I spent a couple of days at SCWS World 2016 last week. Here is a write-up of some of the briefings I had, plus a few of the conference sessions I was able to see.
Overall I would say a few key themes seemed to dominate this year, and these were:
1. virtualisation (including MEC) and thinking in general about how work done to date within small cells can enable 5G.
2. real world deployment challenges – including the sort of challenges thrown up by 5G.
3 enabling shared or neutral host models.
4. and with that in mind MuLTEFire was also a topic that was hot (extending out from last year’s LTE-U/LAA focus).
In addition to these notes below, here are some extra links to stories and content rounding up our presence at the event last week.
1. “Huawei demos 1Gbps indoor performance with Vodafone”
2. “Orange radio network head identifies gap in small cell market”
3. We filmed a short video with Rajesh Mishra, CTO of newcomer Parallel Wireless, to understand how the company is doing things differently in the small cell space.
4. Not only that but TMN also published its annual Small Cells Market Update at the event, and you can access that here.
Time for small cells experts to be given louder voice
A final note – as well as chairing the first half of the second day’s conference session on 5G and IoT, I was also a judge at the Small Cell Forum Awards. It’s clear to me there’s a lot of expertise within the small cell community that is going to be increasingly relevant as mobile operators plan for both denser LTE-A networks and the architecture required in 5G.
It’s time for operators to stop thinking of small cells deployments as “special projects” and to start learning from what those designing and deploying small cells have already learnt. Of course some operators are indeed getting more integrated and this is very much something the Small Cell Forum is keen to promote – but certainly there would only be benefits to both sides if the expertise that small cells specialists have could be further exported into the main network strategy and standards bodies right now.
With that in mind, here then are the results of some of the discussions being had (at least with me) at Small Cells World Summit 2016.
SIGN UP FOR WEEKLY NEWSLETTER TO GET EARLY ACCESS
NOTE: Subscribers to the InsideTMN weekly newsletter could read the reports below last week. If you would like early access to reports such as these in future, sign up here.)
1. Nokia:
What most people want to know is how the company is plugging together the two radio units that came together at merger – Alcatel-Lucent and Nokia Networks – into one whole. In small cells the company already gave an answer back at MWC when it said that it would be keeping hold of the residential, SoHo and enterprise class small cells from Alu, and also going forward with its FlexZone distributed small cell solution.
Stephane Daeuble, Head of Smalls and HetNet Product Marketing at Nokia, said that in his view Alu’s solutions are the more classic “femtocells” – standalone small cells – where Al-Lu has had much more success than Nokia. These products are the ones being deployed by T-Mobile and AT&T, for example, and are built on the Qualcomm FSM99xx chipset.
Above these femto products sits Nokia’s FlexiZone range – which is a distributed small cell solution that combines small cells that offer macro-feature parity managed by a controller node. The controller node manages clusters of small cells locally – it could be hundreds – and presents them back to the core as if they were one macro cell. The latest big win for the FlexiZone architecture is in Chile, a deployment Nokia released details of during the Summit. Here, there are around 150 units across a shopping mall, 30 of them outdoor, serving 10,000 active users. Another recent deployment example that Nokia is proud of is at the Chinese Grand Prix, where the company this year deployed 95 small cells at 10-15 meter intersite distances, and served multi-camera live video locally from an MEC platform hosted at the site to race goers.
So what now? Well Daeuble pointed to increased multifrequency support, and noted that the FlexiZone product can support three bands, where one could be an unlicensed band for LWA or LAA support. The company is also a founder member of the MuLTEFire Alliance. Another product Daeuble said is attracting attention following its launch at MWC is the mini-macro, which is essentially a small cell form factor with the output power of a macro cell (2x20W). This is attracting use as an “outside-to-in” coverage solution, and in rural applications, Daeuble said.
2. Nokia’s FlexiZone is not the only distributed small cell solution addressing the large venue space. CommScope’s OneCell is another, and the company was talking up the capability its controller could provide to deliver non-interfering but overlapping sectors within a venue by building sectors within a dedicated slice (Physical Resource Block) of an LTE channel.
This means that a user or users could be served by 2-3 radio nodes creating a virtual sector operator at very low power. Essentially this is spectrum re-use to the max. The OneCell product still has just the one reference customer, announced last year, but Marketing lead Josh Adelson was positive that the solution was making good momentum in the market with operators interested in OneCell. The company’s standalone cell, the S1000, is being deployed by Sprint as a point solution, starting with deployment in its own retail stores, of which it has 4,500 in the US.
3. SpiderCloud is perhaps the best known of the distributed small cell providers, certainly one that has made progress within the operator landscape – notably with Verizon and with Vodafone in the UK and Netherlands. Art King, Director of Enterprise at SpiderCloud Wireless, said that Verizon had really torn into the deployment challenges, developing an “industrial” scale process so that all systems integrators are given the same design parameters, configuration and commissioning definitions. King said that Verizon is also looking at certifying the small cell clip on product that SpiderCloud and Cisco have collaborated on.
The company is also fully committed to the LTE-U, LAA and MuLTEFire approach, and King said that he saw MuLTEFire as a real option for multi-operator deployments – with the potential for a box to have multiple connections to an operator’s core network. The company is looking at trials in Q3 this year of and LTE-U product, and in fact specifications may have caught up sufficiently for that trial also to include LAA.
4. With so much talk of centralised deployments and C-RAN, Parallel Wireless, which won two awards at the SCF Awards, was expounding the benefit of a different architecture. If fronthaul latency and capacity are such an issue, then why not ride Moore’s Law and keep baseband capacity at the cell itself, it asks. This is what it is doing, and hopes to keep doing, allying low power small cells with its Het Net Gateway that acts as a host for virtualised network gateway and SON functions. “We don’t believe BBU and radio separation is needed,” CTO Rajesh Mishra says. “It creates additional overhead”.
Interestingly, the company is producing a variety of form factors, from an indoor WiFi/LTE solution to an in-vehicle small cell to outdoor, ruggedised versions.
5. Radisys announced its TOTALeNodeB+ platform during the show, and Tom McQuade,vice president and general manager, CellEngine and Trillium Software, said that the platform is principally designed to offer extensions to LTE-A features, from 3GPP R11-13. These included enhanced carrier aggregation, such as LTE-U/LAA, with 20 + 20 MHz features on single silicon. The product is due to ship this quarter, McQuade said. There are already 10 Chinese vendors sampling the software and test product in operator labs, he added.
Also in the pipeline from Radisys is support for virtualised RAN platforms, such as that proposed by ASOCs, a partner that Radisys announced a year ago. McQuade said there would be more announcements in this area this year, proving Radisys’ ability to “disassemble” its product into a vRAN cloud product.
McQuade also name checked support for further LTE-A features such as eICIC (enhanced Inter-cell interference coordination) and also (outside of 3GPP) for MuLTEFire. McQuade said the company would also be able to demo NB-IoT capability from Q3 of this year on Intel chips.
6. MuLTEFire is the standalone version of LTE Unlicensed. That means that a MuLTEFire cell is an LTE cell that can operate in unlicensed spectrum without needing an “anchor” in licensed spectrum. So… cable operators could offer LTE services, MVNOs and private enterprise operators get a means of controllable, high quality wireless access with mobility. Vendors get a new market to address.
But what’s in it for actual mobile operators that do hold spectrum licenses?
Neville Meijers, of Qualcomm and also representing the Alliance at this event, said that operators are usually initially sceptical but have grasped the opportunities. These are: the opportunity to access a genuine neutral host solution, and the ability to add a “secondary access” capability to their own networks, especially in the sort of building where it is not economical for an operator to deploy a dedicated system. Meijers said that mobile operators already exploit WiFi as a secondary access, so there is not big philosophical leap to make – with the advantage of decent QoS and mobility.
Meijers added that the Alliance is also investigating the use of the technology to access other unlicensed bands, he said, such as the 900MHz ISM bands – which could provide support for IoT devices.
For those who watch these sort of things, there was also a hint that in time MuLTEFire may end up being incorporated within 3GPP’s programmes. Of course the real reason the Alliance was established was because 3GPP decided not to go ahead with a stand alone LAA use case within R13. That left the technology falling between the two stools of IEEE and 3GPP specification. So the Alliance stepped in, but Meijers said he expected that “at some point” we would see this technology put into 3GPP.
7. Quortus CEO Andy Odgers said that the company was delivering on its vision of edge based intelligence that sits somewhere between the centralised cloud and the end user app – the “community centric” platform.
Although the company has joined the ETSI MEC group and is messaging quite heavily around MEC, you can think of the way Quortus designs edge intelligence as a sort of MEC+. “We are not non-standard,” Odgers said, “it’s just that there is no standard yet so we will push out and then feed that back into ETSI.”
One drawback of the typical MEC architecture that Odgers identifies is that MEC is currently targeted only at looking at packets on the User Plane (UP) – what he terms a MEC “bump in the wire” on the GTP (GPRS Tunnelling Protocol). Quortus’ edge core technology includes visibility to the the Control Plane (CP) via the MEC appliance, so that it can page devices, manipulate bearers, establish sessions etc. This allows it to do things like allow VoLTE voice, with offload to a local or hosted PBX. It also, by splitting the UP/CP architecture aligns it with the direction of travel on SDN. An integrator like ACS, with whom Quortus partners, then provides the orchestration and control via a controller appliance.
Still on Odger’s to-do list is establishing mobility between MEC appliances, adding WiFi integration and being able to combine the MEC appliance within an enterprise eNodeB, building a single all-in-one smart cell.
8. OPERATOR KEYNOTING:
(i) Telefonica O2 CTO Brendan O’Reilly didn’t offer many pointers to the audience in terms of specifics, but he did say that the biggest challenge he faces is delivering capacity in dense urban environments. But the inhibitors are not about technology, he said, instead he needs deployment solutions – for example to get 28GHz backhaul deployed, and to get small cells onto the street. He also said that the sector needs a dark fibre market to deliver on the Het Net promise.
Alongside O’Reilly on a panel, Tele2 CTIO Ervin Kampans said that talk of customer-centric networks was introducing a need for a new, application-specific KPI – for instance a Facebook KPI. Neither man was particularly sold on small cells as a foundation of neutral host models. Kampans said that DAS remained the answer here, and O’Reilly said that designing a small cell system to the needs of different operators would lead to a compromise solution that would only be worth it where “something is better than nothing”.
(ii) On the second day Vodafone Director of Innovation and Architecture Matt Beal delivered the soundbite of the conference, no doubt pleasing Vodafone colleague and Small Cell Forum chair Alan Law in the process, but stating that “All of 5G looks a lot like a small cell”.
The reasons for this are well practised, low latency demands, along with high throughputs achieved by the use of higher band spectrum, require a densely deployed Het Net. But getting there is – and this is also familiar – going to be a deployment and logistical challenge as much as anything. Beal said that 5G would actually require new deployment techniques, and also made clear that packet backhaul over dark fibre would be essential.
(iii) Dr Chih-Lin I, from China Mobile’s Research Institute said that the company would have live trials of a virtual RAN, featuring a software defined air interface, by 2018. Dr I revealed that CMRI’s recommendation, made 4-5 years ago, for China Mobile’s LTE rollout was that, from an air interface point of view, C-RAN should be the recommended architecture. However, central office was less convinced: with one 20MHz carrier with 8 antennas requiring 9.8Gbps fronthaul, China Mobile simply didn’t have the dark fibre that you’d need.
However, Dr I said that with new WDM technology extracting more performance from available fibre, C-RAN is now no longer about the fronthaul challenge, and offered TCO savings.
