5G Asia: Spectrum is still the hardest part of going 5G

Image credit: Nerthuz / Shutterstock.com

Amid the rich variety of discussion during the keynotes and tracks at 5G Asia, spectrum was a recurring theme – namely, which 5G frequency bands the mobile industry will be able to harmonize around, whether they can get enough of it, and how implementing cellular connectivity on millimeter-wave bands is really hard.

Fumio Watanabe, chairman of KDDI Research, remarked during one panel session that the biggest challenge in rolling out 5G is getting the spectrum license, although the difficulty for Japanese operators specifically is waiting for the regulator to finalize the technical conditions for new 5G spectrum bands under consideration – 3.7 GHz, 4.5 GHz and 28 GHz – to include how much spectrum to allocate for each band.

“This is scheduled to be finalized by the middle of 2018, so the earliest we can get the license is the end of 2018,” he said.

That said, Watanabe added in a presentation later that another spectrum challenge has to do with propagation of radio signals in the 28-GHz band. For example, a common issue is compensating for multipath and NLOS issues, as buildings, cars and even human bodies that get in the way of the signal can cause propagation problems.

KDDI found that a 28-GHz signal that encounters a human body or even a tree will experience degradation 10dB higher than a signal broadcasting at 2 GHz. The more people or trees in the way, the higher the signal loss. That naturally has an impact on coverage, which means operators must deploy more sites to compensate. Takehiro Nakamura, VP and general manager of NTT DoCoMo’s 5G Lab, said his company had explored similar issues with 28 GHz.

Such issues aren’t insurmountable, of course – they’re just engineering challenges that cellcos and vendors will eventually overcome. Even so, Richard Tan, advisor for Sinar Mas Group (which owns Indonesian TDD operator Smartfren), said that he preferred sub-1 GHz bands for 5G in Indonesia.

“The higher up the band you go, there are significant challenges to utilizing it,” he said during a panel discussion on spectrum. “Also, in Indonesia the 2.5 GHz band is classified as S-band, which is used by satellite operators for broadcast services.”

However, he did add that the regulator will ultimately decide what bands cellcos will get, “so we have to use what we’re dealt.”

5G spectrum panel
Lots to talk about at the 5G Asia spectrum panel (from left): Chee Kheong Foong, head of Regulatory Affairs at Axiata Group; Dr. KY-Leng, deputy director general of Cambodia’s Ministry of Post and Telecommunications; Dr. Patrick Tsie, senior director of Technical Marketing at Qualcomm; Dr PS Tang, steering committee member and chairman of the Business Model and Funding for TDD Operator Task Force at GTI; and Richard Tan, advisor for Sinar Mas Group

Spectrum prices and sleepless nights

In any case, technological challenges aside, there’s little doubt that 28 GHz is shaping up to be a key band for 5G. Jake Saunders, MD and VP for Asia-Pacific and Advisory Services at ABI Research, noted that while official designations won’t happen until WRC-19, operators and regulators do seem to be settling on the mmWave bands at 24-28 GHz as a starting point for 5G, followed by the 3.5, 3.6 and 3.8-GHz bands.

Saunders added that ABI is also seeing quite an amount of trial activity at 70 MHz, which is essentially WiGig territory.

One issue facing operators is, of course, having enough spectrum to enable 5G to do what it promises on the tin (10-Gbps speeds, 1ms latency, massive connectivity of things, etc). Saunders pointed out that while the Global mobile Suppliers Association (GSA) recommends that operators ideally be allocated 100 MHz of contiguous spectrum for 5G, they’re unlikely to get it. Meanwhile, 5G standardization efforts are focused on deployment scenarios utilizing anywhere from 20 MHz to 100 MHz worth of spectrum.

Consequently, he said, carrier aggregation will be a crucial technology for helping cellcos deploy 5G services, as well as massive MIMO to squeeze as much capacity out of each band as possible.

A related issue is spectrum pricing – even if regulators are willing to allocate massive amounts of spectrum, operators in highly competitive, low-ARPU markets may not be able to afford it, depending on governement spectrum policies. As an example of how not to do it, India set the reserve prices for last year’s 900-MHz auction so high that cellcos refused to buy any.

Tan of Sinar Mas said Indonesia faces a similar problem. “The situation in Indonesia is that operators are selling data at 40 cents to 80 cents per GB. That is clearly not sustainable. If you look at the situation in emerging markets, we need bandwidth to be able to invest in network capex. It’s not just equipment – the cost of rolling out is also significant. I was just reviewing the cost of rolling out small cell networks in a building, with the planning and rolling out the wiring – it’s not cheap. Whatever the situation is, we can’t roll this out and continue to sell data at 50 cents per GB unless everyone starts chewing up 100 or 200 GB a month.”

Under market conditions, he said, operators will be in no mood to bid for 5G spectrum. “Operators have to work with the regulator under some kind of partnership model, where maybe it’s revenue sharing or something. But you need to have dialogue with the regulator.”

“Pricing is the one thing that keeps me and my team awake,” commented Chee Kheong Foong, head of regulatory affairs at Axiata Group. “Some regulators like in Cambodia can strike the right balance – others will set the reserve price at unreasonable levels, and it’s doomed to fail.”

Be the first to comment

What do you think?

This site uses Akismet to reduce spam. Learn how your comment data is processed.