5G: a peek under the hood

Dimitris Mavrakis/Ovum
22 Jul 2015
00:00

5G is a hot topic, but confusion reigns on all of its aspects: technology, commercial opportunity, application in verticals, and overall timeline for deployment. This confusion is accentuated by the fact that most mobile operators still haven’t devised a successful way to monetize their LTE networks.

But if we look to the past, can we identify what the next step towards 5G should be?

A new mobile network generation typically refers to a completely new architecture, which has traditionally been identified by its radio access: from analog to TDMA (GSM) to CDMA (UMTS) and finally to OFDM (LTE). Clearly, 5G will require a new technology and a new standard to address current subscriber demands that previous technologies cannot answer.

However, given current trends in traffic growth, 5G necessitates a complete network overhaul that cannot be achieved organically. Software-driven architectures, fluid networks that are extremely dense, higher frequency and wider spectrum, billions of devices and gigabits of capacity are a few of the requirements that cannot be achieved by LTE and LTE-Advanced.

5G technology candidates

The requirements of a 5G network have been identified and are outlined by the 5G Infrastructure Public-Private Partnership (5G-PPP). Current technology developments and user demands merely provide a glimpse of the nature of 5G networks. At the moment, cost is not a major driver of 5G technology discussions, allowing a wide list of candidate technologies to be considered. Rather than addressing cost, this article outlines the radical, disruptive, and required changes to the existing network.

Clearly a new air interface will be needed. ZTE has gone as far as to suggest that a 5G network will consist of several air interfaces coexisting in the same network. From a theoretical perspective, this is ideal (e.g. OFDM does not lend itself to small cells and hetnets, but other interfaces do), but from an operational and economic perspective, this would mean significant development costs and deployment effort.

Let’s examine a selection of technology candidates.

Extreme densification: As soon as 3G networks became congested, mobile operators realized the need to introduce either new cells or more sectors into the system. This has evolved to include many flavors of small cells, which essentially move the access point much closer to the end user. There is simply no other way to increase the overall system capacity of a mobile network significantly. 5G networks are likely to consist of the several layers of connectivity that hetnets currently suggest: a macro layer for lower data speed connectivity, a highly granular layer for very high data speeds, and many layers in between. Network deployment and coordination are major challenges to be addressed here, as they increase exponentially with the number of network layers.

Multi-network association: Several networks currently provide connectivity for end-user devices: cellular, Wi-Fi, millimeter-wave, and device-to-device are a few examples. 5G systems are likely to tightly coordinate the integration of these domains to provide an uninterrupted user experience. However, bringing these different domains together is a considerable challenge. Hotspot 2.0 and Next Generation Hotspot are perhaps the first examples of cellular/Wi-Fi integration. Whether a 5G device will be able to connect to several connectivity domains remains to be seen, and a major challenge is the ability to successfully switch from one to another.

Full duplex: All existing mobile communication networks rely on a duplex mode to manage the uplink and downlink. There are frequency duplex or FDD schemes (such as LTE, where uplink and downlink are separated in frequency) and time duplex or TDD schemes (where the transmitter and receiver transmit at different points in time, as in TD-LTE). A duplex mode is necessary to coordinate uplink and downlink, but full-duplex technologies are now being discussed. In these schemes, a device transmits and receives at the same time, thus achieving almost double the capacity of a FDD or TDD system.

This approach would entail major technology challenges - requiring what is essentially self-interference cancellation - and major changes in both networks and devices. However, the potential increase in overall capacity is substantial.

Related content

Comments
No Comments Yet! Be the first to share what you think!