Close

HPE Aruba Networking Blogs

Turn Off Your APs??? You Should Try It

By Jonathan Davis, Contributor

Every tech conference is responsible for providing a wireless network for a large number of devices that attendees tote with them. When that conference is put on by Aruba keeping the network up and usable is an absolute priority.

Some companies live and die by the uptime metric maintaining a hands-off approach unless something is broken. To them, the network is proof that they can meet their customers every expectation, real or imagined. At Atmosphere, Aruba takes a different approach. Instead, they see the conference network as a big, real-life lab, and they like to experiment.

During the day two keynote Partha Narasimhan, Aruba's CTO, spent a few minutes talking about one of those experiments. The experiment asked the question: Can we dynamically turn off access points when unneeded for capacity? Aruba used data provided by their product NetInsight along with the conference schedule to make those decisions. It made for a compelling story, which is even more compelling once numbers are applied.

Many wireless engineers understand the vast difference between providing adequate coverage and adequate capacity. In carpeted office space, that may only be a two- or three-fold increase in the number of required APs. However, in many large public venues (LPV), that increase may be six or more times.

This conundrum is further exacerbated by the fact that many wireless networks only need capacity coverage for a portion of each day. In carpeted office space, that may be 12 hours each day. In LPV, the capacity coverage may only be needed for 12 hours each week!

Currently, our wireless networks are incapable of adjusting to these needs, and that is where Aruba's experiment comes into play.

Each access point pulls approximately 20 watts of AC power after accounting for PoE loss in Ethernet cables, power backup, cooling, and AC/DC conversion. Twenty watts of power isn't a lot. But once you consider the number of APss in many enterprise organizations or LPVs, it adds up quickly.

Each AP using 20 watts of power consumes 175kWh of energy each year. The average US commercial power rate is $0.1047 per kWh, which brings the per AP power cost to $18.32 annually.

The national average CO2 footprint of each kWh is 1.222 lbs. Thus, each AP is responsible for 214lbs of CO2 annually.

To further explore the implications, let's walk through two examples.

First, a large enterprise environment with 500 APs at locations spread across the US. During the evening, after most employees have left for the day, half of the APs can be powered down for 12 hours, while still providing full coverage of each facility. Powering down 250 APs for 12 hours each day saves 21,900kWh of electricity.

In this scenario, our annual run-time costs reduce by $2,290. Our annual CO2 footprint reduces by 26,750 lbs.

Our second use case involves a multi-use LPV. It has 450 access points needed for full capacity, which occurs for 20 hours each week. During non-capacity hours, coverage for all occupied spaces requires 100 active APs. By powering down APs, the venue saves 53,872kWh each year. The run-time costs reduce by $5,640 and the carbon footprint by 65,831 lbs of CO2.

Aside from the financial and ecological savings, the simple elegance of the solution requires mentioning. Aruba plans to use NetInsight to enable those APss automatically anytime they are needed for capacity, which simplifies scheduling and ensures the facilities manager or network team have one less thing to consider.

Furthermore, without the signal attenuation provided by guest and attendees, a wireless network built for capacity is a highly non-optimal coverage design. As NetInsight disables APs across a facility during non-peak times, co-channel contention and SNR metrics will improve. That ensures users who are on the coverage based network have the most available airtime and enhances their user experience, proving the mantra, "Less is more."

Disabling APs based on demand was one of the many experiments which Aruba ran on the network during Atmosphere 2018. It might not be as flashy or as technically significant as some of the other items which Partha discussed. But, it has the potential to make a considerable impact on run-time cost, carbon footprint, and the user experience of many wireless networks.

Sources:
Aruba AP340 Specifications http://www.arubanetworks.com/assets/ds/DS_AP340Series.pdf
Average Commercial Energy Price for January 2018 https://www.eia.gov/electricity/monthly/epm_table_grapher.php?t=epmt_5_6_a
Average CO2 emissions per KWHhttps://www.epa.gov/sites/production/files/2015-10/documents/egrid2012_summarytables_0.pdf