The role of 5G is to “disrupt existing industries and create new ones”, Sandra Rivera, Senior VP of the Network Platforms Group, Intel, said at the company’s 5G Summit prior to MWC Americas. She was talking about the consumers of 5G services – the media companies, industrial manufacturers and so on who will be challenged by and spurred on by 5G capabilities. But if this vision is true, it is one that cuts back on those who would supply the platforms for this disruption.
Take Intel itself. As Rivera presented the end-to-end nature of Intel’s engagement around 5G, she ran through Intel’s presence in the network, through the edge and core cloud, to the client. Across this ecosystem Intel hopes to be the enabling foundational technology partner for those building the 5G platforms of disruption. At the event Intel presented with Nokia and Ericsson, noting its presence in Nokia’s AirScale and AirFrame platforms, for instance. But note too that it partners with “new” entrants such as Mavenir, utilising its FlexRAN Reference Architecture – as loudly as it does with an established vendor. It works with and contributes to Open Source projects and it partners with a customised end-to-end solution provider such as Nokia and its Reef Shark chipset. Disrupting old businesses, creating new ones.
Rivera runs through it: from the device, through the network to the datacentre, cloud and AI.
Intel plays in the distributed platforms that host virtualised network functions and the edge AI and analytics platforms that 5G could utilise. Rivera : “By moving the compute much closer to the source of data creation and consumption. By moving it closer, businesses are able to perform more intelligent analytics and turn all that aggregation of data into valuable insights.” It is in the virtualised RAN – witness the market-making it is engaged in via its aforementioned FlexRAN reference platform – and it is in the centralised data centres.
And it is also in the devices themselves.
And here it is worth acknowledging that Intel has one big problem when it comes to 5G. It is being outgunned by Qualcomm – at least in terms of messaging – around market readiness for a 5G device modem and chipset. Intel would not acknowledge this, of course, but from the outside in terms of market announcements and perception, it is true.
Intel is sticking to its line that it will have a market-ready device modem for 5G NR by mid 2019. Until then, it is making its FPGA-based mobile test platform (MTP) available to operators and device makers to test against. On its face that puts it behind Qualcomm, which is claiming it already has a chipset that can power test devices at smartphone form factor, and that will ship within commercial product in early 2019.
So what does this divergence tell us about how operators themselves are approaching 5G, and how does it impact upon Intel’s wider 5G story?
Rob Topol, GM for 5G & Technology, Communications & Devices, Intel, pointed out to TMN that Intel sees the 5G opportunity as a wider than just the device market, and more than just the eMBB case. This is mirrored by the way in which operators will migrate to 5G. Some will go early with limited use cases, whilst others (especially in Europe, perhaps) see 5G as more of an opportunity to create new B2B and B2B2C business models.
“We know there will be a wave to be first” Topol said, “But others are working on a broader scale that goes beyond a smartphone transition.”
Topol also reminded TMN that Intel did “bring up the first single mode baseband chipset for early 5G, in a pre-standards iteration”, but that it had decided to keep the flexibility of the MTP because its strategy is “not just about bringing up R15 and eMBB for 5G”. That goes back to 5G being a platform for other use cases, such as low latency communications, so Intel wanted the flexibility into R16 and 17 to “keep the corridor of wireless evolution happening.”
“The reason there is no chipset today is that we wanted to commercialise a solution where 2-5 modes were integrated a single integrated platform.” And Topol says this maps to likely scale deployments of 5G. “The plan is to target the timing for broader deployments -. primarily based on customer feedback for when to scale.”
Qualcomm, for its part, disputes that it is closing itself off to future evolution by moving to an initial ASIC-based mmWave and sub-6GHz solution.
Danny Tseng, Technical Marketing, Qualcomm, said, “This is productising R15 for the smartphone form factor and power budget. We will continue to develop the FPGA platform in parallel, keeping it flexible and adding new resources before R16 is productised.”
Mohammed Al-Khairy, a senior product marketing manager in Qualcomm’s chipset team, said that the importance of moving to a solution that can fit in a smartphone form factor is that it enables device manufacturers and operators to test the performance of mmWave and 5G NR in a more standard sized device.
“The most important aspect of the smartphone form factor is that people said there was no way mmWave could happen on a mobile platform, or that it needed LOS and so on. In the X50 the baseband modem connects up to four mmWave modules – one on each side of a phone – solving propagation challenges sized into a module that can go into a smartphone.”
In essence then, Qualcomm claims the lead on the device chipset and mmWave module integration and says it is crucial to get that worked out in a reasonably real world manner.
But Intel (whilst still claiming the first 5G modem in any case) says that its platform will be ready for when the bulk of operators are addressing mass market services.
Most importantly its strategy is that it will play in a far wider way, in network and in cloud core and edge – enabling the full range of 5G use cases and driving that creative disruption that 5G is meant for. It will have to ride out the early spate of launch announcements before its device modem is shipping commercially. Can its end-to-end message cut through?
