The companies that sell equipment to test wireless networks are typically down in the trenches during wireless network buildouts. And that means they are privy to the unexpected challenges that might occur when operators deploy a new technology.
5G is no exception. From measuring 5G to 4G handovers to testing signal distortion, executives from several top network-testing firms say that 5G is so dramatically different from LTE that operators are spending a lot more time testing their networks than in previous network upgrades.
Interestingly, test equipment makers say that operators launching the non-standalone (NSA) version of 5G in millimeter-wave spectrum such as 28GHz and 39GHz are now considering moving to the standalone (SA) 5G version more quickly because of the complexity of non-standalone 5G. In the US, Verizon, T-Mobile and AT&T are all using millimeter wave (mmWave) spectrum and NSA for their initial 5G deployments.
"In the short-term, NSA gets you up and running faster," said Paul Denisowski, product management engineer at Rohde and Schwarz North America, one of the industry's biggest vendors for testing and measurement equipment. "But then it becomes a crazy, convoluted thing to support."
That sentiment was echoed by Sameh Yamany, CTO of test and monitoring firm Viavi. "We are seeing an acceleration in standalone," he said.
One of the big problems with NSA is the handoff from 4G to 5G because it introduces latency to the network. "A lot of carriers are realizing that in urban areas with high density, they will have handoff issues," Yamany said. The reason for the latency is that 5G networks that use mmWave spectrum have trouble maintaining a stable signal. If the network can’t find a stable 5G signal, it will push that connection to 4G, and that introduces latency. Indeed, those difficulties were recently highlighted by a report on 5G in South Korea by IHS' RootMetrics, which found that "upswitching" from 4G to 5G was adding significant latency to some users' connections.
And while the handoff issue with NSA may be fixable, Denisowski said that manufacturers are not sure how much time to spend working on NSA if operators are going to migrate quickly to standalone 5G. "As an engineer, NSA offends me because it is overly complicated," he said. "The other part is that people would like to get to standalone 5G because it is pure 5G."
Already T-Mobile has signaled it plans to launch standalone 5G as early as next year, and Verizon could launch SA 5G in a similar time frame. Such a move would pave the way for them to eventually shut off LTE networks; NSA 5G requires an LTE network to provide things like authentication, while standalone 5G does not require a supporting LTE network.
Another big issue with 5G are the spectrum bands that carriers are initially using for it. 3G and 4G wireless networks were always deployed in low-band spectrum, so wireless network operators don’t have as much experience working with spectrum in the mmWave bands where today's 5G sits. That may be causing some of the handoff and latency issues.
Russel Lindsay, senior product manager at Anritsu, another testing and measurement company, said that wireless networks deployed in millimeter wave tend to have intermodulation issues. For example, simple things like nails or unused wires can create signal distortion. Lindsay added that because of these intermodulation issues, operators need to figure out new modulation schemes for their 5G networks. "We try to work with them to give them as much information as possible about their signal," Lindsay said.
He added that different tools are required for 5G testing because 5G networks have very different requirements. For example, LTE networks could be tested more easily because an LTE radio sends its signal in a wide arc. With 5G, though, the signal can behave differently in different directions because it can use beam-forming technology, which directs connections to specific targets. "I could be 10 yards from the radio and not be able to measure it," Lindsay said.
This difference in how the 5G signal radiates also means that drive testing -- the decades-old method of testing the signal strength of a network in a neighborhood or city by driving around in a car -- may not be as important of a tool as it was in the past. Instead, network testing will likely be done by technicians wearing backpacks filled with test gear. "We home-brewed a backpack solution," Lindsay explained.
That situation is also due to the propagation characteristics of transmissions in mmWave spectrum. Such signals only travel a few thousand feet, unlike signals in low-band spectrum that can often travel several miles.
Another possibility is using drones to do network testing. Yamany said drones can be used to test signal penetration in high buildings that can’t be tested on the ground. “Testing the network in higher bands becomes more complicated,” he acknowledged.
— Sue Marek, special to Light Reading. Follow her @suemarek.