now), the AMR codec family, the GSM codec family, and PCM. Potentially, Opus will be added in order to support interworking with WebRTC, which declares Opus as a "mandatory-to-implement" codec for the technology. If Sprint was to do VoLTE now, it'd have to use the AMR codec family to provide a high-quality voice service.

According to s4gru's sources, Sprint has tested VoLTE with both AMR-WB & EVRC-WB. By now they've probably also run some tests with EVS. If the upcoming 3GPP spec does not officially allow for EVRC to run over VoLTE, then I imagine they'll wait for EVS, rather than trying to go outside/expand the standard. I just don't see Sprint deploying any version of AMR, when few (or none?) of their CCA/RRPP partners have.

 

FWIW, I'd much rather have calls that drop down to PCM on CDMA 70-80% of the time, rather than have the call be guaranteed to drop altogether every time I lose LTE. If the resulting call quality really turns out to be that terrible, one can hang up and start a new call (which presumably would default to EVRC-(W)B over CDMA if the scanned LTE signal strength is weak or non-existent), which you'd have to do anyway if the call dropped.

[Conan Kudo]

 

According to s4gru's sources, Sprint has tested VoLTE with both AMR-WB & EVRC-WB. By now they've probably also run some tests with EVS. If the upcoming 3GPP spec does not officially allow for EVRC to run over VoLTE, then I imagine they'll wait for EVS, rather than trying to go outside/expand the standard. I just don't see Sprint deploying any version of AMR, when few (or none?) of their CCA/RRPP partners have.

Verizon Wireless, AT&T, T-Mobile US, and all other VoLTE network operators (like VTel, et al) are using the AMR-WB codec now, so anyone who wants to connect to those systems will need to have the codec anyway. So the AMR codec family absolutely will be supported.

 

FWIW, I'd much rather have calls that drop down to PCM on CDMA 70-80% of the time, rather than have the call be guaranteed to drop altogether every time I lose LTE. If the resulting call quality really turns out to be that terrible, one can hang up and start a new call (which presumably would default to EVRC-(W)B over CDMA if the scanned LTE signal strength is weak or non-existent), which you'd have to do anyway if the call dropped.

 

PCM at 8kbps is pretty horrible. Too many artifacts to make sense of it, most of the time. Regardless of that, the success rate is too low relative to SRVCC between GSM/WCDMA and LTE and call continuity expectations by mobile voice users.

 

A 70% success rate is terrible, given that the current standard of success is well over 98%. Even a 85-90% success rate is considered unacceptable, and that's what we have now with CSFB from LTE to CDMA. The only reason it's even being done is because Sprint can't afford SV-LTE for Spark devices (because that makes them blatantly extremely custom devices).

[gnoj]

Verizon Wireless, AT&T, T-Mobile US, and all other VoLTE network operators (like VTel, et al) are using the AMR-WB codec now, so anyone who wants to connect to those systems will need to have the codec anyway. So the AMR codec family absolutely will be supported.

I know the other Tier 1 operators are all using AMR-WB for their VoLTE networks. I was referring only to interoperability with the bulk of Sprint's CCA partners, who have not yet deployed VoLTE. At this point, I believe they'll just move to EVS. It wouldn't be the first time Sprint goes their own way (see: EVRC-WB, domestic carrier device locking), but this time, it'll be the other T1 carriers who would be a step behind in the standards roadmap.

 

PCM at 8kbps is pretty horrible. Too many artifacts to make sense of it, most of the time. Regardless of that, the success rate is too low relative to SRVCC between GSM/WCDMA and LTE and call continuity expectations by mobile voice users.

 

A 70% success rate is terrible, given that the current standard of success is well over 98%. Even a 85-90% success rate is considered unacceptable, and that's what we have now with CSFB from LTE to CDMA. The only reason it's even being done is because Sprint can't afford SV-LTE for Spark devices (because that makes them blatantly extremely custom devices).

I realize that PCM voice would probably suck, especially compared to what we're now accustomed to with HD Voice. My point is that unless it's actually painful or completely incomprehensible, it's still better than nothing. A 70% success rate is indeed horrible, but it's still greater than 0%. This would only apply in those few areas where B26 can't quite reach, and there's no Wi-Fi to hand off to, but there's still a weak yet usable CDMA signal (robust as 1x is) to grab on to. It would allow someone to at least finish up their call politely, rather than cutting it short unceremoniously. Depending on the final success rate, Sprint probably shouldn't advertise that handover will work, but it can be a small bonus and surprise for customers when it does.

 

Also, lower device cost is not the only benefit of eCSFB, but the various pros and cons of that have already been covered elsewhere.

[Conan Kudo]

I know the other Tier 1 operators are all using AMR-WB for their VoLTE networks. I was referring only to interoperability with the bulk of Sprint's CCA partners, who have not yet deployed VoLTE. At this point, I believe they'll just move to EVS. It wouldn't be the first time Sprint goes their own way (see: EVRC-WB, domestic carrier device locking), but this time, it'll be the other T1 carriers who would be a step behind in the standards roadmap.

It's not just Tier 1 U.S. operators. It's everyone who is deploying VoLTE. That includes SoftBank, DoCoMo, KDDI, SK Telecom, LG U+, KT, Hutchison Telecom (d/b/a "3"), T-Mobile International, Vodafone, O₂, and others. They are also all preparing for EVS. It'll be easy for them to support the codec, but that won't matter if handsets don't quickly adopt it. It does help that EVS will work on GSM, WCDMA, and LTE, though.

 

I realize that PCM voice would probably suck, especially compared to what we're now accustomed to with HD Voice. My point is that unless it's actually painful or completely incomprehensible, it's still better than nothing. A 70% success rate is indeed horrible, but it's still greater than 0%. This would only apply in those few areas where B26 can't quite reach, and there's no Wi-Fi to hand off to, but there's still a weak yet usable CDMA signal (robust as 1x is) to grab on to. It would allow someone to at least finish up their call politely, rather than cutting it short unceremoniously. Depending on the final success rate, Sprint probably shouldn't advertise that handover will work, but it can be a small bonus and surprise for customers when it does.

 

Also, lower device cost is not the only benefit of eCSFB, but the various pros and cons of that have already been covered elsewhere.

 

PCM sucks compared to members of the AMR and EVRC codec families used over a decade ago. PCM is only just better than the first-generation codecs used in the earliest digital network systems. It is an awful codec to fall down to.

 

And actually, 70% is worse than 0%, because it reinforces an expectation that it should work. If it was 0%, then no one expects it to work, and thus, Sprint's churn rates can't get even worse from voice related problems.

[WiWavelength]

PCM at 8kbps is pretty horrible.

 

Let us define what you mean by "PCM."  I doubt that you really mean PCM at 8 kbps.  What bit depth?  What sampling rate?  POTS quality PCM is 64 kbps -- 8 bit, 8 kHz. 

 

AJ

[Conan Kudo]

Let us define what you mean by "PCM."  I doubt that you really mean PCM at 8 kbps.  What bit depth?  What sampling rate?  POTS quality PCM is 64 kbps -- 8 bit, 8 kHz. 

 

AJ

I meant 8 bit+8 kHz. I simply forgot that it wasn't also 8 Kbps. Still pretty crappy audio, though.

[S4GRU]

According to the information we have about Sprint's VoLTE plan there will be VoLTE to CDMA handoff.

 

"The Sprint VoLTE network will be designed to hand off calls to the existing Sprint CDMA network, including HD Voice calls, via the EVRC-NW codec"

 

"VoLTE calls will not be given QoS Priority on LTE initially. Should LTE capacity constraints be experienced during a VoLTE call, the call will be handed over to the 1x network"

 

http://s4gru.com/index.php?/blog/1/entry-368-sprint-is-proceeding-with-a-volte-network-that-focuses-on-interoperability-with-domestic-and-international-volte-carriers/

 

That is a schematic document.  A schematic plan or program plan is where all the stakeholders put together a list of all the things they want to achieve.  They may not all be possible as the program statement evolves into actual testing and an FIT.

[utiz4321]

That is a schematic document. A schematic plan or program plan is where all the stakeholders put together a list of all the things they want to achieve. They may not all be possible as the program statement evolves into actual testing and an FIT.

So if sprint can't make cdma handoffs work I think they are going to lag on volte. Given their spectrum i would think they would need to densify their network dramatically.

[S4GRU]

So if sprint can't make cdma handoffs work I think they are going to lag on volte. Given their spectrum i would think they would need to densify their network dramatically.

 

Yeah, you're probably right.  And that's my fear.  They have to pull the band-aid eventually, though.

[cdk]

Is there any way to hand off calls from lte to cdma?

 

Not without some really messy breakages. Essentially, because the two systems use totally different core network platforms (as opposed to GSM/UMTS+LTE systems), making the two systems establish seamless voice handover is so technically challenging that it is not practical to make it work.

 

That said, some work was done two years ago to establish a prototype mechanism for it. It works by treating the CDMA network as an visitor network and using the roaming handover mechanism to transfer the call. By doing so, however, the call drops to the lowest common codec: PCM. So HD Voice and all the other stuff is gone. What's worse is that the prototype didn't work very well in testing, with extraordinarily high failure rates (>20-30% of all calls failed to transfer). In the end, the CDMA operators who were driving the development of the idea killed it because of the unacceptable quality. Verizon and KDDI went on to just roll out VoLTE without any form of interconnection with the CDMA network layer.

 

Maybe I am confusing things I have read around here in the past, but isn't there a way to run a CDMA network on a GSM core? If there is, would that allow the seamless handovers?

[Conan Kudo]

Maybe I am confusing things I have read around here in the past, but isn't there a way to run a CDMA network on a GSM core? If there is, would that allow the seamless handovers?

You're referring to GSM1x. It could work, but then Sprint would have to turn off all EvDO, since it wouldn't be supported by it. Handover between the GSM1x system to the CDMA core to enable EvDO would create the same problems.

 

It also would force Sprint to move to SIM based authentication for its CDMA handsets, and older handsets would break, since the GSM1x system uses 3GPP authentication schemes.

[cdk]

You're referring to GSM1x. It could work, but then Sprint would have to turn off all EvDO, since it wouldn't be supported by it. Handover between the GSM1x system to the CDMA core to enable EvDO would create the same problems.

 

It also would force Sprint to move to SIM based authentication for its CDMA handsets, and older handsets would break, since the GSM1x system uses 3GPP authentication schemes.

 

At least my memory works, but that certainly doesn't sound like a workable solution for the problem...

 

 

On further reflection its not a short term solution. Maybe if they where at the point where they where shutting down all the EVDO carriers and leaving only LTE and 1x carriers this might be a long term way to do it.

Edited January 12, 2015 by cdk

[payturr]

Reading Conan's facts behind this, I say to hell with the handoff then. Sprint would be better off spending the money on expanding LTE and firing up VoLTE then trying to make a handoff to 1x perfect. CDMA is a dying technology anyway, might as well forget about it and look to the future.

[utiz4321]

Reading Conan's facts behind this, I say to hell with the handoff then. Sprint would be better off spending the money on expanding LTE and firing up VoLTE then trying to make a handoff to 1x perfect. CDMA is a dying technology anyway, might as well forget about it and look to the future.

The problem is the cost of running such a network. Unless sprint picks up some 600 at the much delayed auction the cell site density required to match the reliability of their cmda 1900 network seems like it would make it prohibitively expensive.

[payturr]

The problem is the cost of running such a network. Unless sprint picks up some 600 at the much delayed auction the cell site density required to match the reliability of their cmda 1900 network seems like it would make it prohibitively expensive.

Sprint doesn't need to densify, they just need to focus on band 26 rollout, amplify the signal at band 25 only sites and add additional band 25 carriers to increase capacity. They can make what they have work, 600MHz isn't as much of a necessity as some people believe. The only carrier who truly needs it is T-Mobile since they have very little sub-1GHz spectrum.

[travismheim]

Sprint doesn't need to densify, they just need to focus on band 26 rollout, amplify the signal at band 25 only sites and add additional band 25 carriers to increase capacity. They can make what they have work, 600MHz isn't as much of a necessity as some people believe. The only carrier who truly needs it is T-Mobile since they have very little sub-1GHz spectrum.

Amplifying b25 won't help your handset maintain its uplink connection and would likely lead to greater interference. They do need to density if they are going to an all LTE network or there will be a reduction in voice coverage and reliability.

 

Sent from my SM-T217S using Tapatalk

[maxsilver]

Sprint doesn't need to densify, they just need to focus on band 26 rollout, amplify the signal at band 25 only sites and add additional band 25 carriers to increase capacity. They can make what they have work, 600MHz isn't as much of a necessity as some people believe.

 

Sprint needs to densify in most markets, to roll out a successful VoLTE service in the near future (unless new spectrum / technology breakthroughs / etc occur)

 

They usually can't just "amplify the signal" or crank up the power on sites to improve coverage -- it's just not that simple, for a bunch of reasons.

 

Your right in part -- 600MHz or other low-band spectrum isn't required, and I believe they could "make what they have work". And I agree with your strategy, VoLTE is the future, spending cash on LTE is good, and spending new money rolling out old 1x service is extremely poor circumstance.

 

However, for Sprint to not need low band spectrum and still service everyone with reliable LTE (the kind of reliability that maintains calls without dropping) the only alternative (using current LTE technology / spectrum holdings) is extra density.

 

Without extra density, VoLTE-only would make the usable call coverage footprint drop drastically. You'd be asking Sprint to accept call failures / call drops increases of 10 to 30+% in various markets, and similar numbers in reduced coverage footprints. The streets would run red with failed RootMetrics call/text reports.

 

Until Sprint is ready (from a financial and technical standpoint) to densify, relying on 1x is the right decision.

 

EDIT : Your user card says your in NYC. NYC is unique, they already have some of the highest Sprint cell density in the nation. VoLTE could potentially be rolled out using just the existing sites / spectrum they have today. However, most markets are not NYC, and have only a fraction of that density. 

[utiz4321]

Sprint doesn't need to densify, they just need to focus on band 26 rollout, amplify the signal at band 25 only sites and add additional band 25 carriers to increase capacity. They can make what they have work, 600MHz isn't as much of a necessity as some people believe. The only carrier who truly needs it is T-Mobile since they have very little sub-1GHz spectrum.

They don't have enough 800 to have all there voice traffic run across it. Further that makes their voice coverage around what 1900 is and not even close to 800 1x

[Paynefanbro]

Sprint doesn't need to densify, they just need to focus on band 26 rollout, amplify the signal at band 25 only sites and add additional band 25 carriers to increase capacity. They can make what they have work, 600MHz isn't as much of a necessity as some people believe. The only carrier who truly needs it is T-Mobile since they have very little sub-1GHz spectrum.

 

 

EDIT : Your user card says your in NYC. NYC is unique, they already have some of the highest Sprint cell density in the nation. VoLTE could potentially be rolled out using just the existing sites / spectrum they have today. However, most markets are not NYC, and have only a fraction of that density. 

 

Yeah, like maxsilver said, Sprint's site density in NYC is more or less an anomaly when compared to other cities. In some areas, namely Midtown Manhattan, Sprint has a site on nearly every other block. In other areas like Crown Heights Brooklyn, Sprint could use an extra site to fill in coverage. Realistically, Sprint does need to increase cell site density if they plan on doing a VoLTE release with no call handoff to 1x.

[payturr]

 

Sprint needs to densify in most markets, to roll out a successful VoLTE service in the near future (unless new spectrum / technology breakthroughs / etc occur)

 

They usually can't just "amplify the signal" or crank up the power on sites to improve coverage -- it's just not that simple, for a bunch of reasons.

 

Your right in part -- 600MHz or other low-band spectrum isn't required, and I believe they could "make what they have work". And I agree with your strategy, VoLTE is the future, spending cash on LTE is good, and spending new money rolling out old 1x service is extremely poor circumstance.

 

However, for Sprint to not need low band spectrum and still service everyone with reliable LTE (the kind of reliability that maintains calls without dropping) the only alternative (using current LTE technology / spectrum holdings) is extra density.

 

Without extra density, VoLTE-only would make the usable call coverage footprint drop drastically. You'd be asking Sprint to accept call failures / call drops increases of 10 to 30+% in various markets, and similar numbers in reduced coverage footprints. The streets would run red with failed RootMetrics call/text reports.

 

Until Sprint is ready (from a financial and technical standpoint) to densify, relying on 1x is the right decision.

 

EDIT : Your user card says your in NYC. NYC is unique, they already have some of the highest Sprint cell density in the nation. VoLTE could potentially be rolled out using just the existing sites / spectrum they have today. However, most markets are not NYC, and have only a fraction of that density.

I don't mean shut down CDMA today and do VoLTE only. I mean forget the handoff, launch VoLTE, reduce overall # of people on CDMA while densifying and then once the LTE network is dense enough, shut down CDMA.

 

I said amplify the signal because here in Staten Island (which is far less cell site dense compared to Manhattan and Brooklyn), we have a small gap in coverage in specific areas. Amplifying the signal there would dissolve the gap, and I'm sure there's plenty of spots like that nationwide. I don't mean amplify it to the point of miles but a block or two may help.

[Paynefanbro]

I said amplify the signal because here in Staten Island (which is far less cell site dense compared to Manhattan and Brooklyn), we have a small gap in coverage in specific areas. Amplifying the signal there would dissolve the gap, and I'm sure there's plenty of spots like that nationwide. I don't mean amplify it to the point of miles but a block or two may help.

 

But increasing power is useless because it is more liable to cause interference on the handset side. If there is not enough power on the receiving end (handset), then turning power up at the site won't do much at all.

 

 

I don't mean shut down CDMA today and do VoLTE only. I mean forget the handoff, launch VoLTE, reduce overall # of people on CDMA while densifying and then once the LTE network is dense enough, shut down CDMA.

 

Still a bad idea. Sprint's dropped call rate would skyrocket to the point where their reputation would be damaged even further. Sprint's best bet is to densify now (which they're doing in some markets already) and once they reach a sufficient density then they turn it on alongside 1x. You want users to have a seamless experience on both networks. Slowly transition folks over to VoLTE, and once you reach a low enough amount of people on 1x, then you can make devices with only LTE. However, that won't be for years.

[maxsilver]

I don't mean shut down CDMA today and do VoLTE only. I mean forget the handoff.

 

No worries, I already assumed that's what you meant when I wrote the post.

 

I assume you meant "VoLTE-only" would mean no CDMA to rely on. Doesn't really matter whether Sprint towers still broadcast CDMA or not, since a user's hypothetical "VoLTE-only" phone couldn't use it (presumably by definition) then it wouldn't make any difference to them. 

 

I mean forget the handoff, launch VoLTE, reduce overall # of people on CDMA while densifying and then once the LTE network is dense enough, shut down CDMA.

I would agree with this new statement, if "launch VoLTE" meant "release devices that have VoLTE + CDMA support, but only use CDMA by default" similar to what Verizon has today with iPhones, until that time in the future "once the LTE network is dense enough"

 

But if you remember, that's a big shift from what you wrote just this morning. You originally said "Sprint doesn't need to densify" for VoLTE, which is why I wrote all the stuff about why it's necessary (in most but not necessarily all markets).