A proposal to speed up the adoption of specifications for one architectural variant of 5G’s new air interface has been adopted by the 3GPP RAN Plenary meeting, according to a report from Signal Research.
The proposal, reported here and here by TMN, was for a Non StandAlone (NSA) version of 5G New Radio (5GNR) to be adopted earlier than the full StandAlone version. This would allow vendors and operators interested in getting to market earlier to do so with a version of 5G radio that roots core control in the existing LTE EPC.
Opponents were concerned that it might lead to incompatibility with finalised standards and that it would not take into account the architecture of the 5G core.
However, at the meeting being held this week, the acceleration of this option seems to have been adopted – and it seems that this was always likely to be the case. TMN was told last week by AT&Ts Vice President for RAN & Device Design, Gordon Mansfield, that he expected to see several names hitherto notable by their absence in public documents sign up for the agreement. Nokia, for example, was playing a dual game last week – signing the original proposal but keeping its name off the press release.
Yesterday Signal Research’s Michael Thelander wrote, “After a bit of back and forth haggling on the text used in the joint proposal, everyone came together and endorsed a way forward.” He added:
“Ultimately, the “real list” of endorsing companies – per the 3GPP submission – also includes Apple, Broadcom, Cisco, Nokia, Samsung, and Verizon. Two of those names – Samsung and Verizon – are especially noteworthy since they were staunch opponents of the accelerated schedule. ZTE’s name was included in the press release, but it was also against the accelerate schedule until now.”
So what difference will it make to 5G developments to have an interim partial-freeze on one element? Some are saying, not much, after all. Thelander adds that the “accelerated schedule” merely puts in writing the implied Release 15 schedule that was needed all along to meet the full completion of the Release 15 specification in time.
This “alls well that ends well” view doesn’t quite explain the nine month delay since the original proposal – nor the pretty clear prior opposition from Telefonica, Verizon and Samsung to accelerating NSA. Nor the weirdness of certain companies adding their name to the original 3GPP proposal, but asking not to be on the public press release.
AT&T’s Mansfield said that since the June proposal for acceleration, the pace of development has picked up, making the 3 month early deadline more widely acceptable. “Work has progressed tremendously,” he said. “Others have been equally engaged, but not necessarily named for certain reasons in the PR. At the March meeting we will see formal alignment.”
Dan Warren, who heads up 5G radio research at Samsung, a company that had opposed the early NSA option, said that the argument had always been about balancing the three month acceleration against a “quality” standard.
Warren said, “The acceleration is certainly a reflection of a need in the industry to put a stake in ground so people can develop product against the spec. The balance that needs to be struck is between the quality of the spec you get in the timeline versus the additional work that needs to be done after the line is drawn.”
Warren said one aspect of debate was “how you extend what comes after that line to add what was not included in that first spec.”
“The question was whether putting that line in the sand three months earlier still allowed enough time to put a quality spec in at the end of the day. If everyone is committed then quality should be quite good but it is a race to get out the door quickly balanced against quality – and moving the line back three months runs the risk of potentially reducing the quality of what comes out.”
Thelander reports the ability to extend back into the NSA specs, even after the supposed December 2017 freeze was something both “sides” were keen to see. In his words:
“One haggling point involved the matter of backward compatibility – both sides wanted to allow for backward incompatible change requests (CRs) post March 2018 without implying that there would likely be an incompatible version of Release 15 (NSA only) in December 2018 [sic – I think Thelander means December 2017 here].”
So was the early NSA option all a storm in a tea cup that changes little? After all, the eventual June 2018 5GNR deadline is still the same, and even with the NSA acceleration its December 2017 spec will be perhaps more of a semi-freeze than a permafrost – as it will allow for some change requests after that date (per Thelander).
The fact remains that the NSA acceleration proposal did expose a fault line between operators – notably AT&T and Verizon who have both made publicly barbed comments at several conferences over the past months – and between some vendors. That it has taken nine months to cover over tells its own story. Despite that, by 2020 the fact that there was an early architectural variant on the table in 2018, three months earlier than another StandAlone option, is unlikely to matter much, unless the two specs are so far apart as to mean commercially deployed NSA versions cannot progress forward to work with the eventual 5G core architecture.