This Q&A was originally published on webscalenetworking.com. For more on the future of 5G networks and fiber, watch Brian Lavallée’s webinar here.

I keep hearing that 5G will require millions of small cells. Why is that?

Brian:  Higher frequencies expected to be used by 5G do not travel as far as lower frequencies used today, and also have a harder time penetrating walls, bridges, trees, or anything physical. To get around that, MNOs will deploy more cells that cover smaller areas serving fewer people albeit at much higher speeds, which will drive the proliferation of small cells.

Will 5G smartphones support 10Gb/s?

Brian: Initially, probably not. 5G is expected to support a minimum of at least 100Mb/s yet be capable of speeds as high as 10Gb/s. The mobile infrastructure won’t be able to support hundreds of millions of people accessing the network at this top speed, at least early on and this is fine in most situations. Today, for example, watching video is the most bandwidth-intensive smartphone application, but it doesn’t consume anywhere near 10Gb/s because there are only so many pixels in a smartphone screen to populate. With that bandwidth, who knows what future applications will be possible.

Will the market for 5G backhaul services be the same as it is for 4G?

Brian:  In North America today, most 4G MNOs purchase backhaul from third-party service providers. This market will likely continue for 5G, albeit at much higher rates. I believe it will be increasingly done using dark fiber rather than packet-optical-based services, at least for the higher supported rates.

Will the Internet of Things (IoT) generate video traffic?

Brian:  It depends on what you classify as a “thing”. If it includes HD surveillance camera, for example, that runs on a cellular network, then yes, and lots of it. Smaller things, such as temperature sensors, will generate far less traffic, but there may be billions of them deployed, so it adds up quickly. For the most part, the challenge of IoT will likely be about the number of individual services, not capacity.

What’s driving the spike in mobile video traffic?

Brian: The answer is probably right in front of you! It’s mobile devices with HD cameras for one, that are feeding content into YouTube and Facebook, which are, in turn streamed to mobile devices more than anywhere else. When Facebook enabled auto-play on its videos it drove a massive increase in mobile bandwidth consumption overnight. Other sites have replicated this model, so I don’t see the trend slowing.

How will 5G change the wireline architecture that currently supports 4G mobile backhaul?

Brian:  That will depend on MNO’s rollout patterns, how many users are on the network, and what limits they’ll set on speeds. Today MNOs typically feed a 1 Gb/s Ethernet line to a 4G cell sites, of which an average of 200-300 Mb/s is being used. With 5G, that same macro tower may need 10’s to 100’s of Gbps, which will require an enormous capacity upgrades to those towers.

Will fiber deployment patterns change when 5G is deployed?

Brian:  Direct communication between cells is planned for 5G networks, skipping the connection to the packet core wherever possible, which would offload a lot of traffic. Those cells might be wired together directly in the shortest path, but I think traffic will mostly home back to a central location. 

Should a service provider now planning to start a 4G deployment hold off until 5G?

Brian: Yes, as there will be significant and time-consuming upgrades to the existing 4G wireline network that will ease 5G deployments, as soon as the latter technology arrives. There are many things that MNOs can do today, such as deploy more fiber to small and macro cells, that are compatible with both 4G networks of today, and 5G network of tomorrow.

Will the 5G backbone be physically and virtually separate from the 4G backbone?

Brian:  Not physically. If, for example, a service provider has a 4G cell and wants to add 5G radios to that macro towers, they would likely end up sharing the aggregation and core network back to the data center, perhaps over different wavelengths or different parts of the network, L2/L3 VPNs, or Optical VPNs. Rolling out completely separate networks for both it would become cost prohibitive quickly and much harder to get to ROI. Some parts of the network will be only for 5G, and some shared.

Will there be a relation between 5G and WiFi?

Brian:  Yes. The Heterogeneous Network (HetNet) that is expected to achieve the require geographic coverage for 5G will include a wide variety of cells such as pico, femto, micro, small, and even WiFi, which all serve a smaller area than traditional macro cells via spectrum reuse. People will still want to use WiFi to avoid affecting their cellular usage caps, with many of those coffee-shop WiFi access points being owned by MNOs.

How will 5G latency be lower if processing is being done in the cloud?

Brian: Latency in the case of 5G refers to over-the-air latency, between the user device and the radio access network radio. To reduce overall latency, service providers will need to put data much closer to end-users, which is why some MNOs are considering leasing space at their cell sites to host data and applications for content providers, particularly in large urban centers. Ultimately, the placement of data and processing will be determined by the supported 5G application. The Mobile Edge Computing (MEC) initiative will play a major role in addressing this important challenge.

What impact will 5G have on backhaul networks?

Brian: Data received at the 5G radio located in a cell site won’t go anywhere without fiber. That’s why many large MNOs are buying large fiber footprints. They know that fiber and its geographic footprint will ultimately dictate the performance and commercial success of their 5G services and applications. 5G will need fiber, and lots of it.