with Mike Murphy, CTO of Nokia Networks in North America
Murphy (Source: Nokia)
Mike Murphy, chief technology officer for Nokia Networks' business in North America, is an industry veteran whose experience includes research and development in WCDMA at Nortel Networks. His tenure at Nokia began in 2006, when he served as country manager in Japan, later becoming head of technology for Asia Pacific before taking on an expanded role in the Asia, Middle East & Africa region. He was heavily involved in the first LTE deployments in South Korea and Japan. Early this year, he landed in the United States, making this his eighth country of residence. He recently talked with FierceWirelessTech Editor Monica Alleven about 5G, Wi-Fi, small cells and the disruptive nature of the open cloud environment. Following is an excerpt from the conversation. It has been edited for clarity and brevity.
FierceWirelessTech: Can you talk a little about your experiences overseas and your observations on technology rollouts in general?
Murphy: The sequence we see is Korea starts things out--call them an early adopter so to speak--and then Japan is a very close second, and I would have said before [this conversation], the U.S. follows, but now I think the U.S. is picking up a lot and we're seeing a lot of aggression in the U.S. in terms of technology advancements.
FierceWirelessTech: Where are you seeing it in the U.S., in terms of what kinds of technology?
Murphy: For example, Sprint has a lot of spectrum, 150 megahertz for TD in 2.5, and by nature of having that, they're able to do more combinations for carrier aggregation, like 2 x 20, or 3 x 20 and eventually, 4 x 20. They'll probably push the envelope there, perhaps even globally in carrier aggregation for TD, and they also have FD, and it remains to be seen, but they may be one of the first in the world to do TD-FD aggregation in some kind of scale. AT&T pushes the boundaries on cloud computing. By a long shot, they are the most aggressive on the whole movement toward cloud, so the U.S. is starting to lead in a few places. Now if we had this conversation in Korea or Japan, they might think differently (laughs).
FierceWirelessTech: When will 5G technology hit the U.S. market and how will it get rolled out?
Murphy: There's no perfect answer to that question. It's still early days, but there's a few data points along the way that give us some hints. For example, the two clearest ones are the 2018 Olympics in Korea and the 2020 Olympics in Japan. Especially given my background, there's absolutely zero doubt in my mind that there will be some showcase of 5G in both cases. Particularly, the Korean government, there's an element of showcasing the country and national pride, and because they have the handsets--No. 2 and No. 3 in the world, Samsung and LG--they have the ability to do some things perhaps faster than is otherwise possible.
There will be some element of 5G. I was talking to analysts a few weeks ago and we had a big debate, do you call it 5G, is it really 5G? I guess it remains to be seen… I'd say 2018 is the extremely aggressive scenario, 2020 is the more realistic [scenario], partly because of two things. One is the standards bodies, in the 3GPP, R-15 is seen as the first release of the 5G spec, and that comes out around 2019, and then there's the WRC 19, and WRC 19 is seen as the time when 5G recommendations for global spectrum usage will come out, and that's also 2019. That's kind of how the dates line up on a global basis.
Then you come to the U.S., the FCC is doing a Notice of Inquiry, asking for input from the industry on spectrum usage, and the guess is between two to three years after they get the input, which I think is due sometime in December, we would see that the FCC would start allocating spectrum for 5G. These dates are quite loose, it's hard to pinpoint. You see all the dates hovering around this 2018, 2022 kind of range. I've heard some other vendors say 2022, but the truth is nobody can really pinpoint that date. There are a lot of variables, but broadly speaking, it's in the 2020 range.
FierceWirelessTech: How should Wi-Fi be incorporated into operators' macro networks? What is the role of Wi-Fi?
Murphy: Wi-Fi plays a role, always has and always will. It's another air interface. Today, Wi-Fi from the Ciscos and Arubas, HP, is just using unlicensed Wi-Fi as a stand-alone, and there's a new standard coming out called LTE Unlicensed, which is used for something called supplemental downlink, which is basically, you have normal cellular spectrum, and then you add Wi-Fi to get more downlink to improve your capacity by merging the two together. That will come, sometime around the end of next year perhaps, or early '16.
Wi-Fi is going to continue playing a role and the way to use it is evolving a little bit too, so this is one evolution, bringing it in as a complement, let's say, to LTE, and then there's HotSpot 2.0, which will facilitate authentication, which is probably the biggest trouble at the moment, logging into an SSID and having a password, so there will be HotSpot 2.0 aiming to simplify all that, so you kind of bring roaming, including global roaming, to Wi-Fi just like what you have on cellular right now. It will get easier and maybe more [similar] to cellular, so it's not going away anytime soon.
The other thing is in some of the new spectrum auctions, different non-telco players are pushing a little bit more for shared spectrum. For example, in the 3.5, Google and others are being a little more aggressive to say, well, why don't we make that shared or even unlicensed in some of the other frequencies, and you could say those are potential Wi-Fi candidates too, maybe, for the future unlicensed spectrum.
FierceWirelessTech: What about small cells?
For small cells, you have to separate, at least in my mind, there's a bit of a distinction between indoor and outdoor. In outdoor, it's largely to fill coverage holes or add capacity to an existing network and let's call that carrier small cells, for lack of a better word at the moment. Here, there's been a lot of large forecasts in the past, [but] it's proving more challenging than everybody thought for permits, for the cost of a site--and in particular in the cost part--getting the backhaul connectivity, so all of that has kind of slowed it down from what a lot of analysts and ourselves have forecast in the past.
We need to kind of get over these hurdles to make that go faster for the outdoor, which is one of the reasons we bought SAC, who kind of specializes in streamlining this whole activity. Just to give you one data point, we're deploying small cells in Chicago, and the guys were saying it was taking up to 120 days just to get the permit, and then in another case we're looking at the fiber backhaul for small cells, and the numbers were equaling the costs of macro sites almost, so it's very challenging until the whole industry solves this.
Small cells indoor is kind of a completely different beast. You could probably divide indoor into two parts as well. One is just enterprises like Nokia, like IBM, just the enterprise itself, and most enterprises have Wi-Fi today, so bringing in cellular is to add, in particular voice coverage, and maybe data as well, and in that case, then let's say you have a Wi-Fi AP sitting in your office. The ideal is you can put up a new small cell and have both your cellular and Wi-Fi together in one new box, and it's still a little bit the early days for that as well because you have to look at how you integrate that new box into the whole enterprise's system. Most enterprises have security and firewalls and authentication, so on the enterprise side, we have to help streamline that part of it.
The other indoor part is not so much enterprise, but public indoor, stadiums or Starbucks and so on, and perhaps that's a little bit more straightforward. So there's a few different use cases for small cells, each with their own challenges, and I think everybody's kind of working on these, and perhaps we'll overcome them.
FierceWirelessTech: Operators say they want non-proprietary open solutions. Is it difficult moving from a proprietary-based solutions world to this open cloud environment?
Murphy: In the past, in the core network, we'd sell the hardware, we sell the software, we do the … soup to nuts, end-to-end--this all comes from the vendors like us. When you go to cloud, the current view is that most operators will buy the hardware themselves, so we don't provide the hardware anymore. Then there's virtualization software, like from VMware or various OpenStack versions, and it may be that the customer buys that themselves as well, so it buys the hardware, then a base layer software, and then they come to us for the VNFs, which is the main telco application, so all of this is enabled by that NFV framework. That's normally what we think about when we talk about going from proprietary to open, and it's open if you follow the NFV framework architecture, and if you don't follow it, you're not open.
Coming back to your question, is it difficult? Certainly, it's difficult, it's a new way of doing things and it's very disrupting for both the customer and the vendor. In our case, we go away from perhaps delivering the hardware, then you have to understand what are the performance requirements the customer has? On the customer side, if they're buying these things piece parts, then they have to be very wary of how do they do systems integration, how do they keep their SLAs (service level agreements) and KPIs (key performance indicators). So it's quite challenging for the industry and absolutely, I think we're at the early stages where not everybody fully understands the impact of going this way. When you did proprietary end-to-end, you had one person to go to [when] you have a problem with performance. The term I use is going from a vertical structure to a horizontal structure. When you go to a horizontal structure, you have more players and it's more difficult to control the end result.