At 5G World the exhibition floor was free from your usual suspects such as Huawei, Ericsson, Nokia, Qualcomm and Intel. Bad news for InformaTech’s sales bookers perhaps, but good news for the sort of vendors that normally make up the “supporting cast” at these events. A brief walking tour showed that for many of these companies, 5G is very much at the go-to-market phase.
One of those that took the chance to play up its part was Mavenir, which combined the largest amount of floor space with an analyst event the day before the Conference and a couple of speaking slots.
As you may expect, Mavenir was here to talk vRAN and the path to Cloud Native in the core of the network. One interesting solution it was able to talk about was a vRAN-based in-building design that can support multiple operators on the same physical infrastructure.
Mavenir has been working with BT Enterprise’s wholesale unit on the offer for the past year. It says that the issue of getting coverage indoors is increasingly important to MNOs and will only become more so as 5G introduces higher spectrum bands. BT’s estimate is that operators may need four indoor small cells for every outdoor macro – meaning each operator could be looking at 160-180,000 indoor small cells by 2025. That means that a small cell has to come down to about 10% of the price of an installed macro – hence the reason to offer a wholesale offer that shares infrastructure.
The solution uses virtualised Central Units (BBUs) connected to small cells (acting as Distributed Units) from Sercomm, both running Mavenir radio software. The small cells can operate on different frequencies per operator, or can create a single frequency network with shared instances.
The radio function split between Central and Distributed Units is based on the 3GPP-defined Split 2 Option. Split 2 is a high level split that keeps quite a bit of the radio functionality in the distributed unit. This means that the latency budget on the fronthaul between the CU and the DU is not as taxing as lower layer splits – where more of the processing is centralised. BT’s Vickery said that running this split gives 10x the fronthaul efficiency from the radio to BBU, meaning BT can aggregate multi-operators on the link, giving it more efficient transport.
John Vickery, from the CTO Office in BT Enterprise, said that the latency budget enabled by the CU-DU Split 2 meant that CU could be present in just eight locations in BT’s network estate. That would give it enough presence to cover the network nationally, at launch. The solution can be built using BT’s Network Cloud, for which it just announced Juniper as a key supporting vendor.
The idea of the solutions is that because the CUs are virtualised, operators are able to apply their own QoS and security parameters, as well as run different versions of Mavenir’s radio software independent of each other. That means that operators can run different upgrade paths to each other, which means that they can have different go to market strategies even over shared infrastructure. Failures are also fenced off, so that if one operator’s service is interrupted for any reason, other operators on the infrastructure are not affected.
Speaking of mmWave, one company offering a new RF chip (RFIC) and mmWave antennas was SiverSima. You can think of the company as sort of a boutique competitor to Qualcomm. So how is it going to take on the market? It’s betting on offering an optimised performance to vendors looking to build very high quality product. Operators want to be able to support the maximum number of users from an access point, the company said, and therefore having high performance CPE means that they improve the economics of each site.
A live demo between two prototype units running the chip – 16T/R beam forming transceivers – was achieving 2Gbps data rates at 28GHz, with 64QAM and 12 khz subcarrier spacing.
Test company Mentor Graphics, part of Siemens, was there to talk about testing the fronthaul link between the BBU and RRH, or the midhaul link in the split BBU CU-DU version, and finally between the radio and the MIMO antenna. The aim is to give the network equipment vendors the ability to test new CPRI, eCPRI and O-RAN protocols as they bring new product to market. Test kit like this being demo’d at large-ish events is always a good momentum indicator.
Another test company, Viavi, was showing off a nice three stage progression path for 5G testing. It moves from lab, to in-field commissioning and acceptance, to ongoing monitoring and assurance.
The lab test on demo was a “surround” test of an AMF – a new 5G Core function. Viavi emulates the base station end and the core network, introducing errors and stress testing the function under high loads.
The integrated, three-step approach, is a result of an ongoing acquisition policy that has seen it merge together smarts from different vendors, from test kit to SON engines, and there’s now a new one to add to the list – 3Z.
3Z markets base station test equipment that, amongst other things, enable installers to make sure they have got their installation correct. One a separate booth, the company was showing off an antenna alignment tool that uses GPS and an on-board camera to tell an operative if their antenna is aligned in the right direction. It also generates an acceptance report back to the operator. It’s all part of a drive to get the overall cost of deployment down, in this case by reducing errors and re-visits, but also speeding up the acceptance process.
Another company you could bracket under that heading is Vricon. Indeed its booth imagery asked, “How do you plan large-scale networks while keeping down costs?” It’s marketing very high resolution data and 3D imagery of buildings to those building out network plans using the usual RF network planning software tools. Two of those companies, Ranplan and iBwave were also in the room. Hopefully they all met for a coffee.
Start up ORI is putting together an “edge enablement platform” that will let developers get their applications into the network edge, independent of who owns, operates and defines that edge. The idea is to act as a kind of bridge between the network platform and the application developer, giving developers a service creation, deployment and orchestration environment that will work across hybrid network edges. Operators will not all run the same technology stack at the edge, Ori said, and that gives developers an issue.