For decades, as data traffic soared, Mobile Operators deployed increasing quantities of spectrum to sustain network performance (bps/km2) required to deliver competitive consumer experience. Throughout this time, the spectrum has shifted from lower center frequencies (700MHz-850MHz) to higher center frequencies (2.6GHz-3.5GHz). In addition, collocating the high bands on the same towers as the low band has remained a network financing reality. Because higher frequencies propagate shorter distances than the lower bands, coverage “gaps” become an issue. To compensate, more power is required to offset these high frequency coverage gaps.
Spectrum deployments have also focused on a second operational reality – delivering enough peak capacity required for a quality consumer experience during moments of high traffic demand and cell site congestion. However, these traffic peaks are highly stochastic and momentary in nature coming and going as consumers move in and out of cells using a plethora of different apps. Yet today, regardless of the actual traffic demand required at any moment in time, all spectrum bands are powered 24×7.
In summary, the decades long evolution of wireless networks has led to sites having between 9 and 18+ cells of predominantly high band spectrum, powered continually to meet temporal traffic peaks.
With 5G rollouts underway this challenge is compounded with 3.5GHz where coverage becomes much more challenging, and operators look to massive MIMO to cover the widening gaps. Massive MIMO doubles or triples energy consumption depending on configuration. With these spectrum and network realities coupled with soaring energy costs, cell site electrical consumption and sustainable networks have become the most important industry subject.
A long-standing leader of ML-based network performance improvement solutions, Opanga Networks introduces Joule ML: a portfolio of rapidly deployable ML software solutions that reduces RAN energy costs by 20% – 40%. Joule deployments are rapidly scaled on 250Gbps capable 2U COTS servers and are suited to bare metal or virtualized environments.
Energized 24/7 Joule ML Optimization