20Gbps Wireless Connectivity – Today

By Dudy Cohen

In the last couple of blog posts, we have described business and network scenarios in which ultra-high- capacity wireless connectivity is required. These scenarios include: connecting a large enterprise to an ISP service (this could be the primary connection or a fiber-backup connection); connecting aggregated customer traffic to a data center (again, as a primary or backup connection); and implementing fronthaul in a C-RAN environment where speed of deployment and cost-effectiveness call for a wireless connection.

Those scenarios are gathering traction as a result of two main trends. First is the massive move of enterprise businesses from on-premise servers and applications towards an XaaS model over a cloud infrastructure. The second trend is moving towards 5G services and use cases in which massive network densification is required and a cost-effective C-RAN scheme is mandatory. The near-term capacity target, therefore, is higher than 10Gbps and should be set at 20Gbps. It is obvious that this kind of capacity requires wide channels – much wider than those typically used in microwave links. This calls for the use of millimeter-wave communications – specifically E-Band.

In E-Band, wide channels are available, with channel spacing ranging from 62.5MHz (in some countries, even lower than that) up to 2000MHz, or 2GHz. The fact we cannot use channels wider than 2GHz in E-Band limits most current applications to 10Gbps, which could be achieved using a single 2GHz channel. While in the long run the use of higher frequency bands will allow us to achieve 100Gbps, the current need calls for accommodating 20Gbps over the E-Band spectrum.

This target is achievable!

Having said that, in order to make it a reality, it requires three specific capabilities. 

1. The implementation of cross-polarization interference cancellation (XPIC) over E-band. This allows reusing the 2GHz E-Band channel over the same link but in a different polarization, in order to double the link’s capacity. While it is possible to use two different 2GHz channels over the same link without implementing XPIC, this approach may result in the lack of spectrum in urban areas. 

2. A layer-1 carrier bonding engine. In order to distribute the 20Gbps traffic between the two 2GHz carriers efficiently, a layer-1 mechanism is required. An alternative “default” mechanism (e.g. LAG) will result in underutilization of the link due to uneven distribution of the traffic across the two carriers. 

3.A high-speed interface (25Gbps, 40Gbps or 100Gbps). This will create a single physical connection to the end device (e.g. router or switch), eliminating the need to distribute the traffic across multiple interfaces that may result, again, in underutilization.

At MWC 2019, in less than a month, Ceragon will demonstrate a new solution that does just that! The demo will show two radios, running traffic across a 2GHz E-Band link, at a rate of 20Gbps. One of the radios is connected to the end device over a 40GbE connection and distributes the 20Gbps traffic across its own carrier and the second radio carrier, at layer-1, resulting in a fully utilized 20Gbps link.

Watch our webinar to learn more about Disaggregated Wireless Backhaul

 Watch now