Network slicing is the technology of the future … right?
In Western markets, everyone talks about network slicing as a 5G feature that has not happened yet. It’ll be a big deal when we get the automation worked out and operators make inroads with enterprise customers, but that level of nirvana remains in the future.
In China, however, the picture is very different. Huawei and ZTE are implementing hundreds of network slices every month for China Mobile, China Telecom and China Unicom. More than 9,000 enterprises have some kind of special service today that support private use cases with SLAs that define their service.
A few customers outside of China are taking advantage of this also. Hutchison Drei Austria has implemented FWA in their 5G C-band network, using the same core network as their mobile service, with a “slice” that dedicates resources to each application. Drei now offers guaranteed throughput to their FWA customers and they believe that they can set up additional slices with SLAs for local enterprises within minutes. That’s a winner!
Purists can send their hate mail to [email protected] Yes, I understand that not all of these thousands of “slices” fit the vision laid out in your 2018 PowerPoint slide. Some of them rely on excess capacity in the RAN with a dedicated carrier or dedicated resource blocks, which is a crude and simple implementation.
My point is that enterprises don’t care whether your network slice is configured in the most efficient way. They simply want the assurance that they’ll get a guaranteed outcome.
In China, we now have time-sensitive networks for electrical grids and wide-area private networks for mines. Hundreds of manufacturing sites are using 5G now. Some of these involve on-premises radio deployment, but many do not. The vendors offer slimmed-down core networks that handle the private traffic on-premises so that even when the local tower is handling the traffic, everything gets routed back into the enterprise locally for fast latency.
There are still drawbacks. As I have reported extensively, Huawei is hamstrung by American sanctions on semiconductors and as a result, operators in China are deploying narrowband FDD base stations for 5G.
They’re giving the ridiculous argument that FDD is better than TDD because the latency can be lower. This is technically true, but clearly, the higher capacity of TDD will be much more important than saving 0.2 ms in the radio link.
Western network vendors will catch up and I believe that, in the end, the Darwinian approach to development is likely to surpass this early success in China.
In China, we have a simple model where the (state-owned) operators handle all of the traffic for the enterprise and route the traffic to on-prem equipment in some cases. It’s a rigid model that won’t fit most enterprises in the West.
Instead, some bright people will develop a simple private network with integral edge computing and the key will be open RAN software that makes the network cheap and easy to adapt for 5,000 different unique business apps.
I have heard many times that Chinese RAN equipment is superior to Western equipment. But when you get into the details, it becomes clear that the rigid top-down approach does not result in the best outcome for the end users — or even for the telcos.
Here’s an example: China will deploy 2.2 million 5G radios this year, but will only add 190 Tbps to their radio network at a cost of $8 billion. Western operators will add 850 Tbps of capacity for $13 billion in RAN spending this year, so I conclude that Western telcos are roughly 3X more efficient than China Inc. in 5G network deployment.
The West uses a hard-nosed Darwinian approach to technology choices, and it results in better solutions in the end. You can see the details in our strategic forecast of the RAN.
This race is important. Where the “Race to 5G” was essentially meaningless because China’s operators don’t compete with Western operators, the “Race to Private Networks” is the key to a future trillion-dollar market. The Chinese have an early lead, but I’m betting that a Western startup company will emerge with the best approach in the end.
Original article can be seen at: