Software defined power architecture for 5G cellular

March 30, 2015 //By Jessica Lipsky, EE Times
Next-generation cellular communications will require a drastic reduction in energy consumption, something researchers at Ericsson are calling “ultra-lean transmission.”

“Ultra lean transmission is not only the coolest thing we do for energy performance in 5G, but one of the coolest things we’ll do for energy performance at all,” explained Ylva Jading, a senior researcher at Ericsson. “We need to separate how we think around control and data plane, and really make sure the control part becomes more scalable. This will become absolutely essential to get the lean framework for 5G,” he said.

Macro base stations for 5G cellular, for example, will need to reduce power by a factor of 10. To do so while supporting increased use, engineers must focus on improving signal transmission. Rather than continuously transmitting signals once per millisecond in an “always on” format, a base station should transmit once per 100 milliseconds.

This could be achieved by increasing sleep modes or using advanced beamforming techniques to reduce interference and focus energy on specific users, Jading said. Additionally, a focus on power management software can reduce much of the fixed cost associated with power modules.

“I heard Ulf [Ewaldsson, Ericsson CTO] say that at least we should not grow, and that is pretty amazing considering what we need to pack into 5G,” Jading added. “Normally we would need to add infrastructure but with visualization and so forth we can actually port this off on much less general purpose hardware, than we would have with our traditional design principles.”

The mobile industry has made great strides in power management, a trend moving into infrastructure equipment such as base stations, servers and routers. Highly power efficient processors have driven power management within radio systems, which couldn’t adjust power profiles to traffic until recently, according to Ericsson’s Patrick Le Fevre.

Today’s microprocessors and ASICs with built-in energy optimization are leading to routers that can deliver twice the data with half as much energy. Le Fevre believes traffic management software is the next step. Virtualizing this management layer will allow for more directed bursts of energy.

The idea is that there is something called a traffic manager that can get information about the network where each macro cell and small cell is connected, then create map of how traffic is moving. It can identify different scenarios from almost non-moving traffic to peak traffic. So these algorithms from the traffic manager are connected to traffic grid.

“Building energy management intelligence into their system, that’s where they’re going to get closer and closer to software defined architecture in the router and data center industry,” Le Fevre added.

Design category: