When Informa told me they were rebranding their popular LTE World Series into 5GWorld, I wondered if this was perhaps too early. However this first event, held in London, retained tracks discussing today’s practical issues alongside showcasing emerging 5G technology.
On arrival at Olympia, which had several parallel events in progress, you might have been forgiven for thinking you were entering a car show. The connected car section took up a fair bit of floor space, demonstrating the level of technology embedded in vehicles. It’s an eclectic show, featuring up to seven concurrent presentation tracks, making it difficult to choose which to attend. The quality of presentations, even keynotes, was variable. For example, NTT DoCoMo gave some clear technical leadership of the path they plan to take rather than making bland assurances of being customer driven.
The problem remains that there are still just too many competing options and standards. This is true not just for 5G (which is still in very early stages), but for IoT (both indoor and wide area) and RAN architecture (Cloud, Network Slicing or standalone small cells). This lack of clarity holds back industry investment.
5G standards far from being finalised…
3GPP won’t even agree 5G requirements until March 2017, with technical standards published mid 2018 at the earliest. Four major operators have taken a lead by choosing spectrum for their 5G program. Adrian Scrase, ETSI CTO, said this was an unusual excursion from previous generations where spectrum had been aligned first through WARC, but thought this could be worked around.
3GPP have given a name to their current project to develop 5G use cases and requirements - SMARTER = New Services and Markets Technology Enabler. The timetable presented is shown below:
When challenged by one delegate as to whether the spectrum issue was fragmenting the 3GPP standards process, given other examples of LTE-U and MulteFire, Adrian commented that those were accelerators. The 3GPP only has capacity to work on a limited number of projects, and these projects didn’t have universal support. But once developed, he expected them to return and become adopted within future standards releases. He was open to finding another route to agree spectrum allocation if that was what the industry wanted.
…which didn’t stop plenty of 5G vendor demonstrations
All of the major RAN vendors featured 5G on their booths. The term they are using is 3GPP 5G NR (New Radio) which isn’t limited to above 6GHz bands. Ericsson, Nokia, Huawei and ZTE all highlighted very high speed data rates, multiple gigabits, with themes of massive MIMO antenna arrays, lots of radio heads (which may or may not be called small cells). Beam forming using SU-MIMO is key, with each user being given their own “virtual cell”. But to constrain the size of handsets, most of these massive MIMO antenna arrays will be at the cellsites.
Ericsson's theme was "out of the lab and into the neighbourhood". It still seems quite early days to me, but they've signed 25 contracts, expect standards by end 2017 (which doesn't match 3GPP's position above) with commercial deployment in 2020 although some early adopters may launch in 2018.
Huawei spoke of how data was growing much more quickly indoors than out, and so indoor solutions would be needed. They forecast 33x indoor vs 20x outdoor data growth by 2020, a ratio of 1.65:1, and that investment for indoor systems wil be greater than for outdoor by as soon as 2017.
ZTE explained that Chinese trials were in the 15GHz band and thought that 5G might initially operate similar to LAA, extending LTE by coupling an existing LTE connection with very high speed datapipe. They saw a two part evolution which included sub-6GHz bands serving IoT and Emergency Services alongside higher frequencies for superfast speeds.
BT said “just faster” is important, but actually the other benefits, QoS, low latency, security etc. have to be user driven. It’s no good to “just build it and they will come”.
But LTE can achieve 1Gbps already
TDC Denmark explained how they relatively quickly moved from being worst in their country for cellular network performance to best (and now remain in the top 10 worldwide). They swapped out all 3,400 of their basestations in 12 months for Huawei, refarmed spectrum to LTE and made many related RAN optimisations. With plenty of spectrum at 800, 1800 and 2600 they’ve demonstrated 1Gbps data rates using carrier aggregation – the only reason you can’t use this is the lack of terminals today.
While their pretty flat terrain countrywide makes this easier to plan for, it still shows what’s feasible with today’s technology.
One delegate commented that the industry is in a holding pattern at the moment. Operators know they need to continue to increase capacity through densification, but haven’t strategically decided to adopt a particular direction/solution. Both NTT DoCoMo and South Korean operators gave some excellent and well thought out strategies which suit their own markets. I’d guess there is a herd mentality here and the global industry direction isn’t yet clear. As an indicator, eBlink, who sell wireless fronthaul solutions, said they were finding plenty of opportunities where single remote radio heads have been tactically deployed as a stop-gap measure.
The pace of VoLTE rollout
The pace of VoLTE (voice over LTE) adoption remains relatively slow. The primary commercial benefit is the ability to use sub 1GHz spectrum for LTE, extending the range and in-building penetration from existing cellsites. I don’t think anyone really believes VoLTE will increase or substitute voice revenues, when WhatsApp and Skype are taking so much traffic.
I sat through a somewhat frightening presentation from BICS (a wholesale interconnect carrier owned by Belgacom, Swisscom and one other) about just how complex and difficult it can be to commission VoLTE roaming, either through a central hub or directly. There are a couple of different architectures, you have to decide if you want to share your ENUM database with others, and what to do in case of fallback to 3G (outside LTE coverage) - do you stick with VoIP on 3G (SRVCC) or switch to circuit switched 2G or 3G. The architecture differs radically between these two, either a dumb datapipe or SIP proxy. Of 550 LTE commercial networks worldwide, BICS IPX connects to 300 for data roaming. If the 55 VoLTE commercial networks worldwide, hardly any are fully VoLTE interconnected today. This is likely to appear first within the large global network operator groups, who will have standardised their platforms in different countries, before later rolling out between 3rd party networks. Meanwhile Skype, WhatsApp and other OTT solutions work well with adequate capacity and are growing as credible competitors.
This has some implications for Small Cells, and especially Neutral Hosts, who may either operate as if just another RAN (the simple method) or through roaming.
It also questions the importance of voice services to network operators overall. Perhaps the “green call button” on our smartphones will no longer remain their preserve, as we discussed with Apple’s forthcoming iOS10 which opens it up to any voice calling App.
IoT deployments growing
One of TEOCO’s activities is to test new devices being launched onto the networks. They told me that the proportion of new IoT cellular devices now outweighs traditional smartphones by 3:1 for one large operator. Smartphones are ever more complex to test, partly due to the huge numbers of frequency bands/modes and carrier aggregation options. IoT devices are complex because of the sheer variety in the number of vendors, physical formats and operating modes. Even the choice of which cellular generation to use isn’t clear – 2G or 3G or 4G.
Aeris, whose IoT platform enables operators and MVNOs to support their own and third party IoT services, is growing fast – doubling in size annually. They see huge opportunity and variety in cellular IoT, wanting to enable many vertical market sectors to create their own solutions. I see this in a similar way to enabling App developers. Indoors, Aeris expect many residential IoT to connect to a home hub using one of the lower power/short range wireless protocols rather than cellular, but the trouble is again there are just too many competing standards.
Several vendors spoke in depth about growing security and signalling concerns, especially with the rapid onslaught from these M2M and IoT devices. Speakers highlighted growing issues of security and signalling complexity. While data traffic might grow by 8x by 2020, signalling traffic (no doubt for all those IoT devices) might grow by 25x.
Small cells on show
Nokia were demonstrating a standalone solar powered urban LTE small cell. While they wouldn’t state the RF output power, the concept used wireless backhaul and included a battery for overnight operation. Perhaps feasible for sunny climes, I think it might struggle to achieve de-minimis planning approval in some urban areas – it’s still quite bulky and somewhat fragile looking.
Software vendor Athonet were operating eMBMS (LTE Broadcast) using a Nokia basestation. Their suggestion was that a large venue, such as a stadium, might deploy their own mini-network equipped with eMBMS and Athonet’s core network. Multiple video streams could be shared by tens of thousands efficiently, so that anyone could select any one of many camera angles throughout. It will need compatible handsets which aren’t yet widely available. Athonet's PriMo highly scaleable core network software runs on off-the-shelf hardware and supports both 3G and LTE with standard 3GPP interfaces to the RAN.
Panel sessions discuss the today’s practical issues
A good selection of knowledgeable and sometimes forthright panellists discussed today’s issues and future concerns.
I asked if Wi-Gig (Wi-Fi at 60GHz) due to be commercially launched later this year would be a threat to 5G. Sami Susiaho, Head of Edge Techonologoes at Wi-Fi operator The Cloud, said he thought it was more relevant for in-room device-to-device connection, such as throwing video from your smartphone or tablet up onto a nearby TV screen. Vodafone commented that it may take some of the residential/consumer indoor use cases. Most speakers thought that 5G will focus on reliability, security, resilience etc. However, when looking at drive-by demo (5G in a moving van), the data rate can be quite variable and works best when is direct line of sight.
The event coincided with Wi-Fi Global Congress in Liverpool and GSMA in China, and attracted a reasonable but not very large number of delegates. The major RAN vendors were visible here, but almost no small cell vendors. Although 5G was a leading theme, with plenty of RF demos, there is growing awareness of horizontal applications. Connected Cars was the most visible (and sexy), but there are clearly many IoT applications and services being introduced at a rapid rate.
5G can’t be just another faster radio interface, which is what’s really being demonstrated at the show. LTE can already achieve very good performance and service from existing macrocell sites. VoLTE remains a complex issue to resolve, especially for roaming. In-building is important and should be the next step for the industry. Superfast 5G risks being restricted to urban areas only – it’s not a nationwide capability – and we can see operators looking more carefully at investments to ensure they address higher value use cases rather than just deliver ever faster speeds.
- Informa have also rebranded their conference division to KNect365. They have very good logistics and execution throughout. You could even pre-print your badge before arrival. My only gripe was lack of easy access to presentations after the event.
- Europe’s GPS equivalent Gallileo will launch commercially in November 2016. A key differentiator is authentication to avoid spoofing (and at a higher level for a fee). This can be used for sync, timing as well as navigation, but may require higher capacity chips to track both GPS and Galileo systems in parallel.
- 5G isn’t just about very high frequencies. Qualcomm just released a sub-6GHz trial platform based on OFDM in addition to their 28GHz millimetre wave prototype.
- Samsung has established a Next Generation business team, predicting 5G trials by early 2018 using millimetre wave, high speed, low latency radio interface.
- FirstNet, the national US first responder network, was given $7 Billion and 20MHz of 700MHz nationwide spectrum to build out a new shared LTE network. RFP submissions are being assessed with contract award is expected to be announced early November. FirstNet will allow others to share/use their network (at lower priority) to help finance ongoing operation. Designed to handle peak emergencies, I would expect it to have lots of spare capacity most of the time.