Prior to the iPhone launch in mid 2007, AT&T hurried to rollout a major upgrade of its 3G mobile data service, in anticipation of a tenfold increase in network traffic. Appetite for mobile data and the number of smartphones and data-centric devices has only increased since then. In fact, data traffic over cellular networks is expected to grow almost 40-fold till 2015, and UK firm Coda Research anticipates that in the US, mobile video will account for over 60 percent of all mobile data usage. But can backhaul networks support such staggering capacities?
Network operators, utility companies, public safety organizations and enterprises are all struggling to meet the immense demands for data that are required by today’s applications. At the same time, they must plan and prepare for the continued growth.
Fiber Where You Can – Microwave Everywhere Else
When examining the wireless vs. wireline alternatives, fiber’s nearly unlimited capacity immediately stands out. However, the fiber option is not always practical. Whether due to deployment restrictions (rough geographical terrain) or regulatory restrictions (dense metropolitan areas) laying out fiber infrastructure may be too costly and time consuming. In such cases, wireless, or more accurately - wireless backhaul based on point-to-point microwave, emerges as the best solution.
Microwave backhaul solutions are capable of delivering high bandwidth, carrier-grade Ethernet and TDM services. Microwave is suitable for all capacities up to several Gbps over a single link - and may be scaled up to multiple Gbps using aggregated links techniques. Unlike fiber, wireless solutions can be set up quickly and are much more cost efficient on a per-bit basis from day one.
The End of the Fat Pipe Era
Today microwave backhaul offers much more than fat pipes connecting two endpoints. Microwave has evolved over the years and has accumulated advanced service features such as service and network topology awareness - features that until recently were only available using expensive external boxes for switching and routing traffic.
As networks become more complex, and user experience and quality of service become major differentiators between operators, the role of microwave systems within those networks is changing. Microwave backhaul should be much more than a “dumb pipe.” It must be smart.
The capacity characteristic of today’s traffic requires different attributes from the legacy install based backhaul. One example could be the peak-to-average ratio. A 4:1 peak-to-average traffic ratio is not uncommon in backhaul networks. The microwave solution must be traffic aware in order to manage multiple applications with differentiated quality of service levels. For instance, voice calls consume relatively low bandwidths and require high priority with minimal latency. Web browsing or ftp downloads on the other hand, require high data volumes, but the user’s quality of experience is less affected by latency issues.
Moreover, to ensure quality of service, a smart microwave backhaul solution should also be aware of network topology. It should integrate carrier Ethernet functionality and be able to independently re-route traffic in case of network failure.
Can Microwave Really Support 4G Backhaul Capacities?
The answer to the question whether microwave can cope with future capacity requirements is a simple “yes.” Microwave can support up to multiple Gigabit Ethernet over a single link, and real-life performance really depends on the available frequency resources, as depicted in the table below. Advanced microwave supports high spectral efficiency and can better utilize the available spectrum. This translates into much more capacity at a given channel. Other, more developed microwave systems offer advanced lossless compression techniques, as illustrated in the table below, that allow even more capacity over a given wireless link with additional support for burst peak-to-average issues.
In our non-perfect world, most backhaul networks are not “greenfield” cases, but rather backhaul networks that are evolving. This evolution requires a smooth and risk-free migration plan from legacy networks to next-generation, packet-based communications. This is paramount for network operators, but also common with electrical companies implementing smart-grid applications. Replacing legacy TDM networks with IP-based networks must be carefully planned as it involves a gradual process, with a hybrid network having to provide simultaneous support of TDM and IP/Ethernet communications.
David Patterson, known for his pioneering research that led to RAID, clusters and more, is part of a team at UC Berkeley that recently made its RISC-V processor architecture an open source hardware offering. We talk with Patterson and one of his colleagues behind the effort about the opportunities they see, what new kinds of designs they hope to enable and what it means for today’s commercial processor giants such as Intel, ARM and Imagination Technologies.